Global analysis of the relationships among a vegetation index, precipitation and land surface temperature

Abstract Abstract. This study is designed to examine the spatial variability of the relationships among global NDVI (Normalized Difference Vegetation Index) data, remotely-sensed land surface temperature data, and gridded station precipitation data as well as to investigate the potential for the combined use of NDVI and temperature data for global bioclimate monitoring. The relationships among the three variables are examined using single and multiple temporal correlations and the analysis is augmented by the computation of the first annual harmonic of each parameter. In addition, the global variability of growing season liming is analysed using a proxy for the onset and conclusion of the growing season, based upon slopes of the NDVI time series. The NDVI data set as processed for this study has significant sources of systematic error, which include aerosol and cloud contamination, orbital drift, and instrument degradation. This analysis provides insight into the manner in which the relationships among NDVI, precipitation and remotely-sensed land surface temperature vary geographically, in spite of the data noise. Due to excessive systemic error, anomalies of this NDVI data set are not highly correlated with precipitation, or multiply correlated with temperature to precipitation. Greater immediate promise for interannual bioclimate monitoring is contained in the proxies presented here for the growing season onset and length.