Search all library resources

A hydrologic and slope stability analysis of a steep, tropical watershed : an integrated modeling approach

The abundance of water lies at the heart of many of the risks facing the tropics, including hazards such as floods and landslides. This region is subject to some of the highest intensities of land cover disturbance through forest removal and is also particularly vulnerable to climate change. However, there are few detailed hydrologic studies of tropical environments relative to their temperate counterparts, and the extent to which knowledge gleaned from detailed studies of temperate watersheds can be applied to the tropics remains unclear. We use ParFlow.CLM, a three-dimensional, integrated hydrologic model to advance our understanding of the hydrologic and slope-stability response of watersheds in the Northern Range of Trinidad, a tropical island in the Caribbean. Despite a low frequency of extreme storms, this location has one of the highest incidences of reported landslides. This dissertation includes a collection of sensitivity experiments at the scale of a single hillslope, as well as the watershed scale. We investigate under what conditions a series of short-duration rainfall events leads to slope instability. Even though a fully coupled approach is computationally expensive to implement, our results demonstrate that using a simple, 1D hydrologic model without lateral flow and heterogeneity to simulate slope failures can lead to non-conservative results, particularly in the case of multiple short duration rainfall events. We also perform a collection of experiments to analyze the effect of model discretization, hydraulic conductivity heterogeneity, and land cover change on hydrologic model response. We show that coarsening the spatial discretization at grid cell sizes less than 100 m results in sizable differences in the magnitude of the evapotranspiration flux. Changes in subsurface storage and surface runoff obtained with ParFlow.CLM are of the same order of magnitude as HRU based models such as SWAT which have become the de facto standard for analyzing the effects of land-cover changes on the water budget.