| 25 Aug 2026
WIX: a wakeness index for preliminary assessment of regional wind farm wake interactions
Abstract. The rapid clustering of large offshore wind farms in the North Sea and other coastal areas in the world raises concerns about compounded wake losses that propagate for tens of kilometres. Simulating these regional wakes with wind farm parameterisations in numerical weather prediction models is however computationally prohibitive for multi-scenario planning, as each change in layout or turbine specification demands a new simulation. We showcase the Wakeness Index (WIX), a geometry-based indicator that produces rapid, first-order estimates of climatological wake footprints from only three inputs: turbine coordinates and rotor diameters, wind direction frequency statistics, and a single tuneable decay factor df. The WIX is here calibrated against year-long simulations using the Weather Research and Forecasting (WRF) model with the Fitch wind farm parameterisation for current, and projected 2030 and 2050 North Sea scenarios, using the footprints' intersection-over-union (IoU) at the 92, 95, and 98 % wind speed recovery levels and the pixel-wise coefficient of determination R2 for the velocity deficit field. The footprint optimum over all scenarios (df = 2.3, mean IoU = 0.75) is adopted as the operating point. The future, higher-density layouts are reproduced skilfully and share common optima (IoU up to 0.87, R2 up to 0.92), demonstrating generalisability across comparable turbine populations. The WIX transfers to the more stable regime of the Baltic Sea without region-specific tuning, and wind direction frequencies from the New European Wind Atlas (NEWA) dataset can substitute the WRF-derived wind direction frequency statistics with near-perfect agreement, removing the dependency on output from localized simulations. Computing the WIX over the entire North Sea is reduced to a few seconds on 32 CPUs, against the CPU-weeks required by the WRF model, the WIX is well suited as a lightweight pre-feasibility screening tool for early-stage wind farm and marine spatial planning.
Competing interests: Two of the (co-)authors are members of the editorial board of Wind Energy Science.
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. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.
This paper proposes the Wind-farm Impact indeX (WIX) as a simplified method for estimating climatological wind-farm wake footprints. Developing a computationally inexpensive screening tool is potentially useful. However, I do not think the manuscript in its present form provides a sufficiently clear, well-defined, or independently validated methodology. The principal problem is that it is difficult to determine exactly what WIX is intended to represent. At different points in the paper, it is treated as an index, a surrogate for NWP wake fields, a model of wind-speed deficit, and a pre-feasibility planning tool. These are different objectives and require different standards of calibration and validation. The manuscript should more precisely define what quantity WIX predicts, and the scientific or engineering use claimed for it.
This ambiguity is reflected directly in the calibration procedure. The decay factor is optimized using two objectives: spatial overlap of wake footprints and pixel-wise agreement in wind-speed deficit. These yield substantially different optimal values for the decay factor. The authors then state that the appropriate value is application-dependent. This raises a basic question about the formulation: if the principal model parameter does not have a unique calibration, what precisely is the proposed WIX model? A model intended to predict wake extent is not necessarily the same model as one intended to predict deficit magnitude.
There is also a significant problem with the distinction between calibration and validation. WIX is calibrated against WRF-derived wind-speed-deficit fields, and WRF-derived wind-direction frequencies are initially used in constructing the index. Its performance is then assessed primarily by comparison with those same WRF-derived wake fields. Replacing the WRF wind-direction climatology with NEWA is a useful sensitivity test, but it does not provide independent validation of the wake-decay formulation itself. The manuscript therefore demonstrates that WIX can be fitted to reproduce features of the WRF-Fitch solution, but this is not equivalent to demonstrating that it predicts physical wind-farm wakes independently.
The claimed generalizability is also overstated. The authors argue that the similar optimum parameters obtained for the Y2030 and Y2050 scenarios demonstrate generalizability across layouts of comparable rotor-diameter range and density. These are related scenarios from the same modeling framework and geographical setting. Agreement between them is useful, but it is a limited test of transferability rather than a demonstration of general applicability. The Baltic Sea exercise is more useful in this respect because the North Sea calibration is transferred to another region. However, the method performs less well for smaller and less dense wind farms, and the manuscript attributes this largely to the number and density of turbines used to construct the WIX field. This appears to be more than a minor limitation. If the skill of the method depends strongly on turbine density, then this may be an inherent property of the WIX construction and should be treated as a fundamental restriction on its applicability.
I am also not convinced by the pre-feasibility demonstration using hypothetical wind-farm clusters. No independent wind-speed-deficit data exist for these cases, yet the resulting WIX fields are described as physically meaningful because their wake footprints align with the prevailing wind directions. Since wind-direction frequency is an explicit input to the formulation, such alignment is expected by construction. It does not demonstrate that the predicted wake extent or deficit magnitude is physically correct. More generally, the manuscript does not distinguish clearly enough between three different levels of agreement: reproducing the shape of a WRF wake footprint, reproducing the magnitude of the WRF wind-speed deficit, and providing a physically meaningful estimate for a new wind-farm layout where no reference calculation exists. The calibration results themselves show that the first two objectives do not yield the same optimal model parameter, yet the discussion tends to move between these objectives as though they were interchangeable.
There is also a more basic issue concerning the physical meaning of the reference quantity. The WRF wake deficit used for calibration is the annual-average difference between simulations with and without the Fitch wind-farm parameterization at a specified height. WIX is therefore constructed primarily as an empirical reduced representation of a particular mesoscale model response. That is not necessarily the same thing as a physical model of wind-farm wakes. The manuscript needs to state this distinction explicitly and limit its claims accordingly.
In its present form, the study appears to contain a potentially useful empirical screening concept, but the model objective, calibration strategy, validation logic, and domain of applicability are not sufficiently well-defined. The broader claims concerning generalizability and use for pre-feasibility assessment are not supported by the evidence presented. Resolving these issues would require more than editorial clarification. The paper would need a sharper definition of the quantity being modeled, a consistent calibration strategy, substantially more independent validation, and a clearer demonstration of when and why the method can be expected to work.