Modeling the ecohydrologic role of solar radiation on catchment development in semi-arid ecosystems.

Abstract

The role of solar radiation on ecohydrologic fluxes, vegetation dynamics, species composition, and landscape morphology has long been documented in field studies. However, these studies miss the value offered by a numerical modeling approach that integrates a range of ecohydrologic and geomorphic processes in exploring the landscape response to multiple controlling factors. This study represented flood generation and solar-radiation-driven echydrologic dynamics in a landscape evolution model (LEM) to investigate how ecohydrologic differences caused by differential irradiance on opposing hillslopes manifest themselves on the organization of modeled topography, soil moisture, and plant biomass. We use the CHILD (Channel-Hillslope Integrated Landscape Development) LEM equipped with a spatially-distributed solar-radiation component, leading to spatial patterns of soil moisture; a vegetation dynamics component that explicitly tracks above- and below-ground biomass; and a runoff component that allows for runoff-runon processes along landscape flow paths. This study starts with data analysis, and then followed by a modeling part. In the first part, the relationship between land surface properties (e.g. soil, vegetation, and lithology) and landscape morphology quantified by the catchment descriptors: the slope-area (S-A) relation, curvature-area (C-A) relation, and the cumulative area distribution (CAD), in two semiarid basins in central New Mexico. All three land surface properties were found to have significant influences on the S-A and C-A relations, while the power-law exponents of the CADs for these properties did not show any significant deviations from the narrow range of universal scaling exponents reported in the literature. Among the three different surface properties we investigated, vegetation had the most profound impact on the catchment descriptors. Following data analysis, the role of solar radiation on landscape morphology was investigated in the second part with a numerical model framework that integrated a range of ecohydrologic and geomorphic processes. Modeled spatial patterns of soil moisture confirmed empirical observations at the landscape scale as well as other hydrologic modeling studies. The spatial variability in soil moisture was controlled by aspect prior to the wet season (North American Monsoon, NAM), and by the hydraulic connectivity of the flow network during the NAM. Aspect and network connectivity signatures were also manifested on plant biomass with typically denser vegetation cover on north-facing slopes than south-facing slopes. Over the long-term, CHILD gave slightly steeper and less dissected north-facing slopes, more dissected south-facing slopes, and overall asymmetry in the modeled morphology of valleys. Aspect influence on hillslope asymmetry was enhanced with greater uplift rates. Model simulations showed how subtle differences in biomass and soil moisture dynamics at annual scales lead to distinct geomorphic differences at both hillslope and catchment scales. The controls of latitude and mean annual precipitation (MAP) on the development of hillslope asymmetry were investigated in the third part by using the CHILD LEM. In simulations the mean slope of north-facing slopes was steepened towards the poles, while south-facing slopes became gentler toward the poles. As a result of this inverse pattern, the relative differences between north- verses south-facing slopes become larger toward the poles. The model outcomes, which are compatible with field observations, show north-facing slopes to be steeper (shallower) than south-facing slopes in the northern (southern) hemisphere. Our results underscore the influence of solar radiation as a global control on the development of hillslope asymmetry. Variations in MAP at the same latitude have little impact on hillslope asymmetry in comparison to variations in latitude at the same MAP. In the last part, the observed spatial patterns in erosion rates caused by aspect-driven microclimatic and ecohydrologic conditions are examined with the CHILD LEM forced with a uniform uplift rate obtained by averaging the erosion estimates from the study site. Climate represented in the model ranges from simple to more realistic. The climate forcing is simulated by: (1) stationary climate represents the recent climate that prevails in the study site; (2) cyclic climate represents the late Pleistocene climate that prevailed in the region; (3) paleo-constructed climate based on paleoclimate proxies. Recent field study in central New Mexico shows that long-term erosion rates (~10,000 years) on south-facing slopes are faster than opposing north-facing slopes. However, CHILD simulations show that the discrepancy in erosion rates on opposing hillslopes is not sustainable over the long-term. Depending on the climate forcing or internal dynamics of erosion mechanism, either north- or south-facing slopes can be more erosive than their counterparts. Over the long-term, however, the fluctuations in spatial erosion rates are averaging out. Hence, under a given uniform uplift, erosion rates on opposing hillslopes are found to be the same.