Abstract
In forest areas, seasonal ecosystem changes may affect soil hydraulic properties and ultimately alter the pattern of soil moisture dynamics. A National Ecological Station in a tropical rainforest of Xishuangbanna located in southern China collected in-situ measurements of meteorological forcing, soil moisture content, litterfall, leaf area index (LAI), and water and carbon flux from 2003 to 2008, which provided an ideal dataset for a detailed investigation on how the soil hydraulic properties would vary in response to seasonal ecosystem changes.
Methodologically, the Marquardt Levenberg optimization algorithm was integrated in an unsaturated flow model to inversely estimate the soil hydraulic parameters based on unvaried, seasonally varied, and monthly varied parameterization strategies, respectively. The results indicated that the simulated soil moisture using seasonally varied and monthly varied parameters were more accurate than that with unvaried parameters.
Moreover, the statistical analysis of monthly varied soil hydraulic parameters suggested that the temporal variations in soil hydraulic properties were tightly related to the seasonal ecosystem changes in terms of LAI, litterfall, and heterotrophic respiration. In deciduous periods, a significant reduction of LAI was accompanied with abundant litterfall that replenished organic matter to the topsoil.
In contrast, during the hot rainy periods, the high heterotrophic respiration rate indicated that the soil organic matter underwent intensive decomposition under the condition of high air temperature and wet soil moisture- leading to strong seasonal cycles in soil hydraulic properties. The seasonal ecosystem change led to a significant increase in saturated hydraulic conductivity K(s) from less than 100 cm/day in hot rainy season to over 150 cm/day in cool dry season.
Moreover, the inversely estimated soil hydraulic parameters of saturated water content θ(s) and field capacity θ(f) at the topsoil layer were also featured with clear seasonal cycles (ranging from 0.32 to 0.42 and 0.23 to 0.27 respectively). In comparison with unvaried soil hydraulic parameter, inclusion of varied soil hydraulic properties provided more accurate simulations of soil moisture dynamics, which in turn, benefited the hydro-ecological modeling in vegetated areas.