By Aries Purwanto – (Thesis M.Si Sriwijaya University, 2013). Supervised by Prof. Charlotte de Fraiture Ph.D., M.Sc. (UNESCO-IHE), Prof. Ir. Robiyanto H. Susanto, M.Agr.Sc., Ph.D (Sriwijaya University), F.X. Suryadi, Ph.D., M.Sc. (UNESCO-IHE).

Soil erosion is one of the major environmentally problems in tropical and semi-arid area both in lowland moreover in upland areas. Its long-term effects will cause soil degradation, lowering soil productivity and off course sustainability of the land as a whole. Erosion is a three-step process: detachment followed by transport and deposition. Recently, Revised Universal Soil Loss Equation (RUSLE) is becoming widely implemented in predicting annual average soil erosion. By considering five factors influenced soil erosion by water; rainfall erosivity (R), soil erodibility (K), slope-length (LS), cropping management (C) and erosion control practices (P) factor and also combined with Geographic Information System (GIS) techniques, the prediction results of potential annual soil erosion will be more comprehensive.

This study was conducted in Lambidaro Sub-Watershed, which is located in the upper part of Musi River that passing through the city of Palembang, South Sumatra Province, Indonesia. The objectives are: to predict the level of erosion level using GIS-RUSLE method, to determine the different potential soil erosion (A) due to the changes of C factor and P factor value, to analyze the characteristics of the sedimentation process and to recommend some measures to reduce soil erosion.

Five scenarios have been developed related to the implementation of erosion control practices and the possibility of land use change. P factor value will be adjusted based on the method applied, while C factor value will change due to the possibility of land use change in Lambidaro sub-watershed. Scenario I is defined as the existing condition for R, K, LS, and C factor value, while P factor value is defined as 1 for the entire area. It is assumed that there is no erosion control practices implemented. For Scenario II, III and IV, three mechanical erosion control measures were applied. Contouring method is implemented in Scenario II with P factor values varied from 0.55 – 0.85 based on the slope value. Scenario III is applying level retention bench terrace with P factor value equal to 0.01 for the entire area. In addition, in Scenario IV contour bunds method is applied with P factor value ranging from 0.275-0.425. Furthermore, other factors such as R, K, LS, and C factor values for those three scenarios are based on the existing conditions. Dissimilar assumption is applied in Scenario V. The value of R, K, LS and P factor are the same as the existing condition in the Scenario I, however there is a change in C factor value due to the changing of swampy area and paddy field (C factor = 0.01) into settlements (C factor = 0.5).

From the computation results of the first scenario, the Annual average soil erosion (A) values are dominated by 0-0.5 t/ha/yr (50% of total area), 0.5-1 (8%) and 1-10 (36%). In addition, other values are ranging from 10-20 (7%) and 20-250 (<1%). Assuming that the average soil depth in Lambidaro sub-watershed is 60-90 cm from the surface (moderate), the level of erosion can be predicted as follows: low level (0-15 t/ha/yr) is 97% of area, moderate level (15-60 t/ha/yr) is <3% of area, high level (60-180 t/ha/yr) is 0.05% of area and very high level (180-480 t/ha/yr) is only 0.0001% of area. Other results show that Scenario III has the highest reduction of annual soil erosion with almost all 100% of area has A value only 0-0.5 t/ha/yr. Level retention bench terrace is mostly effective to reduce potential soil erosion in lowland areas, recommended only on slopes up to 4.5o. Whereas for the slopes from 7o – 30o, bench terraces method are more appropriate to decrease soil erosion. Land use change will increase C factor value in the converted area. It shows in Scenario V, where the conversion of swampy areas and paddy field to settlement will increase potential soil erosion.

Based on the field measurement data and calculation in the dry season, the existing condition of artificial canal in Lambidaro sub-watershed is not stable and there is bed load transport. The total sediment yield (Y) is 2675 tons/year, while the potential soil erosion (A) is 19,501 tons/year. Therefore, Sediment Delivery Ratio (SDR) in Lambidaro sub-watershed will be 0.14. All the calculation results related to the sedimentation processes is limited by the assumption that the tidal fluctuation does not affect the results. In addition, the results are only valid in the dry season. However, by assuming that in the rainy season there will be banks-full volume, where the water depth (h) = 3.5 m and the water discharge (QW) in the rainy season will reach 59.5m3/s, so the sediment discharge (Qs) will be 55,537 tons/year.

Efforts to control soil erosion in the upstream part are more important to do instead of waiting for the response to sedimentation in the downstream part. Economically, the total cost needed is relatively low compared to the potential negative impacts that occur due to significant and continuous soil erosion in the future. Level retention bench terrace is mostly effective to reduce potential soil erosion in lowland areas, recommended only on slopes up to 4.5o. Furthermore, for the slopes 7o-30o, bench terraces method is more appropriate to decrease soil erosion.

RUSLE method is only predicting sheet and rill-inter rill soil erosion without considering steam banks erosion that might be occurred at the canals in the watershed. Therefore, the further research related to the stream banks erosion to determine the source of soil erosion and sedimentation accurately. In addition, further research is also needed to determine the influence of tidal fluctuation and the effect of rainy season to the characteristic of the sediment and also sedimentation process, so that the result will be more comprehensive.

By obtaining all the data sources and following all the procedures in this RUSLE-GIS Modelling we can compute and model the potential annual soil erosion in other watershed or sub-watershed in Indonesia.

Keywords: GIS-Based Model, Land Use, Lowlands, RUSLE Method, Sedimentation, Soil Erosion.