Spatio-temporal evolution and driving forces of rainfall use efficiency in land restoration

Returning farmland to forest (grass), named Grain for Green Project, is one of the major ecological land restoration in China. In the context of the global climate change, the study of the impact of precipitation patterns on the productivity of ecosystems become an important means to evaluate the use efficiency of returning farmland to forests (grass) for the ecological restoration. Baoji region was used to this research in order to provide a specific theoretical reference for the improvement and implementation of the subsequent national ecological restoration policy, and the corresponding vegetation growth season was selected as March-November. Based on the standardized rainfall use efficiency Z(RUE) model and integrated climate, land use/cover and optical remote sensing, this present study explored the influence of the temporal and spatial evolution characteristics of Z(RUE) and driving forces by the use of isolated natural and human factors to monitor the inter-annual and seasonal growth seasons. The result shows: 1) In 2001 to 2017, after two rounds of returning farmland to forests (grass) in Baoji area, Z(RUE) showed an overall improvement trend, especially in the area of pixels that showed a significant increase after the first round. The highest increase was 65.69%. The whole region was changed from the artificial intervention region to the reduction of human intervention under the first round of the project implementation; 2) The spring Z(RUE) change was the smallest among the inter-annual differences, and the transition point appeared in 2009, which was the period of 2001 to 2009 and 2009 to 2017 were positive and negative growth distribution, respectively. Z(RUE) in Summer increased most significantly, together with the largest annual difference in Autumn. The Z(RUE) mean of each year on the inter-annual and spring scales was the highest in the grassland and cultivated land, respectively, and the Z(RUE) of urban and rural land was the highest in Summer and Autumn; 3) Different type of vegetation, slope, direction of slope angle, change of Z(RUE) were in the peak distribution of "three high and one low". The much higher peaks occurred in April, June, and November (maximum), whereas the low peaks occurred in September (minimum), corresponding to greening (April), harvesting (June), and sowing (September); 4) Illumination duration become the main driving factor for the Z(RUE) variations in Baoji area. The main driving factors were temperature, sunshine hours and per capita GDP. Since the implementation of the project of returning farmland to forests (grass), the grassland improvement trend in the growing season was the best, indicating consistent with the increasing trend of Z(RUE) on the grassland. Except Fufeng, Linyou County and Fengxian, the remaining counties (districts) have much more contributed to the variation of Z(RUE) than human activities in Baoji area of China.