The climatic controls on dune mobility, especially the relative importance of wind strength, remain incompletely understood. This is a key research problem in semi-arid northern China, both for interpreting past dune activity as evidence of paleoclimate and for predicting future environmental change. Potential eolian sand transport, which is approximately proportional to wind power above the threshold for sand entrainment, has decreased across much of northern China since the 1970s. Over the same period, effective moisture (ratio of precipitation to potential evapotranspiration) has not changed significantly. This "natural experiment" provides insight on the relative importance of wind power as a control on dune mobility in three dunefields of northern China (Mu Us, Otindag, and Horqin), although poorly understood and potentially large effects of human land use complicate interpretation. Dune forms in these three regions are consistent with sand transport vectors inferred from weather station data, suggesting that wind directions have remained stable and the stations adequately represent winds that shaped the dunes. The predicted effect of weaker winds since the 1970s would be dune stabilization, with lower sand transport rates allowing vegetation cover to expand. Large portions of all three dunefields remained stabilized by vegetation in the 1970s despite high wind power. Since the 1970s, trends in remotely sensed vegetation greenness and change in mobile dune area inferred from sequential Landsat images do indicate widespread dune stabilization in the eastern Mu Us region. On the other hand, expansion of active dunes took place farther west in the Mu Us dunefield and especially in the central Otindag dunefield, with little overall change in two parts of the Horqin dunes. Better ground truth is needed to validate the remote sensing analyses, but results presented here place limits on the relative importance of wind strength as a control on dune mobility in the study areas. High wind power alone does not completely destabilize these dunes. A large decrease in wind power either has little short-term effect on the dunes, or more likely its effect is sufficiently small that it is obscured by human impacts on dune stability in many parts of the study areas. ?? 2008 Elsevier B.V. All rights reserved.