The panel approach with fixed effects and nonlinear weather effects has become a popular method to uncover weather impacts on economic outcomes, but its ability to capture long-run climatic adaptation remains unclear. Building upon a framework proposed by McIntosh and Schlenker (2006), this paper identifies empirical conditions under which the nonlinear panel approach can approximate a long-run response to climate. When these conditions fail, the obtained relationship may still be interpretable as a weighted average of underlying short-run and long-run responses. We use this decomposition to revisit recently published climate impact estimates. For spatially large panels, the estimated temperature-outcome relationship mostly reflects the long-run climatic response; this is not so for precipitation. We find some evidence of long-run climatic adaptation for crop yield outcomes in the United States and France.