Modelling the hydrological responses of a headwater catchment under shifting cultivation in upland Eastern Madagascar

Land cover in catchments undergoing shifting cultivation typically represents a mosaic of agricultural fields, fallows in different stages of regrowth, remnant forest, and degraded grasslands. Although runoff responses of the respective land-cover types are expected to differ, there is little quantitative information on how such a mosaic of land covers affects rainfall-runoff responses at the catchment scale.

From February 2015 to February 2016, we measured rainfall, streamflow, plus plot-scale (saturation) overland flow (SOF), soil water, and shallow groundwater dynamics under different land covers in the 31.7 ha Marolaona catchment in Eastern Madagascar. The catchment has undergone shifting agriculture for over 70 years which has resulted in a mosaic of vegetation at different stages of regrowth. Plot-scale hydrological responses varied between land covers and topographic positions, but catchment stormflow responses were generally small and dominated by pre-event water, suggesting that most storm-runoff was generated in the valley-bottom. However, for events exceeding an antecedent soil moisture storage plus rainfall threshold value, stormflow increased considerably, indicating contributions from the hillslopes as well. Despite lower rainfall in 2015/16, stormflow totals and annual water yield were higher than values reported for the driest years in the 1960s[1]. This is thought to reflect a deterioration of soil physical properties by repeated burning, cropping, and associated loss of topsoil during the intervening years, which has reduced the depth to an impeding layer and increases in amount and frequency of SOF.

To gain further insight into the runoff-responses across the catchment, we applied the RoGeR_Dyn model[2] using field-measured soil physical parameters for the respective land covers, topographic data, and climatic inputs. The simulations thus far highlighted greater deep percolation under tree fallows, while SOF was more common during the early stages of regrowth. Amounts of subsurface stormflow and percolation to deeper layers were highest in concave areas where flows converged. Soil moisture contents were lowest under tree-based land covers during the dry season. For nearly the entire dry season, rainfall events supplied the minimum amount of water needed to maintain soil moisture contents above critical levels for transpiration. This agrees with measured soil moisture and tree transpiration rates for nearby sites.

Although the hillslope modelling results are not expected to provide sufficient evidence to determine the effect of land cover on catchment-scale ecosystem services such as streamflow regulation, the lower runoff and higher deep percolation found along hillslopes under older fallows and forest suggest that these may be beneficial for flood mitigation and dry-season water provisioning to downstream areas. Next, RoGeR_Dyn will be used to determine the effects of regional changes in hillslope land cover and climate change on streamflow dynamics.

[1] Bailly, C., et al. (1974). Étude de l'influence du couvert naturel et de ses modifications á Madagascar. Expérimentations en bassins versants élémentaires. Cahiers Scientifiques, 4. Centre Scientifique Forestier Tropical, Nogent-sur-Marne, France, 114 pp.

[2] Steinbrich, A., Leistert, H., Weiler, M. (2021). RoGeR – ein bodenhydrologisches Modell für die Beantwortung  einer Vielzahl hydrologischer Fragen. In Korrespondenz Wasserwirtschaft, 14. Jahrgang, Heft Nr. 2, Feb-2021.  https://doi.org/10.3243/kwe2021.02.004