<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE article PUBLIC "-//NLM//DTD Journal Publishing with OASIS Tables v3.0 20080202//EN" "https://jats.nlm.nih.gov/nlm-dtd/publishing/3.0/journalpub-oasis3.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:oasis="http://docs.oasis-open.org/ns/oasis-exchange/table" xml:lang="en" dtd-version="3.0" article-type="research-article">
  <front>
    <journal-meta><journal-id journal-id-type="publisher">WES</journal-id><journal-title-group>
    <journal-title>Wind Energy Science</journal-title>
    <abbrev-journal-title abbrev-type="publisher">WES</abbrev-journal-title><abbrev-journal-title abbrev-type="nlm-ta">Wind Energ. Sci.</abbrev-journal-title>
  </journal-title-group><issn pub-type="epub">2366-7451</issn><publisher>
    <publisher-name>Copernicus Publications</publisher-name>
    <publisher-loc>Göttingen, Germany</publisher-loc>
  </publisher></journal-meta>
    <article-meta>
      <article-id pub-id-type="doi">10.5194/wes-11-4349-2026</article-id><title-group><article-title>Fast models for blockage and gravity-wave effects on wind-farm power output</article-title><alt-title>Fast gravity-wave blockage models for wind-farm power prediction</alt-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author" corresp="yes" rid="aff1">
          <name><surname>Devesse</surname><given-names>Koen</given-names></name>
          <email>koen.devesse@kuleuven.be</email>
        <ext-link>https://orcid.org/0000-0003-2404-6444</ext-link></contrib>
        <contrib contrib-type="author" corresp="no" rid="aff1">
          <name><surname>Meyers</surname><given-names>Johan</given-names></name>
          
        <ext-link>https://orcid.org/0000-0002-2828-4397</ext-link></contrib>
        <aff id="aff1"><label>1</label><institution>Department of Mechanical Engineering, KU Leuven, Leuven, Belgium</institution>
        </aff>
      </contrib-group>
      <author-notes><corresp id="corr1">Koen Devesse (koen.devesse@kuleuven.be)</corresp></author-notes><pub-date><day>17</day><month>November</month><year>2026</year></pub-date>
      
      <volume>11</volume>
      <issue>11</issue>
      <fpage>4349</fpage><lpage>4365</lpage>
      <history>
        <date date-type="received"><day>16</day><month>January</month><year>2026</year></date>
           <date date-type="rev-request"><day>13</day><month>February</month><year>2026</year></date>
           <date date-type="rev-recd"><day>25</day><month>September</month><year>2026</year></date>
           <date date-type="accepted"><day>26</day><month>October</month><year>2026</year></date>
      </history>
      <permissions>
        <copyright-statement>Copyright: © 2026 Koen Devesse</copyright-statement>
        <copyright-year>2026</copyright-year>
      <license license-type="open-access"><license-p>This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this licence, visit <ext-link ext-link-type="uri" xlink:href="https://creativecommons.org/licenses/by/4.0/">https://creativecommons.org/licenses/by/4.0/</ext-link></license-p></license></permissions><self-uri xlink:href="https://wes.copernicus.org/articles/11/4349/2026/wes-11-4349-2026.html">This article is available from https://wes.copernicus.org/articles/11/4349/2026/wes-11-4349-2026.html</self-uri><self-uri xlink:href="https://wes.copernicus.org/articles/11/4349/2026/wes-11-4349-2026.pdf">The full text article is available as a PDF file from https://wes.copernicus.org/articles/11/4349/2026/wes-11-4349-2026.pdf</self-uri>
      <abstract><title>Abstract</title>

      <p id="d2e88">Large offshore wind farms can trigger atmospheric gravity waves, with the associated hydrostatic blockage effect impacting their energy yield. Unfortunately, to date no tools exist that can model this wind-farm–gravity-wave interaction and blockage at a computational cost that is not drastically higher than conventional engineering wake models. To address this, this paper develops ways of speeding up atmospheric perturbation models (APMs) by an order of magnitude by applying insights from the recently proposed two-scale momentum (2SM) theory (<xref ref-type="bibr" rid="bib1.bibx24" id="altparen.1"/>, A2). This leads to two different models, with one using pre-computed farm-level coefficients to compute the turbine forces and power output, while the other relies on repeated wake model evaluations. The core 2SM hypothesis of scale separation and both developed models are validated using a prior LES dataset of a large rectangular wind farm with regular spacing and constant thrust coefficients. In this idealized setup, both fast 2SM–APMs perform well, predicting blockage-corrected farm power at a computational cost that is only a factor 35 slower than a standalone wake model.</p>
  </abstract>
    
<funding-group>
<award-group id="gs1">
<funding-source>Federale Overheidsdienst Economie, K.M.O., Middenstand en Energie</funding-source>
<award-id>-</award-id>
</award-group>
<award-group id="gs2">
<funding-source>HORIZON EUROPE Climate, Energy and Mobility</funding-source>
<award-id>101084205</award-id>
</award-group>
<award-group id="gs3">
<funding-source>Agentschap Innoveren en Ondernemen</funding-source>
<award-id>HBC.2022.0549</award-id>
</award-group>
</funding-group>
</article-meta>
  </front>
<body>
      

<sec id="Ch1.S1" sec-type="intro">
  <label>1</label><title>Introduction</title>
      <p id="d2e103">As offshore wind turbines and farms continue to grow in both size and installed capacity, they reach the point where they form an obstacle to the entire atmospheric boundary layer (ABL) <xref ref-type="bibr" rid="bib1.bibx5" id="paren.2"/>. Large eddy simulation (LES) studies have shown that in conventionally neutral boundary layers (CNBLs), where a neutral ABL is topped by a capping inversion and a stratified upper atmosphere, wind farms can trigger gravity waves, creating upstream pressure gradients and blockage effects <xref ref-type="bibr" rid="bib1.bibx1 bib1.bibx2 bib1.bibx6 bib1.bibx16 bib1.bibx17 bib1.bibx19 bib1.bibx20 bib1.bibx21 bib1.bibx32" id="paren.3"/>. As these mesoscale effects affect turbine power output, they are becoming more important to farm operation and planning.</p>
      <p id="d2e112">Currently, conventional engineering tools rely on the superposition of individual turbine effects, thereby implicitly assuming that their aggregate effect is not larger than the sum of its parts <xref ref-type="bibr" rid="bib1.bibx26" id="paren.4"/>. These <italic>wake models</italic> (WMs) remain widely used because of their low computational cost, which makes them well-suited for farm planning, control, and annual energy production estimates. In contrast, computational fluid dynamics (CFD) methods like Reynolds-averaged Navier–Stokes (RANS), LES, or mesoscale climate models are able to capture most of the relevant flow phenomena. However, their high cost prevents them from being widely used for fast farm power predictions <xref ref-type="bibr" rid="bib1.bibx22" id="paren.5"/>.</p>
      <p id="d2e124">Installing and operating the next generation of offshore wind farms will therefore require new models that can simulate the generation of blockage and gravity waves, while keeping the computational cost orders of magnitude below conventional CFD methods. One promising recently developed tool that fits these requirements is the atmospheric perturbation model (APM) <xref ref-type="bibr" rid="bib1.bibx3 bib1.bibx8 bib1.bibx31" id="paren.6"/>. Building on earlier work by <xref ref-type="bibr" rid="bib1.bibx29" id="text.7"/>, this model simulates large-scale wind-farm–atmosphere interactions explicitly, while coupling to conventional wake models to account for turbine-scale flow effects. In order to capture the local feedback effects between the turbine forces and the mesoscale flow, the governing system of equations is solved repeatedly in a fixed-point iteration. This constitutes the main cost of the model, which is low compared to conventional CFD tools but still several hundred to a thousand times higher than standard wake models <xref ref-type="bibr" rid="bib1.bibx9" id="paren.8"/>.</p>
      <p id="d2e136">One way to further simplify and speed up the modeling of wind-farm–atmosphere interactions is the two-scale momentum (2SM) framework introduced by <xref ref-type="bibr" rid="bib1.bibx24" id="text.9"/>. At its core, this approach assumes a separation between the turbine- and farm-scale flows. As a result of this hypothesis, variations in farm power output due to mesoscale effects are fully determined by the associated changes in the average velocity over the farm. Based on high-fidelity LES results, this hypothesis has been confirmed for large wind farms in idealized conditions <xref ref-type="bibr" rid="bib1.bibx13 bib1.bibx14" id="paren.10"/>.</p>
      <p id="d2e146">Other recent studies have proposed wind farm models relying on similar ideas of scale separation. <xref ref-type="bibr" rid="bib1.bibx35" id="text.11"/> and <xref ref-type="bibr" rid="bib1.bibx33" id="text.12"/> proposed actuator-farm models for RANS and LES, respectively, allowing these methods to simulate wind farms without resolving the turbine-scale flow. Similarly, <xref ref-type="bibr" rid="bib1.bibx11" id="text.13"/> developed a wind-farm parametrization for mesoscale weather models, which due to their kilometer-scale resolution cannot represent individual turbines, based on matching the average velocity in the farm between the mesoscale model and a wake model. However, since these studies are not as explicit in their theoretical treatment of this scale separation, we will refer to the overall idea as the two-scale momentum approach, following <xref ref-type="bibr" rid="bib1.bibx24" id="text.14"/>.</p>
      <p id="d2e161">Our goal is to develop fast models for predicting wind-farm power output that account for blockage and gravity-wave effects. To that end, the 2SM framework can be useful, as it implies that flow solvers simulating the farm-scale flow do not have to resolve the turbine-scale flow. This would allow for coarser grid resolutions and reduced computational costs. However, even when using coarser grids, conventional CFD tools are still multiple orders of magnitude slower than engineering wake models <xref ref-type="bibr" rid="bib1.bibx25" id="paren.15"/>. In contrast, the APM already uses a coarse resolution, while its main cost lies in capturing the local feedback effects between the turbine forces and the mesoscale flow. Therefore, we focus on combining the 2SM and APM approaches, leveraging insights from the former to speed up and simplify the latter.</p>
      <p id="d2e167">Since, according to the 2SM hypothesis, a farm's power output only depends on the average velocity in the farm, the APM may neglect local changes in the turbine force distribution within the farm induced by gravity-wave effects. Instead, this force distribution can be scaled uniformly over the farm to match the total farm force as predicted by 2SM theory. Due to the linear nature of the APM, this transforms the model's governing equations into a system of scalar equations, drastically lowering its computational cost. We use this to develop two versions of the sped-up APM, one which uses farm thrust and power coefficients to model the farm, following earlier 2SM studies and one which is directly coupled to a standard wake model. Both models are validated using the LES dataset of a large wind farm operating in idealized CNBL conditions from <xref ref-type="bibr" rid="bib1.bibx19" id="text.16"/>.</p>
      <p id="d2e173">The remainder of this paper is outlined as follows. Section <xref ref-type="sec" rid="Ch1.S2"/> provides an overview of 2SM theory and demonstrates its core hypothesis using the LES results of <xref ref-type="bibr" rid="bib1.bibx19" id="text.17"/>. Section <xref ref-type="sec" rid="Ch1.S3"/> then introduces the APM's governing equations and how they are solved. Combining these two, Sect. <xref ref-type="sec" rid="Ch1.S4"/> shows how a 2SM–APM can cheaply calculate a farm's power output if its farm-level thrust and power coefficients are known. Building on this, Sect. <xref ref-type="sec" rid="Ch1.S5"/> develops an alternative method that couples the farm-averaged APM output directly to a wake model, thereby eliminating the need for these farm coefficients but requiring repeated wake model evaluations. Section <xref ref-type="sec" rid="Ch1.S6"/> compares the performance, advantages, and limitations of both models. Finally, Sect. <xref ref-type="sec" rid="Ch1.S7"/> gives an overview of the results and an outlook for future research.</p>
</sec>
<sec id="Ch1.S2">
  <label>2</label><title>Two-scale momentum theory</title>
      <p id="d2e200">Recently, <xref ref-type="bibr" rid="bib1.bibx14" id="text.18"/> demonstrated that the 2SM theory developed by <xref ref-type="bibr" rid="bib1.bibx24" id="text.19"/> can be used to simplify the interaction between mesoscale atmospheric flows and wind farm operation. The core hypothesis of 2SM theory is a separation of internal and external effects and the associated spatial scales. Internal effects, such as turbine setpoints and wakes, determine the ratios of the individual turbine disk velocities <inline-formula><mml:math id="M1" display="inline"><mml:mrow><mml:msub><mml:mi>S</mml:mi><mml:mi>k</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> to the average streamwise velocity <inline-formula><mml:math id="M2" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> in the farm region. External effects, such as large-scale momentum advection, entrainment above the farm, or gravity-wave-induced pressure gradients, act on the farm scale and affect the average velocity within the farm <inline-formula><mml:math id="M3" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>. The total farm power <inline-formula><mml:math id="M4" display="inline"><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> can then be expressed as

          <disp-formula id="Ch1.E1" content-type="numbered"><label>1</label><mml:math id="M5" display="block"><mml:mtable class="split" rowspacing="0.2ex" displaystyle="true" columnalign="right left"><mml:mtr><mml:mtd/><mml:mtd><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:mi mathvariant="italic">ρ</mml:mi><mml:msubsup><mml:mi>U</mml:mi><mml:mi mathvariant="normal">f</mml:mi><mml:mn mathvariant="normal">3</mml:mn></mml:msubsup></mml:mrow><mml:mn mathvariant="normal">2</mml:mn></mml:mfrac></mml:mstyle><mml:msub><mml:mi>A</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mo>,</mml:mo><mml:mspace width="1em" linebreak="nobreak"/><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:munderover><mml:mo movablelimits="false">∑</mml:mo><mml:mrow><mml:mi>k</mml:mi><mml:mo>=</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow><mml:mrow><mml:msub><mml:mi>N</mml:mi><mml:mi mathvariant="normal">t</mml:mi></mml:msub></mml:mrow></mml:munderover><mml:msubsup><mml:mi>C</mml:mi><mml:mrow><mml:mi mathvariant="normal">p</mml:mi><mml:mo>,</mml:mo><mml:mi>k</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msubsup><mml:mi>S</mml:mi><mml:mi>k</mml:mi><mml:mn mathvariant="normal">3</mml:mn></mml:msubsup></mml:mrow><mml:mrow><mml:msubsup><mml:mi>U</mml:mi><mml:mi mathvariant="normal">f</mml:mi><mml:mn mathvariant="normal">3</mml:mn></mml:msubsup></mml:mrow></mml:mfrac></mml:mstyle><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mi>A</mml:mi><mml:mi>k</mml:mi></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>A</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>,</mml:mo></mml:mrow></mml:mtd></mml:mtr><mml:mtr><mml:mtd/><mml:mtd><mml:mrow><mml:msub><mml:mi>A</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:munderover><mml:mo movablelimits="false">∑</mml:mo><mml:mrow><mml:mi>k</mml:mi><mml:mo>=</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow><mml:mrow><mml:msub><mml:mi>N</mml:mi><mml:mi mathvariant="normal">t</mml:mi></mml:msub></mml:mrow></mml:munderover><mml:msub><mml:mi>A</mml:mi><mml:mi>k</mml:mi></mml:msub><mml:mo>,</mml:mo></mml:mrow></mml:mtd></mml:mtr></mml:mtable></mml:math></disp-formula>

        where <inline-formula><mml:math id="M6" display="inline"><mml:mi mathvariant="italic">ρ</mml:mi></mml:math></inline-formula> is the air density, <inline-formula><mml:math id="M7" display="inline"><mml:mrow><mml:msub><mml:mi>A</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> is the total rotor area of the farm turbines, <inline-formula><mml:math id="M8" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> is the internal farm power coefficient, <inline-formula><mml:math id="M9" display="inline"><mml:mrow><mml:msub><mml:mi>N</mml:mi><mml:mi mathvariant="normal">t</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> is the number of turbines, and <inline-formula><mml:math id="M10" display="inline"><mml:mrow><mml:msub><mml:mi>A</mml:mi><mml:mi>k</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M11" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mrow><mml:mi mathvariant="normal">p</mml:mi><mml:mo>,</mml:mo><mml:mi>k</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> are the turbine disk areas and disk-based power coefficients, respectively, with <inline-formula><mml:math id="M12" display="inline"><mml:mi>k</mml:mi></mml:math></inline-formula> indicating the turbine number.</p>
      <p id="d2e485"><xref ref-type="bibr" rid="bib1.bibx24" id="text.20"/> make an analogous argument for the farm thrust, so that

          <disp-formula id="Ch1.E2" content-type="numbered"><label>2</label><mml:math id="M13" display="block"><mml:mrow><mml:mi>F</mml:mi><mml:mo>=</mml:mo><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msubsup><mml:mi>U</mml:mi><mml:mrow><mml:mspace width="-0.125em" linebreak="nobreak"/><mml:mi mathvariant="normal">f</mml:mi></mml:mrow><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup></mml:mrow><mml:mn mathvariant="normal">2</mml:mn></mml:mfrac></mml:mstyle><mml:msub><mml:mi>A</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mo>,</mml:mo><mml:mspace width="1em" linebreak="nobreak"/><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:munderover><mml:mo movablelimits="false">∑</mml:mo><mml:mrow><mml:mi>k</mml:mi><mml:mo>=</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow><mml:mrow><mml:msub><mml:mi>N</mml:mi><mml:mi mathvariant="normal">t</mml:mi></mml:msub></mml:mrow></mml:munderover><mml:msubsup><mml:mi>C</mml:mi><mml:mrow><mml:mi mathvariant="normal">T</mml:mi><mml:mo>,</mml:mo><mml:mi>k</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msubsup><mml:mi>S</mml:mi><mml:mi>k</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup></mml:mrow><mml:mrow><mml:msubsup><mml:mi>U</mml:mi><mml:mi mathvariant="normal">f</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup></mml:mrow></mml:mfrac></mml:mstyle><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mi>A</mml:mi><mml:mi>k</mml:mi></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>A</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

        where <inline-formula><mml:math id="M14" display="inline"><mml:mi>F</mml:mi></mml:math></inline-formula> is the total farm force (scaled with air density), <inline-formula><mml:math id="M15" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> is the farm thrust coefficient, and <inline-formula><mml:math id="M16" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mrow><mml:mi mathvariant="normal">T</mml:mi><mml:mo>,</mml:mo><mml:mi>k</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> are the disk-based turbine thrust coefficients. Like <inline-formula><mml:math id="M17" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>, the internal farm thrust coefficient <inline-formula><mml:math id="M18" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> is assumed to only depend on the internal problem.</p>
      <p id="d2e666"><xref ref-type="bibr" rid="bib1.bibx14" id="text.21"/> investigated this assumption using LES results of large wind farms with uniform turbine densities in below-rated conditions, so that the <inline-formula><mml:math id="M19" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mrow><mml:mi mathvariant="normal">T</mml:mi><mml:mo>,</mml:mo><mml:mi>k</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> are constant over the farm. They found that the internal power coefficient <inline-formula><mml:math id="M20" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> as defined in Eq. (<xref ref-type="disp-formula" rid="Ch1.E1"/>) is a good heuristic for internal effects, and remained constant over cases with different mesoscale flow behavior. This then implies that <inline-formula><mml:math id="M21" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> can be calculated using small-scale simulations of the same farm or engineering wake models, while the problem of including mesoscale stratification effects in farm power calculations is simplified to solving only the external problem, i.e., estimating <inline-formula><mml:math id="M22" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mrow><mml:mspace width="-0.125em" linebreak="nobreak"/><mml:mi mathvariant="normal">f</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula>.</p>
      <p id="d2e730">We demonstrate this by taking <inline-formula><mml:math id="M23" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> from a neutral flow simulation and <inline-formula><mml:math id="M24" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mrow><mml:mspace width="-0.125em" linebreak="nobreak"/><mml:mi mathvariant="normal">f</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula> from different CNBL simulations with different meso-scale stratification effects but the same farm layout and geostrophic wind speed. We then estimate the power produced in the CNBL cases using Eq. (<xref ref-type="disp-formula" rid="Ch1.E1"/>) and compare it against the actual production. For this, we use the prior LES dataset from <xref ref-type="bibr" rid="bib1.bibx19" id="text.22"/>, which consists of 36 simulations of the same large wind farm in idealized steady CNBL conditions <xref ref-type="bibr" rid="bib1.bibx18" id="paren.23"/>, which was also used by <xref ref-type="bibr" rid="bib1.bibx14" id="text.24"/>. This farm has 160 10 MW turbines, arranged uniformly in a rectangular staggered layout, operating at below-rated conditions, with the disk-based thrust coefficients fixed to <inline-formula><mml:math id="M25" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mrow><mml:mi mathvariant="normal">T</mml:mi><mml:mo>,</mml:mo><mml:mi>k</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:mn mathvariant="normal">1.94</mml:mn></mml:mrow></mml:math></inline-formula>. Each simulation has a different combination of boundary layer heights <inline-formula><mml:math id="M26" display="inline"><mml:mi>H</mml:mi></mml:math></inline-formula>, capping inversion strengths <inline-formula><mml:math id="M27" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:mrow></mml:math></inline-formula>, and atmospheric lapse rates <inline-formula><mml:math id="M28" display="inline"><mml:mi mathvariant="normal">Γ</mml:mi></mml:math></inline-formula>. The different cases are identified with these parameters, so that <inline-formula><mml:math id="M29" display="inline"><mml:mi>H</mml:mi></mml:math></inline-formula>500-<inline-formula><mml:math id="M30" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:mrow></mml:math></inline-formula>5-<inline-formula><mml:math id="M31" display="inline"><mml:mrow><mml:mi mathvariant="normal">Γ</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:mrow></mml:math></inline-formula> denotes the case with the precursor profile initialized with these values. For each boundary layer height, the dataset also includes simulations of a wind-farm in truly neutral flow and of a stand-alone turbine with <inline-formula><mml:math id="M32" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi><mml:mo>=</mml:mo><mml:mn mathvariant="normal">5</mml:mn></mml:mrow></mml:math></inline-formula> K and <inline-formula><mml:math id="M33" display="inline"><mml:mrow><mml:mi mathvariant="normal">Γ</mml:mi><mml:mo>=</mml:mo><mml:mn mathvariant="normal">4</mml:mn></mml:mrow></mml:math></inline-formula> K km<sup>−1</sup>. A detailed description of the LES setup and results is given by <xref ref-type="bibr" rid="bib1.bibx19" id="text.25"/>.</p>
      <p id="d2e887">We calculate <inline-formula><mml:math id="M35" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mrow><mml:mspace linebreak="nobreak" width="-0.125em"/><mml:mi mathvariant="normal">f</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula> as a straightforward average over the wind farm, so that

          <disp-formula id="Ch1.E3" content-type="numbered"><label>3</label><mml:math id="M36" display="block"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mrow><mml:mspace linebreak="nobreak" width="-0.125em"/><mml:mi mathvariant="normal">f</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:munderover><mml:mo movablelimits="false">∫</mml:mo><mml:mn mathvariant="normal">0</mml:mn><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:munderover><mml:munder><mml:mo movablelimits="false">∬</mml:mo><mml:mrow><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:munder><mml:mi>u</mml:mi><mml:mspace width="0.33em" linebreak="nobreak"/><mml:mtext>d</mml:mtext><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mtext>d</mml:mtext><mml:mi>z</mml:mi><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

        where <inline-formula><mml:math id="M37" display="inline"><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> is the height of the 2SM averaging volume, and <inline-formula><mml:math id="M38" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> the planform area of the wind farm. We take <inline-formula><mml:math id="M39" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> to be a rectangle around the farm edge, thereby including the outer edges of the turbine disks at the sides. We set the height <inline-formula><mml:math id="M40" display="inline"><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mrow><mml:mspace width="-0.125em" linebreak="nobreak"/><mml:mi mathvariant="normal">f</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula> of the volume to twice the turbine hub height <inline-formula><mml:math id="M41" display="inline"><mml:mrow><mml:msub><mml:mi>z</mml:mi><mml:mi mathvariant="normal">h</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, as this corresponds to the setup of the APM (see Sect. <xref ref-type="sec" rid="Ch1.S4.SS1"/>). This height is above the turbine tip height but below the ABL height for all the cases in the dataset of <xref ref-type="bibr" rid="bib1.bibx19" id="text.26"/>. While some of the <inline-formula><mml:math id="M42" display="inline"><mml:mrow><mml:mi>H</mml:mi><mml:mn mathvariant="normal">150</mml:mn></mml:mrow></mml:math></inline-formula> cases have lower unperturbed ABL heights, strong vertical displacements of the capping inversions over the wind farm ensure that even there the ABL height is above the 2SM control volume. The exact dimensions of the resulting control volume are listed in Table <xref ref-type="table" rid="T1"/>. This simple approach is possible because the analyzed farm has a convex and uniform layout. More realistic farms might require a more complex horizontal averaging method such as the one proposed by <xref ref-type="bibr" rid="bib1.bibx35" id="text.27"/>, where the wind speed is weighted by Gaussian distributions centered around the turbines. Alternatively, a farm could be split in two or more homogeneous regions. Further investigation of such approaches is beyond the scope of this study and left as promising future work.</p>

