Ecosystem Net Primary Production Responses to Changes in Precipitation Using an Annual Integrated MODIS EVI

Abstract

In this study, the relationship of above-ground net primary productivity (ANPP) with precipitation using the enhanced vegetation index (EVI) from satellite data as surrogate for ANPP was assessed. To use EVI as a proxy for ANPP we extracted the satellite data from areas with uniform vegetation in a 2x2 km area for the multi-site approach.In the multi-site analysis in the United States our results showed a strong exponential relationship between iEVI and annual precipitation across the sites and climate regimes studied. We found convergence of all sites toward common and maximum rain use efficiency under the water-limited conditions represented by the driest year at each site. Measures of inter-annual variability in iEVI with rainfall variation across biomes were similar to that reported by Knapp and Smith (2001) in which the more herbaceous dominant sites were found to be most sensitive to interannual variations in precipitation with no relationships found in woodland sites.The relationship was also evaluated in the southern hemisphere using a multi-site analysis with information from satellite TRMM for precipitation and MOD13Q1 from MODIS for EVI values at calendar and hydrologic year periods. The tested sites were located across the 6 major land cover types inAustralia, obtained from MODIS MCD12Q1 product and used to compare the relationship across different biomes. The results showed significant agreement between the annual iEVI and annual precipitation across the biomes involved in this study showing non-significant differences between the calendar and hydrologic years for the 24 sites across different climatic conditions.At the regional scale we also assessed the ANPP-precipitation relationship across all of Australia. Precipitation data from TRMM was obtained at 0.25x0.25 degrees spatial resolution and monthly temporal resolution and EVI values were obtained from the CGM (Climate Grid Modeling) MOD13C1-16-days and 5.6km temporal and spatial resolutions, respectively. Our results were in fair agreement with those from our first two studies and previous research and provided specific insights regarding the use iEVI as a proxy for productivity over extended regions as well as its combination with data sets from TRMM sensor for precipitation data.