Lecturer Central China Normal University, China (People's Republic)
Abstract: As one of the most vulnerable ecosystems, semi-arid grasslands provide ecosystem functions and services under the risk of climate change. While it remains uncertain about the mechanism of environmental factors influencing semi-arid grasslands’ productivity, a drier climatic trend with a higher frequency of extreme events is threatening these important ecosystems. In this study, we are trying to quantify the influence of water-balance on a semi-arid grassland’s temporal stability (TS) of productivity with remote sensing and modeling data.
Our study area is the typical semi-arid grassland located in Inner Mongolia, China. It is part of the largest steppe belt – the Eurasian steppe belt. Datasets used in this study include a model-tree-ensembled model to estimate gross primary productivity (MTE-GPP), remote sensing product (MOD13A1-EVI), gridded climate forcing data (CRU TS4.03), and a state-of-the-art drought index Standard Precipitation-Evapotranspiration Index (SPEI). We estimated the spatial distribution of TS during 2001-2015.The TS was calculated as the mean of productivity divided by the standard deviation (SD) of productivity. We used simple linear correlation to detect which operator, mean or SD, is more relevant to the TS. Partial correlation analysis was used to quantify the correlation between SPEI and productivity, despite the collinearity of temperature and precipitation.
As a result, the TS of GPP (TSGPP) illustrates a latitudinal zonality. TSGPP is increasing as the latitude is lower, from TS=5 at 50°N to TS=15 at 34°N. Contrastingly, the TS of EVI (TSEVI) shows a smaller variation. Most of the TSEVI values are close to 10, from 34°N to 50°N. Both TSGPP and TSEVI are significantly correlated to the SD of productivity, rather than the mean value (p = 0.00). Last but not least, both TSEVI and TSGPP are significantly correlated to SPEI, while the seasonal scale SPEI (SPEI03) is positively correlated with TSEVI (r=0.046, p=0) and TSGPP (r=0.214, p=0), but the annual scale SPEI (SPEI12) is negatively correlated with TSEVI (r=-0.237, p=0) and TSGPP (r=-0.376, p=0).
In conclusion, these results suggest that the semi-arid grassland in Inner Mongolia is sensitive to variations of water balance. The influence of drought at different timescale might be diverse. These results will support vegetation dynamical modeling and rangeland management.