<table-wrap id="T1"><label>Table 1</label><caption><p id="d2e1052">Specification of the 2SM farm averaging volume for the cases of <xref ref-type="bibr" rid="bib1.bibx19" id="text.28"/>.</p></caption><oasis:table frame="topbot"><oasis:tgroup cols="2">
     <oasis:colspec colnum="1" colname="col1" align="left"/>
     <oasis:colspec colnum="2" colname="col2" align="left"/>
     <oasis:thead>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">Dimension</oasis:entry>
         <oasis:entry colname="col2">Value</oasis:entry>
       </oasis:row>
     </oasis:thead>
     <oasis:tbody>
       <oasis:row>
         <oasis:entry colname="col1">Height <inline-formula><mml:math id="M43" display="inline"><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">238 m</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Length</oasis:entry>
         <oasis:entry colname="col2">14.85 km</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Width</oasis:entry>
         <oasis:entry colname="col2">9.603 km</oasis:entry>
       </oasis:row>
     </oasis:tbody>
   </oasis:tgroup></oasis:table></table-wrap>

      <p id="d2e1119">To calculate <inline-formula><mml:math id="M44" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>, we use the 3D time-averaged velocity fields of case <inline-formula><mml:math id="M45" display="inline"><mml:mi>H</mml:mi></mml:math></inline-formula>500-<inline-formula><mml:math id="M46" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:mrow></mml:math></inline-formula>0-<inline-formula><mml:math id="M47" display="inline"><mml:mi mathvariant="normal">Γ</mml:mi></mml:math></inline-formula>0. After averaging the streamwise velocity field to obtain <inline-formula><mml:math id="M48" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mrow><mml:mspace linebreak="nobreak" width="-0.125em"/><mml:mi mathvariant="normal">f</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula>, we obtain <inline-formula><mml:math id="M49" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> from Eq. (<xref ref-type="disp-formula" rid="Ch1.E1"/>). Subsequently, the farm power outputs for the 36 CNBL cases are calculated by averaging their <inline-formula><mml:math id="M50" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mrow><mml:mspace width="-0.125em" linebreak="nobreak"/><mml:mi mathvariant="normal">f</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula> and combining it with the prior <inline-formula><mml:math id="M51" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> through Eq. (<xref ref-type="disp-formula" rid="Ch1.E1"/>). In order to avoid issues with the power calculation in LES related to the used actuator-disk method and grid size <xref ref-type="bibr" rid="bib1.bibx27" id="paren.29"/>, the resulting power outputs are scaled with the power outputs <inline-formula><mml:math id="M52" display="inline"><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mn mathvariant="normal">0</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> of standalone turbines operating at the same inflow wind speeds, as was done by <xref ref-type="bibr" rid="bib1.bibx19" id="text.30"/>. The results are shown in Fig. <xref ref-type="fig" rid="F1"/>. The root-mean square error (RMSE) on <inline-formula><mml:math id="M53" display="inline"><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> is 1 % of the farm installed capacity, which shows that a <inline-formula><mml:math id="M54" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> value calculated from a neutral simulation remains applicable over a wide range of CNBL flows. We conclude that the 2SM assumption, i.e.,  that <inline-formula><mml:math id="M55" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M56" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> are solely determined by turbine-level setpoints and wake positions, holds under the idealized conditions tested here.</p>

      <fig id="F1"><label>Figure 1</label><caption><p id="d2e1289">A comparison of farm power <inline-formula><mml:math id="M57" display="inline"><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> as calculated by LES (<inline-formula><mml:math id="M58" display="inline"><mml:mi>x</mml:mi></mml:math></inline-formula> axis) and Eq. (<xref ref-type="disp-formula" rid="Ch1.E1"/>) (<inline-formula><mml:math id="M59" display="inline"><mml:mi>y</mml:mi></mml:math></inline-formula> axis) for all 36 CNBL cases from the dataset by <xref ref-type="bibr" rid="bib1.bibx19" id="text.31"/>, scaled with the power outputs of standalone turbine simulations <inline-formula><mml:math id="M60" display="inline"><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mn mathvariant="normal">0</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>. The power coefficient <inline-formula><mml:math id="M61" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> is calculated using a neutral flow simulation (i.e., case <inline-formula><mml:math id="M62" display="inline"><mml:mi>H</mml:mi></mml:math></inline-formula>500-<inline-formula><mml:math id="M63" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:mrow></mml:math></inline-formula>0-<inline-formula><mml:math id="M64" display="inline"><mml:mi mathvariant="normal">Γ</mml:mi></mml:math></inline-formula>0 from the same dataset).</p></caption>
        <graphic xlink:href="https://wes.copernicus.org/articles/11/4349/2026/wes-11-4349-2026-f01.png"/>

      </fig>


</sec>
<sec id="Ch1.S3">
  <label>3</label><title>Atmospheric perturbation model</title>
      <p id="d2e1387">If <inline-formula><mml:math id="M65" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">P</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M66" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> are known, calculating the farm power output with the 2SM approach then comes down to finding the average farm velocity <inline-formula><mml:math id="M67" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>. The APM is well-suited for this, as it is a fast model for mesoscale atmosphere–farm interaction. Therefore, this section presents a brief overview of the APM.</p>
      <p id="d2e1427">The APM is a simplified model, and treats the ABL as two layers of vertically averaged flow, separated by a pliant surface following the flow streamlines <xref ref-type="bibr" rid="bib1.bibx3" id="paren.32"/>. In each of these layers, the continuity and horizontal momentum equations are solved explicitly for the average velocities and layer thicknesses. The wind-farm forces are added in the momentum equations for the lower layer, also called the wind-farm layer, while only affecting the upper layer via interactions through the pliant surface. The resulting model structure is visualized in Fig. <xref ref-type="fig" rid="F2"/>.</p>

      <fig id="F2" specific-use="star"><label>Figure 2</label><caption><p id="d2e1437">Schematic representation of the atmospheric perturbation model. Figure reproduced from <xref ref-type="bibr" rid="bib1.bibx3" id="text.33"/>.</p></caption>
        <graphic xlink:href="https://wes.copernicus.org/articles/11/4349/2026/wes-11-4349-2026-f02.png"/>

      </fig>

      <p id="d2e1450">The main advantage of the APM is that the vertical averaging drastically lowers the computational cost of the model compared to conventional CFD or mesoscale simulations, since it essentially reduces the vertical resolution to the ABL height. The horizontal resolution is likewise reduced by applying a Gaussian filter with a kernel size of 1 km <xref ref-type="bibr" rid="bib1.bibx3 bib1.bibx8" id="paren.34"/>.</p>
      <p id="d2e1456">Mathematically, the APM can be derived by applying these filtering and height-averaging operators to the flow. For example, the lower-layer velocity is defined as <xref ref-type="bibr" rid="bib1.bibx8" id="paren.35"/>

          <disp-formula id="Ch1.E4" content-type="numbered"><label>4</label><mml:math id="M68" display="block"><mml:mrow><mml:mo>〈</mml:mo><mml:mover accent="true"><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mo mathvariant="normal" stretchy="true">̃</mml:mo></mml:mover><mml:msub><mml:mo>〉</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>=</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mrow><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:munderover><mml:mo movablelimits="false">∫</mml:mo><mml:mn mathvariant="normal">0</mml:mn><mml:mrow><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:munderover><mml:munder><mml:mo movablelimits="false">∬</mml:mo><mml:mrow><mml:msub><mml:mi>L</mml:mi><mml:mi>x</mml:mi></mml:msub><mml:mo>×</mml:mo><mml:msub><mml:mi>L</mml:mi><mml:mi>y</mml:mi></mml:msub></mml:mrow></mml:munder><mml:msub><mml:mi>G</mml:mi><mml:mi mathvariant="normal">ℓ</mml:mi></mml:msub><mml:mo>(</mml:mo><mml:mi>x</mml:mi><mml:mo>-</mml:mo><mml:msup><mml:mi>x</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mo>,</mml:mo><mml:mi>y</mml:mi><mml:mo>-</mml:mo><mml:msup><mml:mi>y</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mo>)</mml:mo><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mo>(</mml:mo><mml:msup><mml:mi>x</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mo>,</mml:mo><mml:msup><mml:mi>y</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mo>,</mml:mo><mml:mi>z</mml:mi><mml:mo>)</mml:mo><mml:mi mathvariant="normal">d</mml:mi><mml:msup><mml:mi>x</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mi mathvariant="normal">d</mml:mi><mml:msup><mml:mi>y</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mi mathvariant="normal">d</mml:mi><mml:mi>z</mml:mi><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

        where <inline-formula><mml:math id="M69" display="inline"><mml:mi>x</mml:mi></mml:math></inline-formula>, <inline-formula><mml:math id="M70" display="inline"><mml:mi>y</mml:mi></mml:math></inline-formula>, and <inline-formula><mml:math id="M71" display="inline"><mml:mi>z</mml:mi></mml:math></inline-formula> denote the streamwise, spanwise, and vertical coordinates, respectively, the subscript 1 indicates the lower layer, the angled brackets and tilde denote the height-averaging and filtering, respectively, <inline-formula><mml:math id="M72" display="inline"><mml:mi mathvariant="bold-italic">u</mml:mi></mml:math></inline-formula> is the horizontal velocity vector, <inline-formula><mml:math id="M73" display="inline"><mml:mrow><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> is the lower-layer thickness, <inline-formula><mml:math id="M74" display="inline"><mml:mrow><mml:msub><mml:mi>L</mml:mi><mml:mi>x</mml:mi></mml:msub><mml:mo>×</mml:mo><mml:msub><mml:mi>L</mml:mi><mml:mi>y</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> is the APM domain size, and <inline-formula><mml:math id="M75" display="inline"><mml:mrow><mml:msub><mml:mi>G</mml:mi><mml:mi mathvariant="normal">ℓ</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> is a Gaussian kernel with a fixed filter length <inline-formula><mml:math id="M76" display="inline"><mml:mrow><mml:mi mathvariant="normal">ℓ</mml:mi><mml:mo>=</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:math></inline-formula> km. The pliant surface <inline-formula><mml:math id="M77" display="inline"><mml:mrow><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> is defined as following the streamlines of the horizontally filtered flow, so that <xref ref-type="bibr" rid="bib1.bibx3" id="paren.36"/>

          <disp-formula id="Ch1.E5" content-type="numbered"><label>5</label><mml:math id="M78" display="block"><mml:mrow><mml:mover accent="true"><mml:mi>w</mml:mi><mml:mo stretchy="true" mathvariant="normal">̃</mml:mo></mml:mover><mml:mo>(</mml:mo><mml:mi>x</mml:mi><mml:mo>,</mml:mo><mml:mi>y</mml:mi><mml:mo>,</mml:mo><mml:msub><mml:mi>z</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>)</mml:mo><mml:mo>=</mml:mo><mml:mover accent="true"><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mo stretchy="true" mathvariant="normal">̃</mml:mo></mml:mover><mml:mo>(</mml:mo><mml:mi>x</mml:mi><mml:mo>,</mml:mo><mml:mi>y</mml:mi><mml:mo>,</mml:mo><mml:msub><mml:mi>z</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>)</mml:mo><mml:mo>⋅</mml:mo><mml:msub><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi></mml:msub><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

        where <inline-formula><mml:math id="M79" display="inline"><mml:mi>w</mml:mi></mml:math></inline-formula> is the vertical velocity component and <inline-formula><mml:math id="M80" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> is the horizontal del operator. The flow is then further linearized around a uniform background state, so that for the lower layer

          <disp-formula id="Ch1.E6" content-type="numbered"><label>6</label><mml:math id="M81" display="block"><mml:mrow><mml:mo>〈</mml:mo><mml:mover accent="true"><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mo mathvariant="normal" stretchy="true">̃</mml:mo></mml:mover><mml:msub><mml:mo>〉</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>=</mml:mo><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup><mml:mo>,</mml:mo><mml:mspace width="1em" linebreak="nobreak"/><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>=</mml:mo><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

        where <inline-formula><mml:math id="M82" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M83" display="inline"><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> are the unperturbed layer velocity and thickness in the absence of the wind farm and <inline-formula><mml:math id="M84" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M85" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> are the perturbations to this state. For the upper layer, the notation is analogous.</p>
      <p id="d2e1881">The governing equations of the model can then be found by applying the above operators to the incompressible Reynolds-averaged steady-state Navier–Stokes equations and linearizing the result. Thorough descriptions of this procedure can be found in <xref ref-type="bibr" rid="bib1.bibx3" id="text.37"/> and <xref ref-type="bibr" rid="bib1.bibx8" id="text.38"/>. The resulting set of equations is

              <disp-formula specific-use="gather" content-type="numbered"><mml:math id="M86" display="block"><mml:mtable displaystyle="true"><mml:mlabeledtr id="Ch1.E7"><mml:mtd><mml:mtext>7</mml:mtext></mml:mtd><mml:mtd><mml:mrow><mml:mstyle class="stylechange" displaystyle="true"/><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>⋅</mml:mo><mml:msub><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi></mml:msub><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:msub><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi></mml:msub><mml:mo>⋅</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:mn mathvariant="normal">0</mml:mn><mml:mo>,</mml:mo></mml:mrow></mml:mtd></mml:mlabeledtr><mml:mlabeledtr id="Ch1.E8"><mml:mtd><mml:mtext>8</mml:mtext></mml:mtd><mml:mtd><mml:mrow><mml:mstyle class="stylechange" displaystyle="true"/><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mo>⋅</mml:mo><mml:msub><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi></mml:msub><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:msub><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi></mml:msub><mml:mo>⋅</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mn mathvariant="normal">2</mml:mn><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:mn mathvariant="normal">0</mml:mn><mml:mo>,</mml:mo></mml:mrow></mml:mtd></mml:mlabeledtr><mml:mlabeledtr id="Ch1.E9"><mml:mtd><mml:mtext>9</mml:mtext></mml:mtd><mml:mtd><mml:mrow><mml:mstyle displaystyle="true" class="stylechange"/><mml:mtable rowspacing="0.2ex" class="split" displaystyle="true" columnalign="right left"><mml:mtr><mml:mtd><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>⋅</mml:mo><mml:msub><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi></mml:msub><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo></mml:mrow></mml:mtd><mml:mtd><mml:mrow><mml:mo>-</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mi mathvariant="italic">ρ</mml:mi></mml:mfrac></mml:mstyle><mml:msub><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi></mml:msub><mml:msup><mml:mi>p</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mo>+</mml:mo><mml:msub><mml:mi>f</mml:mi><mml:mi mathvariant="normal">c</mml:mi></mml:msub><mml:mi mathvariant="bold">J</mml:mi><mml:mo>⋅</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup><mml:mo>+</mml:mo><mml:msub><mml:mi mathvariant="italic">ν</mml:mi><mml:mrow><mml:mi>t</mml:mi><mml:mo>,</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msub><mml:msubsup><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:mtd></mml:mtr><mml:mtr><mml:mtd/><mml:mtd><mml:mrow><mml:mo>+</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msup><mml:mi mathvariant="bold">D</mml:mi><mml:mo>′</mml:mo></mml:msup></mml:mrow><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>⋅</mml:mo><mml:msubsup><mml:mi mathvariant="normal">Δ</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup><mml:msup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mo>-</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msup><mml:mi mathvariant="bold">C</mml:mi><mml:mo>′</mml:mo></mml:msup></mml:mrow><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>⋅</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup><mml:mo>-</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">T</mml:mi><mml:mrow><mml:mi mathvariant="normal">h</mml:mi><mml:mn mathvariant="normal">3</mml:mn><mml:mo>,</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msub><mml:mo>-</mml:mo><mml:msub><mml:mi mathvariant="bold-italic">T</mml:mi><mml:mrow><mml:mi mathvariant="normal">h</mml:mi><mml:mn mathvariant="normal">3</mml:mn><mml:mo>,</mml:mo><mml:mn mathvariant="normal">0</mml:mn></mml:mrow></mml:msub></mml:mrow><mml:mrow><mml:msubsup><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup></mml:mrow></mml:mfrac></mml:mstyle><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mtd></mml:mtr><mml:mtr><mml:mtd/><mml:mtd><mml:mrow><mml:mo>+</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mover accent="true"><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mo stretchy="true" mathvariant="normal">̃</mml:mo></mml:mover><mml:mi mathvariant="normal">wf</mml:mi></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mover accent="true"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi mathvariant="bold-italic">τ</mml:mi></mml:mrow><mml:mo stretchy="true" mathvariant="normal">̃</mml:mo></mml:mover><mml:mi mathvariant="normal">wf</mml:mi></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>+</mml:mo><mml:msub><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi></mml:msub><mml:mo>⋅</mml:mo><mml:msub><mml:mfenced close="〉" open="〈"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">τ</mml:mi><mml:mrow><mml:mi mathvariant="normal">d</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">hh</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:mfenced><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>,</mml:mo></mml:mrow></mml:mtd></mml:mtr></mml:mtable></mml:mrow></mml:mtd></mml:mlabeledtr><mml:mlabeledtr id="Ch1.E10"><mml:mtd><mml:mtext>10</mml:mtext></mml:mtd><mml:mtd><mml:mrow><mml:mstyle class="stylechange" displaystyle="true"/><mml:mtable rowspacing="0.2ex" class="split" displaystyle="true" columnalign="right left"><mml:mtr><mml:mtd><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mo>⋅</mml:mo><mml:msub><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi></mml:msub><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mn mathvariant="normal">2</mml:mn><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo></mml:mrow></mml:mtd><mml:mtd><mml:mrow><mml:mo>-</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mi mathvariant="italic">ρ</mml:mi></mml:mfrac></mml:mstyle><mml:msub><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi></mml:msub><mml:msup><mml:mi>p</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mo>+</mml:mo><mml:msub><mml:mi>f</mml:mi><mml:mi mathvariant="normal">c</mml:mi></mml:msub><mml:mi mathvariant="bold">J</mml:mi><mml:mo>⋅</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mn mathvariant="normal">2</mml:mn><mml:mo>′</mml:mo></mml:msubsup><mml:mo>+</mml:mo><mml:msub><mml:mi mathvariant="italic">ν</mml:mi><mml:mrow><mml:mi>t</mml:mi><mml:mo>,</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:msub><mml:msubsup><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mn mathvariant="normal">2</mml:mn><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:mtd></mml:mtr><mml:mtr><mml:mtd/><mml:mtd><mml:mrow><mml:mo>-</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msup><mml:mi mathvariant="bold">D</mml:mi><mml:mo>′</mml:mo></mml:msup></mml:mrow><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>⋅</mml:mo><mml:msubsup><mml:mi mathvariant="normal">Δ</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup><mml:msup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mo>+</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">T</mml:mi><mml:mrow><mml:mi mathvariant="normal">h</mml:mi><mml:mn mathvariant="normal">3</mml:mn><mml:mo>,</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msub></mml:mrow><mml:mrow><mml:msubsup><mml:mi>H</mml:mi><mml:mn mathvariant="normal">2</mml:mn><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup></mml:mrow></mml:mfrac></mml:mstyle><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mo>-</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mover accent="true"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi mathvariant="bold-italic">τ</mml:mi></mml:mrow><mml:mo stretchy="true" mathvariant="normal">̃</mml:mo></mml:mover><mml:mi mathvariant="normal">wf</mml:mi></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>,</mml:mo></mml:mrow></mml:mtd></mml:mtr></mml:mtable></mml:mrow></mml:mtd></mml:mlabeledtr></mml:mtable></mml:math></disp-formula>

        where the subscript <inline-formula><mml:math id="M87" display="inline"><mml:mi>h</mml:mi></mml:math></inline-formula> indicates horizontal vector components, <inline-formula><mml:math id="M88" display="inline"><mml:mrow><mml:msup><mml:mi>p</mml:mi><mml:mo>′</mml:mo></mml:msup></mml:mrow></mml:math></inline-formula> is the pressure perturbation induced by gravity waves, <inline-formula><mml:math id="M89" display="inline"><mml:mi mathvariant="italic">ρ</mml:mi></mml:math></inline-formula> is the air density, <inline-formula><mml:math id="M90" display="inline"><mml:mrow><mml:msub><mml:mi>f</mml:mi><mml:mi mathvariant="normal">c</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> is the Coriolis parameter, <inline-formula><mml:math id="M91" display="inline"><mml:mi mathvariant="bold">J</mml:mi></mml:math></inline-formula> is the two-dimensional rotation dyadic, and <inline-formula><mml:math id="M92" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">ν</mml:mi><mml:mi>t</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> are height-averaged eddy viscosities. Furthermore, <inline-formula><mml:math id="M93" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">T</mml:mi><mml:mrow><mml:mi mathvariant="normal">h</mml:mi><mml:mn mathvariant="normal">3</mml:mn><mml:mo>,</mml:mo><mml:mn mathvariant="normal">0</mml:mn></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M94" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">T</mml:mi><mml:mrow><mml:mi mathvariant="normal">h</mml:mi><mml:mn mathvariant="normal">3</mml:mn><mml:mo>,</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula> are the unperturbed turbulent shear stresses at the ground and the layer interface, respectively, and <inline-formula><mml:math id="M95" display="inline"><mml:mrow><mml:msup><mml:mi mathvariant="bold">C</mml:mi><mml:mo>′</mml:mo></mml:msup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M96" display="inline"><mml:mrow><mml:msup><mml:mi mathvariant="bold">D</mml:mi><mml:mo>′</mml:mo></mml:msup></mml:mrow></mml:math></inline-formula> are the Jacobians of these stresses to the layer velocities. Finally, <inline-formula><mml:math id="M97" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mi mathvariant="normal">wf</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M98" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi mathvariant="bold-italic">τ</mml:mi><mml:mi mathvariant="normal">wf</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, and <inline-formula><mml:math id="M99" display="inline"><mml:mrow><mml:mo>〈</mml:mo><mml:msub><mml:mi mathvariant="bold-italic">τ</mml:mi><mml:mrow><mml:mi mathvariant="normal">d</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">hh</mml:mi></mml:mrow></mml:msub><mml:msub><mml:mo>〉</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> are the perturbing effects of the wind farm, with <inline-formula><mml:math id="M100" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mi mathvariant="normal">wf</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> denoting the turbine forces (scaled with air density), <inline-formula><mml:math id="M101" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi mathvariant="bold-italic">τ</mml:mi><mml:mi mathvariant="normal">wf</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> representing the shear stresses between the layers due to the added turbulence intensity caused by the farm, and <inline-formula><mml:math id="M102" display="inline"><mml:mrow><mml:mo>〈</mml:mo><mml:msub><mml:mi mathvariant="bold-italic">τ</mml:mi><mml:mrow><mml:mi mathvariant="normal">d</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">hh</mml:mi></mml:mrow></mml:msub><mml:msub><mml:mo>〉</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> denoting dispersive stresses due to turbine-scale flow heterogeneity. For the sake of simplicity, the latter two are not included in this study.</p>
      <p id="d2e2671">The pressure perturbation in the ABL <inline-formula><mml:math id="M103" display="inline"><mml:mrow><mml:msup><mml:mi>p</mml:mi><mml:mo>′</mml:mo></mml:msup></mml:mrow></mml:math></inline-formula> is caused by the gravity waves on the capping inversion and in the free atmosphere. These waves are themselves triggered by the displacement of the capping inversion <inline-formula><mml:math id="M104" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mi>t</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>. Their pressure feedback can be related to this displacement in Fourier space, so that <xref ref-type="bibr" rid="bib1.bibx12 bib1.bibx29" id="paren.39"/>

          <disp-formula id="Ch1.E11" content-type="numbered"><label>11</label><mml:math id="M105" display="block"><mml:mrow><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msup><mml:mi>p</mml:mi><mml:mo>′</mml:mo></mml:msup></mml:mrow><mml:mi mathvariant="italic">ρ</mml:mi></mml:mfrac></mml:mstyle><mml:mo>=</mml:mo><mml:msup><mml:mi>g</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mi>t</mml:mi></mml:msub><mml:mo>+</mml:mo><mml:msup><mml:mi mathvariant="script">F</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup><mml:mfenced close=")" open="("><mml:mi mathvariant="normal">Φ</mml:mi></mml:mfenced><mml:mo>*</mml:mo><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mi>t</mml:mi></mml:msub><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

        where <inline-formula><mml:math id="M106" display="inline"><mml:mrow><mml:msup><mml:mi>g</mml:mi><mml:mo>′</mml:mo></mml:msup></mml:mrow></mml:math></inline-formula> is the reduced gravity in the capping inversion, <inline-formula><mml:math id="M107" display="inline"><mml:mrow><mml:msup><mml:mi mathvariant="script">F</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> denotes an inverse Fourier transform, and <inline-formula><mml:math id="M108" display="inline"><mml:mo>*</mml:mo></mml:math></inline-formula> denotes a convolution. Continuing, <inline-formula><mml:math id="M109" display="inline"><mml:mi mathvariant="normal">Φ</mml:mi></mml:math></inline-formula> are the stratification coefficients, which determine the pressure feedback of the internal gravity waves in the free atmosphere. For uniform barotropic free atmospheres, these coefficients are calculated in Fourier space as <xref ref-type="bibr" rid="bib1.bibx3" id="paren.40"/>

