Spatiotemporal Patterns of Production Can Be Used to Detect State Change Across an Arid Landscape

Methods to detect and quantify shifts in the state of ecosystems are increasingly important as global change drivers push more systems toward thresholds of change. Temporal relationships between precipitation and aboveground net primary production (ANPP) have been studied extensively in arid and semiarid ecosystems, but rarely has spatial variation in these relationships been investigated at a landscape scale, and rarely has such information been viewed as a resource for mapping the distribution of different ecological states. We examined the broad-scale effects of a shift from grassland to shrubland states on spatiotemporal patterns of remotely sensed ANPP proxies in the northern Chihuahuan Desert. We found that the normalized difference vegetation index (NDVI), when averaged across an eight-year period, did not vary significantly between these states, despite changes in ecosystem attributes likely to influence water availability to plants. In contrast, temporal relationships between precipitation and time-integrated NDVI (NDVI-I) modeled on a per-pixel basis were sensitive to spatial variation in shrub canopy cover, a key attribute differentiating ecological states in the region. The slope of the relationship between annual NDVI-I and 2-year cumulative precipitation was negatively related to, and accounted for 71% of variation in, shrub canopy cover estimated at validation sites using high spatial resolution satellite imagery. These results suggest that remote sensing studies of temporal precipitation–NDVI relationships may be useful for deriving shrub canopy cover estimates in the region, as well as for mapping other ecological state changes characterized by shifts in long-term ANPP, plant functional type dominance, or both.