T171 Historical Range of Variability in Landscapes

Tuesday, 2 November 2010
8:00 AM - 12:00 PM
Room 406

INVESTIGATING LAVAKA (GULLY) EROSION IN MADAGASCAR: LITHOLOGIC CONTROLS

Paper number: T171-7
Program: Technical Sessions
Time: 9:25 AM – 9:40 AM

Authors

Abstract

Madagascar's central highlands are characterized by a particular form of gully erosion. Malagasy people call it "lavaka" or "kady" or "hady", but "lavaka" became its international name. Lavaka are environmental hazards for Madagascar. They destroy substantial infrastructure: roads, bridges, and agricultural land. Some previous studies have argued that the bush fires responsible for the denudation of certain areas of the country are a principal factor of the origin of the lavaka, but we disagree because we can find many areas where people burn (sometimes two or three times on a year) on the flanks of mountains before cultivating rice or another plants, but there are few or no lavaka. In contrast, some areas have no people living there and yet there are many gullies. We know that badly conceived and badly studied constructions of road, hydraulic stopping, and drainage channels can cause formation of the lavaka. But geological controls are very important in lavaka origin. Heterogeneous petrography may be a strong factor controlling the origin and evolution of lavaka.

Lavaka form only in thick saprolites that developed on crystalline bedrock, and other workers have shown that ground water flow is important in lavaka formation. We measured the proportion of quartz in lavaka saprolites and find it is higher than in non-lavakaed bed rock. Petrographic heterogeneity of the initial rocks (especially small variation of the proportion of quartz) from surface towards depth may cause variation in flow paths for ground water and help localise lavaka formation. GIS analysis shows that lavaka line up along topographic ridges; but if you follow the concentration of lavaka they follow also the geologic foliation. Projection of lavaka locations onto the geologic map shows a concentration along the contact between two units. Field observations suggest that monolithologic hills (e.g. all amphibolite or all gabbro) are less likely to have lavakas but that mixed-lithology slopes (e.g. a migmatite with some granite or gabbro) will develop more lavakas.

Understanding the connection between lithology and lavaka formation may help Malagasy people predict where lavakas are more likely to form and so perhaps they can plan better erosion remediation.