          <disp-formula id="Ch1.E12" content-type="numbered"><label>12</label><mml:math id="M110" display="block"><mml:mrow><mml:mi mathvariant="normal">Φ</mml:mi><mml:mo>=</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:mi>i</mml:mi><mml:mfenced open="(" close=")"><mml:mrow><mml:msubsup><mml:mi>N</mml:mi><mml:mi mathvariant="normal">g</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup><mml:mo>-</mml:mo><mml:msup><mml:mi mathvariant="normal">Ω</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msup></mml:mrow></mml:mfenced></mml:mrow><mml:mi>m</mml:mi></mml:mfrac></mml:mstyle><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

        where <inline-formula><mml:math id="M111" display="inline"><mml:mrow><mml:msub><mml:mi>N</mml:mi><mml:mi mathvariant="normal">g</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:msqrt><mml:mrow><mml:mstyle displaystyle="false"><mml:mfrac style="text"><mml:mi>g</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:mfrac></mml:mstyle><mml:mstyle displaystyle="false"><mml:mfrac style="text"><mml:mrow><mml:mi mathvariant="normal">d</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:mrow><mml:mrow><mml:mi mathvariant="normal">d</mml:mi><mml:mi>z</mml:mi></mml:mrow></mml:mfrac></mml:mstyle></mml:mrow></mml:msqrt></mml:mrow></mml:math></inline-formula> is the Brunt–Väisälä frequency in the free atmosphere and <inline-formula><mml:math id="M112" display="inline"><mml:mi mathvariant="normal">Ω</mml:mi></mml:math></inline-formula> is the intrinsic wave frequency. The vertical wavenumber <inline-formula><mml:math id="M113" display="inline"><mml:mi>m</mml:mi></mml:math></inline-formula> is given by the dispersion relation <xref ref-type="bibr" rid="bib1.bibx28" id="paren.41"/>:

          <disp-formula id="Ch1.E13" content-type="numbered"><label>13</label><mml:math id="M114" display="block"><mml:mrow><mml:msup><mml:mi>m</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msup><mml:mo>=</mml:mo><mml:mfenced open="(" close=")"><mml:mrow><mml:msup><mml:mi>k</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msup><mml:mo>+</mml:mo><mml:msup><mml:mi>l</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msup></mml:mrow></mml:mfenced><mml:mfenced open="(" close=")"><mml:mrow><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msubsup><mml:mi>N</mml:mi><mml:mi mathvariant="normal">g</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup></mml:mrow><mml:mrow><mml:msup><mml:mi mathvariant="normal">Ω</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msup></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:mfenced><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

        where the sign of <inline-formula><mml:math id="M115" display="inline"><mml:mi>m</mml:mi></mml:math></inline-formula> has to be chosen to ensure that the perturbation dies out with height if <inline-formula><mml:math id="M116" display="inline"><mml:mrow><mml:msup><mml:mi>m</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msup><mml:mo>&lt;</mml:mo><mml:mn mathvariant="normal">0</mml:mn></mml:mrow></mml:math></inline-formula> or that energy radiates outwards if <inline-formula><mml:math id="M117" display="inline"><mml:mrow><mml:msup><mml:mi>m</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msup><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">0</mml:mn></mml:mrow></mml:math></inline-formula>.</p>
      <p id="d2e2981">The grid spacing of the APM is significantly larger than the individual turbine diameters. Therefore, the turbine forces are considered to be point forces before they are filtered onto the numerical grid using the same filter as in Eq. (<xref ref-type="disp-formula" rid="Ch1.E4"/>), so that the total force is simply a sum of Gaussians <xref ref-type="bibr" rid="bib1.bibx3" id="paren.42"/>:

          <disp-formula id="Ch1.E14" content-type="numbered"><label>14</label><mml:math id="M118" display="block"><mml:mrow><mml:msub><mml:mover accent="true"><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mo mathvariant="normal" stretchy="true">̃</mml:mo></mml:mover><mml:mi mathvariant="normal">wf</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:munderover><mml:mo movablelimits="false">∑</mml:mo><mml:mi>k</mml:mi><mml:mrow><mml:msub><mml:mi>N</mml:mi><mml:mi mathvariant="normal">t</mml:mi></mml:msub></mml:mrow></mml:munderover><mml:msub><mml:mi>G</mml:mi><mml:mrow><mml:mi mathvariant="normal">ℓ</mml:mi><mml:mo>,</mml:mo><mml:mi>k</mml:mi></mml:mrow></mml:msub><mml:msub><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mi>k</mml:mi></mml:msub><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

        where each <inline-formula><mml:math id="M119" display="inline"><mml:mrow><mml:msub><mml:mi>G</mml:mi><mml:mrow><mml:mi mathvariant="normal">ℓ</mml:mi><mml:mo>,</mml:mo><mml:mi>k</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula> is a Gaussian kernel centered around turbine <inline-formula><mml:math id="M120" display="inline"><mml:mi>k</mml:mi></mml:math></inline-formula>, and the turbine thrusts <inline-formula><mml:math id="M121" display="inline"><mml:mrow><mml:msub><mml:mi>f</mml:mi><mml:mi>k</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> are <xref ref-type="bibr" rid="bib1.bibx3" id="paren.43"/>

          <disp-formula id="Ch1.E15" content-type="numbered"><label>15</label><mml:math id="M122" display="block"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mi>k</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mn mathvariant="normal">2</mml:mn></mml:mfrac></mml:mstyle><mml:msubsup><mml:mi>C</mml:mi><mml:mrow><mml:mi mathvariant="normal">T</mml:mi><mml:mo>,</mml:mo><mml:mi>k</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:msub><mml:mi>A</mml:mi><mml:mi>k</mml:mi></mml:msub><mml:msubsup><mml:mi>S</mml:mi><mml:mi>k</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup><mml:msub><mml:mi mathvariant="bold-italic">e</mml:mi><mml:mi>k</mml:mi></mml:msub><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

        where <inline-formula><mml:math id="M123" display="inline"><mml:mrow><mml:msub><mml:mi>S</mml:mi><mml:mi>k</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M124" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">e</mml:mi><mml:mi>k</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> are the turbine disk velocities and orientations, respectively, and <inline-formula><mml:math id="M125" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mrow><mml:mi mathvariant="normal">T</mml:mi><mml:mo>,</mml:mo><mml:mi>k</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> can be evaluated using <inline-formula><mml:math id="M126" display="inline"><mml:mrow><mml:msub><mml:mi>S</mml:mi><mml:mi>k</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> from the turbine thrust coefficient curve. Note that this paper uses the disk-averaged velocities and thrust coefficients, as this simplifies the comparison against LES. In practical application, their free-stream counterparts can also be used, as is commonly done by wake models. Because the APM's resolution is too low to resolve the turbine-scale flow, these quantities are calculated using a wake model. To capture the dependence of the turbine forces on the local mesoscale flow, past studies developed ways to calculate the background velocity used by the wake model based on the APM state <xref ref-type="bibr" rid="bib1.bibx3 bib1.bibx31 bib1.bibx8 bib1.bibx9" id="paren.44"/>.</p>
      <p id="d2e3176">Since the pressure feedback of the internal gravity waves is mostly easily expressed in Fourier space, the system of Eqs. (<xref ref-type="disp-formula" rid="Ch1.E7"/>)–(<xref ref-type="disp-formula" rid="Ch1.E11"/>) is solved with a Fourier–Galerkin method. All terms in the equations except the turbine forces <inline-formula><mml:math id="M127" display="inline"><mml:mrow><mml:msub><mml:mover accent="true"><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mo mathvariant="normal" stretchy="true">̃</mml:mo></mml:mover><mml:mi mathvariant="normal">wf</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> then decouple per wavenumber, so that for a given <inline-formula><mml:math id="M128" display="inline"><mml:mrow><mml:msub><mml:mover accent="true"><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mo stretchy="true" mathvariant="normal">̃</mml:mo></mml:mover><mml:mi mathvariant="normal">wf</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> the system can be easily solved through matrix inversion. The turbine forces are included through a fixed-point iteration, where at each step the linear APM equations are solved for a given forcing. The resulting APM state is used to infer a background velocity for the wake model, which is then used to re-calculate the turbine forces <inline-formula><mml:math id="M129" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mi>k</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>. This process is repeated until convergence. For details, see <xref ref-type="bibr" rid="bib1.bibx8" id="text.45"/>.</p>
      <p id="d2e3227">The main numerical costs of the classical APM lie in solving Eqs. (<xref ref-type="disp-formula" rid="Ch1.E7"/>)–(<xref ref-type="disp-formula" rid="Ch1.E11"/>) and re-calculating the wake model background velocity at each step of the fixed point iteration <xref ref-type="bibr" rid="bib1.bibx9" id="paren.46"/>. These repeated calculations were needed to resolve the local coupling between the mesoscale APM state <inline-formula><mml:math id="M130" display="inline"><mml:mrow><mml:mo>(</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup><mml:mo>,</mml:mo><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>,</mml:mo><mml:msup><mml:mi>p</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mo>)</mml:mo></mml:mrow></mml:math></inline-formula> and the turbine forces. However, 2SM theory allows us to neglect this local coupling, and consider a farm-averaged approach instead. This is the focus of the following sections.</p>
</sec>
<sec id="Ch1.S4">
  <label>4</label><title>Combining 2SM and APM</title>
      <p id="d2e3276">The 2SM approach simplifies the turbine-scale problem by relating the turbine forces and power outputs to only the average velocity in the farm through <inline-formula><mml:math id="M131" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">P</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M132" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>. The most straightforward way of combining this with the APM is then to use the APM as the external model to find <inline-formula><mml:math id="M133" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, while using <inline-formula><mml:math id="M134" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M135" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> to represent the turbine forces in the APM. This section focuses on developing such coupling and on leveraging the simplifications of 2SM theory to speed up the APM.</p>
      <p id="d2e3342">The goal of this section is use the APM to solve the external problem as cheaply as possible. Therefore, we treat the farm coefficients <inline-formula><mml:math id="M136" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">P</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M137" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> as known inputs for now. However, their computation still represents an additional cost, associated with calculating the farm-averaged velocity in a prior (neutral) wake model simulation. This will be addressed in Sect. <xref ref-type="sec" rid="Ch1.S5.SS2"/>, using insights from the methods developed in this section.</p>
      <p id="d2e3373">We will build up the combination of the 2SM and APM approaches step by step. First, Sect. <xref ref-type="sec" rid="Ch1.S4.SS1"/> discusses the differences between the mathematical averaging operators used by 2SM theory and in the derivation of the APM and bridges the gap between them. Then, Sect. <xref ref-type="sec" rid="Ch1.S4.SS2"/> sets up a simplified version of the APM focused on finding <inline-formula><mml:math id="M138" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, thereby leveraging the insights of 2SM theory to speed up the model. Sect. <xref ref-type="sec" rid="Ch1.S4.SS3"/> validates the developed 2SM–APM with the same LES dataset from <xref ref-type="bibr" rid="bib1.bibx19" id="text.47"/> that was used in Sect. <xref ref-type="sec" rid="Ch1.S2"/>. Finally, Sect. <xref ref-type="sec" rid="Ch1.S4.SS4"/> investigates the use of wake models to calculate <inline-formula><mml:math id="M139" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">P</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M140" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>.</p>
<sec id="Ch1.S4.SS1">
  <label>4.1</label><title>Relating the average APM and 2SM velocities</title>
      <p id="d2e3435">Combining the 2SM and APM approaches requires us to relate the average farm velocity used in the 2SM equations to the average lower-layer APM velocity within the farm. We defined the former as a straightforward average over the wind farm, so that

            <disp-formula id="Ch1.E16" content-type="numbered"><label>3</label><mml:math id="M141" display="block"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:munderover><mml:mo movablelimits="false">∫</mml:mo><mml:mn mathvariant="normal">0</mml:mn><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:munderover><mml:munder><mml:mo movablelimits="false">∬</mml:mo><mml:mrow><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:munder><mml:mi>u</mml:mi><mml:mspace linebreak="nobreak" width="0.33em"/><mml:mtext>d</mml:mtext><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mtext>d</mml:mtext><mml:mi>z</mml:mi><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

          where <inline-formula><mml:math id="M142" display="inline"><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> is the height of the 2SM averaging volume and <inline-formula><mml:math id="M143" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> the planform area of the wind farm. As noted in Sect. 2, modeling wind farms with irregular turbine layouts may require more complex horizontal averaging operators.</p>
      <p id="d2e3525">In contrast, the APM predicts a horizontally filtered, height-averaged velocity field:

            <disp-formula id="Ch1.E17" content-type="numbered"><label>4</label><mml:math id="M144" display="block"><mml:mtable class="split" rowspacing="0.2ex" displaystyle="true" columnalign="right left"><mml:mtr><mml:mtd><mml:mrow><mml:mo>〈</mml:mo><mml:mover accent="true"><mml:mi>u</mml:mi><mml:mo stretchy="true" mathvariant="normal">̃</mml:mo></mml:mover><mml:msub><mml:mo>〉</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>=</mml:mo></mml:mrow></mml:mtd><mml:mtd><mml:mrow><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:munderover><mml:mo movablelimits="false">∫</mml:mo><mml:mn mathvariant="normal">0</mml:mn><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:munderover><mml:munder><mml:mo movablelimits="false">∬</mml:mo><mml:mrow><mml:msub><mml:mi>L</mml:mi><mml:mi>x</mml:mi></mml:msub><mml:mo>×</mml:mo><mml:msub><mml:mi>L</mml:mi><mml:mi>y</mml:mi></mml:msub></mml:mrow></mml:munder><mml:msub><mml:mi>G</mml:mi><mml:mi mathvariant="normal">ℓ</mml:mi></mml:msub><mml:mo>(</mml:mo><mml:mi>x</mml:mi><mml:mo>-</mml:mo><mml:msup><mml:mi>x</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mo>,</mml:mo><mml:mi>y</mml:mi><mml:mo>-</mml:mo><mml:msup><mml:mi>y</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mo>)</mml:mo></mml:mrow></mml:mtd></mml:mtr><mml:mtr><mml:mtd/><mml:mtd><mml:mrow><mml:mi>u</mml:mi><mml:mo>(</mml:mo><mml:msup><mml:mi>x</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mo>,</mml:mo><mml:msup><mml:mi>y</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mo>,</mml:mo><mml:mi>z</mml:mi><mml:mo>)</mml:mo><mml:mi mathvariant="normal">d</mml:mi><mml:msup><mml:mi>x</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mi mathvariant="normal">d</mml:mi><mml:msup><mml:mi>y</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mi mathvariant="normal">d</mml:mi><mml:mi>z</mml:mi><mml:mo>.</mml:mo></mml:mrow></mml:mtd></mml:mtr></mml:mtable></mml:math></disp-formula>

          Averaging the APM variables over the farm can therefore only be done horizontally, so that

            <disp-formula id="Ch1.E18" content-type="numbered"><label>16</label><mml:math id="M145" display="block"><mml:mrow><mml:msub><mml:mover accent="true"><mml:mrow><mml:mo>〈</mml:mo><mml:mover accent="true"><mml:mi>u</mml:mi><mml:mo stretchy="true" mathvariant="normal">̃</mml:mo></mml:mover><mml:mo>〉</mml:mo></mml:mrow><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>=</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mrow><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:munder><mml:mo movablelimits="false">∬</mml:mo><mml:mrow><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:munder><mml:mo>〈</mml:mo><mml:mover accent="true"><mml:mi>u</mml:mi><mml:mo stretchy="true" mathvariant="normal">̃</mml:mo></mml:mover><mml:msub><mml:mo>〉</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mtext>d</mml:mtext><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mo>,</mml:mo><mml:mspace linebreak="nobreak" width="1em"/><mml:msubsup><mml:mover accent="true"><mml:mi>u</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mrow><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:munder><mml:mo movablelimits="false">∬</mml:mo><mml:mrow><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:munder><mml:msubsup><mml:mi>u</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup><mml:mtext>d</mml:mtext><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

          where the bars indicate horizontally averaging over the farm.</p>
      <p id="d2e3794">Clearly, the averages defined by Eqs. (<xref ref-type="disp-formula" rid="Ch1.E3"/>) and (<xref ref-type="disp-formula" rid="Ch1.E18"/>) are not equivalent. Since the internal 2SM problem has to remain constant, the volume over which the farm average is defined should be independent of the mesoscale perturbation. However, the layer interface in the APM follows the flow streamlines, and its vertical displacement <inline-formula><mml:math id="M146" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> is part of the model state. This results in two different control volumes, as sketched in Fig. <xref ref-type="fig" rid="F3"/>. Moreover, the APM velocity is filtered horizontally before being averaged, while the 2SM velocity is not.</p><fig id="F3"><label>Figure 3</label><caption><p id="d2e3817">Sketch of the farm control volume (black) and lower APM layer (red) for a rectangular farm. Note that in contrast to the 2SM farm control volume, the upper boundary of the APM layer is not rigid, as it follows the streamlines.</p></caption>
          <graphic xlink:href="https://wes.copernicus.org/articles/11/4349/2026/wes-11-4349-2026-f03.png"/>

        </fig>

      <p id="d2e3826">In this work, we neglect the filtering effects, thus presuming

            <disp-formula id="Ch1.E19" content-type="numbered"><label>17</label><mml:math id="M147" display="block"><mml:mrow><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mrow><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:munder><mml:mo movablelimits="false">∬</mml:mo><mml:mrow><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:munder><mml:mo>〈</mml:mo><mml:mover accent="true"><mml:mi>u</mml:mi><mml:mo stretchy="true" mathvariant="normal">̃</mml:mo></mml:mover><mml:msub><mml:mo>〉</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mtext>d</mml:mtext><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mo>≈</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mrow><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:munder><mml:mo movablelimits="false">∬</mml:mo><mml:mrow><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:munder><mml:mo>〈</mml:mo><mml:mi>u</mml:mi><mml:msub><mml:mo>〉</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mtext>d</mml:mtext><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mo>.</mml:mo></mml:mrow></mml:math></disp-formula>

          Relating the 2SM and APM averages then only requires accounting for the discrepancy in the control volumes. Choosing <inline-formula><mml:math id="M148" display="inline"><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> makes these volumes as similar as possible, so that the remaining difference is solely due to the streamline displacement <inline-formula><mml:math id="M149" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>. The velocity in this region depends on the relative strength of the turbulent entrainment and gravity-wave-induced pressure gradients. Here, we take the simplest possible approach by assuming that the average velocity perturbation in the part of the APM layer above the 2SM volume scales with the velocity deficit within it, so that

            <disp-formula id="Ch1.E20" content-type="numbered"><label>18</label><mml:math id="M150" display="block"><mml:mrow><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mrow><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">η</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:munderover><mml:mo movablelimits="false">∫</mml:mo><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">η</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:munderover><mml:mo>(</mml:mo><mml:mover accent="true"><mml:mi>u</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mo>-</mml:mo><mml:msub><mml:mi>U</mml:mi><mml:mn mathvariant="normal">0</mml:mn></mml:msub><mml:mo>)</mml:mo><mml:mtext>d</mml:mtext><mml:mi>z</mml:mi><mml:mo>≈</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mi mathvariant="italic">α</mml:mi><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:munderover><mml:mo movablelimits="false">∫</mml:mo><mml:mn mathvariant="normal">0</mml:mn><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:munderover><mml:mo>(</mml:mo><mml:mover accent="true"><mml:mi>u</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mo>-</mml:mo><mml:msub><mml:mi>U</mml:mi><mml:mn mathvariant="normal">0</mml:mn></mml:msub><mml:mo>)</mml:mo><mml:mtext>d</mml:mtext><mml:mi>z</mml:mi><mml:mo>=</mml:mo><mml:mi mathvariant="italic">α</mml:mi><mml:msubsup><mml:mi>u</mml:mi><mml:mi mathvariant="normal">f</mml:mi><mml:mo>′</mml:mo></mml:msubsup><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

          where <inline-formula><mml:math id="M151" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mn mathvariant="normal">0</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> is the unperturbed velocity profile, <inline-formula><mml:math id="M152" display="inline"><mml:mi mathvariant="italic">α</mml:mi></mml:math></inline-formula> is a tuning parameter between 0 (no velocity deficit in the region <inline-formula><mml:math id="M153" display="inline"><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>&lt;</mml:mo><mml:mi>z</mml:mi><mml:mo>&lt;</mml:mo><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">η</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>) and 1 (perfectly mixed APM layer), and <inline-formula><mml:math id="M154" display="inline"><mml:mrow><mml:msubsup><mml:mi>u</mml:mi><mml:mi mathvariant="normal">f</mml:mi><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> is the 2SM-farm-averaged velocity deficit:

            <disp-formula id="Ch1.E21" content-type="numbered"><label>19</label><mml:math id="M155" display="block"><mml:mrow><mml:msubsup><mml:mi>u</mml:mi><mml:mi mathvariant="normal">f</mml:mi><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:msub><mml:mi>U</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mo>-</mml:mo><mml:munderover><mml:mo movablelimits="false">∫</mml:mo><mml:mn mathvariant="normal">0</mml:mn><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:munderover><mml:msub><mml:mi>U</mml:mi><mml:mn mathvariant="normal">0</mml:mn></mml:msub><mml:mtext>d</mml:mtext><mml:mi>z</mml:mi><mml:mo>=</mml:mo><mml:msub><mml:mi>U</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mo>-</mml:mo><mml:msub><mml:mi>U</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>.</mml:mo></mml:mrow></mml:math></disp-formula>

          Taking the vertical integral up to <inline-formula><mml:math id="M156" display="inline"><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">η</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, we then find

            <disp-formula id="Ch1.E22" content-type="numbered"><label>20</label><mml:math id="M157" display="block"><mml:mrow><mml:mo>(</mml:mo><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">η</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>)</mml:mo><mml:msubsup><mml:mover accent="true"><mml:mi>u</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup><mml:mo>≈</mml:mo><mml:mo>(</mml:mo><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:mi mathvariant="italic">α</mml:mi><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">η</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>)</mml:mo><mml:msubsup><mml:mi>u</mml:mi><mml:mi mathvariant="normal">f</mml:mi><mml:mo>′</mml:mo></mml:msubsup><mml:mo>.</mml:mo></mml:mrow></mml:math></disp-formula>

          The 2SM-farm-averaged velocity can then be calculated from the APM state as

            <disp-formula id="Ch1.E23" content-type="numbered"><label>21</label><mml:math id="M158" display="block"><mml:mrow><mml:msubsup><mml:mi>u</mml:mi><mml:mi mathvariant="normal">f</mml:mi><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:msubsup><mml:mover accent="true"><mml:mi>u</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">η</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:mi mathvariant="italic">α</mml:mi><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">η</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>,</mml:mo><mml:mspace linebreak="nobreak" width="1em"/><mml:msub><mml:mi>U</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:msub><mml:mi>U</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msubsup><mml:mover accent="true"><mml:mi>u</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">η</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:mi mathvariant="italic">α</mml:mi><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">η</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>.</mml:mo></mml:mrow></mml:math></disp-formula>

          We tune <inline-formula><mml:math id="M159" display="inline"><mml:mi mathvariant="italic">α</mml:mi></mml:math></inline-formula> using the NBL case <inline-formula><mml:math id="M160" display="inline"><mml:mi>H</mml:mi></mml:math></inline-formula>500-<inline-formula><mml:math id="M161" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:mrow></mml:math></inline-formula>0-<inline-formula><mml:math id="M162" display="inline"><mml:mi mathvariant="normal">Γ</mml:mi></mml:math></inline-formula>0 from the dataset of <xref ref-type="bibr" rid="bib1.bibx19" id="text.48"/>. We calculate the APM state (<inline-formula><mml:math id="M163" display="inline"><mml:mrow><mml:msubsup><mml:mi>u</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M164" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>) directly from the LES data by horizontally filtering the time-averaged flow field, advecting a set of points to find the displacement of the APM layer <inline-formula><mml:math id="M165" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, and height-averaging the flow. A detailed description of this procedure can be found in <xref ref-type="bibr" rid="bib1.bibx8" id="text.49"/>. By relating the LES-based APM state set up this way to the <inline-formula><mml:math id="M166" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> calculated directly from the LES using Eq. (<xref ref-type="disp-formula" rid="Ch1.E3"/>), we find <inline-formula><mml:math id="M167" display="inline"><mml:mrow><mml:mi mathvariant="italic">α</mml:mi><mml:mo>=</mml:mo><mml:mn mathvariant="normal">0.16</mml:mn></mml:mrow></mml:math></inline-formula>. Note that this basic approach does not capture the significant differences in velocity profiles between NBL, CNBL, and realistic flows, or the effects of wind farm layout and thrust distribution. Instead, it is chosen for its simplicity.</p>
</sec>
<sec id="Ch1.S4.SS2">
  <label>4.2</label><title>Basic 2SM–APM coupling</title>
      <p id="d2e4512">We now focus on a fast 2SM–APM variant that will require solving the APM Eqs. (<xref ref-type="disp-formula" rid="Ch1.E7"/>)–(<xref ref-type="disp-formula" rid="Ch1.E11"/>) only once, rather than at every iteration, thereby significantly speeding up the model.</p>
      <p id="d2e4519">We begin by simplifying the notation. We can write the APM as a set of linear operators that transform the turbine forces to an APM state, so that

