Sediment budgets are useful in informing catchment management, however the resources and data required to create a sediment budget may be prohibitive, particularly in developing countries. This paper presents a multidisciplinary approach to an investigation of sediment sources and the creation of a first-order sediment budget for a catchment in eastern Indonesia: a data-poor region in the wet-dry tropics with high sediment production rates and a population largely dependent on subsistence agriculture. The approach integrates results from geospatial analysis and key informant interviews and radionuclide tracers. Free software and imagery were used to demonstrate that geospatial analysis can be achieved without high costs. Surface soil erosion rates were mapped using the Revised Universal Soil Loss Equation, and subsoil sediment sources were digitised from high-resolution imagery. Key informant interviews identified additional gully erosion, not detected through spatial analysis. Radionuclide tracers 137Cs, 210Pb(ex) and 239Pu were used to determine the relative contributions of surface soil to sediment. The main sediment sources in the Kambaniru catchment were surface soils (31%), channel change (22%), gully erosion (8%), and landslides (1%), with an estimated annual sediment load of 1440000tyr-1 at the weir. This sediment budget showed subsoils were a major source of sediment, which contradicts assumptions that underpin Indonesian catchment management policies. The geospatial, interview and field methods were effective in identifying, mapping and quantifying subsoil sediment sources, and can readily be applied in areas where data are scarce and technical skills are low. Radionuclide tracers provide essential information but are expensive.