ABSTRACT. In Madagascar, soil erosion is significant even when it is compared to world averages. A resulting special geomorphic feature is a form of gully erosion known as lavaka that appears in the highlands of the country. Lavakization (the generation and development of these features) is due to rather unique multifactorial environmental conditions. Among many factors (geology, soil composition, human activities, etc.)

Mathematical models are effective tools for evaluating the impact of predicted climate change on agricultural production, but it is difficult to test their applicability to future weather conditions. We applied the SWAP model to assess its applicability to climate conditions, differing from those, for which the model was developed. We used a database obtained from a winter wheat drought stress experiment. Winter wheat was grown in six soil columns, three having optimal water supply (NS), while three were kept under drought-stressed conditions (S). The SWAP model was successfully calibrated against measured values of potential evapotranspiration (PET), potential evaporation (PE) and total amount of water (TSW) in the soil columns. The Nash-Sutcliffe model efficiency coefficient (N-S) for TWS for the stressed columns was 0.92. For the NS treatment, we applied temporally variable soil hydraulic properties because of soil consolidation caused by regular irrigation. This approach improved the N-S values for the wetting-drying cycle from -1.77 to 0.54. We concluded that the model could be used for assessing the effects of climate change on soil water regime. Our results indicate that soil water balance studies should put more focus on the time variability of structuredependent soil properties.

No abstract