Reply to Badger and Volker: Correctly estimating wind resources at large scales requires more than simple extrapolation.

Badger and Volker (1) claim that our paper (2) incorrectly estimates the wind energy resource and underestimates it bymore than 50%. However, a detailed reading of their work (3) reveals that their estimates are consistent with ours (and others; table 2 of ref. 2), yet their extrapolation to larger scales is flawed and so are their implications. In contrast to ref. 1, we used an atmospheric general circulation model (GCM) that explicitly simulates the dynamics of generating and dissipating kinetic energy associated with the large-scale atmospheric circulation to derive the wind resource. These dynamics are very well understood and documented, and our GCM reproduces these energetics (4–6). While some smaller-scale meteorological phenomena such as land–sea and mountain– valley breezes may be important for local wind resources and may not be represented in ref. 2, we are not aware of any studies that suggest these small-scale effects add substantial kinetic energy to the global atmosphere. Furthermore, the estimates by Badger and Volker (1) are consistent with ours (2, 7). However, these estimates cannot be extrapolated to larger scales, as in ref. 1, because regional wind-speed reductions are not accounted for, thus resulting in overestimates. They cited a generation of 0.69 MWe km −2 for a wind farm in the US Midwest with an installed capacity of 2.8 MWi km −2 (1, 3). This is essentially the same as our estimate in Kansas of 0.68 MWe km −2 for 2.5 MWi km −2 (7), which is substantially influenced by reduced wind speeds within the wind-farm region (7). It is also below the 0.8–1.1 MWe km −2 generation limit for this region (2, 7). This is an unusually windy region within the United States, and such a large wind farm would leave a substantial wake in the downwind region (7). Such estimates can thus not simply be extrapolated as six isolated wind farms and compared with the US electricity demand as done by Badger and Volker (1) because regional wind speed reductions are not accounted for. For the North Sea, we estimate a large-scale limit of 1.3–1.5 MWe km −2 (2), which is lower than the 2.5 MWe km −2 estimate for a 340-km wind farm with 6.4 MWi km −2 that Badger and Volker quote (1). When wind farms over 103,000 km (18% of the North Sea) are considered, this generation rate drops to 1.6 MWe km −2