skip to primary navigation skip to content

Department of Geography

 

Palaeoflood hydrology in the sub-humid tropics: the Mae Chaem Catchment, Thailand

Renée Kidson

Flood frequency analyses are often limited by the relatively short records from which the extreme value distribution is estimated, and from which extrapolation is then attempted. However, historical records and sedimentary evidence may be used to extend the record beyond the gauged period. In the Mae Cheam catchment is the Ob Luang gorge, the site of a National Park. Gorges are ideal locations for the preservation of such sedimentary records; slackwater sediments accumulate in gully entrances, caves, and cracks in the rock, at high elevations above the normal river level because flood flows in such confined reaches are very deep. In 2000 an extreme flood – the highest in living memory in the area – left continuous evidence of its peak water levels, in the form of trash lines, sand deposits and silt drapes.

The Ob Luang Gorge

The Ob Luang Gorge

This project is modelling the water level of the 2000 flood using HEC-RAS, a 1-D step-backwater curve model, from the gauging station P14 about 10km below the gorge, upstream into the gorge. This provides an excellent opportunity to evaluate the problems of calibrating the limited evidence of much older flood levels. In fitting the water surface evidence, it is possible to optimise parameters such as the roughness coefficient. When using step-backwater models to simulate the water surfaces of older floods that left higher surface indicators, there is little evidence to use, and a Manning coefficient has to be selected rather than optimised. This is obviously an area of uncertainty.

As part of this investigation, an expert survey is being conducted in which those with experience of selecting roughness coefficients are asked to use their judgement in defining coefficients for the study reach. Please go to www.srcf.ucam.org/~rlk23/Manning/ and complete the simple, brief web-based questionnaire located there.

Marks of the 2000 flood on pillars of a shelter in the National Park, and a stranded log in a tree opposite

Marks of the 2000 flood on pillars of a shelter in the National Park, and a stranded log in a tree opposite

This project will also include attempts to improve the incorporation of palaeoflood evidence in flood frequency analysis by (a) dating the evidence of previous flood levels; (b) investigating regional climatic and land use changes at the relevant dates; (c) undertaking regional flood frequency analyses to adjust the flood frequency for these changes; and (d) examining the use of a catchment rainfall-runoff model as an alternative means of estimating the flood frequencies.