            <disp-formula id="Ch1.E24" content-type="numbered"><label>22</label><mml:math id="M168" display="block"><mml:mrow><mml:msubsup><mml:mi>u</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi>u</mml:mi></mml:msub><mml:mfenced close=")" open="("><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mi mathvariant="normal">wf</mml:mi></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle></mml:mfenced><mml:mo>,</mml:mo><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>=</mml:mo><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi mathvariant="italic">η</mml:mi></mml:msub><mml:mfenced close=")" open="("><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mi mathvariant="normal">wf</mml:mi></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle></mml:mfenced><mml:mo>,</mml:mo><mml:msup><mml:mi>p</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mo>=</mml:mo><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi>p</mml:mi></mml:msub><mml:mfenced open="(" close=")"><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mi mathvariant="normal">wf</mml:mi></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle></mml:mfenced><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

          where <inline-formula><mml:math id="M169" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi>u</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M170" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi mathvariant="italic">η</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, and <inline-formula><mml:math id="M171" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi>p</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> are linear operators that follow from the APM Eqs. (<xref ref-type="disp-formula" rid="Ch1.E7"/>)–(<xref ref-type="disp-formula" rid="Ch1.E11"/>) and <inline-formula><mml:math id="M172" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mi mathvariant="normal">wf</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> is the turbine force. Note that this assumes that the direct turbine forces are the only perturbing forcing in the ABL. A more general approach would have to include the dispersive stresses and the increase in momentum flux as well. A discussion on such an extension is given at the end of this section.</p>
      <p id="d2e4667">When using the APM with conventional wake model coupling methods, one of the largest computational costs is applying these operators <inline-formula><mml:math id="M173" display="inline"><mml:mi mathvariant="script">A</mml:mi></mml:math></inline-formula>, which involves a large linear system solve, at each step of the fixed-point iteration <xref ref-type="bibr" rid="bib1.bibx9" id="paren.50"/>. In the classical APM, this iteration is needed because at each point in space, the local height-averaged farm force depends on the local velocity and layer thickness. While these local values impact individual turbine performance, they are not relevant to the 2SM approach, as it only requires the average velocity over the farm to calculate the farm power output. Since the goal of the 2SM–APM is to compute this farm-averaged velocity as cheaply as possible, including local variations is an unnecessary cost. Therefore, we will assume that the distribution of the force in space, normalized by the total force <inline-formula><mml:math id="M174" display="inline"><mml:mrow><mml:mi>F</mml:mi><mml:mo>/</mml:mo><mml:msub><mml:mover accent="true"><mml:mi>h</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, is independent of the APM state <inline-formula><mml:math id="M175" display="inline"><mml:mrow><mml:mo>(</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup><mml:mo>,</mml:mo><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>,</mml:mo><mml:msup><mml:mi>p</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mo>)</mml:mo></mml:mrow></mml:math></inline-formula>, so that

            <disp-formula id="Ch1.E25" content-type="numbered"><label>23</label><mml:math id="M176" display="block"><mml:mrow><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mi mathvariant="normal">wf</mml:mi></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>≈</mml:mo><mml:msub><mml:mi mathvariant="bold">Π</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mi>F</mml:mi><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">η</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

          where <inline-formula><mml:math id="M177" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold">Π</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> is the geometric footprint of the height-averaged force, and <inline-formula><mml:math id="M178" display="inline"><mml:mi>F</mml:mi></mml:math></inline-formula> is the total farm thrust (see Eq. <xref ref-type="disp-formula" rid="Ch1.E2"/>). If this footprint is indeed independent of the APM state, the model is reduced to a set of scalar equations:

            <disp-formula id="Ch1.E26" content-type="numbered"><label>24</label><mml:math id="M179" display="block"><mml:mrow><mml:msubsup><mml:mover accent="true"><mml:mi>u</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:msub><mml:mi>a</mml:mi><mml:mi>u</mml:mi></mml:msub><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mi>F</mml:mi><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">η</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>,</mml:mo><mml:mspace linebreak="nobreak" width="1em"/><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">η</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>=</mml:mo><mml:msub><mml:mi>a</mml:mi><mml:mi mathvariant="italic">η</mml:mi></mml:msub><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mi>F</mml:mi><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">η</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

          where the coefficients <inline-formula><mml:math id="M180" display="inline"><mml:mrow><mml:msub><mml:mi>a</mml:mi><mml:mi>u</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M181" display="inline"><mml:mrow><mml:msub><mml:mi>a</mml:mi><mml:mi mathvariant="italic">η</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> can be pre-computed:

            <disp-formula id="Ch1.E27" content-type="numbered"><label>25</label><mml:math id="M182" display="block"><mml:mtable rowspacing="0.2ex" class="split" displaystyle="true" columnalign="right left"><mml:mtr><mml:mtd/><mml:mtd><mml:mrow><mml:msub><mml:mi>a</mml:mi><mml:mi>u</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mrow><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:munder><mml:mo movablelimits="false">∬</mml:mo><mml:mrow><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:munder><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi>u</mml:mi></mml:msub><mml:mfenced close=")" open="("><mml:mrow><mml:msub><mml:mi mathvariant="bold">Π</mml:mi><mml:mi>f</mml:mi></mml:msub></mml:mrow></mml:mfenced><mml:mtext>d</mml:mtext><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mo>,</mml:mo></mml:mrow></mml:mtd></mml:mtr><mml:mtr><mml:mtd/><mml:mtd><mml:mrow><mml:msub><mml:mi>a</mml:mi><mml:mi mathvariant="italic">η</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mrow><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:munder><mml:mo movablelimits="false">∬</mml:mo><mml:mrow><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:munder><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi mathvariant="italic">η</mml:mi></mml:msub><mml:mfenced open="(" close=")"><mml:mrow><mml:msub><mml:mi mathvariant="bold">Π</mml:mi><mml:mi>f</mml:mi></mml:msub></mml:mrow></mml:mfenced><mml:mtext>d</mml:mtext><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mo>.</mml:mo></mml:mrow></mml:mtd></mml:mtr></mml:mtable></mml:math></disp-formula>

          We then arrive at a closed system of equations by computing the farm force <inline-formula><mml:math id="M183" display="inline"><mml:mi>F</mml:mi></mml:math></inline-formula> with Eq. (<xref ref-type="disp-formula" rid="Ch1.E2"/>), while relating <inline-formula><mml:math id="M184" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> to the average APM perturbations using Eq. (<xref ref-type="disp-formula" rid="Ch1.E23"/>):

                <disp-formula specific-use="gather" content-type="numbered"><mml:math id="M185" display="block"><mml:mtable displaystyle="true"><mml:mlabeledtr id="Ch1.Ex1"><mml:mtd><mml:mtext>2</mml:mtext></mml:mtd><mml:mtd><mml:mrow><mml:mstyle displaystyle="true" class="stylechange"/><mml:mi>F</mml:mi><mml:mo>=</mml:mo><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msubsup><mml:mi>U</mml:mi><mml:mi mathvariant="normal">f</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup></mml:mrow><mml:mn mathvariant="normal">2</mml:mn></mml:mfrac></mml:mstyle><mml:msub><mml:mi>A</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mo>,</mml:mo></mml:mrow></mml:mtd></mml:mlabeledtr><mml:mlabeledtr id="Ch1.Ex2"><mml:mtd><mml:mtext>21</mml:mtext></mml:mtd><mml:mtd><mml:mrow><mml:mstyle displaystyle="true" class="stylechange"/><mml:msub><mml:mi>U</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:msub><mml:mi>U</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msubsup><mml:mover accent="true"><mml:mi>u</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">η</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:mi mathvariant="italic">α</mml:mi><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">η</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>.</mml:mo></mml:mrow></mml:mtd></mml:mlabeledtr></mml:mtable></mml:math></disp-formula>

          The linear system of Eqs. (<xref ref-type="disp-formula" rid="Ch1.E7"/>)–(<xref ref-type="disp-formula" rid="Ch1.E11"/>) now has to be inverted only once when calculating <inline-formula><mml:math id="M186" display="inline"><mml:mrow><mml:msub><mml:mi>a</mml:mi><mml:mi>u</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M187" display="inline"><mml:mrow><mml:msub><mml:mi>a</mml:mi><mml:mi mathvariant="italic">η</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>. This effectively reduces the cost of the model to a single step of the fixed-point iteration used by the classical APM while still incorporating the overall feedback between the total farm force and the blockage effects.</p>
      <p id="d2e5174">The closer the footprint <inline-formula><mml:math id="M188" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold">Π</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> is to the actual normalized force distribution, the more accurate we expect the coefficients <inline-formula><mml:math id="M189" display="inline"><mml:mrow><mml:msub><mml:mi>a</mml:mi><mml:mi>u</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M190" display="inline"><mml:mrow><mml:msub><mml:mi>a</mml:mi><mml:mi mathvariant="italic">η</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> to be. However, this actual footprint is not known a priori. Instead, we use a standalone wake model to estimate the footprint cheaply. The wake model can predict the turbine disk velocities <inline-formula><mml:math id="M191" display="inline"><mml:mrow><mml:msub><mml:mi>S</mml:mi><mml:mi>k</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, which can be combined with Eqs. (<xref ref-type="disp-formula" rid="Ch1.E14"/>)–(<xref ref-type="disp-formula" rid="Ch1.E15"/>) to find the farm forcing, so that

            <disp-formula id="Ch1.E28" content-type="numbered"><label>26</label><mml:math id="M192" display="block"><mml:mrow><mml:msub><mml:mover accent="true"><mml:mi mathvariant="bold">Π</mml:mi><mml:mo mathvariant="normal" stretchy="true">̃</mml:mo></mml:mover><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msubsup><mml:mo>∑</mml:mo><mml:mi>k</mml:mi><mml:mrow><mml:msub><mml:mi>N</mml:mi><mml:mi mathvariant="normal">t</mml:mi></mml:msub></mml:mrow></mml:msubsup><mml:mfenced close=")" open="("><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mrow><mml:mi mathvariant="normal">T</mml:mi><mml:mo>,</mml:mo><mml:mi>k</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">WM</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:msub><mml:mi>A</mml:mi><mml:mi>k</mml:mi></mml:msub><mml:msubsup><mml:mi>S</mml:mi><mml:mrow><mml:mi>k</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">WM</mml:mi></mml:mrow><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup><mml:msub><mml:mi>G</mml:mi><mml:mrow><mml:mi mathvariant="normal">ℓ</mml:mi><mml:mo>,</mml:mo><mml:mi>k</mml:mi></mml:mrow></mml:msub><mml:msub><mml:mi mathvariant="bold-italic">e</mml:mi><mml:mi>x</mml:mi></mml:msub></mml:mrow></mml:mfenced></mml:mrow><mml:mrow><mml:msubsup><mml:mo>∑</mml:mo><mml:mi>k</mml:mi><mml:mrow><mml:msub><mml:mi>N</mml:mi><mml:mi mathvariant="normal">t</mml:mi></mml:msub></mml:mrow></mml:msubsup><mml:mfenced close=")" open="("><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mrow><mml:mi mathvariant="normal">T</mml:mi><mml:mo>,</mml:mo><mml:mi>k</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">WM</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:msub><mml:mi>A</mml:mi><mml:mi>k</mml:mi></mml:msub><mml:msubsup><mml:mi>S</mml:mi><mml:mrow><mml:mi>k</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">WM</mml:mi></mml:mrow><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup></mml:mrow></mml:mfenced></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

          where the subscript <inline-formula><mml:math id="M193" display="inline"><mml:mi mathvariant="normal">WM</mml:mi></mml:math></inline-formula> indicates the variables are calculated using a standalone wake model. This approximates the footprint of a farm in neutral conditions at a low computational cost.</p>
      <p id="d2e5363">We briefly discuss how this 2SM–APM coupling relates to the coupling methods between the APM and underlying wake models developed in previous work. The most similar conventional method is the velocity matching (VM) approach developed by <xref ref-type="bibr" rid="bib1.bibx8" id="text.51"/>, which calculates a background velocity that ensures that the wake model and the APM predict the same mesoscale velocity field at each point in the farm. This 2SM coupling imposes the same condition but only on a farm average, so that the average farm velocity is the same in the internal and external problems <xref ref-type="bibr" rid="bib1.bibx24" id="paren.52"/>. Therefore, the 2SM–APM coupling can be seen as the VM method with just one degree of freedom for the background velocity.</p>
      <p id="d2e5372">Including additional wind-farm perturbing forces, such as the induced dispersive and turbulent stresses, can be done by following the same steps as above. Since the APM operators are linear, the responses to different inputs can simply be summed up. If the perturbing forces scale with <inline-formula><mml:math id="M194" display="inline"><mml:mrow><mml:mi>F</mml:mi><mml:mo>/</mml:mo><mml:msub><mml:mover accent="true"><mml:mi>h</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, as can be assumed for the dispersive stresses, this results in a modification of <inline-formula><mml:math id="M195" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold">Π</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> when calculating <inline-formula><mml:math id="M196" display="inline"><mml:mrow><mml:msub><mml:mi>a</mml:mi><mml:mi>u</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M197" display="inline"><mml:mrow><mml:msub><mml:mi>a</mml:mi><mml:mi mathvariant="italic">η</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>. However, including perturbing forces that scale differently, like the turbulent momentum flux in the upper layer, requires calculating additional coefficients for the scalar system of equations. We leave these complications to future work.</p>
</sec>
<sec id="Ch1.S4.SS3">
  <label>4.3</label><title>Comparison to LES</title>
      <p id="d2e5434">To test the 2SM–APM approach, we validate the model with the CNBL LES results of <xref ref-type="bibr" rid="bib1.bibx19" id="text.53"/> that were used in Sect. <xref ref-type="sec" rid="Ch1.S2"/>. Like before, the farm coefficients are calculated using the neutral <inline-formula><mml:math id="M198" display="inline"><mml:mi>H</mml:mi></mml:math></inline-formula>500-<inline-formula><mml:math id="M199" display="inline"><mml:mi>C</mml:mi></mml:math></inline-formula>0-<inline-formula><mml:math id="M200" display="inline"><mml:mi>G</mml:mi></mml:math></inline-formula>0 case, which results in <inline-formula><mml:math id="M201" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:mn mathvariant="normal">1.15</mml:mn></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M202" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:mn mathvariant="normal">0.92</mml:mn></mml:mrow></mml:math></inline-formula>. This ensures that any errors introduced will come from the 2SM–APM coupling and not from the estimation of the farm coefficients. The unperturbed background APM state is based on the precursor simulation driving each of the wind-farm simulations. We leave out the <inline-formula><mml:math id="M203" display="inline"><mml:mi>H</mml:mi></mml:math></inline-formula>150 cases, since the three-layer APM can not handle very shallow upper-layer heights <inline-formula><mml:math id="M204" display="inline"><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> for the time being <xref ref-type="bibr" rid="bib1.bibx8" id="paren.54"/>.</p>
      <p id="d2e5519">The numerical setup is the same as described in <xref ref-type="bibr" rid="bib1.bibx8" id="text.55"/>. As discussed in the previous section, the APM operators <inline-formula><mml:math id="M205" display="inline"><mml:mi mathvariant="script">A</mml:mi></mml:math></inline-formula> are solved using a Fourier–Galerkin method. To prevent the perturbations recycling through the periodic boundaries imposed by this method, we use the very large domain length of <inline-formula><mml:math id="M206" display="inline"><mml:mrow><mml:msub><mml:mi>L</mml:mi><mml:mi>x</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:mn mathvariant="normal">10</mml:mn></mml:mrow></mml:math></inline-formula> 000 km in the streamwise direction. In the spanwise direction, we take <inline-formula><mml:math id="M207" display="inline"><mml:mrow><mml:msub><mml:mi>L</mml:mi><mml:mi>y</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:mn mathvariant="normal">30</mml:mn></mml:mrow></mml:math></inline-formula> km in order to follow the LES results as closely as possible. The grid resolution is half of the horizontal filter length, so that <inline-formula><mml:math id="M208" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi>x</mml:mi><mml:mo>=</mml:mo><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi>y</mml:mi><mml:mo>=</mml:mo><mml:mn mathvariant="normal">500</mml:mn></mml:mrow></mml:math></inline-formula> m.</p>
      <p id="d2e5582">The resulting predictions for the farm power <inline-formula><mml:math id="M209" display="inline"><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> are shown in Figs. <xref ref-type="fig" rid="F4"/> and <xref ref-type="fig" rid="F5"/>, and the corresponding error statistics are listed in Table <xref ref-type="table" rid="T2"/>. The fast 2SM–APM seems to perform well overall, with the general trend being clearly visible. The root-mean square error (RMSE) across all analyzed cases is <inline-formula><mml:math id="M210" display="inline"><mml:mrow><mml:mn mathvariant="normal">4.8</mml:mn><mml:msub><mml:mi>P</mml:mi><mml:mn mathvariant="normal">0</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> (where <inline-formula><mml:math id="M211" display="inline"><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mn mathvariant="normal">0</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> is the power output of an isolated turbine), or 3 % of the farm's installed capacity. As shown in Table <xref ref-type="table" rid="T2"/>, this RMSE is mostly due to the variance in the model errors, as the mean bias is comparatively small. Looking at Fig. <xref ref-type="fig" rid="F5"/>, there seems to be no consistent relation between the model error and the varying atmospheric parameters <inline-formula><mml:math id="M212" display="inline"><mml:mi>H</mml:mi></mml:math></inline-formula>, <inline-formula><mml:math id="M213" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi mathvariant="normal">Θ</mml:mi></mml:mrow></mml:math></inline-formula>, or <inline-formula><mml:math id="M214" display="inline"><mml:mi mathvariant="normal">Γ</mml:mi></mml:math></inline-formula>. Instead, Fig. <xref ref-type="fig" rid="F4"/> shows that there is a small but consistent positive bias and that the scatter on the model's predictions is larger at lower farm power outputs, i.e., for cases with stronger blockage effects.</p>

      <fig id="F4"><label>Figure 4</label><caption><p id="d2e5661">A comparison of farm power <inline-formula><mml:math id="M215" display="inline"><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> as calculated by LES and the fast 2SM–APM model for 27 different CNBL cases from <xref ref-type="bibr" rid="bib1.bibx19" id="text.56"/>. The footprint of the height-averaged turbine forces is assumed to be constant, so that the linear system of APM equations only has to be solved once. The axes are normalized with the power output of a standalone turbine <inline-formula><mml:math id="M216" display="inline"><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mn mathvariant="normal">0</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>.</p></caption>
          <graphic xlink:href="https://wes.copernicus.org/articles/11/4349/2026/wes-11-4349-2026-f04.png"/>

        </fig>

      <fig id="F5" specific-use="star"><label>Figure 5</label><caption><p id="d2e5697">The farm power <inline-formula><mml:math id="M217" display="inline"><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> as calculated by LES (black, <inline-formula><mml:math id="M218" display="inline"><mml:mo>□</mml:mo></mml:math></inline-formula>) and the fast 2SM–APM model (blue, <inline-formula><mml:math id="M219" display="inline"><mml:mo>∘</mml:mo></mml:math></inline-formula>) for the different cases from the dataset by <xref ref-type="bibr" rid="bib1.bibx19" id="text.57"/>. All values are normalized by the power output of a standalone turbine <inline-formula><mml:math id="M220" display="inline"><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mn mathvariant="normal">0</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>. The dashed lines separate the cases with different ABL heights.</p></caption>
          <graphic xlink:href="https://wes.copernicus.org/articles/11/4349/2026/wes-11-4349-2026-f05.png"/>

        </fig>

<table-wrap id="T2" specific-use="star"><label>Table 2</label><caption><p id="d2e5748">The root-mean square (RMSE), mean, and standard deviation (STD) of the error of farm power <inline-formula><mml:math id="M221" display="inline"><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> of all the models analyzed in the current paper for the LES dataset from <xref ref-type="bibr" rid="bib1.bibx19" id="text.58"/>. The “Section” column lists which sections of the paper describe the various models. All values are scaled with the power output of a standalone turbine. Note that the three error metrics are related, as <inline-formula><mml:math id="M222" display="inline"><mml:mrow><mml:msup><mml:mi mathvariant="normal">RMSE</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msup><mml:mo>=</mml:mo><mml:msup><mml:mi mathvariant="normal">Mean</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msup><mml:mo>+</mml:mo><mml:msup><mml:mi mathvariant="normal">STD</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msup></mml:mrow></mml:math></inline-formula>.</p></caption><oasis:table frame="topbot"><oasis:tgroup cols="5">
     <oasis:colspec colnum="1" colname="col1" align="left"/>
     <oasis:colspec colnum="2" colname="col2" align="right"/>
     <oasis:colspec colnum="3" colname="col3" align="right"/>
     <oasis:colspec colnum="4" colname="col4" align="right"/>
     <oasis:colspec colnum="5" colname="col5" align="right"/>
     <oasis:thead>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">Model</oasis:entry>
         <oasis:entry colname="col2">Section</oasis:entry>
         <oasis:entry colname="col3">RMSE [–]</oasis:entry>
         <oasis:entry colname="col4">Mean error [–]</oasis:entry>
         <oasis:entry colname="col5">Error STD [–]</oasis:entry>
       </oasis:row>
     </oasis:thead>
     <oasis:tbody>
       <oasis:row>
         <oasis:entry colname="col1"><inline-formula><mml:math id="M223" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> from LES, <inline-formula><mml:math id="M224" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M225" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> from NBL LES</oasis:entry>
         <oasis:entry colname="col2"><xref ref-type="sec" rid="Ch1.S2"/></oasis:entry>
         <oasis:entry colname="col3">1.6</oasis:entry>
         <oasis:entry colname="col4">  0.8</oasis:entry>
         <oasis:entry colname="col5">1.4</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">2SM–APM, <inline-formula><mml:math id="M226" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M227" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> from NBL LES</oasis:entry>
         <oasis:entry colname="col2"><xref ref-type="sec" rid="Ch1.S4.SS2"/></oasis:entry>
         <oasis:entry colname="col3">4.8</oasis:entry>
         <oasis:entry colname="col4">  2.5</oasis:entry>
         <oasis:entry colname="col5">4.1</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">2SM–APM, <inline-formula><mml:math id="M228" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M229" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> from WM</oasis:entry>
         <oasis:entry colname="col2"><xref ref-type="sec" rid="Ch1.S4.SS4"/></oasis:entry>
         <oasis:entry colname="col3">5.9</oasis:entry>
         <oasis:entry colname="col4"><inline-formula><mml:math id="M230" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>4.8</oasis:entry>
         <oasis:entry colname="col5">3.5</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Farm-averaged PB</oasis:entry>
         <oasis:entry colname="col2"><xref ref-type="sec" rid="Ch1.S5.SS2"/></oasis:entry>
         <oasis:entry colname="col3">5.1</oasis:entry>
         <oasis:entry colname="col4"><inline-formula><mml:math id="M231" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>3.5</oasis:entry>
         <oasis:entry colname="col5">3.7</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Standalone WM</oasis:entry>
         <oasis:entry colname="col2"><xref ref-type="sec" rid="Ch1.S4.SS4"/></oasis:entry>
         <oasis:entry colname="col3">5.2</oasis:entry>
         <oasis:entry colname="col4"><inline-formula><mml:math id="M232" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>1.1</oasis:entry>
         <oasis:entry colname="col5">5.1</oasis:entry>
       </oasis:row>
     </oasis:tbody>
   </oasis:tgroup></oasis:table></table-wrap>

      <p id="d2e6026">To investigate the model's performance further, Fig. <xref ref-type="fig" rid="F6"/> shows velocity deficits <inline-formula><mml:math id="M233" display="inline"><mml:mrow><mml:msubsup><mml:mi>u</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> as predicted by the fast 2SM–APM model for three different cases, and compares them to the LES results. The three cases are <inline-formula><mml:math id="M234" display="inline"><mml:mrow><mml:mi>H</mml:mi><mml:mn mathvariant="normal">500</mml:mn></mml:mrow></mml:math></inline-formula>-<inline-formula><mml:math id="M235" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:mrow></mml:math></inline-formula>5-<inline-formula><mml:math id="M236" display="inline"><mml:mi mathvariant="normal">Γ</mml:mi></mml:math></inline-formula>4 (left), <inline-formula><mml:math id="M237" display="inline"><mml:mrow><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1000</mml:mn></mml:mrow></mml:math></inline-formula>-<inline-formula><mml:math id="M238" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:mrow></mml:math></inline-formula>2-<inline-formula><mml:math id="M239" display="inline"><mml:mi mathvariant="normal">Γ</mml:mi></mml:math></inline-formula>1 (middle), and <inline-formula><mml:math id="M240" display="inline"><mml:mrow><mml:mi>H</mml:mi><mml:mn mathvariant="normal">300</mml:mn></mml:mrow></mml:math></inline-formula>-<inline-formula><mml:math id="M241" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:mrow></mml:math></inline-formula>8-<inline-formula><mml:math id="M242" display="inline"><mml:mi mathvariant="normal">Γ</mml:mi></mml:math></inline-formula>1 (right), which are the cases where the model performs best, median, and worst, respectively. In general, accurate predictions of the average velocity are in part due to a cancellation of errors, as the model over- and underpredicts the velocity at the front and back of the farm, respectively. While the model predicts the wrong location for the resonant lee waves for some of the <inline-formula><mml:math id="M243" display="inline"><mml:mrow><mml:mi>H</mml:mi><mml:mn mathvariant="normal">300</mml:mn></mml:mrow></mml:math></inline-formula> cases, as can be seen in Fig. <xref ref-type="fig" rid="F6"/> (right), this is not a consistent source of error. These errors in the local velocity profiles are not due to the 2SM simplifications, as earlier conventional versions of the APM have similar behavior <xref ref-type="bibr" rid="bib1.bibx8" id="paren.59"/>.</p>

