Rainfall intensity and soil erosion by water : limitations of current erosion models and implications for erosion model-based studies under future climates
Choi, Daniel Mintae
PublisherUniversity of Oxford
MetadataShow full item record
Existing simulation studies of the effects of future climate change upon erosionindicate that, under land usages that leave the soil unprotected, even minor increasesin rainfall amounts are likely to result in disproportionately large increasesin erosion, but commonly make the simplifying assumption that distributionsof future rainfall intensities remain unchanged from the present. This researchaims to determine implications of rainfall -intensity changes on soil erosion usingcomputerised models. Thus, this thesis is a step towards the ultimate goal ofpredicting future rates of soil erosion caused by future rainfall intensity changes.Three soil erosion models, WEPP, EUROSEM, and RillGrow are employed toinvestigate impacts of various rainfall intensities on runoff and soil loss rates. Twoextreme daily rainfall events in summer and autumn are subjectively selected fromthe tipping-bucket rainfall data, and runoff and soil losses are simulated using threeerosion models. Estimated runoff and soil loss rates with high resolution rainfalldata are greater than those with low temporal resolution rainfall data. Within-StormIntensity Patterns (WSIPs) affect soil erosion amount, although runoff wasnot much affected. An additional daily rainfall event with Within-Storm Gaps(WSGs) is also selected to investigate effects of WSG removals on soil erosion.For a given amount of rainfall, events with constant low intensity (constant WSIP)produced dramatically less erosion: thus it appears that assuming a constant (oraveraged) intensity throughout a storm does not provide a good representationof a real rainfall with its continuously varying intensity. Analyses of outputsfrom WEPP simulations revealed a problem that WEPP modifies original rainfallintensity and, thus, simulates erroneous runoff and erosion rates. Future soilerosion rates are estimated using WEPP and CLIGEN data. 30 year-long weather isgenerated by CLIGEN. Likely future rainfall frequency and intensity are anticipatedby changing the mean maximum 30 minutes peak intensity also known as MX.SP.No fu ture rainfall amount change is assumed. WEPP simulation results suggestthat where mean maximum 30-min peak intensity of the wet months increases soilerosion increases at a greater rate than runoff. This research assists in improvingthe performance of erosion models with respect to changes of rainfall intensityby highlighting where current problem exists. In conclusion, greater knowledgefound here will, once future changes in rainfall intensity become better known andappropriate rainfall data become available, improve our ability to estimate futurerates of erosion.