Comparison of temporal trends from multiple soil moisture data sets and precipitation: The implication of irrigation on regional soil moisture trend

Abstract

In this study, soil moisture trend during 1996–2010 in China was analyzed based on three soil moisture data sets, namely microwave-based multi-satellite surface soil moisture product released from European Space Agency’s Climate Change Initiative (ESA CCI), ERA-Interim/Land reanalysis, and in-situ measurements collected from the nationwide agro-meteorological network. Taking the in-situ soil moisture as reference, it is found that ESA CCI generally captured soil moisture trend more accurately than ERA-Interim/Land did. From the spatial distribution of trend analysis results, it is seen that significant decreasing trend for summer soil moisture in northwestern China and northern Inner Mongolia, as well as the significant increasing trend for autumn soil moisture in northern China were identified by both ESA CCI and ERA-Interim/Land. This is in alignment with results from gauge-based precipitation provided by Institute of Geographic Sciences and Natural Resources Research (IGSNRR) and satellite-based precipitation from Tropical Rainfall Measuring Mission (TRMM). However, disagreements in derived trends between ESA CCI, ERA-Interim/Land and IGSNRR were observed in the southwest and north of China, especially in major irrigation regions, such as the oases in northern Xinjiang and large areas in Sichuan province. Prominent difference between soil moisture and precipitation exhibited in the extensively irrigated Huang-Huai-Hai Plain. The spatial coincidence between significantly wetting areas (identified by ESA CCI) and heavily irrigated areas, as well as the grid-based Student's t-test sampling from various irrigation levels revealed that the observed discrepancy was caused by massive anthropogenic interference in this region. Results indicate that, for regions with great magnitude of human interference, modules considering actual irrigation practice are crucial for successful modeling of soil moisture and capturing the long-term trend. Furthermore, results could provide insights on hindcast of historical irrigation areas using satellite-based precipitation and soil moisture data sets.