      <fig id="F6" specific-use="star"><label>Figure 6</label><caption><p id="d2e6145">Lower-layer velocity deficit through the center of the farm as predicted by LES (black) and the fast 2SM–APM (blue, orange) for the cases <inline-formula><mml:math id="M244" display="inline"><mml:mrow><mml:mi>H</mml:mi><mml:mn mathvariant="normal">500</mml:mn></mml:mrow></mml:math></inline-formula>-<inline-formula><mml:math id="M245" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:mrow></mml:math></inline-formula>5-<inline-formula><mml:math id="M246" display="inline"><mml:mi mathvariant="normal">Γ</mml:mi></mml:math></inline-formula>4 (left), <inline-formula><mml:math id="M247" display="inline"><mml:mrow><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1000</mml:mn></mml:mrow></mml:math></inline-formula>-<inline-formula><mml:math id="M248" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:mrow></mml:math></inline-formula>2-<inline-formula><mml:math id="M249" display="inline"><mml:mi mathvariant="normal">Γ</mml:mi></mml:math></inline-formula>1 (middle), and <inline-formula><mml:math id="M250" display="inline"><mml:mrow><mml:mi>H</mml:mi><mml:mn mathvariant="normal">300</mml:mn></mml:mrow></mml:math></inline-formula>-<inline-formula><mml:math id="M251" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:mrow></mml:math></inline-formula>8-<inline-formula><mml:math id="M252" display="inline"><mml:mi mathvariant="normal">Γ</mml:mi></mml:math></inline-formula>1 (right). The dotted lines denote the wind farm region. The blue and orange lines showing the 2SM–APM predictions were obtained with <inline-formula><mml:math id="M253" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> values calculated using LES and a wake model, respectively.</p></caption>
          <graphic xlink:href="https://wes.copernicus.org/articles/11/4349/2026/wes-11-4349-2026-f06.png"/>

        </fig>

      <p id="d2e6249">Lastly, we discuss the computational cost of the 2SM–APM. As desired, the model achieves a relatively low computational cost, with each calculation taking roughly 1.3 CPU seconds on an Intel Xeon Platinum 8360Y 2.4 GHz CPU. This is only approximately 35 times longer than a wake model evaluation for the same farm, and drastically faster than the conventional APM with local coupling <xref ref-type="bibr" rid="bib1.bibx9" id="paren.60"/>. The main computational gains are achieved by eliminating the costs of re-calculating the turbine forces and of repeatedly solving the APM Eqs. (<xref ref-type="disp-formula" rid="Ch1.E7"/>)–(<xref ref-type="disp-formula" rid="Ch1.E11"/>). While the former cost is shifted to the calculation of the farm coefficients <inline-formula><mml:math id="M254" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M255" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>, as discussed in the next section, the latter is fully removed.</p>
</sec>
<sec id="Ch1.S4.SS4">
  <label>4.4</label><title>Using a wake model to compute <inline-formula><mml:math id="M256" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M257" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula></title>
      <p id="d2e6319">When using the 2SM–APM in practical settings, it is not feasible to rely on LES to calculate the farm thrust and power coefficients <inline-formula><mml:math id="M258" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M259" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>, as these simulations are too expensive. Therefore, this section aims to briefly demonstrate how a basic wake model can be used to calculate these coefficients instead of LES. We merely wish to show that this approach is conceptually possible and point out some potential issues.</p>
      <p id="d2e6348">In essence, the process is the same as before, i.e., the standalone wake model is used to estimate <inline-formula><mml:math id="M260" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mrow><mml:mspace width="-0.125em" linebreak="nobreak"/><mml:mi mathvariant="normal">f</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M261" display="inline"><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, and <inline-formula><mml:math id="M262" display="inline"><mml:mi>F</mml:mi></mml:math></inline-formula> for a given farm, so that <inline-formula><mml:math id="M263" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M264" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> can be calculated using Eqs. (<xref ref-type="disp-formula" rid="Ch1.E1"/>) and (<xref ref-type="disp-formula" rid="Ch1.E2"/>). As shown in Sect. <xref ref-type="sec" rid="Ch1.S2"/>, the resulting coefficients capture the turbine-scale flow characteristics and are valid across a range of atmospheric conditions as long as the turbine setpoints and wake shapes remain the same. Once they are calculated, the 2SM–APM can use them to predict the farm power output in different mesoscale flow conditions.</p>
      <p id="d2e6415">To calculate the farm power and thrust, we use a Gaussian wake model <xref ref-type="bibr" rid="bib1.bibx4" id="paren.61"/> with the wake merging method of <xref ref-type="bibr" rid="bib1.bibx15" id="text.62"/> and the turbulence-intensity model of <xref ref-type="bibr" rid="bib1.bibx23" id="text.63"/>. The wake model is not tuned and is run with the parameters reported by the above authors. The turbines are mirrored to account for ground effects, as this significantly reduces the error of the predicted power output and farm-averaged velocity. This conflicts with the results from <xref ref-type="bibr" rid="bib1.bibx4" id="text.64"/>, who found that mirroring the turbines had little effect on the WM's predictions. This difference with the original papers might be because we use a turbine type with a higher ratio of <inline-formula><mml:math id="M265" display="inline"><mml:mrow><mml:mi>D</mml:mi><mml:mo>/</mml:mo><mml:msub><mml:mi>z</mml:mi><mml:mi>h</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, leading to the rotors being relatively closer to the ground.</p>
      <p id="d2e6446">Accurately calculating <inline-formula><mml:math id="M266" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mrow><mml:mspace linebreak="nobreak" width="-0.125em"/><mml:mi mathvariant="normal">f</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula> with a wake model requires more modeling effort than <inline-formula><mml:math id="M267" display="inline"><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M268" display="inline"><mml:mi>F</mml:mi></mml:math></inline-formula> <xref ref-type="bibr" rid="bib1.bibx8" id="paren.65"/>. Since wake models typically only need to estimate the velocity at downstream turbine locations, they are not always optimized for velocity estimations in the near-wake and induction regions. If not properly accounted for, this can lead to large errors in the average velocity predictions. Therefore, we follow <xref ref-type="bibr" rid="bib1.bibx8" id="text.66"/> and include a turbine-induction model when estimating <inline-formula><mml:math id="M269" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mrow><mml:mspace width="-0.125em" linebreak="nobreak"/><mml:mi mathvariant="normal">f</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula> <xref ref-type="bibr" rid="bib1.bibx34" id="paren.67"/>.</p>
      <p id="d2e6503">The farm-averaged velocity <inline-formula><mml:math id="M270" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mrow><mml:mspace width="-0.125em" linebreak="nobreak"/><mml:mi mathvariant="normal">f</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula> is calculated by evaluating the wake model on a dense, evenly spaced grid over the whole 2SM averaging volume. This brute-force approach is computationally expensive, and a more efficient placement of the evaluation points would likely improve it. However, as this section simply aims to demonstrate that this type of wake model coupling can be done in principle, we leave further optimizations of how to estimate <inline-formula><mml:math id="M271" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mrow><mml:mspace width="-0.125em" linebreak="nobreak"/><mml:mi mathvariant="normal">f</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula> to future work.</p>
      <p id="d2e6532">The resulting <inline-formula><mml:math id="M272" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M273" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> are 1.05 and 0.79, respectively. Both values are lower than those calculated from the neutral LES (1.15 and 0.92). This is mostly due to an error in <inline-formula><mml:math id="M274" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mrow><mml:mspace width="-0.125em" linebreak="nobreak"/><mml:mi mathvariant="normal">f</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula>, as even with an induction model, the farm-averaged velocity is overestimated <xref ref-type="bibr" rid="bib1.bibx8" id="paren.68"/>.</p>
      <p id="d2e6577">With these coefficients, we now repeat the earlier validation using the LES dataset from <xref ref-type="bibr" rid="bib1.bibx19" id="text.69"/>. Figure <xref ref-type="fig" rid="F7"/> and Table <xref ref-type="table" rid="T2"/> show the resulting farm power predictions and error statistics, respectively, for both the 2SM–APM and the standalone wake model. Using the wake model to calculate <inline-formula><mml:math id="M275" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M276" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> leads to a higher errors, mainly due to the consistent negative bias. Interestingly, the standalone wake model has a similar RMSE as the 2SM–APM despite not capturing any of the power variations in the dataset, mostly because its mean error is very small. In comparison, the 2SM–APM have a higher bias but a significantly lower variance.</p>

      <fig id="F7"><label>Figure 7</label><caption><p id="d2e6615">A comparison of farm power <inline-formula><mml:math id="M277" display="inline"><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> as calculated by LES (<inline-formula><mml:math id="M278" display="inline"><mml:mi>x</mml:mi></mml:math></inline-formula> axis), the fast 2SM–APM model (orange circles), and a standalone wake model (black crosses) for 27 different CNBL cases from <xref ref-type="bibr" rid="bib1.bibx19" id="text.70"/>. The farm coefficients <inline-formula><mml:math id="M279" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M280" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> used by the 2AM–APM are calculated using the wake model. The axes are normalized with the power output of a standalone turbine <inline-formula><mml:math id="M281" display="inline"><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mn mathvariant="normal">0</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, as calculated using LES (horizontal) and the wake model (vertical).</p></caption>
          <graphic xlink:href="https://wes.copernicus.org/articles/11/4349/2026/wes-11-4349-2026-f07.png"/>

        </fig>

      <p id="d2e6684">To investigate further, Fig. <xref ref-type="fig" rid="F6"/> shows the velocity perturbations predicted by the 2SM–APM. As a result of the lower <inline-formula><mml:math id="M282" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> value, the model predicts smaller <inline-formula><mml:math id="M283" display="inline"><mml:mrow><mml:msubsup><mml:mi>u</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> than before, leading to higher <inline-formula><mml:math id="M284" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>. However, this is counteracted by the lower <inline-formula><mml:math id="M285" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>, so that the resulting farm power predictions have a consistent negative bias.</p>
      <p id="d2e6739">The above discussion shows that in principle wake models can serve the role of a prior CFD simulation in determining <inline-formula><mml:math id="M286" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M287" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> but that this requires the wake model to accurately represent the flow in the induction and near-wake regions. The associated computational cost can be significantly reduced by future research, as we have made no attempts to minimize it. Furthermore, constructing lookup tables for <inline-formula><mml:math id="M288" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M289" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> as a function of wind speed and wake model inputs would allow this upfront cost to be shared by many 2SM–APM cases when performing multiyear AEP estimates, among other things. Finally, the coefficients could potentially also be inferred directly from operational data. Instead of investigating these approaches further, the following section develops a second version of the 2SM–APM which is directly coupled to a wake model, eliminating the need for an explicit calculation of <inline-formula><mml:math id="M290" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M291" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>.</p>
</sec>
</sec>
<sec id="Ch1.S5">
  <label>5</label><title>Direct wake model coupling</title>
      <p id="d2e6831">The previous section treated <inline-formula><mml:math id="M292" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M293" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> as known inputs from a prior simulation or wake model and focused on improving the APM's cost and performance. However, calculating these coefficients with a wake model places requirements on said wake model's performance and comes at a potentially significant computational cost. Therefore, we now apply the constant-footprint technique developed in the previous section to the pressure-based (PB) wake model coupling developed by <xref ref-type="bibr" rid="bib1.bibx31" id="text.71"/>, resulting in a blockage-corrected model for farm power that does not require the explicit calculation of <inline-formula><mml:math id="M294" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M295" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>.</p>
      <p id="d2e6889">This section first briefly derives the pressure-based wake model coupling technique. Afterwards, it applies the techniques developed in the previous section to eliminate the repeated APM operators. The resulting model is again validated using the CNBL simulations of <xref ref-type="bibr" rid="bib1.bibx19" id="text.72"/>.</p>
<sec id="Ch1.S5.SS1">
  <label>5.1</label><title>Derivation of the pressure-based coupling method</title>
      <p id="d2e6902">This section derives the equations for the pressure-based coupling method developed by <xref ref-type="bibr" rid="bib1.bibx31" id="text.73"/>, by interpreting the APM as a correction model for CNBL effects. A similar but less thorough derivation has been made before by <xref ref-type="bibr" rid="bib1.bibx9" id="text.74"/>. A graphical representation of the coupling and the derivation is given in Fig. <xref ref-type="fig" rid="F8"/>, and an overview of the shorthand notations for the different operators used is given in Table <xref ref-type="table" rid="T3"/>.</p>

      <fig id="F8" specific-use="star"><label>Figure 8</label><caption><p id="d2e6917">A conceptual sketch of the pressure-based coupling method. We argue that the APM velocity perturbation can be split up into wake and pressure components and that the latter can serve as the change in background velocity for the wake model due to blockage and gravity-wave effects. Figure based on <xref ref-type="bibr" rid="bib1.bibx31" id="text.75"/>.</p></caption>
          <graphic xlink:href="https://wes.copernicus.org/articles/11/4349/2026/wes-11-4349-2026-f08.png"/>

        </fig>

<table-wrap id="T3"><label>Table 3</label><caption><p id="d2e6932">Overview of the shorthand notations for the different operators used in this paper, the sets of equations they represent, and the variables that serve as their in- and outputs.</p></caption><oasis:table frame="topbot"><oasis:tgroup cols="4">
     <oasis:colspec colnum="1" colname="col1" align="left"/>
     <oasis:colspec colnum="2" colname="col2" align="right"/>
     <oasis:colspec colnum="3" colname="col3" align="left"/>
     <oasis:colspec colnum="4" colname="col4" align="left"/>
     <oasis:thead>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">Operator</oasis:entry>
         <oasis:entry colname="col2">Equations</oasis:entry>
         <oasis:entry colname="col3">Input</oasis:entry>
         <oasis:entry colname="col4">Output</oasis:entry>
       </oasis:row>
     </oasis:thead>
     <oasis:tbody>
       <oasis:row>
         <oasis:entry colname="col1"><inline-formula><mml:math id="M296" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi>u</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M297" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi mathvariant="italic">η</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M298" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi>p</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">(<xref ref-type="disp-formula" rid="Ch1.E7"/>)–(<xref ref-type="disp-formula" rid="Ch1.E11"/>)</oasis:entry>
         <oasis:entry colname="col3"><inline-formula><mml:math id="M299" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mi mathvariant="normal">wf</mml:mi></mml:msub><mml:mo>/</mml:mo><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col4"><inline-formula><mml:math id="M300" display="inline"><mml:mrow><mml:msubsup><mml:mi>u</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M301" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M302" display="inline"><mml:mrow><mml:msup><mml:mi>p</mml:mi><mml:mo>′</mml:mo></mml:msup></mml:mrow></mml:math></inline-formula></oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"><inline-formula><mml:math id="M303" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mi>u</mml:mi><mml:mi mathvariant="normal">n</mml:mi></mml:msubsup></mml:mrow></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">(<xref ref-type="disp-formula" rid="Ch1.E37"/>)–(<xref ref-type="disp-formula" rid="Ch1.E38"/>)</oasis:entry>
         <oasis:entry colname="col3"><inline-formula><mml:math id="M304" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mi mathvariant="normal">wf</mml:mi></mml:msub><mml:mo>/</mml:mo><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col4"><inline-formula><mml:math id="M305" display="inline"><mml:mrow><mml:msubsup><mml:mi>u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">w</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula></oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"><inline-formula><mml:math id="M306" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mi>u</mml:mi><mml:mi mathvariant="normal">c</mml:mi></mml:msubsup></mml:mrow></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">(<xref ref-type="disp-formula" rid="Ch1.E40"/>)–<xref ref-type="disp-formula" rid="Ch1.E42"/>)</oasis:entry>
         <oasis:entry colname="col3"><inline-formula><mml:math id="M307" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mi mathvariant="normal">wf</mml:mi></mml:msub><mml:mo>/</mml:mo><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col4"><inline-formula><mml:math id="M308" display="inline"><mml:mrow><mml:msubsup><mml:mi>u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula></oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"><inline-formula><mml:math id="M309" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mrow><mml:mi>u</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">P</mml:mi></mml:mrow><mml:mi mathvariant="normal">c</mml:mi></mml:msubsup></mml:mrow></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">(<xref ref-type="disp-formula" rid="Ch1.E40"/>)–(<xref ref-type="disp-formula" rid="Ch1.E41"/>)</oasis:entry>
         <oasis:entry colname="col3"><inline-formula><mml:math id="M310" display="inline"><mml:mrow><mml:msup><mml:mi>p</mml:mi><mml:mo>′</mml:mo></mml:msup></mml:mrow></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col4"><inline-formula><mml:math id="M311" display="inline"><mml:mrow><mml:msubsup><mml:mi>u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula></oasis:entry>
       </oasis:row>
     </oasis:tbody>
   </oasis:tgroup></oasis:table></table-wrap>

      <p id="d2e7254">A wake model estimates the turbine-scale velocity field <inline-formula><mml:math id="M312" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mi mathvariant="normal">micro</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> by superposing turbine wake functions on a given background velocity <inline-formula><mml:math id="M313" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, so that

            <disp-formula id="Ch1.E29" content-type="numbered"><label>27</label><mml:math id="M314" display="block"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mi mathvariant="normal">micro</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:mi mathvariant="normal">WM</mml:mi><mml:mo>(</mml:mo><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub><mml:mo>)</mml:mo><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

          where WM is the wake model operator.</p>
      <p id="d2e7305">The background velocity <inline-formula><mml:math id="M315" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> is defined as being the velocity field that will result in the correct <inline-formula><mml:math id="M316" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mi mathvariant="normal">micro</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> when the wakes are superimposed on it. In the absence of blockage and gravity-wave effects, this is simply the unperturbed velocity in the absence of the wind farm, so that <inline-formula><mml:math id="M317" display="inline"><mml:mrow><mml:mo>〈</mml:mo><mml:mover accent="true"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub></mml:mrow><mml:mo mathvariant="normal" stretchy="true">̃</mml:mo></mml:mover><mml:msub><mml:mo>〉</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>=</mml:mo><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>. In contrast, when coupling a wake model to the APM, <inline-formula><mml:math id="M318" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> has to ensure that horizontally filtering and height-averaging the turbine-scale velocity field results in the APM velocity, so that

            <disp-formula id="Ch1.E30" content-type="numbered"><label>28</label><mml:math id="M319" display="block"><mml:mrow><mml:mo>〈</mml:mo><mml:mover accent="true"><mml:mrow><mml:mi mathvariant="normal">WM</mml:mi><mml:mo>(</mml:mo><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub><mml:mo>)</mml:mo></mml:mrow><mml:mo mathvariant="normal" stretchy="true">̃</mml:mo></mml:mover><mml:msub><mml:mo>〉</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>=</mml:mo><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup><mml:mo>.</mml:mo></mml:mrow></mml:math></disp-formula>

          <xref ref-type="bibr" rid="bib1.bibx8" id="text.76"/> call this the <italic>velocity matching</italic> (VM) equation. This condition gives an implicit equation for <inline-formula><mml:math id="M320" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> and is related to using the pre-computed farm coefficients <inline-formula><mml:math id="M321" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M322" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> when simplifying the APM–WM coupling using two-scale momentum theory, as discussed in the previous section. Similar to computing these farm coefficients, solving the VM equation for <inline-formula><mml:math id="M323" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> is relatively expensive <xref ref-type="bibr" rid="bib1.bibx9" id="paren.77"/>.</p>
      <p id="d2e7475">In order to find a computationally cheaper way to estimate <inline-formula><mml:math id="M324" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, we decompose the filtered, height-averaged wake model operator into a wake and a correction component, so that

            <disp-formula id="Ch1.E31" content-type="numbered"><label>29</label><mml:math id="M325" display="block"><mml:mrow><mml:mo>〈</mml:mo><mml:mover accent="true"><mml:mrow><mml:mi mathvariant="normal">WM</mml:mi><mml:mo>(</mml:mo><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub><mml:mo>)</mml:mo></mml:mrow><mml:mo mathvariant="normal" stretchy="true">̃</mml:mo></mml:mover><mml:msub><mml:mo>〉</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>=</mml:mo><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">w</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>+</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

          where the wake component <inline-formula><mml:math id="M326" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">w</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> is the filtered and height-averaged velocity deficit corresponding to the turbine wakes, as predicted by the wake model, and the correction component <inline-formula><mml:math id="M327" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> accounts for the remainder of the total velocity perturbation <inline-formula><mml:math id="M328" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>. This splitting is shown conceptually on the right-hand side of Fig. <xref ref-type="fig" rid="F8"/>. For sufficiently small flow perturbations, we can then assume that subtracting the wake component <inline-formula><mml:math id="M329" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">w</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> from the above equation corresponds to not applying the wake model operator WM, so that

            <disp-formula id="Ch1.E32" content-type="numbered"><label>30</label><mml:math id="M330" display="block"><mml:mrow><mml:mo>〈</mml:mo><mml:mover accent="true"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub></mml:mrow><mml:mo mathvariant="normal" stretchy="true">̃</mml:mo></mml:mover><mml:msub><mml:mo>〉</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>=</mml:mo><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>.</mml:mo></mml:mrow></mml:math></disp-formula>

          We can therefore interpret the correction component <inline-formula><mml:math id="M331" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> as representing the changes in the background velocity that we are trying to estimate.</p>
      <p id="d2e7687">Instead of deriving an expression for <inline-formula><mml:math id="M332" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> directly, we will derive it as the difference between the APM state <inline-formula><mml:math id="M333" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and the wake component <inline-formula><mml:math id="M334" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">w</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>. Using Eq. (<xref ref-type="disp-formula" rid="Ch1.E30"/>), we find

            <disp-formula id="Ch1.E33" content-type="numbered"><label>31</label><mml:math id="M335" display="block"><mml:mrow><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mo>′</mml:mo></mml:msubsup><mml:mo>-</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">w</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>.</mml:mo></mml:mrow></mml:math></disp-formula>

          Therefore, the next step is to estimate <inline-formula><mml:math id="M336" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">w</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>, i.e., the part of the APM velocity perturbation that corresponds to the wake deficit predicted by the wake model. Rather than obtaining this by averaging over the microscale velocity predicted by the wake model, we aim at finding an APM operator that allows us to make a faster direct estimation of <inline-formula><mml:math id="M337" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">w</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>. Since conventional standalone wake models are designed to estimate the response of a neutral atmosphere to turbine forcing, we can reasonably take

            <disp-formula id="Ch1.E34" content-type="numbered"><label>32</label><mml:math id="M338" display="block"><mml:mrow><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">w</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mi>u</mml:mi><mml:mi mathvariant="normal">n</mml:mi></mml:msubsup><mml:mfenced open="(" close=")"><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mi mathvariant="normal">wf</mml:mi></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle></mml:mfenced><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

          where <inline-formula><mml:math id="M339" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mi>u</mml:mi><mml:mi mathvariant="normal">n</mml:mi></mml:msubsup></mml:mrow></mml:math></inline-formula> is the APM operator for neutral conditions. The component <inline-formula><mml:math id="M340" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> can then be seen as a correction to neutral flow due to blockage effects and is defined as

            <disp-formula id="Ch1.E35" content-type="numbered"><label>33</label><mml:math id="M341" display="block"><mml:mrow><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mi>u</mml:mi><mml:mi mathvariant="normal">c</mml:mi></mml:msubsup><mml:mfenced close=")" open="("><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mi mathvariant="normal">wf</mml:mi></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle></mml:mfenced><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

          where <inline-formula><mml:math id="M342" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mi>u</mml:mi><mml:mi mathvariant="normal">c</mml:mi></mml:msubsup></mml:mrow></mml:math></inline-formula> is the APM operator for the difference between CNBL and neutral flow:

            <disp-formula id="Ch1.E36" content-type="numbered"><label>34</label><mml:math id="M343" display="block"><mml:mrow><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mi>u</mml:mi><mml:mi mathvariant="normal">c</mml:mi></mml:msubsup><mml:mo>=</mml:mo><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi>u</mml:mi></mml:msub><mml:mo>-</mml:mo><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mi>u</mml:mi><mml:mi mathvariant="normal">n</mml:mi></mml:msubsup><mml:mo>.</mml:mo></mml:mrow></mml:math></disp-formula>

