Coupling water cycle processes with water demand routes of vegetation using a cascade causal modeling approach in arid inland basins

Abstract

Vegetation degradation is the key cause of land desertification in arid areas. Water stress is one of the most critical factors leading to vegetation degradation. The water needed for vegetation growth is inseparable from the water cycle processes. It is a new scope to reveal the vegetation water demand mechanisms from the water cycle processes. Water cycle processes in arid inland basins can be conceptually separated as RFA (runoff formation area) and RCA (runoff consumption area). In this study, both the water demand mechanisms of natural vegetation and farmland were discovered by creatively constructing the vegetation water demand route model. The TRB (Tarim River Basin), a typical arid inland basin system that RFA is separated from RCA, is considered as the study area. The tendency and relevance of water cycle factors and NDVI were detected. The dominant factors of vegetation growth were identified. According to the interaction causality of water cycle factors and vegetation, the PLS-SEM (partial least squares structural equation models) were constructed in RFA and RCA. Results displayed that SMroot (root-zone soil moisture), groundwater and precipitation were the dominant water sources for natural vegetation in RFA. The water demand for natural vegetation in RCA mainly came from SMroot and that for farmland mainly came from SMsurf (surface soil moisture). New findings showed that blue and green water circulations were more active in RFA than in RCA. Natural vegetation had better adaptability and resilience to water shortages compared with farmland. The higher effect of vegetation on AET (actual evapotranspiration) denoted the better growth status. It is contributed to the rational utilization of water resources in arid basins.

Keywords: Farmland; Natural vegetation; Partial least squares structural equation model; Runoff formation and consumption areas; Vegetation water demand.