Soil erosion and nitrogen leaching in northern Vietnam: expression and modelling

Keywords:   Soil degradation, remote sensing, watershed, soil erosion model, paddy fields, terraces, water balance model, nitrogen balance model, geostatistics, rice-based systems, spatial dynamic model.Two years research has been conducted in Tam Duong district, Vinh Phuc province, upstream in the Red River Delta in northernVietnam, comprising three geographical regions, i.e. mountainous areas, terraces in hilly land and flat land. The extent of soil degradation in the district was delineated, using the Red/Near-Infrared band ratio of satellite images, calibrated on the basis of soil maps and field checks. Results showed strong dynamics in soil degradation with the largest area of degraded soil of 3280 ha in 1992, associated with a substantial reduction in forest cover and expansion of the agricultural area from the mid-1980s. Subsequently (1996),re-forestation, particularly planting of eucalyptus plantations, led to a reduction, followed (2000) by expansion again, as a consequence of harvesting of production forests.In the mountainous and hilly areas, soil erosion is the dominant degradation process, very severe at individual plot scale, but far less at sub-watershed and watershed scales (i.e. measured soil losses over eight events, with the same total rainfall, were 1360 and 773 kg ha -1 in a cassava and an eucalyptus plot, while it was 45 in the sub-watershed and 125 in the main watershed). Performance of the soil erosion model LISEM was evaluated in the study area; results showed differences in performance at different spatial scales. In the main watershed, simulated total runoff and soil loss were underestimated, because of storage and release in terraces and rice fields. In the upland sub-watershed, total soil loss was overestimated due to overestimation of sediment concentration, as a result of high detachment and transport capacity.In addition to soil erosion from the terraces, nitrogen leaching from the root zone in these sandy soils contributes to negative environmental impact of agriculture. A nitrogen balance model was developed, and calibrated on the basis of measured soil nitrogen concentrations in different cropping systems. Results from the model showed increased nitrogen leaching with increasing fertilizer doses, i.e.in a rice-rice-maize rotation, the traditional land use pattern in the district, annual nitrogen leaching losses varied from 52 to 60 kg ha -1 , 56 to 114 kg ha -1 , and 58 to 154 kg ha -1 for low, intermediate, and high fertilizer inputs, respectively.In the lowland area, agriculture has diversified and intensified, and high value crops are cultivated with very high doses of fertilizer. In these cropping systems, nitrogen leaching is particularly high. Annual leaching losses were calculated of up to 193 kg ha -1 in flowers and 115 kg ha -1 in vegetables compared to about 50 kg ha -1 in rice.From a set of point measurements, spatial distributions of nitrate- and ammonium nitrogen for a small region were predicted, using regression block kriging. The method was based on a stepwise backward linear regression, combined with expert knowledge, taking into account the weighted influences of various explanatory variables, i.e. elevation, soil type, and land use.Temporal variability in nitrogen concentrations in the groundwater were mainly the result of variations in rainfall and in land use, characterized by different irrigation and fertilizer regimes.For exploration of the consequences of introduction of alternative land use types and crop management, a spatial dynamic model was developed to simulate nitrogen dynamics at the scale of a sub-region, consisting of a mosaic of different soil and land use types. The model was calibrated and validated on the basis of two years of measured mineral nitrogen (bothNO 3 and NH 4 ) concentrationsunder different land use types with different irrigation and fertilizer regimes. Simulated results showed annual leaching losses varying from 88 to 122 kg N ha -1 in flowers, 64 to 82 in vegetables of the 'cabbage group'(i.e. paprika, cabbage, eggplant, kohlrabi), 51 to 76 in chili, 56 to 75 in vegetables of the 'squash group'(i.e. cucumber, tomato, pumpkin), and 36 to 55 in rice. The model, after further calibration and validation, is a useful tool for regional environmental assessment, and management support.The study has indicated that current agricultural developments in Tam Duong district, that are strongly influenced by (short-term) economic drivers, present a serious threat for the quality of the natural resource base, particularly soil and water and thus for the sustainability of the production systems. The obvious conflicts between the various objectives need to be addressed through integrated analysis in which the various stakeholders are involved.