Measuring broad scale responses of agro-ecosystems to climate change and CO2

Projections of broad scale ecosystem responses to climate change typically rely on up-scaling models developed at relatively fine scales. This represents a valuable approach but one that is subject to uncertainties in the scaling process that are not easily estimated. A complementary approach is to develop empirical models using observations made at the broad scales of interest. We present recent work that utilizes climate, satellite, and agricultural census data to estimate the response of national and global scale crop yields to changes in climate and CO₂. Global scale data exhibit surprisingly strong relationships between average temperatures and crop yields that are generally consistent with results from up-scaled modeling approaches. These relationships are then used to estimate the net impact of past climate trends on crop yields. The effects of CO₂ changes on yields are relatively poorly constrained with existing datasets. General strengths and limitations of these empirical approaches are discussed.