Attribution analyses of evapotranspiration and gross primary productivity changes in Ziya-Daqing basins, China during 2001–2015

Evapotranspiration (ET a ) and gross primary productivity (GPP) are two of the key flux indices in a basin system. This study investigated the evolutions of ET a and GPP and the contributions of climate variables and vegetation index over multiple land cover types, simulated by the VIP model (Vegetation Interfaces Processes, RS version) with Terra-Modis 250 m Normalized Difference Vegetation Index (NDVI) dataset in the Ziya-Daqing basins (ZDB), China, during 2001–2015. The nonparametric Mann-Kendall test was used to analyze the changes in ET a and GPP. Contributions to the variations of ET a and GPP were explored using a differential equation method. The analyses revealed that ET a and GPP were high in the eastern plain, and spatial distributions were mainly associated with distributions of the major land-cover types. ET a and GPP showed positive trends in the western Taihang Mountains and negative trends in the eastern plain region. The annual ET a and GPP increased from 2001 to 2015 in most land-use types except cropland and urban. The decreases in ET a and GPP were primarily attributed to solar dimming in cropland and urban. The increases in ET a and GPP mainly occurred in the natural land-cover types (e.g., mixed forest and grassland) as a tradeoff between the positive effect of greening and the negative effect of declined radiation. In ZDB, the contributions of net radiation, air temperature, wind speed, and leaf area index to the change in ET a were − 0.681%, − 0.379%, − 0.032%, and 1.442%, respectively. The contributions of net short-wave radiation, air temperature, relative humidity, and fraction of photosynthetically active radiation to the change in GPP were − 2.411%, 0.085%, − 0.020%, and 10.224%, respectively. Thus, the “greening” was the major reason for ET a and GPP increases in natural vegetations, whereas solar dimming was the major reason for ET a and GPP decreases in cropland.