Patterns of floristic richness in vegetation communities of California: regional scale analysis with multi‐temporal NDVI

Abstract

ABSTRACTAim Applying water‐energy dynamics and heterogeneity theory to explain species richness via remote sensing could allow for the regional characterization and monitoring of vegetation community assemblages and their environment. We assess the relationship of multi‐temporal normalized difference vegetation index (NDVI) to plant species richness in vegetation communities.Location California, USA.Methods Sub‐regions containing species inventories for chaparral, coastal sage scrub, foothill woodland, and yellow pine forest communities were intersected with a vegetation community map and an AVHRR NDVI time series for 1990, 1991, 1992, 1995 and 1996. Principal components analysis reduced the AVHRR data to three variables representing the sum and temporal trajectories of NDVI within each community. A fourth variable representing heterogeneity was tested using the standard deviation of the first component. Quadratic forms of these variables were also tested. Species richness was analysed by stepwise regression.Results Chaparral, coastal sage scrub, and yellow pine forest had the best relationships between species richness and NDVI. Richness of chaparral was related to NDVI heterogeneity and spring greenness (r2 varied between 0.26 and 0.62 depending on year of NDVI data). Richness of coastal sage scrub was nonlinearly related to annual NDVI and heterogeneity (r2 0.63–0.81), with peak richness at intermediate values. Foothill woodland richness was related to heterogeneity in a monotonic curvilinear fashion (r2 0.28–0.35). Yellow pine forest richness was negatively related to spring greenness and positively related to heterogeneity (r2 0.40–0.46).Main Conclusions While NDVI's relationship to species richness varied, the selection of NDVI variables was generally consistent across years and indicated that spatial variability in NDVI may reflect important patterns in water‐energy use that affect plant species richness. The principal component axis that should correspond closely with annual mean NPP showed a less prominent role. We conclude that plant species richness for coarse vegetation associations can be characterized and monitored at a regional scale and over long periods of time using relatively coarse resolution NDVI data.