The relationship between erosion measured in a standardised ‘erosion plot’ or predicted by amodel calibrated from such experiments and the magnitude of sediment yield in rivers can beoverstated due to the intercepting effect of filter strips/border strips and hedgerows that capturesediment from erosion higher up slope.There appears to be a major gap between results of plot level erosion studies and studiesbased on stream flow. The whole matrix of field borders in which the plots are set appears to havebeen ignored in most previous researches. To properly link between the on-site erosion and sedimentyield in rivers, factors, such as field border strip and the ‘hot spots’ that directly contribute sedimentto streams e.g. from footpaths should be taken in to consideration.The first part of this research emphasises the measurement of sediment transfer across varioustypes of filter strips. The working hypothesis is that transfer beyond filter strips is small, regardless ofland use, despite significant soil loss within fields for some types of land use.Nevertheless, those measurements are considered as point measurements. At landscape scale,the spatial configuration of mentioned filter strips will greatly influence the linkage between on-siteand off-site effect. The configuration of the filter strips in a catchment scale has not been taken intoconsideration in the first part of this research. Therefore, those point measurements need to betransferred into broader scale. To transfer into a broader scale a spatial model is needed.An existing spatial model, the so called ANSWERS model is considered as an appropriatetool to simulate the sediment transfer. The application of the ANSWERS to simulate the sedimenttransfer across filter strips is the second emphasis of this research.The third part of this research deals with analysis of the flow pattern of the Way Besai river inSumber Jaya (Lampung, Sumatra). We analysed these flow patterns and their association to rainfalland land use change over the past 25 years.