          If this correction operator <inline-formula><mml:math id="M344" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mi>u</mml:mi><mml:mi mathvariant="normal">c</mml:mi></mml:msubsup></mml:mrow></mml:math></inline-formula> is known, it can be used to directly predict <inline-formula><mml:math id="M345" display="inline"><mml:mrow><mml:msubsup><mml:mi>u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> (and correspondingly <inline-formula><mml:math id="M346" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>), without explicitly solving the velocity matching equation. To find <inline-formula><mml:math id="M347" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mi>u</mml:mi><mml:mi mathvariant="normal">c</mml:mi></mml:msubsup></mml:mrow></mml:math></inline-formula>, we first construct the neutral APM operator <inline-formula><mml:math id="M348" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mi>u</mml:mi><mml:mi mathvariant="normal">n</mml:mi></mml:msubsup></mml:mrow></mml:math></inline-formula>, since they are related through the above equation.</p>
      <p id="d2e8054">To find the neutral APM operator <inline-formula><mml:math id="M349" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mi>u</mml:mi><mml:mi mathvariant="normal">n</mml:mi></mml:msubsup></mml:mrow></mml:math></inline-formula>, we re-derive the APM for neutral flow. Doing so, the continuity and momentum equations remain unaltered. The only equation that changes is the pressure closure equation. In the absence of gravity waves, the pressure perturbations triggered by the farm are solely due to hydrodynamic effects, which are small compared to the buoyancy effects at typical wind farm length scales. This is demonstrated by the LES data of <xref ref-type="bibr" rid="bib1.bibx19" id="text.78"/>, where the pressure perturbations in the neutral cases are negligible compared to the CNBL simulations. Therefore, we take <inline-formula><mml:math id="M350" display="inline"><mml:mrow><mml:msubsup><mml:mi>p</mml:mi><mml:mi mathvariant="normal">n</mml:mi><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:mn mathvariant="normal">0</mml:mn></mml:mrow></mml:math></inline-formula>.</p>
      <p id="d2e8090">We then easily find the model for the neutral component as Eqs. (<xref ref-type="disp-formula" rid="Ch1.E7"/>)–(<xref ref-type="disp-formula" rid="Ch1.E11"/>) without the pressure term. For the sake of brevity, we only write the equations for the lower layer:

                <disp-formula specific-use="gather" content-type="numbered"><mml:math id="M351" display="block"><mml:mtable displaystyle="true"><mml:mlabeledtr id="Ch1.E37"><mml:mtd><mml:mtext>35</mml:mtext></mml:mtd><mml:mtd><mml:mrow><mml:mstyle displaystyle="true" class="stylechange"/><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>⋅</mml:mo><mml:msub><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi></mml:msub><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">n</mml:mi></mml:mrow></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:msub><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi></mml:msub><mml:mo>⋅</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">n</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:mn mathvariant="normal">0</mml:mn><mml:mo>,</mml:mo></mml:mrow></mml:mtd></mml:mlabeledtr><mml:mlabeledtr id="Ch1.E38"><mml:mtd><mml:mtext>36</mml:mtext></mml:mtd><mml:mtd><mml:mrow><mml:mstyle class="stylechange" displaystyle="true"/><mml:mtable class="split" rowspacing="0.2ex" displaystyle="true" columnalign="right left"><mml:mtr><mml:mtd><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mtd><mml:mtd><mml:mrow><mml:mo>⋅</mml:mo><mml:msub><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi></mml:msub><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">n</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:msub><mml:mi>f</mml:mi><mml:mi mathvariant="normal">c</mml:mi></mml:msub><mml:mi mathvariant="bold">J</mml:mi><mml:mo>⋅</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">n</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>+</mml:mo><mml:msub><mml:mi mathvariant="italic">ν</mml:mi><mml:mrow><mml:mi>t</mml:mi><mml:mo>,</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msub><mml:msubsup><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">n</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>+</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msup><mml:mi mathvariant="bold">D</mml:mi><mml:mo>′</mml:mo></mml:msup></mml:mrow><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>⋅</mml:mo><mml:msubsup><mml:mi mathvariant="normal">Δ</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mi mathvariant="normal">n</mml:mi><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:mtd></mml:mtr><mml:mtr><mml:mtd/><mml:mtd><mml:mrow><mml:mo>-</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msup><mml:mi mathvariant="bold">C</mml:mi><mml:mo>′</mml:mo></mml:msup></mml:mrow><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>⋅</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">n</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>-</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">T</mml:mi><mml:mrow><mml:mi mathvariant="normal">h</mml:mi><mml:mn mathvariant="normal">3</mml:mn><mml:mo>,</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msub><mml:mo>-</mml:mo><mml:msub><mml:mi mathvariant="bold-italic">T</mml:mi><mml:mrow><mml:mi mathvariant="normal">h</mml:mi><mml:mn mathvariant="normal">3</mml:mn><mml:mo>,</mml:mo><mml:mn mathvariant="normal">0</mml:mn></mml:mrow></mml:msub></mml:mrow><mml:mrow><mml:msubsup><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup></mml:mrow></mml:mfrac></mml:mstyle><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">n</mml:mi></mml:mrow></mml:msub><mml:mo>+</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mover accent="true"><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mo mathvariant="normal" stretchy="true">̃</mml:mo></mml:mover><mml:mi mathvariant="normal">wf</mml:mi></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">n</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>,</mml:mo></mml:mrow></mml:mtd></mml:mtr></mml:mtable></mml:mrow></mml:mtd></mml:mlabeledtr><mml:mlabeledtr id="Ch1.E39"><mml:mtd><mml:mtext>37</mml:mtext></mml:mtd><mml:mtd><mml:mrow><mml:mstyle displaystyle="true" class="stylechange"/><mml:msubsup><mml:mi>p</mml:mi><mml:mi mathvariant="normal">n</mml:mi><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:mn mathvariant="normal">0</mml:mn><mml:mo>.</mml:mo></mml:mrow></mml:mtd></mml:mlabeledtr></mml:mtable></mml:math></disp-formula></p>
      <p id="d2e8425">The equations for the correction operator <inline-formula><mml:math id="M352" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mi>u</mml:mi><mml:mi mathvariant="normal">c</mml:mi></mml:msubsup></mml:mrow></mml:math></inline-formula> can then be found by simply subtracting the above equations for the neutral model from the original APM continuity and momentum equations. Doing this results in

                <disp-formula specific-use="gather" content-type="numbered"><mml:math id="M353" display="block"><mml:mtable displaystyle="true"><mml:mlabeledtr id="Ch1.E40"><mml:mtd><mml:mtext>38</mml:mtext></mml:mtd><mml:mtd><mml:mrow><mml:mstyle displaystyle="true" class="stylechange"/><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>⋅</mml:mo><mml:msub><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi></mml:msub><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:msub><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi></mml:msub><mml:mo>⋅</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:mn mathvariant="normal">0</mml:mn><mml:mo>,</mml:mo></mml:mrow></mml:mtd></mml:mlabeledtr><mml:mlabeledtr id="Ch1.E41"><mml:mtd><mml:mtext>39</mml:mtext></mml:mtd><mml:mtd><mml:mrow><mml:mstyle class="stylechange" displaystyle="true"/><mml:mtable class="split" rowspacing="0.2ex" displaystyle="true" columnalign="right left"><mml:mtr><mml:mtd><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mtd><mml:mtd><mml:mrow><mml:mo>⋅</mml:mo><mml:msub><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi></mml:msub><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:msub><mml:mi>f</mml:mi><mml:mi mathvariant="normal">c</mml:mi></mml:msub><mml:mi mathvariant="bold">J</mml:mi><mml:mo>⋅</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>+</mml:mo><mml:msub><mml:mi mathvariant="italic">ν</mml:mi><mml:mrow><mml:mi>t</mml:mi><mml:mo>,</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msub><mml:msubsup><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>+</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msup><mml:mi mathvariant="bold">D</mml:mi><mml:mo>′</mml:mo></mml:msup></mml:mrow><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>⋅</mml:mo><mml:msubsup><mml:mi mathvariant="normal">Δ</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mi mathvariant="normal">c</mml:mi><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:mtd></mml:mtr><mml:mtr><mml:mtd/><mml:mtd><mml:mrow><mml:mo>-</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msup><mml:mi mathvariant="bold">C</mml:mi><mml:mo>′</mml:mo></mml:msup></mml:mrow><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>⋅</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>-</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">T</mml:mi><mml:mrow><mml:mi mathvariant="normal">h</mml:mi><mml:mn mathvariant="normal">3</mml:mn><mml:mo>,</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msub><mml:mo>-</mml:mo><mml:msub><mml:mi mathvariant="bold-italic">T</mml:mi><mml:mrow><mml:mi mathvariant="normal">h</mml:mi><mml:mn mathvariant="normal">3</mml:mn><mml:mo>,</mml:mo><mml:mn mathvariant="normal">0</mml:mn></mml:mrow></mml:msub></mml:mrow><mml:mrow><mml:msubsup><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup></mml:mrow></mml:mfrac></mml:mstyle><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow></mml:msub><mml:mo>-</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mi mathvariant="italic">ρ</mml:mi></mml:mfrac></mml:mstyle><mml:msub><mml:mi mathvariant="normal">∇</mml:mi><mml:mi mathvariant="normal">h</mml:mi></mml:msub><mml:msubsup><mml:mi>p</mml:mi><mml:mi mathvariant="normal">c</mml:mi><mml:mo>′</mml:mo></mml:msubsup><mml:mo>,</mml:mo></mml:mrow></mml:mtd></mml:mtr></mml:mtable></mml:mrow></mml:mtd></mml:mlabeledtr><mml:mlabeledtr id="Ch1.E42"><mml:mtd><mml:mtext>40</mml:mtext></mml:mtd><mml:mtd><mml:mrow><mml:mstyle class="stylechange" displaystyle="true"/><mml:msubsup><mml:mi>p</mml:mi><mml:mi mathvariant="normal">c</mml:mi><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:msup><mml:mi>p</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mo>=</mml:mo><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi>p</mml:mi></mml:msub><mml:mfenced open="(" close=")"><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mrow><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle></mml:mfenced><mml:mo>,</mml:mo></mml:mrow></mml:mtd></mml:mlabeledtr></mml:mtable></mml:math></disp-formula>

          where we again only write the lower-layer equations for the sake of brevity. Note that for a given forcing <inline-formula><mml:math id="M354" display="inline"><mml:mrow><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mo>/</mml:mo><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, the pressure perturbation <inline-formula><mml:math id="M355" display="inline"><mml:mrow><mml:msup><mml:mi>p</mml:mi><mml:mo>′</mml:mo></mml:msup></mml:mrow></mml:math></inline-formula> follows from Eq. (<xref ref-type="disp-formula" rid="Ch1.E24"/>) (based on the original APM operator <inline-formula><mml:math id="M356" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi>p</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>) and is an input to Eqs. (<xref ref-type="disp-formula" rid="Ch1.E40"/>)–(<xref ref-type="disp-formula" rid="Ch1.E41"/>).</p>
      <p id="d2e8824">We again define a simplified notation that expresses <inline-formula><mml:math id="M357" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> as the solution of the above system of equations:

            <disp-formula id="Ch1.E43" content-type="numbered"><label>41</label><mml:math id="M358" display="block"><mml:mrow><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mrow><mml:mi>u</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">P</mml:mi></mml:mrow><mml:mi mathvariant="normal">c</mml:mi></mml:msubsup><mml:mo>(</mml:mo><mml:msup><mml:mi>p</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mo>)</mml:mo><mml:mo>=</mml:mo><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mrow><mml:mi>u</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">P</mml:mi></mml:mrow><mml:mi mathvariant="normal">c</mml:mi></mml:msubsup><mml:mfenced close=")" open="("><mml:mrow><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi>p</mml:mi></mml:msub><mml:mfenced close=")" open="("><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mi mathvariant="normal">wf</mml:mi></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle></mml:mfenced></mml:mrow></mml:mfenced><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

          where the operator <inline-formula><mml:math id="M359" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mrow><mml:mi>u</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">P</mml:mi></mml:mrow><mml:mi mathvariant="normal">c</mml:mi></mml:msubsup></mml:mrow></mml:math></inline-formula> represents the solution operator of Eqs. (<xref ref-type="disp-formula" rid="Ch1.E40"/>)–(<xref ref-type="disp-formula" rid="Ch1.E41"/>) given <inline-formula><mml:math id="M360" display="inline"><mml:mrow><mml:msup><mml:mi>p</mml:mi><mml:mo>′</mml:mo></mml:msup></mml:mrow></mml:math></inline-formula> as an input. Note that <inline-formula><mml:math id="M361" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mrow><mml:mi>u</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">P</mml:mi></mml:mrow><mml:mi mathvariant="normal">c</mml:mi></mml:msubsup><mml:mo>(</mml:mo><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi>p</mml:mi></mml:msub><mml:mo>)</mml:mo></mml:mrow></mml:math></inline-formula> thus defines <inline-formula><mml:math id="M362" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mi>u</mml:mi><mml:mi mathvariant="normal">c</mml:mi></mml:msubsup></mml:mrow></mml:math></inline-formula> (Eq. <xref ref-type="disp-formula" rid="Ch1.E36"/>).</p>
      <p id="d2e9000">Finally, we have thus arrived at a closed system of equations, as through Eq. (<xref ref-type="disp-formula" rid="Ch1.E43"/>) we can calculate the change in background velocity for a given forcing <inline-formula><mml:math id="M363" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mi mathvariant="normal">wf</mml:mi></mml:msub><mml:mo>/</mml:mo><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>. This allows us to again employ a fixed-point iteration, where <inline-formula><mml:math id="M364" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mi mathvariant="normal">wf</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> is calculated using the wake model, and the corresponding changes in <inline-formula><mml:math id="M365" display="inline"><mml:mrow><mml:msub><mml:mi>h</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M366" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula> are calculated using <inline-formula><mml:math id="M367" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi mathvariant="italic">η</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M368" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mrow><mml:mi>u</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">P</mml:mi></mml:mrow><mml:mi mathvariant="normal">c</mml:mi></mml:msubsup><mml:mo>(</mml:mo><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi>p</mml:mi></mml:msub><mml:mo>)</mml:mo></mml:mrow></mml:math></inline-formula>, respectively. The wake model then re-calculates the forces with the updated background velocity. This process is repeated until convergence.</p>
      <p id="d2e9100">The only remaining issue is that the input to the wake model requires a turbine-scale background velocity that is not height-averaged. The pressure-based coupling method therefore constructs a local background velocity for the wake model based on <inline-formula><mml:math id="M369" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> <xref ref-type="bibr" rid="bib1.bibx31" id="paren.79"/>. Following the linearization of the APM, we write this background velocity <inline-formula><mml:math id="M370" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> as the sum of the unperturbed velocity profile in the absence of the farm <inline-formula><mml:math id="M371" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mn mathvariant="normal">0</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> and a perturbation around it, so that

            <disp-formula id="Ch1.E44" content-type="numbered"><label>42</label><mml:math id="M372" display="block"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub><mml:mo>(</mml:mo><mml:mi>x</mml:mi><mml:mo>,</mml:mo><mml:mi>y</mml:mi><mml:mo>,</mml:mo><mml:mi>z</mml:mi><mml:mo>)</mml:mo><mml:mo>=</mml:mo><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mn mathvariant="normal">0</mml:mn></mml:msub><mml:mo>(</mml:mo><mml:mi>z</mml:mi><mml:mo>)</mml:mo><mml:mo>+</mml:mo><mml:msub><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub><mml:mo>(</mml:mo><mml:mi>x</mml:mi><mml:mo>,</mml:mo><mml:mi>y</mml:mi><mml:mo>)</mml:mo><mml:msub><mml:mi>f</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub><mml:mo>(</mml:mo><mml:mi>z</mml:mi><mml:mo>)</mml:mo><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

          where <inline-formula><mml:math id="M373" display="inline"><mml:mrow><mml:mo>〈</mml:mo><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mn mathvariant="normal">0</mml:mn></mml:msub><mml:msub><mml:mo>〉</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>=</mml:mo><mml:msub><mml:mi mathvariant="bold-italic">U</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, and <inline-formula><mml:math id="M374" display="inline"><mml:mrow><mml:msub><mml:mi>f</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> is a shape function (we typically choose a logarithmic function). The perturbation to the background velocity is then found by matching the height-averaged <inline-formula><mml:math id="M375" display="inline"><mml:mrow><mml:msubsup><mml:mi>u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>, so that

            <disp-formula id="Ch1.E45" content-type="numbered"><label>43</label><mml:math id="M376" display="block"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:msup><mml:mfenced close="]" open="["><mml:mrow><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:munderover><mml:mo movablelimits="false">∫</mml:mo><mml:mn mathvariant="normal">0</mml:mn><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:munderover><mml:msub><mml:mi>f</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub><mml:mi mathvariant="normal">d</mml:mi><mml:mi>z</mml:mi></mml:mrow></mml:mfenced><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup><mml:mo>.</mml:mo></mml:mrow></mml:math></disp-formula></p>
      <p id="d2e9349"><xref ref-type="bibr" rid="bib1.bibx31" id="text.80"/> did not include the background component of the continuity equations (Eq. <xref ref-type="disp-formula" rid="Ch1.E40"/>) in their original coupling. However, the above derivation of the method shows that the whole APM operator should be split. These continuity equations add a parabolic behavior to the correction operator <inline-formula><mml:math id="M377" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mrow><mml:mi>u</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">P</mml:mi></mml:mrow><mml:mi mathvariant="normal">c</mml:mi></mml:msubsup></mml:mrow></mml:math></inline-formula>, so that the periodic boundary conditions imposed by the Fourier–Galerkin spectral method used to solve the original APM equations become inappropriate. However, since the closure equation for the internal gravity waves (Eqs. <xref ref-type="disp-formula" rid="Ch1.E11"/>–<xref ref-type="disp-formula" rid="Ch1.E12"/>) is not a part of the correction operator, we are free to use other numerical methods. Therefore, we solve Eqs. (<xref ref-type="disp-formula" rid="Ch1.E40"/>)–(<xref ref-type="disp-formula" rid="Ch1.E41"/>) with a simple first-order explicit Euler marching scheme, with as inlet boundary conditions <inline-formula><mml:math id="M378" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>,</mml:mo><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">2</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>,</mml:mo><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow></mml:msub><mml:mo>,</mml:mo><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mrow><mml:mn mathvariant="normal">2</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:mn mathvariant="normal">0</mml:mn></mml:mrow></mml:math></inline-formula>, which correspond to there being no pressure-induced change in background velocity at the inlet. For these conditions to be valid, the inlet should be placed far enough upstream of the farm so that <inline-formula><mml:math id="M379" display="inline"><mml:mrow><mml:msup><mml:mi>p</mml:mi><mml:mo>′</mml:mo></mml:msup><mml:mo>≈</mml:mo><mml:mn mathvariant="normal">0</mml:mn></mml:mrow></mml:math></inline-formula>, which for realistic atmospheric conditions is typically less than 100 km <xref ref-type="bibr" rid="bib1.bibx30" id="paren.81"/>. As a result, the operator <inline-formula><mml:math id="M380" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mrow><mml:mi>u</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">P</mml:mi></mml:mrow><mml:mi mathvariant="normal">c</mml:mi></mml:msubsup></mml:mrow></mml:math></inline-formula> can use far smaller grid sizes than the periodic boundary conditions required of the conventional APM. Finally, in order to simplify the marching solver, we neglect the horizontal eddy viscosity, so that there are no second-order derivatives in the equations.</p>
</sec>
<sec id="Ch1.S5.SS2">
  <label>5.2</label><title>Farm-level pressure-based APM–WM coupling</title>
      <p id="d2e9487">We now apply the farm-averaging approach developed in the previous section to the PB coupling method. As discussed above, the background component of the APM velocity perturbation is given by

            <disp-formula id="Ch1.E46" content-type="numbered"><label>41</label><mml:math id="M381" display="block"><mml:mrow><mml:msubsup><mml:mi mathvariant="bold-italic">u</mml:mi><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mrow><mml:mi>u</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">P</mml:mi></mml:mrow><mml:mi mathvariant="normal">c</mml:mi></mml:msubsup><mml:mfenced open="(" close=")"><mml:mrow><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi>p</mml:mi></mml:msub><mml:mfenced close=")" open="("><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mi mathvariant="bold-italic">f</mml:mi><mml:mi mathvariant="normal">wf</mml:mi></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mi mathvariant="italic">η</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle></mml:mfenced></mml:mrow></mml:mfenced><mml:mo>.</mml:mo></mml:mrow></mml:math></disp-formula>

          By plugging in the constant footprint (Eq. <xref ref-type="disp-formula" rid="Ch1.E25"/>), we find that the average streamwise correction component is

            <disp-formula id="Ch1.E47" content-type="numbered"><label>44</label><mml:math id="M382" display="block"><mml:mrow><mml:msubsup><mml:mover accent="true"><mml:mi>u</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:mo>=</mml:mo><mml:msub><mml:mi>a</mml:mi><mml:mrow><mml:mi>u</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow></mml:msub><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mi>F</mml:mi><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">η</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

          where the coefficient can be calculated as

            <disp-formula id="Ch1.E48" content-type="numbered"><label>45</label><mml:math id="M383" display="block"><mml:mrow><mml:msub><mml:mi>a</mml:mi><mml:mrow><mml:mi>u</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mrow><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:munder><mml:mo movablelimits="false">∬</mml:mo><mml:mrow><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:munder><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mrow><mml:mi>u</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">P</mml:mi></mml:mrow><mml:mi mathvariant="normal">c</mml:mi></mml:msubsup><mml:mfenced close=")" open="("><mml:mrow><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi>p</mml:mi></mml:msub><mml:mfenced close=")" open="("><mml:mrow><mml:msub><mml:mi mathvariant="bold">Π</mml:mi><mml:mi>f</mml:mi></mml:msub></mml:mrow></mml:mfenced></mml:mrow></mml:mfenced><mml:mtext>d</mml:mtext><mml:msub><mml:mi mathvariant="normal">Ω</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mo>.</mml:mo></mml:mrow></mml:math></disp-formula>

          The average layer displacement <inline-formula><mml:math id="M384" display="inline"><mml:mrow><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">η</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> can be found through Eq. (<xref ref-type="disp-formula" rid="Ch1.E26"/>), as in the basic 2SM–APM.</p>
      <p id="d2e9699">Following the pressure-based coupling approach, we now use this <inline-formula><mml:math id="M385" display="inline"><mml:mrow><mml:msubsup><mml:mover accent="true"><mml:mi>u</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> to directly correct the background wind speed in a wake model. Since the above results only hold on a farm average, we assume a constant background over the farm:

            <disp-formula id="Ch1.E49" content-type="numbered"><label>46</label><mml:math id="M386" display="block"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub><mml:mo>(</mml:mo><mml:mi>z</mml:mi><mml:mo>)</mml:mo><mml:mo>=</mml:mo><mml:msub><mml:mi>U</mml:mi><mml:mn mathvariant="normal">0</mml:mn></mml:msub><mml:mo>(</mml:mo><mml:mi>z</mml:mi><mml:mo>)</mml:mo><mml:mo>+</mml:mo><mml:msub><mml:mover accent="true"><mml:mi>u</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mi mathvariant="normal">b</mml:mi></mml:msub><mml:msub><mml:mi>f</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub><mml:mo>(</mml:mo><mml:mi>z</mml:mi><mml:mo>)</mml:mo><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>

          where the average background velocity scale <inline-formula><mml:math id="M387" display="inline"><mml:mrow><mml:msub><mml:mover accent="true"><mml:mi>u</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mi mathvariant="normal">b</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> is found by matching <inline-formula><mml:math id="M388" display="inline"><mml:mrow><mml:msubsup><mml:mover accent="true"><mml:mi>u</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>:

            <disp-formula id="Ch1.E50" content-type="numbered"><label>47</label><mml:math id="M389" display="block"><mml:mrow><mml:msub><mml:mover accent="true"><mml:mi>u</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mi mathvariant="normal">b</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:msubsup><mml:mover accent="true"><mml:mi>u</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>,</mml:mo><mml:mi mathvariant="normal">c</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:msup><mml:mfenced open="[" close="]"><mml:mrow><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">η</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:munderover><mml:mo movablelimits="false">∫</mml:mo><mml:mn mathvariant="normal">0</mml:mn><mml:mrow><mml:msub><mml:mi>H</mml:mi><mml:mn mathvariant="normal">1</mml:mn></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mover accent="true"><mml:mi mathvariant="italic">η</mml:mi><mml:mo mathvariant="normal">‾</mml:mo></mml:mover><mml:mn mathvariant="normal">1</mml:mn></mml:msub></mml:mrow></mml:munderover><mml:msub><mml:mi>f</mml:mi><mml:mi mathvariant="normal">b</mml:mi></mml:msub><mml:mi mathvariant="normal">d</mml:mi><mml:mi>z</mml:mi></mml:mrow></mml:mfenced><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup><mml:mo>.</mml:mo></mml:mrow></mml:math></disp-formula>

          The wake model can then be used to calculate <inline-formula><mml:math id="M390" display="inline"><mml:mi>F</mml:mi></mml:math></inline-formula> and <inline-formula><mml:math id="M391" display="inline"><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, closing the system of equations:

