Effect of vegetation on the temporal stability of soil moisture in grass-stabilized semi-arid sand dunes.

Summary Soil moisture is a critical state variable affecting a variety of land surface and subsurface processes. Despite the complex interactions between soil moisture and its controlling factors, the phenomenon of temporal stability of soil moisture ( TS SM ) has been widely observed under natural conditions. In this study, the control of vegetation on TS SM is investigated by artificially manipulating surface vegetation (e.g., vegetated and de-vegetated plots) in a native grassland-stabilized sand dune area with similar soil texture and topography. Soil moisture data were collected at the depths of 30 cm (within the root zone) and 110 cm (below the root zone) over a period of four years. Using soil moisture data from the de-vegetated plots as a baseline, TS SM within the root zone is shown to be mainly affected by vegetation phenology at the study site. Therefore, the control of vegetation on TS SM varies on both seasonal and annual time scales. The change in the interseasonal patterns of TS SM is tightly related to plant phenology and the control of vegetation on the ranking of mean relative difference ( MRD ) of soil moisture significantly weakens during non-growing seasons due to diminished root water uptake. It suggests that the timing of sampling schemes (e.g., growing season vs. non-growing season) may alter TS SM patterns. On annual time scales, TS SM is affected by climatic conditions, as the control of vegetation on TS SM becomes stronger under drier conditions. In particular, vegetation tends to create larger contrasts in soil moisture levels between vegetated and de-vegetated plots in drier years. The soil moisture data also provide evidence that vegetation tends to reduce TS SM and increase spatial variability in soil moisture at the study site. The standard deviation of relative difference ( SDRD ) of soil moisture at the 30 cm depth (within the root zone) is considerably larger in the vegetated plots than those in the de-vegetated plots. As such, the effectiveness of using representative locations for monitoring mean soil moisture conditions in the vegetated plots deteriorates. [ABSTRACT FROM AUTHOR]