INCA-P operates at a daily time step, tracking the stores and fluxes of water, sediment, dissolved and particulate P in both the land and in-stream phases of a river catchment. The model is spatially ‘semi-distributed’: the water course is split into reaches with associated sub-catchments. Two spatial set-ups are possible – the traditional set-up, in which there is a single main stem, or a ‘branched’ set-up. The latter allows in-stream processes, effluent inputs and residence times in tributaries to be simulated, and can be particularly useful in larger catchments or complex river networks. Each sub-catchment is split into landscape classes – there can be as many of these as desired or warranted by the data resolution or needs of the study. Landscape classes are “functional units” and within each unit P inputs, plant P uptake, soils and flow pathways should be similar, though for convenience classes are often based on land use and/or soil type. All land-based processes are calculated for a generic 1 km2 cell for each landscape class within each sub-catchment. Water, sediment, total dissolved P (TDP) and particulate P (PP) outputs from the 1 km2 cell for each land class are multiplied by the land class area, and summed to provide total inputs from the sub-catchment to the reach. These inputs are assumed to enter the stream reach directly, rather than being routed spatially from one land class to another. Reach inputs are therefore from the land phase and from any upstream reaches.