Sea surface warming and ocean-to-atmosphere
feedback driven by large-scale offshore wind farms under seasonally stratified conditions

Hyodae Seo1,2*†, César Sauvage1,2†, Christoph Renkl2,3†,
Julie K. Lundquist4,5†, Anthony Kirincich2†
Offshore wind farms may induce changes in the upper ocean and near-surface atmosphere through coupled ocean- atmosphere feedbacks. Yet, the role of air-sea interactions mediated by offshore wind farms remains poorly under- stood. Using fully coupled ocean-atmosphere-wave model simulations for seasonally stratified conditions along the US East Coast, we show that simulated cumulative reductions in wind stress due to large-scale wind farm clusters lead to sea surface warming of 0.3° to 0.4°C and a shallower mixed layer. This warming drives upward heat fluxes, destabilizing the atmospheric boundary layer and enhancing wind stress, which partially offsets wake-induced wind deficits. These wake-ocean interactions influence near-surface meteorology and air-sea fluxes, suggesting that a coupled modeling approach may be necessary for assessing potential oceanographic impacts of offshore wind de- velopments. However, ocean coupling exerts limited influence on winds at turbine-relevant heights or within down- stream wakes, resulting in minimal impact on long-term energy. These findings suggest that models without ocean coupling may be adequate for wind energy applications.

INTRODUCTION
Offshore wind energy development in US coastal waters has been expanding steadily, with some commercial-scale wind farms now operational off the US East Coast (Fig. 1A). The scope of existing, planned, and proposed offshore wind projects highlights the need for fundamental research on how clusters of wind farms interact with both the meteorological and oceanographic conditions present within wind energy lease areas.
Wind turbines generate electricity by converting the kinetic en- ergy of the wind into electrical energy. Consequently, energy extrac- tions by turbines reduce wind speed and enhance turbulence (1-3). At both small- and large-scale deployments of offshore wind farms, these wind farm wake effects (4) are well documented, not only as reduced power generated within the wind farms (5) but also as altered near-surface meteorology and air-sea fluxes (6-10). These changes may drive oceanic and ecological responses (11, 12).
For example, in the seasonally stratified North Sea and German Bight, a modeling study by Christiansen et al. (13) showed that re- duced wind stress from the fixed-bottom wind farm clusters (7) suppresses vertical mixing in the upper ocean, leading to stronger stratification and a warming of the depth-averaged ocean tempera- ture by -0.1°C. These changes were shown to influence downstream ocean circulation and biogeochemical cycling (14). Similarly, studies along the California coast (15, 16) showed that changes in wind stress profiles associated with floating wind farms may alter wind-driven upwelling circulation, with potential impacts on nutrient delivery and coastal ecosystem dynamics. On the Mid-Atlantic shelf, Miles et al.
(17) reported that large wind farm clusters may affect nearshore

stratification and formation of the Cold Pool (18), a key subsurface water mass supporting regional fisheries and ecosystems.

Despite the first-order effect of wind farm wakes on the ocean, as examined in previous studies, the role of two-way wake-ocean inter- action mediated by offshore wind farms in driving oceanic and marine atmospheric boundary layer (MABL) responses remains poorly under- stood. Most wind wake studies rely on atmosphere-only (19, 20) or atmosphere-wave coupled models (9, 21, 22), where near-surface meteorology and air-sea fluxes do not respond to sea surface tem- perature (SST) changes induced by wind farm wake (6, 7, 10). This contrasts with the aforementioned ocean modeling studies that high- light the dynamic nature ofSST and upper-ocean variability to wake- driven wind forcing (16, 23).

Similarly, marine hydrodynamic and biogeochemical models typically use prescribed wind forcing, which may limit their ability to capture two-way wake-induced feedback effects on air-sea mo- mentum and heat fluxes—key drivers of ocean responses. While air-sea interaction has been generally recognized as important in offshore wind and oceanographic modeling (13, 24, 25), the extent to which ocean-atmosphere coupling is important and the specific conditions under which fully coupled ocean-atmosphere models are necessary remain unclear. Addressing these issues requires the use of fully coupled modeling systems that can resolve wake-induced two-way feedback.

In this study, we use high-resolution simulations with a fully coupled ocean-atmosphere-wave regional model to assess realistic scenarios involving large-scale, high-density, fixed-bottom offshore wind farms along the US East Coast. Focusing on existing wind en- ergy lease areas (Fig. 1A), the simulations evaluate turbine-induced wake effects using the Fitch wind farm parameterization (1) and quantify their potential impacts on upper-ocean processes, MABL dynamics, and energy production. We define the ocean coupling ef- fect as atmospheric responses to SST anomalies induced exclusively by wind farms. Therefore, our analysis focuses on regions near off- shore wind installations. While turbine foundations can generate

hydrodynamic wakes (23, 26, 27) and modify sea state and wind stress (28), such structural effects are not considered in this study.
We examine boreal summer, a period dominated by stable at- mospheric conditions and low background turbulence (20, 29, 30), during which wake effects are pronounced (31, 32). However, un- stable stratification also occurs frequently in summer, allowing assessment across different stability regimes. The study region covers the seasonally stratified shelf off the coasts of Massachusetts/ Rhode Island (MA/RI) and New Jersey (NJ), allowing for com- parison to previous modeling studies of similar oceanographic con- ditions (13, 15–17, 23).
Simulations include fully coupled cases with and without wind wake parameterization, as well as complementary atmosphere-only simulations that exclude ocean coupling (table S1). We also assess sensitivity to the empirical parameter, α, which governs turbulence kinetic energy (TKE) generation in the Fitch scheme (Materials and Methods) (1). The α parameter ranges from 0 to 1.0, reflecting no to full TKE input. While the influence of α on wake characteristics and energy production is well documented (10, 20, 33, 34), its impact on ocean responses remains not well understood. By varying α and accounting for different atmospheric stability regimes, we system- atically evaluate the sensitivity of upper-ocean responses across dif- ferent wake intensities.
We also quantify how wake-induced SST anomalies influence rotor layer and hub height winds, TKE, and energy production. Previous studies have shown that oceanic mesoscale SST anomalies on length scales of Θ (10 to 100 km) modulate air-sea heat and momentum flux- es, affecting wind shear, buoyancy, and MABL height (35, 36). Hence, warmer SSTs relative to overlying air—commonly observed over warm- core eddies or the Gulf Stream—enhance upward heat flux, deepen the MABL, and accelerate near-surface winds via stronger downward mo- mentum transport (37, 38). By comparing coupled and atmosphere- only simulations, we show that SST anomalies generated by offshore wind farms exert similar effects on atmospheric stability and turbulent mixing through anomalous heat flux into the atmosphere. Our findings suggest that this feedback mechanism may be important for modeling upper-ocean and near-surface atmospheric conditions near the wind farms, but it likely has a limited influence on hub-height winds and long-term energy production.