Does within-biome drought sensitivity reflect patterns across biomes?

Abstract

Droughts are expected to increase in severity and frequency with climate change, and it is important to understand why some ecosystems are more sensitive to drought than others. Currently, there is considerable evidence that ecosystem sensitivity to drought, quantified by reductions in aboveground net primary production (ANPP), is negatively related to mean annual precipitation (MAP). Thus, arid ecosystems are more likely than mesic systems to experience dramatic reductions in productivity during drought. However, evidence for this pattern is primarily from studies that span multiple biomes making it difficult to discern if abiotic (MAP) or biotic factors (differences in plant communities) underlie this relationship. To disentangle these, we assessed patterns of drought sensitivity within a single biome, the semiarid shortgrass steppe of the western U.S., where the dominant vegetation varies minimally compared to cross-biomes studies. We used 23 years of satellite derived ANPP proxies, the Normalized Difference Vegetation Index (NDVI) and the Enhanced Vegetation Index (EVI, 157,929 pixels, 1 km² resolution) with gridded precipitation data to assess relationships between drought sensitivity and MAP within this biome. Consistent with previous multi-biome studies, we found a negative, although relatively weaker, relationship between MAP (250-625 mm) and drought sensitivity within the shortgrass steppe. We conclude that while differences in vegetation types almost certainly contribute to patterns of drought sensitivity, the long-term precipitation history of an ecosystem (e.g. MAP) may also play a role in determining differences in drought sensitivity within ecosystems.