                <disp-formula specific-use="gather" content-type="numbered"><mml:math id="M392" display="block"><mml:mtable displaystyle="true"><mml:mlabeledtr id="Ch1.E51"><mml:mtd><mml:mtext>48</mml:mtext></mml:mtd><mml:mtd><mml:mrow><mml:mstyle class="stylechange" displaystyle="true"/><mml:mi>F</mml:mi><mml:mo>=</mml:mo><mml:munderover><mml:mo movablelimits="false">∑</mml:mo><mml:mi>k</mml:mi><mml:mrow><mml:msub><mml:mi>N</mml:mi><mml:mi mathvariant="normal">t</mml:mi></mml:msub></mml:mrow></mml:munderover><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mn mathvariant="normal">2</mml:mn></mml:mfrac></mml:mstyle><mml:msubsup><mml:mi>C</mml:mi><mml:mrow><mml:mi mathvariant="normal">T</mml:mi><mml:mo>,</mml:mo><mml:mi>k</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:msubsup><mml:mi>S</mml:mi><mml:mi>k</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup><mml:msub><mml:mi>A</mml:mi><mml:mi>k</mml:mi></mml:msub><mml:mo>,</mml:mo></mml:mrow></mml:mtd></mml:mlabeledtr><mml:mlabeledtr id="Ch1.E52"><mml:mtd><mml:mtext>49</mml:mtext></mml:mtd><mml:mtd><mml:mrow><mml:mstyle class="stylechange" displaystyle="true"/><mml:msub><mml:mi>P</mml:mi><mml:mi mathvariant="normal">f</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:munderover><mml:mo movablelimits="false">∑</mml:mo><mml:mi>k</mml:mi><mml:mrow><mml:msub><mml:mi>N</mml:mi><mml:mi mathvariant="normal">t</mml:mi></mml:msub></mml:mrow></mml:munderover><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mn mathvariant="normal">2</mml:mn></mml:mfrac></mml:mstyle><mml:mi mathvariant="italic">ρ</mml:mi><mml:msubsup><mml:mi>C</mml:mi><mml:mrow><mml:mi mathvariant="normal">p</mml:mi><mml:mo>,</mml:mo><mml:mi>k</mml:mi></mml:mrow><mml:mo>′</mml:mo></mml:msubsup><mml:msubsup><mml:mi>S</mml:mi><mml:mi>k</mml:mi><mml:mn mathvariant="normal">3</mml:mn></mml:msubsup><mml:msub><mml:mi>A</mml:mi><mml:mi>k</mml:mi></mml:msub><mml:mo>,</mml:mo></mml:mrow></mml:mtd></mml:mlabeledtr></mml:mtable></mml:math></disp-formula>

          where the turbine coefficients and disk velocities are calculated using a wake model with the background velocity defined in Eq. (<xref ref-type="disp-formula" rid="Ch1.E49"/>).</p>
      <p id="d2e10043">We solve the final system of Eqs. (<xref ref-type="disp-formula" rid="Ch1.E47"/>), (<xref ref-type="disp-formula" rid="Ch1.E26"/>), (<xref ref-type="disp-formula" rid="Ch1.E51"/>), (<xref ref-type="disp-formula" rid="Ch1.E52"/>) using a fixed-point iteration solver with a relaxation factor of 0.7, as is done for the classical APM and the 2SM–APM developed in the previous section.</p>
      <p id="d2e10054">The advantages of this approach compared to the 2SM–APM coupling derived in the previous section are similar to the advantages of PB coupling compared to the VM method. It places no requirements on the wake model and does not need to calculate <inline-formula><mml:math id="M393" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M394" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>. Its main disadvantage is that it requires an additional solver for the operator <inline-formula><mml:math id="M395" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mrow><mml:mi>u</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">P</mml:mi></mml:mrow><mml:mi mathvariant="normal">c</mml:mi></mml:msubsup></mml:mrow></mml:math></inline-formula>, corresponding to Eqs. (<xref ref-type="disp-formula" rid="Ch1.E40"/>)–(<xref ref-type="disp-formula" rid="Ch1.E41"/>). However, since it can use a simple first-order marching scheme and a relatively small grid, this solver is straightforward to implement and cheap to run. Finally, we emphasize that while this coupling no longer explicitly uses any of the results from 2SM theory, it implicitly makes the same assumption that farm power calculations only need the average wind speed within a farm.</p>
<sec id="Ch1.S5.SS2.SSSx1" specific-use="unnumbered">
  <title>Comparison to LES</title>
      <p id="d2e10111">We again validate this approach using the LES cases from the dataset by <xref ref-type="bibr" rid="bib1.bibx19" id="text.82"/> used throughout this work. The numerical setup for the <inline-formula><mml:math id="M396" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="script">A</mml:mi><mml:mi>p</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> operator is the same as in Sect. <xref ref-type="sec" rid="Ch1.S4.SS3"/>. The correction operator <inline-formula><mml:math id="M397" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="script">A</mml:mi><mml:mrow><mml:mi>u</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="normal">P</mml:mi></mml:mrow><mml:mi mathvariant="normal">c</mml:mi></mml:msubsup></mml:mrow></mml:math></inline-formula> can use a smaller numerical domain, since its solver does not require periodic boundary conditions. Here, we use the same spanwise <inline-formula><mml:math id="M398" display="inline"><mml:mrow><mml:msub><mml:mi>L</mml:mi><mml:mi>y</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:mn mathvariant="normal">30</mml:mn></mml:mrow></mml:math></inline-formula> km as before but take a shorter streamwise <inline-formula><mml:math id="M399" display="inline"><mml:mrow><mml:msub><mml:mi>L</mml:mi><mml:mi>x</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:mn mathvariant="normal">43</mml:mn></mml:mrow></mml:math></inline-formula> km, placing the solver's inlet at the same upstream distance of the farm as the LES, i.e., 18 km. We use the same wake model as in Sect. <xref ref-type="sec" rid="Ch1.S4.SS4"/>.</p>
      <p id="d2e10181">The results are shown in Fig. <xref ref-type="fig" rid="F9"/>, and the error statistics are listed in Table <xref ref-type="table" rid="T2"/>. The general trend is visible, despite some scatter and a consistent negative bias. Overall, the wake model and the farm-level PB coupled model have the same RMSE for a power of <inline-formula><mml:math id="M400" display="inline"><mml:mrow><mml:mn mathvariant="normal">5.1</mml:mn><mml:msub><mml:mi>P</mml:mi><mml:mn mathvariant="normal">0</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, or 3.2 % of the farm's installed capacity, with the farm-averaged PB model having a lower error variance but larger bias. The computational cost is the same as before, as the pressure component operator and wake model evaluations have a negligible cost.</p>

      <fig id="F9"><label>Figure 9</label><caption><p id="d2e10203">Farm power as calculated by LES (<inline-formula><mml:math id="M401" display="inline"><mml:mi>x</mml:mi></mml:math></inline-formula> axis) and wake model (<inline-formula><mml:math id="M402" display="inline"><mml:mi>y</mml:mi></mml:math></inline-formula> axis), both standalone (black crosses) and with farm-average pressure-based blockage coupling (green circles). The axes are normalized with the power output of a standalone turbine <inline-formula><mml:math id="M403" display="inline"><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mn mathvariant="normal">0</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, as calculated using LES (horizontal) and the wake model (vertical).</p></caption>
            <graphic xlink:href="https://wes.copernicus.org/articles/11/4349/2026/wes-11-4349-2026-f09.png"/>

          </fig>

      <p id="d2e10237">As shown above, the pressure-based coupling can be derived as a correction to an NBL base state. With that in mind, Fig. <xref ref-type="fig" rid="F10"/> shows the difference in farm power between the coupled and standalone wake model, compared to the farm power differences between the CNBL and the neutral LES reference simulation, i.e., <inline-formula><mml:math id="M404" display="inline"><mml:mi>H</mml:mi></mml:math></inline-formula>500-<inline-formula><mml:math id="M405" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:mrow></mml:math></inline-formula>0-<inline-formula><mml:math id="M406" display="inline"><mml:mi mathvariant="normal">Γ</mml:mi></mml:math></inline-formula>0. It is clear that the farm-average coupling performs decently in the role of a blockage-correction model, as it captures the overall trend despite a consistent bias. The error of this power difference is similar to the error of the farm powers shown in Fig. <xref ref-type="fig" rid="F9"/>, which indicates that the farm-averaged PB model's bias is mostly due to an overprediction of blockage losses.</p>

      <fig id="F10"><label>Figure 10</label><caption><p id="d2e10271">Difference between farm power in CNBL and NBL conditions as calculated by LES (<inline-formula><mml:math id="M407" display="inline"><mml:mi>x</mml:mi></mml:math></inline-formula> axis) and wake model with farm-average pressure-based blockage coupling (<inline-formula><mml:math id="M408" display="inline"><mml:mi>y</mml:mi></mml:math></inline-formula> axis). The axes are normalized with the power output of a standalone turbine <inline-formula><mml:math id="M409" display="inline"><mml:mrow><mml:msub><mml:mi>P</mml:mi><mml:mn mathvariant="normal">0</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, as calculated using LES (horizontal) and the wake model (vertical).</p></caption>
            <graphic xlink:href="https://wes.copernicus.org/articles/11/4349/2026/wes-11-4349-2026-f10.png"/>

          </fig>


</sec>
</sec>
</sec>
<sec id="Ch1.S6">
  <label>6</label><title>Discussion and model comparison</title>
      <p id="d2e10317">This section gives a brief overview and comparison of the performance of the models developed in the previous sections. We emphasize that these conclusions are drawn based on a validation campaign using idealized LES results and that more work is needed to investigate both the 2SM hypothesis and the models' performance further using real operational data.</p>
      <p id="d2e10320">When validated with the idealized LES results from <xref ref-type="bibr" rid="bib1.bibx19" id="text.83"/>, both the <inline-formula><mml:math id="M410" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>-based 2SM–APM and the farm-averaged PB method perform similarly well. When compared to a standalone wake model, they have the same RMSEs on power output of 3 % to 4 % of the farm's installed capacity, as summarized in Table <xref ref-type="table" rid="T2"/>. They can better capture the spread in the dataset due to blockage effects but have a larger bias in their power predictions. Their computational cost is 35 times higher than that of the wake model but still orders of magnitude below conventional CFD or APM tools.</p>
      <p id="d2e10341">The coefficient-based 2SM–APM represents the farm using pre-defined <inline-formula><mml:math id="M411" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M412" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> coefficients. Because of this simplicity, it has the advantage of only needing to solve one set of equations. Furthermore, if the farm power and thrust coefficients are known accurately, it has the lowest RMSE out of all the models tested. However, these coefficients, along with the tuned parameter <inline-formula><mml:math id="M413" display="inline"><mml:mi mathvariant="italic">α</mml:mi></mml:math></inline-formula>, are not trivial to estimate. Its applicability is therefore partially dependent on future work developing methods for calculating the <inline-formula><mml:math id="M414" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M415" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> of real wind farms. Progress on estimating these coefficients would also benefit actuator-farm and mesoscale weather models <xref ref-type="bibr" rid="bib1.bibx35 bib1.bibx11" id="paren.84"/>.</p>
      <p id="d2e10407">The farm-averaged PB method captures turbine-scale effects through repeated wake model evaluations. As shown in the previous section, this can be interpreted as correcting neutral-flow results for gravity-wave effects. This approach avoids the need for pre-calculating farm coefficients, making it easier to set up for different farm layouts and operating setpoints. However, it comes at the cost of having to solve an extra set of equations.</p>
</sec>
<sec id="Ch1.S7" sec-type="conclusions">
  <label>7</label><title>Conclusions</title>
      <p id="d2e10418">Two-scale momentum (2SM) theory offers a new way of including mesoscale effects in wind farm power predictions. Using results from <xref ref-type="bibr" rid="bib1.bibx19" id="text.85"/>, its core hypothesis of scale separation between the turbine- and farm-scale flows was validated for steady-state idealized CNBL conditions. Building on that, this paper developed two ways to combine it with an atmospheric perturbation model (APM) to speed up farm power calculations. This speed-up is possible because according to the 2SM hypothesis, only the average velocity within a farm <inline-formula><mml:math id="M416" display="inline"><mml:mrow><mml:msub><mml:mi>U</mml:mi><mml:mrow><mml:mspace linebreak="nobreak" width="-0.125em"/><mml:mi mathvariant="normal">f</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula> is needed to compute the farm's force and power output. Since local fluctuations in the turbine forcing are therefore not relevant, the spatial distribution of the forces is assumed to be constant and independent of the APM state. Combined with the linearity of the APM, this transforms the APM's governing equations into a system of algebraic equations, which are trivial to solve numerically.</p>
      <p id="d2e10437">The first version of the 2SM–APM uses the APM as a mesoscale model to obtain the average velocity within the farm, and uses the 2SM farm coefficients <inline-formula><mml:math id="M417" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M418" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> to calculate the turbine forces and power output. This basic coupling method is similar to the velocity matching method developed in by <xref ref-type="bibr" rid="bib1.bibx8" id="text.86"/>, as it matches the average farm velocity between the meso- and turbine scales. In a validation campaign using the LES results from <xref ref-type="bibr" rid="bib1.bibx19" id="text.87"/>, its RMSE on farm power was only 3 % of the installed capacity when using LES-based values for <inline-formula><mml:math id="M419" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M420" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula>. Furthermore, we addressed the model's reliance on accurately knowing these coefficients by showing how they could in principle be calculated using a standard wake model.</p>
      <p id="d2e10499">The second model developed in this paper avoids the explicit use of <inline-formula><mml:math id="M421" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M422" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> by coupling directly to a wake model. To this end, it applies the farm-averaging approach to the pressure-based wake model coupling method of <xref ref-type="bibr" rid="bib1.bibx31" id="text.88"/>. Based on a rigorous derivation of this method, we show that it can also be interpreted as a model to correct neutral flow results for CNBL effects. The model performs similarly to the first model and can predict the effects of blockage and gravity waves on farm power, albeit with a consistent bias.</p>
      <p id="d2e10531">Both the 2SM–APM model and the farm-averaged pressure-based approach have a similar accuracy and computational cost. Which of them is preferable largely depends on the trade-off between efficiently calculating the farm coefficients <inline-formula><mml:math id="M423" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M424" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> for the former and solving additional equations for the latter. This investigation is left as future work.</p>
      <p id="d2e10561">Both models achieve a lower variance in the error of their power predictions compared to a standalone wake model, at the cost of introducing a larger bias. Tuning the underlying wake models could compensate for this, as both the wake spreading coefficients and the farm power and thrust coefficients <inline-formula><mml:math id="M425" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M426" display="inline"><mml:mrow><mml:msubsup><mml:mi>C</mml:mi><mml:mi mathvariant="normal">p</mml:mi><mml:mo>*</mml:mo></mml:msubsup></mml:mrow></mml:math></inline-formula> can be modified. The low computational cost of the models allows this tuning to directly incorporate large operational datasets. We believe this is a promising area of further research that will facilitate the model's adoption by the wind energy industry and community.</p>
      <p id="d2e10590">As powerful as the tools offered by the 2SM approach are, it remains unclear how valid the underlying assumptions are outside of the idealized conditions they have been tested in. So far, they have only been validated using LES of infinite farms <xref ref-type="bibr" rid="bib1.bibx13" id="paren.89"/> and farms operating in horizontally homogeneous background conditions with uniform turbine layouts and constant thrust coefficients <xref ref-type="bibr" rid="bib1.bibx14" id="paren.90"/>. Future work should investigate whether the farm coefficients remain constant under realistic and transient atmospheric conditions and how to account for irregular turbine spacing and operating setpoints.</p>
</sec>

      
      </body>
    <back><notes notes-type="codeavailability"><title>Code availability</title>

      <p id="d2e10603">The models developed in this paper have been implemented in the open-source code WAYVE <ext-link xlink:href="https://doi.org/10.48804/XMNVVY" ext-link-type="DOI">10.48804/XMNVVY</ext-link> <xref ref-type="bibr" rid="bib1.bibx7" id="paren.91"/>.</p>
  </notes><notes notes-type="dataavailability"><title>Data availability</title>

      <p id="d2e10615">All data used in this paper is publicly available. The LES data can be found at  <ext-link xlink:href="https://doi.org/10.48804/L45LTT" ext-link-type="DOI">10.48804/L45LTT</ext-link> (Lanzilao and Meyers, 2023b). Additionally, a repository containing post-processed data to be used directly as input for WAYVE can be found at <ext-link xlink:href="https://doi.org/10.48804/WL73U5" ext-link-type="DOI">10.48804/WL73U5</ext-link> <xref ref-type="bibr" rid="bib1.bibx10" id="paren.92"/>.</p>
  </notes><notes notes-type="authorcontribution"><title>Author contributions</title>

      <p id="d2e10630">KD and JM jointly derived the 2SM–APM couplings and set up the validation studies. KD performed code implementations and carried out the simulations. KD and JM jointly wrote the article.</p>
  </notes><notes notes-type="competinginterests"><title>Competing interests</title>

      <p id="d2e10636">At least one of the (co-)authors is a member of the editorial board of <italic>Wind Energy Science</italic>. The peer-review process was guided by an independent editor, and the authors also have no other competing interests to declare.</p>
  </notes><notes notes-type="disclaimer"><title>Disclaimer</title>

      <p id="d2e10645">Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. The authors bear the ultimate responsibility for providing appropriate place names. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.</p>
  </notes><ack><title>Acknowledgements</title><p id="d2e10652">The authors thank Frederik Aerts for insightful discussions on combining the 2SM and APM approaches.</p></ack><notes notes-type="financialsupport"><title>Financial support</title>

      <p id="d2e10657">This research has been supported by the Energy Transition Fund of the Belgian Federal Public Service for Economy, SMEs, and Energy (FOD Economie, K.M.O., Middenstand en Energie), by the European Union Horizon Europe Framework programme (HORIZON-CL5-2021-D3-03-04) (grant no. 101084205), and by Project Cloud4Wake, funded by Vlaamse Agentschap Innoveren &amp; Ondernemen (VLAIO) under the Blue Cluster cSBO programme (grant no. HBC.2022.0549).</p>
  </notes><notes notes-type="reviewstatement"><title>Review statement</title>

      <p id="d2e10663">This paper was edited by Sandrine Aubrun and reviewed by three anonymous referees.</p>
  </notes><ref-list>
    <title>References</title>

