AbstractWe examined the response of the normalized difference vegetation index, integrated over the growing season (gNDVI) to mean precipitation, maximum temperature (Tmax), and minimum temperature (Tmin) over an 11‐year period (1990–2000) for six biomes in the conterminous United States. We focused on within‐ and across‐biome variance in long‐term average gNDVI, emphasizing the degree to which this variance is explained by spatial gradients in long‐term average seasonal climate. Since direct measurements of ecosystem function are unavailable at the spatial and temporal scales studied, we used the satellite‐based gNDVI as a proxy for net photosynthetic activity.Forested and nonforested biomes differed sharply in their response to spatial gradients in temperature and precipitation. Gradients in mean spring and fall precipitation totals explained much of the variance in mean annual gNDVI within arid biomes. For forested biomes, mean annual gNDVI was positively associated with mean annual and seasonal Tmin and Tmax. These trends highlight the importance of the seasonal components of precipitation and temperature regimes in controlling productivity, and reflect the influence of these climatic components on water balance and growing‐season length. According to the International Panel on Climate Change (IPCC) (2001) increases in temperature minima and fall precipitation have contributed the dominant components of US increases in temperature and precipitation, respectively. Within the range of conditions observed over the study region, our results suggest that these trends have particularly significant consequences for above‐ground plant productivity, especially for Grassland, Open Shrubland, and Evergreen Needleleaf Forest. If historical climatic trends and the biotic responses suggested in this analysis continue to hold, we can anticipate further increases in productivity for both forested and nonforested ecoregions in the conterminous US, with associated implications for carbon budgets and woody proliferation.
