The sensitivity of vegetation in the lower Tigris basin landscapes to regional and global climate variability

Abstract

This study investigates the lower Tigris basin's the normalized difference vegetation index (NDVI) sensitivity in 2000-2016 to regional climate variability reflected by the monthly precipitation and temperature time series of seven global datasets as well as to four global circulation indices. To examine the effect of climate variability on the different ecosystems, the study area has been classified into 10 smaller natural and anthropogenic landscapes based on landforms and land cover patterns. The preliminary analysis showed that the maximum biological productivity reflected by the NDVI of March and April has the highest correlation (0.5-0.8) to the same cumulative amounts of October-March period total precipitation and January-March period mean temperatures according to all datasets. In addition, this article showed there is a correlation between landscapes' NDVI and global modulation represented by the September-February state of El Nińo - Southern Oscillation (ENSO) (0.55-0.70) and December state of the dipole mode index (DMI) (0.35-0.72). The significant differences in the original precipitation and temperature levels according to the different datasets have urged the use of normalized time series: z-score of temperatures and analogous six-months the standardized precipitation index (SPI). However, the multiple correlation analysis showed that using ERA - Interim and NCEP-CFSR (MERRA-2) based climate factors can explain from 77 to 83% of the NDVI variability on lowland plains (on higher foothills and mountainous lands). We found also that these prediction percentages can be increased by 2.4-7.2% when using time series of precipitation and temperatures derived from different datasets, in addition to 4.5-9.5% increasing rates when using the global circulation indices as additional predictors.