      <ref id="bib1.bibx1"><label>Allaerts and Meyers(2017)</label><mixed-citation>Allaerts, D. and Meyers, J.: Boundary-layer development and gravity waves in conventionally neutral wind farms, J. Fluid Mech., 814, 95–130, <ext-link xlink:href="https://doi.org/10.1017/jfm.2017.11" ext-link-type="DOI">10.1017/jfm.2017.11</ext-link>, 2017.</mixed-citation></ref>
      <ref id="bib1.bibx2"><label>Allaerts and Meyers(2018)</label><mixed-citation>Allaerts, D. and Meyers, J.: Gravity Waves and Wind-Farm Efficiency in Neutral and Stable Conditions, Bound.-Lay. Meteorol., 166, 269–299, <ext-link xlink:href="https://doi.org/10.1007/s10546-017-0307-5" ext-link-type="DOI">10.1007/s10546-017-0307-5</ext-link>, 2018.</mixed-citation></ref>
      <ref id="bib1.bibx3"><label>Allaerts and Meyers(2019)</label><mixed-citation>Allaerts, D. and Meyers, J.: Sensitivity and feedback of wind-farm-induced gravity waves, J. Fluid Mech., 862, 990–1028, <ext-link xlink:href="https://doi.org/10.1017/jfm.2018.969" ext-link-type="DOI">10.1017/jfm.2018.969</ext-link>, 2019.</mixed-citation></ref>
      <ref id="bib1.bibx4"><label>Bastankhah and Porté-Agel(2014)</label><mixed-citation>Bastankhah, M. and Porté-Agel, F.: A new analytical model for wind-turbine wakes, Renew. Energ., 70, 116–123, <ext-link xlink:href="https://doi.org/10.1016/j.renene.2014.01.002" ext-link-type="DOI">10.1016/j.renene.2014.01.002</ext-link>, 2014.</mixed-citation></ref>
      <ref id="bib1.bibx5"><label>Bleeg et al.(2018)Bleeg, Purcell, Ruisi, and Traiger</label><mixed-citation>Bleeg, J., Purcell, M., Ruisi, R., and Traiger, E.: Wind farm blockage and the consequences of neglecting its impact on energy production, Energies, 11, <ext-link xlink:href="https://doi.org/10.3390/en11061609" ext-link-type="DOI">10.3390/en11061609</ext-link>, 2018.</mixed-citation></ref>
      <ref id="bib1.bibx6"><label>Centurelli et al.(2021)Centurelli, Vollmer, Schmidt, Dörenkämper, Schröder, Lukassen, and Peinke</label><mixed-citation>Centurelli, G., Vollmer, L., Schmidt, J., Dörenkämper, M., Schröder, M., Lukassen, L. J., and Peinke, J.: Evaluating Global Blockage engineering parametrizations with LES, J. Phys. Conf. Ser., 1934, 012021, <ext-link xlink:href="https://doi.org/10.1088/1742-6596/1934/1/012021" ext-link-type="DOI">10.1088/1742-6596/1934/1/012021</ext-link>, 2021.</mixed-citation></ref>
      <ref id="bib1.bibx7"><label>Devesse et al.(2023)Devesse, Lanzilao, Allaerts, Jamaer, and Meyers</label><mixed-citation>Devesse, K., Lanzilao, L., Allaerts, D., Jamaer, S., and Meyers, J.: Wind-fArm gravitY-waVe and blockagE code, GitLab [code], <ext-link xlink:href="https://doi.org/10.48804/XMNVVY" ext-link-type="DOI">10.48804/XMNVVY</ext-link>, 2023.</mixed-citation></ref>
      <ref id="bib1.bibx8"><label>Devesse et al.(2024a)Devesse, Lanzilao, and Meyers</label><mixed-citation>Devesse, K., Lanzilao, L., and Meyers, J.: A meso–micro atmospheric perturbation model for wind farm blockage, J. Fluid Mech., 998, A63, <ext-link xlink:href="https://doi.org/10.1017/jfm.2024.868" ext-link-type="DOI">10.1017/jfm.2024.868</ext-link>, 2024a.</mixed-citation></ref>
      <ref id="bib1.bibx9"><label>Devesse et al.(2024b)Devesse, Stipa, Brinkerhoff, Allaerts, and Meyers</label><mixed-citation>Devesse, K., Stipa, S., Brinkerhoff, J., Allaerts, D., and Meyers, J.: Comparing methods for coupling wake models to an atmospheric perturbation model in WAYVE, J. Phys. Conf. Ser., 2767, 092079, <ext-link xlink:href="https://doi.org/10.1088/1742-6596/2767/9/092079" ext-link-type="DOI">10.1088/1742-6596/2767/9/092079</ext-link>, 2024b.</mixed-citation></ref>
      <ref id="bib1.bibx10"><label>Devesse and Meyers(2026)</label><mixed-citation>Devesse, K. and Meyers, J.: Post-Processed Large-Eddy Simulation Data for Validating Atmospheric Perturbation Models, KU Leuven RDR [data set], <ext-link xlink:href="https://doi.org/10.48804/WL73U5" ext-link-type="DOI">10.48804/WL73U5</ext-link>, 2026.</mixed-citation></ref>
      <ref id="bib1.bibx11"><label>Du et al.(2025)Du, Ge, Li, and Liu</label><mixed-citation>Du, B., Ge, M., Li, X., and Liu, Y.: A Meso-Microscale Coupled Wind Farm Parameterization, Bound.-Lay. Meteorol., 191, 35, <ext-link xlink:href="https://doi.org/10.1007/s10546-025-00928-7" ext-link-type="DOI">10.1007/s10546-025-00928-7</ext-link>, 2025.</mixed-citation></ref>
      <ref id="bib1.bibx12"><label>Gill(1982)</label><mixed-citation> Gill, A. E.: Atmosphere-Ocean Dynamics, International geophysics series 30, Academic Press, San Diego, USA, ISBN 0122835204, 1982.</mixed-citation></ref>
      <ref id="bib1.bibx13"><label>Kirby et al.(2022)Kirby, Nishino, and Dunstan</label><mixed-citation>Kirby, A., Nishino, T., and Dunstan, T. D.: Two-scale interaction of wake and blockage effects in large wind farms, J. Fluid Mech., 953, 1–28, <ext-link xlink:href="https://doi.org/10.1017/jfm.2022.979" ext-link-type="DOI">10.1017/jfm.2022.979</ext-link>, 2022.</mixed-citation></ref>
      <ref id="bib1.bibx14"><label>Kirby et al.(2025)Kirby, Nishino, Lanzilao, Dunstan, and Meyers</label><mixed-citation>Kirby, A., Nishino, T., Lanzilao, L., Dunstan, T. D., and Meyers, J.: Turbine- and farm-scale power losses in wind farms: an alternative to wake and farm blockage losses, Wind Energ. Sci., 10, 435–450, <ext-link xlink:href="https://doi.org/10.5194/wes-10-435-2025" ext-link-type="DOI">10.5194/wes-10-435-2025</ext-link>, 2025.</mixed-citation></ref>
      <ref id="bib1.bibx15"><label>Lanzilao and Meyers(2021)</label><mixed-citation>Lanzilao, L. and Meyers, J.: Set-point optimization in wind farms to mitigate effects of flow blockage induced by atmospheric gravity waves, Wind Energ. Sci., 6, 247–271, <ext-link xlink:href="https://doi.org/10.5194/wes-6-247-2021" ext-link-type="DOI">10.5194/wes-6-247-2021</ext-link>, 2021.</mixed-citation></ref>
      <ref id="bib1.bibx16"><label>Lanzilao and Meyers(2022)</label><mixed-citation>Lanzilao, L. and Meyers, J.: Effects of self-induced gravity waves on finite wind-farm operations using a large-eddy simulation framework, J. Phys. Conf. Ser., 2265, <ext-link xlink:href="https://doi.org/10.1088/1742-6596/2265/2/022043" ext-link-type="DOI">10.1088/1742-6596/2265/2/022043</ext-link>, 2022.</mixed-citation></ref>
      <ref id="bib1.bibx17"><label>Lanzilao and Meyers(2023a)</label><mixed-citation>Lanzilao, L. and Meyers, J.: An Improved Fringe-Region Technique for the Representation of Gravity Waves in Large Eddy Simulation with Application to Wind Farms, Bound.-Lay. Meteorol., 186, 567–593, <ext-link xlink:href="https://doi.org/10.1007/s10546-022-00772-z" ext-link-type="DOI">10.1007/s10546-022-00772-z</ext-link>, 2023a.</mixed-citation></ref>
      <ref id="bib1.bibx18"><label>Lanzilao and Meyers(2023b)</label><mixed-citation>Lanzilao, L. and Meyers, J.: A reference database of wind-farm large-eddy simulations for parametrizing effects of blockage and gravity waves, KU Leuven RDR [data set], <ext-link xlink:href="https://doi.org/10.48804/L45LTT" ext-link-type="DOI">10.48804/L45LTT</ext-link>, 2023b.</mixed-citation></ref>
      <ref id="bib1.bibx19"><label>Lanzilao and Meyers(2024)</label><mixed-citation>Lanzilao, L. and Meyers, J.: A parametric large-eddy simulation study of wind-farm blockage and gravity waves in conventionally neutral boundary layers, J. Fluid Mech., 979, A54, <ext-link xlink:href="https://doi.org/10.1017/jfm.2023.1088" ext-link-type="DOI">10.1017/jfm.2023.1088</ext-link>, 2024.</mixed-citation></ref>
      <ref id="bib1.bibx20"><label>Maas(2023a)</label><mixed-citation>Maas, O.: From gigawatt to multi-gigawatt wind farms: wake effects, energy budgets and inertial gravity waves investigated by large-eddy simulations, Wind Energ. Sci., 8, 535–556, <ext-link xlink:href="https://doi.org/10.5194/wes-8-535-2023" ext-link-type="DOI">10.5194/wes-8-535-2023</ext-link>, 2023a.</mixed-citation></ref>
      <ref id="bib1.bibx21"><label>Maas(2023b)</label><mixed-citation>Maas, O.: Large-eddy simulation of a 15 GW wind farm: Flow effects, energy budgets and comparison with wake models, Frontiers in Mechanical Engineering, 9, 1–23, <ext-link xlink:href="https://doi.org/10.3389/fmech.2023.1108180" ext-link-type="DOI">10.3389/fmech.2023.1108180</ext-link>, 2023b.</mixed-citation></ref>
      <ref id="bib1.bibx22"><label>Meyers et al.(2022)Meyers, Bottasso, Dykes, Fleming, Gebraad, Giebel, Göçmen, and van Wingerden</label><mixed-citation>Meyers, J., Bottasso, C., Dykes, K., Fleming, P., Gebraad, P., Giebel, G., Göçmen, T., and van Wingerden, J.-W.: Wind farm flow control: prospects and challenges, Wind Energ. Sci., 7, 2271–2306, <ext-link xlink:href="https://doi.org/10.5194/wes-7-2271-2022" ext-link-type="DOI">10.5194/wes-7-2271-2022</ext-link>, 2022.</mixed-citation></ref>
      <ref id="bib1.bibx23"><label>Niayifar and Porté-Agel(2016)</label><mixed-citation>Niayifar, A. and Porté-Agel, F.: Analytical Modeling of Wind Farms: A New Approach for Power Prediction, Energies, 9, 741, <ext-link xlink:href="https://doi.org/10.3390/en9090741" ext-link-type="DOI">10.3390/en9090741</ext-link>, 2016.</mixed-citation></ref>
      <ref id="bib1.bibx24"><label>Nishino and Dunstan(2020)</label><mixed-citation>Nishino, T. and Dunstan, T. D.: Two-scale momentum theory for time-dependent modelling of large wind farms, J. Fluid Mech., 894, <ext-link xlink:href="https://doi.org/10.1017/jfm.2020.252" ext-link-type="DOI">10.1017/jfm.2020.252</ext-link>, 2020.</mixed-citation></ref>
      <ref id="bib1.bibx25"><label>Ouro et al.(2025)Ouro, Ghobrial, Ali, and Stallard</label><mixed-citation>Ouro, P., Ghobrial, M., Ali, K., and Stallard, T.: Numerical modelling of offshore wind-farm cluster wakes, Renew. Sust. Energ. Rev., 215, 115526, <ext-link xlink:href="https://doi.org/10.1016/j.rser.2025.115526" ext-link-type="DOI">10.1016/j.rser.2025.115526</ext-link>, 2025.</mixed-citation></ref>
      <ref id="bib1.bibx26"><label>Porté-Agel et al.(2020)Porté-Agel, Bastankhah, and Shamsoddin</label><mixed-citation>Porté-Agel, F., Bastankhah, M., and Shamsoddin, S.: Wind-Turbine and Wind-Farm Flows: A Review, Bound.-Lay. Meteorol., 174, 1–59, <ext-link xlink:href="https://doi.org/10.1007/s10546-019-00473-0" ext-link-type="DOI">10.1007/s10546-019-00473-0</ext-link>, 2020.</mixed-citation></ref>
      <ref id="bib1.bibx27"><label>Shapiro et al.(2019)Shapiro, Gayme, and Meneveau</label><mixed-citation>Shapiro, C. R., Gayme, D. F., and Meneveau, C.: Filtered actuator disks : Theory and application to wind turbine models in large eddy simulation, Wind Energy, 1414–1420, <ext-link xlink:href="https://doi.org/10.1002/we.2376" ext-link-type="DOI">10.1002/we.2376</ext-link>, 2019.</mixed-citation></ref>
      <ref id="bib1.bibx28"><label>Smith(1980)</label><mixed-citation>Smith, R. B.: Linear theory of stratified hydrostatic flow past an isolated mountain., Tellus, 32, 348–364, <ext-link xlink:href="https://doi.org/10.3402/tellusa.v32i4.10590" ext-link-type="DOI">10.3402/tellusa.v32i4.10590</ext-link>, 1980.</mixed-citation></ref>
      <ref id="bib1.bibx29"><label>Smith(2010)</label><mixed-citation>Smith, R. B.: Gravity wave effects on wind farm efficiency, Wind Energy, 13, 449–458, <ext-link xlink:href="https://doi.org/10.1002/we.366" ext-link-type="DOI">10.1002/we.366</ext-link>, 2010.</mixed-citation></ref>
      <ref id="bib1.bibx30"><label>Smith(2024)</label><mixed-citation>Smith, R. B.: The wind farm pressure field, Wind Energ. Sci., 9, 253–261, <ext-link xlink:href="https://doi.org/10.5194/wes-9-253-2024" ext-link-type="DOI">10.5194/wes-9-253-2024</ext-link>, 2024.</mixed-citation></ref>
      <ref id="bib1.bibx31"><label>Stipa et al.(2024a)Stipa, Ajay, Allaerts, and Brinkerhoff</label><mixed-citation>Stipa, S., Ajay, A., Allaerts, D., and Brinkerhoff, J.: The multi-scale coupled model: a new framework capturing wind farm–atmosphere interaction and global blockage effects, Wind Energ. Sci., 9, 1123–1152, <ext-link xlink:href="https://doi.org/10.5194/wes-9-1123-2024" ext-link-type="DOI">10.5194/wes-9-1123-2024</ext-link>, 2024a.</mixed-citation></ref>
      <ref id="bib1.bibx32"><label>Stipa et al.(2024b)Stipa, Ajay, Allaerts, and Brinkerhoff</label><mixed-citation>Stipa, S., Ajay, A., Allaerts, D., and Brinkerhoff, J.: TOSCA – an open-source, finite-volume, large-eddy simulation (LES) environment for wind farm flows, Wind Energ. Sci., 9, 297–320, <ext-link xlink:href="https://doi.org/10.5194/wes-9-297-2024" ext-link-type="DOI">10.5194/wes-9-297-2024</ext-link>, 2024b.</mixed-citation></ref>
      <ref id="bib1.bibx33"><label>Stipa et al.(2024c)Stipa, Ajay, and Brinkerhoff</label><mixed-citation>Stipa, S., Ajay, A., and Brinkerhoff, J.: The actuator farm model for large eddy simulation (LES) of wind-farm-induced atmospheric gravity waves and farm–farm interaction, Wind Energ. Sci., 9, 2301–2332, <ext-link xlink:href="https://doi.org/10.5194/wes-9-2301-2024" ext-link-type="DOI">10.5194/wes-9-2301-2024</ext-link>, 2024c.</mixed-citation></ref>
      <ref id="bib1.bibx34"><label>Troldborg and Meyer Forsting(2017)</label><mixed-citation>Troldborg, N. and Meyer Forsting, A. R.: A simple model of the wind turbine induction zone derived from numerical simulations, Wind Energy, 20, 2011–2020, <ext-link xlink:href="https://doi.org/10.1002/we.2137" ext-link-type="DOI">10.1002/we.2137</ext-link>, 2017.</mixed-citation></ref>
      <ref id="bib1.bibx35"><label>van der Laan et al.(2023)van der Laan, García-Santiago, Kelly, Meyer Forsting, Dubreuil-Boisclair, Sponheim Seim, Imberger, Peña, Sørensen, and Réthoré</label><mixed-citation>van der Laan, M. P., García-Santiago, O., Kelly, M., Meyer Forsting, A., Dubreuil-Boisclair, C., Sponheim Seim, K., Imberger, M., Peña, A., Sørensen, N. N., and Réthoré, P.-E.: A new RANS-based wind farm parameterization and inflow model for wind farm cluster modeling, Wind Energ. Sci., 8, 819–848, <ext-link xlink:href="https://doi.org/10.5194/wes-8-819-2023" ext-link-type="DOI">10.5194/wes-8-819-2023</ext-link>, 2023.</mixed-citation></ref>

  </ref-list></back>
    <!--<article-title-html>Fast models for blockage and gravity-wave effects on wind-farm power output</article-title-html>
<abstract-html/>
<ref-html id="bib1.bib1"><label>Allaerts and Meyers(2017)</label><mixed-citation>
      
Allaerts, D. and Meyers, J.: Boundary-layer development and gravity waves in
conventionally neutral wind farms, J. Fluid Mech., 814, 95–130,
<a href="https://doi.org/10.1017/jfm.2017.11" target="_blank">https://doi.org/10.1017/jfm.2017.11</a>, 2017.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib2"><label>Allaerts and Meyers(2018)</label><mixed-citation>
      
Allaerts, D. and Meyers, J.: Gravity Waves and Wind-Farm Efficiency in Neutral
and Stable Conditions, Bound.-Lay. Meteorol., 166, 269–299,
<a href="https://doi.org/10.1007/s10546-017-0307-5" target="_blank">https://doi.org/10.1007/s10546-017-0307-5</a>, 2018.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib3"><label>Allaerts and Meyers(2019)</label><mixed-citation>
      
Allaerts, D. and Meyers, J.: Sensitivity and feedback of wind-farm-induced
gravity waves, J. Fluid Mech., 862, 990–1028,
<a href="https://doi.org/10.1017/jfm.2018.969" target="_blank">https://doi.org/10.1017/jfm.2018.969</a>, 2019.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib4"><label>Bastankhah and Porté-Agel(2014)</label><mixed-citation>
      
Bastankhah, M. and Porté-Agel, F.: A new analytical model for
wind-turbine wakes, Renew. Energ., 70, 116–123,
<a href="https://doi.org/10.1016/j.renene.2014.01.002" target="_blank">https://doi.org/10.1016/j.renene.2014.01.002</a>, 2014.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib5"><label>Bleeg et al.(2018)Bleeg, Purcell, Ruisi, and Traiger</label><mixed-citation>
      
Bleeg, J., Purcell, M., Ruisi, R., and Traiger, E.: Wind farm blockage and the
consequences of neglecting its impact on energy production, Energies, 11,
<a href="https://doi.org/10.3390/en11061609" target="_blank">https://doi.org/10.3390/en11061609</a>, 2018.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib6"><label>Centurelli et al.(2021)Centurelli, Vollmer, Schmidt,
Dörenkämper, Schröder, Lukassen, and
Peinke</label><mixed-citation>
      
Centurelli, G., Vollmer, L., Schmidt, J., Dörenkämper, M.,
Schröder, M., Lukassen, L. J., and Peinke, J.: Evaluating Global
Blockage engineering parametrizations with LES, J. Phys.
Conf. Ser., 1934, 012021, <a href="https://doi.org/10.1088/1742-6596/1934/1/012021" target="_blank">https://doi.org/10.1088/1742-6596/1934/1/012021</a>,
2021.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib7"><label>Devesse et al.(2023)Devesse, Lanzilao, Allaerts, Jamaer, and
Meyers</label><mixed-citation>
      
Devesse, K., Lanzilao, L., Allaerts, D., Jamaer, S., and Meyers, J.: Wind-fArm
gravitY-waVe and blockagE code, GitLab [code], <a href="https://doi.org/10.48804/XMNVVY" target="_blank">https://doi.org/10.48804/XMNVVY</a>, 2023.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib8"><label>Devesse et al.(2024a)Devesse, Lanzilao, and
Meyers</label><mixed-citation>
      
Devesse, K., Lanzilao, L., and Meyers, J.: A meso–micro atmospheric
perturbation model for wind farm blockage, J. Fluid Mech., 998,
A63, <a href="https://doi.org/10.1017/jfm.2024.868" target="_blank">https://doi.org/10.1017/jfm.2024.868</a>, 2024a.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib9"><label>Devesse et al.(2024b)Devesse, Stipa, Brinkerhoff,
Allaerts, and Meyers</label><mixed-citation>
      
Devesse, K., Stipa, S., Brinkerhoff, J., Allaerts, D., and Meyers, J.:
Comparing methods for coupling wake models to an atmospheric perturbation
model in WAYVE, J. Phys. Conf. Ser., 2767, 092079,
<a href="https://doi.org/10.1088/1742-6596/2767/9/092079" target="_blank">https://doi.org/10.1088/1742-6596/2767/9/092079</a>, 2024b.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib10"><label>Devesse and Meyers(2026)</label><mixed-citation>
      
Devesse, K. and Meyers, J.: Post-Processed Large-Eddy Simulation Data for Validating Atmospheric Perturbation Models, KU Leuven RDR [data set], <a href="https://doi.org/10.48804/WL73U5" target="_blank">https://doi.org/10.48804/WL73U5</a>, 2026.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib11"><label>Du et al.(2025)Du, Ge, Li, and Liu</label><mixed-citation>
      
Du, B., Ge, M., Li, X., and Liu, Y.: A Meso-Microscale Coupled Wind Farm
Parameterization, Bound.-Lay. Meteorol., 191, 35,
<a href="https://doi.org/10.1007/s10546-025-00928-7" target="_blank">https://doi.org/10.1007/s10546-025-00928-7</a>, 2025.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib12"><label>Gill(1982)</label><mixed-citation>
      
Gill, A. E.: Atmosphere-Ocean Dynamics, International geophysics series 30,
Academic Press, San Diego, USA, ISBN 0122835204, 1982.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib13"><label>Kirby et al.(2022)Kirby, Nishino, and Dunstan</label><mixed-citation>
      
Kirby, A., Nishino, T., and Dunstan, T. D.: Two-scale interaction of wake and
blockage effects in large wind farms, J. Fluid Mech., 953,
1–28, <a href="https://doi.org/10.1017/jfm.2022.979" target="_blank">https://doi.org/10.1017/jfm.2022.979</a>, 2022.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib14"><label>Kirby et al.(2025)Kirby, Nishino, Lanzilao, Dunstan, and
Meyers</label><mixed-citation>
      
Kirby, A., Nishino, T., Lanzilao, L., Dunstan, T. D., and Meyers, J.: Turbine- and farm-scale power losses in wind farms: an alternative to wake and farm blockage losses, Wind Energ. Sci., 10, 435–450, <a href="https://doi.org/10.5194/wes-10-435-2025" target="_blank">https://doi.org/10.5194/wes-10-435-2025</a>, 2025.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib15"><label>Lanzilao and Meyers(2021)</label><mixed-citation>
      
Lanzilao, L. and Meyers, J.: Set-point optimization in wind farms to mitigate effects of flow blockage induced by atmospheric gravity waves, Wind Energ. Sci., 6, 247–271, <a href="https://doi.org/10.5194/wes-6-247-2021" target="_blank">https://doi.org/10.5194/wes-6-247-2021</a>, 2021.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib16"><label>Lanzilao and Meyers(2022)</label><mixed-citation>
      
Lanzilao, L. and Meyers, J.: Effects of self-induced gravity waves on finite
wind-farm operations using a large-eddy simulation framework, J.
Phys. Conf. Ser., 2265, <a href="https://doi.org/10.1088/1742-6596/2265/2/022043" target="_blank">https://doi.org/10.1088/1742-6596/2265/2/022043</a>,
2022.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib17"><label>Lanzilao and Meyers(2023a)</label><mixed-citation>
      
Lanzilao, L. and Meyers, J.: An Improved Fringe-Region Technique for the
Representation of Gravity Waves in Large Eddy Simulation with Application to
Wind Farms, Bound.-Lay. Meteorol., 186, 567–593,
<a href="https://doi.org/10.1007/s10546-022-00772-z" target="_blank">https://doi.org/10.1007/s10546-022-00772-z</a>, 2023a.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib18"><label>Lanzilao and Meyers(2023b)</label><mixed-citation>
      
Lanzilao, L. and Meyers, J.: A reference database of wind-farm large-eddy
simulations for parametrizing effects of blockage and gravity waves, KU Leuven RDR [data set],
<a href="https://doi.org/10.48804/L45LTT" target="_blank">https://doi.org/10.48804/L45LTT</a>, 2023b.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib19"><label>Lanzilao and Meyers(2024)</label><mixed-citation>
      
Lanzilao, L. and Meyers, J.: A parametric large-eddy simulation study of
wind-farm blockage and gravity waves in conventionally neutral boundary
layers, J. Fluid Mech., 979, A54, <a href="https://doi.org/10.1017/jfm.2023.1088" target="_blank">https://doi.org/10.1017/jfm.2023.1088</a>,
2024.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib20"><label>Maas(2023a)</label><mixed-citation>
      
Maas, O.: From gigawatt to multi-gigawatt wind farms: wake effects, energy budgets and inertial gravity waves investigated by large-eddy simulations, Wind Energ. Sci., 8, 535–556, <a href="https://doi.org/10.5194/wes-8-535-2023" target="_blank">https://doi.org/10.5194/wes-8-535-2023</a>, 2023a.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib21"><label>Maas(2023b)</label><mixed-citation>
      
Maas, O.: Large-eddy simulation of a 15 GW wind farm: Flow effects, energy
budgets and comparison with wake models, Frontiers in Mechanical
Engineering, 9, 1–23, <a href="https://doi.org/10.3389/fmech.2023.1108180" target="_blank">https://doi.org/10.3389/fmech.2023.1108180</a>, 2023b.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib22"><label>Meyers et al.(2022)Meyers, Bottasso, Dykes, Fleming, Gebraad, Giebel,
Göçmen, and van Wingerden</label><mixed-citation>
      
Meyers, J., Bottasso, C., Dykes, K., Fleming, P., Gebraad, P., Giebel, G., Göçmen, T., and van Wingerden, J.-W.: Wind farm flow control: prospects and challenges, Wind Energ. Sci., 7, 2271–2306, <a href="https://doi.org/10.5194/wes-7-2271-2022" target="_blank">https://doi.org/10.5194/wes-7-2271-2022</a>, 2022.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib23"><label>Niayifar and Porté-Agel(2016)</label><mixed-citation>
      
Niayifar, A. and Porté-Agel, F.: Analytical Modeling of Wind Farms: A
New Approach for Power Prediction, Energies, 9, 741,
<a href="https://doi.org/10.3390/en9090741" target="_blank">https://doi.org/10.3390/en9090741</a>, 2016.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib24"><label>Nishino and Dunstan(2020)</label><mixed-citation>
      
Nishino, T. and Dunstan, T. D.: Two-scale momentum theory for time-dependent
modelling of large wind farms, J. Fluid Mech., 894,
<a href="https://doi.org/10.1017/jfm.2020.252" target="_blank">https://doi.org/10.1017/jfm.2020.252</a>, 2020.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib25"><label>Ouro et al.(2025)Ouro, Ghobrial, Ali, and Stallard</label><mixed-citation>
      
Ouro, P., Ghobrial, M., Ali, K., and Stallard, T.: Numerical modelling of
offshore wind-farm cluster wakes, Renew. Sust. Energ. Rev.,
215, 115526, <a href="https://doi.org/10.1016/j.rser.2025.115526" target="_blank">https://doi.org/10.1016/j.rser.2025.115526</a>, 2025.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib26"><label>Porté-Agel et al.(2020)Porté-Agel, Bastankhah, and
Shamsoddin</label><mixed-citation>
      
Porté-Agel, F., Bastankhah, M., and Shamsoddin, S.: Wind-Turbine and
Wind-Farm Flows: A Review, Bound.-Lay. Meteorol., 174, 1–59,
<a href="https://doi.org/10.1007/s10546-019-00473-0" target="_blank">https://doi.org/10.1007/s10546-019-00473-0</a>, 2020.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib27"><label>Shapiro et al.(2019)Shapiro, Gayme, and Meneveau</label><mixed-citation>
      
Shapiro, C. R., Gayme, D. F., and Meneveau, C.: Filtered actuator disks :
Theory and application to wind turbine models in large eddy simulation, Wind
Energy, 1414–1420, <a href="https://doi.org/10.1002/we.2376" target="_blank">https://doi.org/10.1002/we.2376</a>, 2019.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib28"><label>Smith(1980)</label><mixed-citation>
      
Smith, R. B.: Linear theory of stratified hydrostatic flow past an isolated
mountain., Tellus, 32, 348–364, <a href="https://doi.org/10.3402/tellusa.v32i4.10590" target="_blank">https://doi.org/10.3402/tellusa.v32i4.10590</a>, 1980.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib29"><label>Smith(2010)</label><mixed-citation>
      
Smith, R. B.: Gravity wave effects on wind farm efficiency, Wind Energy, 13,
449–458, <a href="https://doi.org/10.1002/we.366" target="_blank">https://doi.org/10.1002/we.366</a>, 2010.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib30"><label>Smith(2024)</label><mixed-citation>
      
Smith, R. B.: The wind farm pressure field, Wind Energ. Sci., 9, 253–261, <a href="https://doi.org/10.5194/wes-9-253-2024" target="_blank">https://doi.org/10.5194/wes-9-253-2024</a>, 2024.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib31"><label>Stipa et al.(2024a)Stipa, Ajay, Allaerts, and
Brinkerhoff</label><mixed-citation>
      
Stipa, S., Ajay, A., Allaerts, D., and Brinkerhoff, J.: The multi-scale coupled model: a new framework capturing wind farm–atmosphere interaction and global blockage effects, Wind Energ. Sci., 9, 1123–1152, <a href="https://doi.org/10.5194/wes-9-1123-2024" target="_blank">https://doi.org/10.5194/wes-9-1123-2024</a>, 2024a.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib32"><label>Stipa et al.(2024b)Stipa, Ajay, Allaerts, and
Brinkerhoff</label><mixed-citation>
      
Stipa, S., Ajay, A., Allaerts, D., and Brinkerhoff, J.: TOSCA – an open-source, finite-volume, large-eddy simulation (LES) environment for wind farm flows, Wind Energ. Sci., 9, 297–320, <a href="https://doi.org/10.5194/wes-9-297-2024" target="_blank">https://doi.org/10.5194/wes-9-297-2024</a>, 2024b.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib33"><label>Stipa et al.(2024c)Stipa, Ajay, and
Brinkerhoff</label><mixed-citation>
      
Stipa, S., Ajay, A., and Brinkerhoff, J.: The actuator farm model for large eddy simulation (LES) of wind-farm-induced atmospheric gravity waves and farm–farm interaction, Wind Energ. Sci., 9, 2301–2332, <a href="https://doi.org/10.5194/wes-9-2301-2024" target="_blank">https://doi.org/10.5194/wes-9-2301-2024</a>, 2024c.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib34"><label>Troldborg and Meyer Forsting(2017)</label><mixed-citation>
      
Troldborg, N. and Meyer Forsting, A. R.: A simple model of the wind turbine
induction zone derived from numerical simulations, Wind Energy, 20,
2011–2020, <a href="https://doi.org/10.1002/we.2137" target="_blank">https://doi.org/10.1002/we.2137</a>, 2017.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib35"><label>van der Laan et al.(2023)van der Laan, García-Santiago, Kelly,
Meyer Forsting, Dubreuil-Boisclair, Sponheim Seim, Imberger, Peña,
Sørensen, and Réthoré</label><mixed-citation>
      
van der Laan, M. P., García-Santiago, O., Kelly, M., Meyer Forsting, A., Dubreuil-Boisclair, C., Sponheim Seim, K., Imberger, M., Peña, A., Sørensen, N. N., and Réthoré, P.-E.: A new RANS-based wind farm parameterization and inflow model for wind farm cluster modeling, Wind Energ. Sci., 8, 819–848, <a href="https://doi.org/10.5194/wes-8-819-2023" target="_blank">https://doi.org/10.5194/wes-8-819-2023</a>, 2023.

    </mixed-citation></ref-html>--></article>
