Spatial and temporal variation in terrestrial net primary productivity on Chinese Loess Plateau and its influential factors

Abstract: Monitoring dynamics of vegetation productivity and its influential factors are important for understanding the mechanisms of carbon cycling in terrestrial ecosystems. This study investigated the variation in vegetation net primary productivity (NPP) and its major influential factors on the Chinese Loess Plateau during 2000-2010. The remotely sensed time-series data MOD17A3 for vegetation collected by a MODIS sensor was analyzed for NPP changes. The climate factors (temperature, precipitation and solar radiation), palmer drought severity index (PDSI) and land use in the year of 2000-2010 were used to explore the effects of natural factors and human activities on NPP change. Linear least square regression was employed to estimate the trends of NPP in each grid point. The correlation between NPP change and various environmental factors was identified with Pearson coefficient. It was demonstrated that changes in NPP were related to factors such as climate, PDSI, crop area, crop yield and area of accumulative vegetation restoration. With multiple regression analysis, the contribution of human and natural forces to NPP shifts was evaluated separately for the whole Loess Plateau and several hot-spot cities within this region, respectively. The results showed that the NPP on the Loess Plateau was between 100-400 g C/(m2?a), and the average annual NPP during the 11 years was 238 g C/(m2?a). The NPP on the Loess Plateau increased by a rate of 4.57 g C/(m2?a) during the past 11 years. The total annual NPP increased from 119 Tg C in year 2000 to 144 Tg C in 2010. About 91% of the study area showed an increasing NPP trend but only 37% of the study area had a significant (P<0.05) increase in NPP including most areas of Shaanxi and Qinghai, southern Gansu and Ningxia. The increase of NPP varied with vegetation types. It was highest in shrub land (5.25 g C/(m2?a)) and crop land (5.26 g C/(m2?a)) but lowest in wet land (1.40 g C/(m2?a)). The areas with significant (P<0.05) increase in crop and grasses were mainly due to the vegetation restoration scheme. But, the areas with the increase of forest coverage appeared to be the smallest. Across the whole Loess Plateau, both natural and human factors were responsible for changes in NPP during the 11 years. However, accumulated area of farmland returned to forest or grassland, PDSI, farmland area and population tended to be dominant influential factors with the standard regression coefficients between NPP and these factors of 0.634, 0.587, 0.187 and 0.062, respectively. The relative contributions of these four factors were 43%, 40%, 13% and 4%, respectively. Land use change resulted from the central government program of Grain for Green may be the reason for the NPP increase in the regions with vegetation restoration (e.g. northern Shaanxi, southwestern Gansu). The relief of drought events (increased PDSI) during last 11 years might be a reason for the rise of NPP in Qinghai and Inner Mongolia. This study could provide valuable information for regional ecological resources management and ecosystem reconstruction on the Chinese Loess Plateau.