WaNuLCAS modelling of runoff and soil loss for different agroforestry scenarios in a catchment in Northern Vietnam

Abstract: A large part of the world arable land is affected by degradation and a global population pressure drivespeople to unsustainable cultivation methods to meet food demands. Water is the most common cause of erosion and soil erosion tends to be especially severe in steep lands in areas with monsoon rains. Agroforestry, meaning plantation of trees in combination with crops, is considered to be a promising approach to sustainable land use because it can meet the demands of food at the same time as of the need of soil and water conservation. This study was carried out as a Minor Field Study within the LUSLOF project (Sustainable Land Use Practices for the Uplands of Vietnam and Laos: Science and Local Knowledge for Food Security). In a Participatory Landscape Analysis made in 2002 in the project field site, the Dong Cao catchment in Hoa Binh province in Northern Vietnam, it was found that the main issue in the area was declining crop yields due to depleted soil. In this study, the model Water, Nutrient and Light Capture in Agroforestry Systems (WaNuLCAS) was used to evaluate three different cropping systems at plot level in terms of water runoff and soil loss amounts for a 5-year period. The cropping systems were (1) monocropping cassava, (2) cropping cassava with hedgerows of Tephrosia candida and (3) cropping cassava with hedgerows of Bambusa Blumeana. The study compares model results of runoff and soil loss in the cropping systems and compares also up-scaled model results at plot level with measurements at catchment level. Before simulations were performed, sensitivity analysis and calibration were made with input and validation data from experimental runoff plots at a nearby site. During calibration of the model, it was shown that runoff and soil loss simulated by the model agreed well with observed total sums of the 5-year period and on yearly basis, but not on event basis. One of the reasons seems to be overestimation of soil loss by the model in time of the year when soil was uncovered. In simulations of different cropping systems, hedgerows of Bambusa Blumeana gave overall the minimum soil loss, while hedgerows of Tephrosia candida gave the minimum runoff. All scenarios with hedgerows showed to prevent runoff better than the monocropping system, whereas generally only hedgerows of Bambusa Blumeana showed to prevent soil loss better than the monocropping system. The up-scaled model outputs showed that simulated runoff was considerably lower than measured runoff, while simulated soil loss varied greatly compared to measured soil loss, depending on which land use that was considered.