Testing agricultural production simulator model (APSIM) with machanga (Embu) long-term manure application experimental data

Abstract

A long-term study was conducted at Machanga, in arid and semi -arid lands (ASALs) of eastern Kenya to assess the mfluence of manure application on soil nut6ent and the resultant crop productivity. T!1edata fi"om a long-term experiment (1989­1002) were also used to validate the perfonnance ofthe Agricultural Production Simulator (APSIM) model in predicting the crop dry matter yield, soil phosphorus (soil­P) and soil organic carbon (SOC) on P deficient soils.The experimental treatments comprised a control (no inputs), jand 10 t ha-' rates of high quality (26% C; 2,0% N; 0.48% P) manure from a single source applied annually in October. Other treatments were residual manure, where manure application ceased from 1993, and a NP fertilizer application in control plots from 1993. The experiment was simulated using the APSIM P-aware maize module even though maize was not the test crop prior to the November 1999 season. The model performance was tested in terms of dry matter yields, soil P and organic carbon. There was a wide agreement between observed and predicted dry matter yields, soil labile P but poor in organic C. For example the observed and predicted organic was not significantly different between 5 ndlOt ha-1, the decline of soil C in the :sidual treatment, and the magnitude of the crease in oil C when fertilizer was applied. The model model prediction ability need to be tested on different environment and crop and soil parameters before being recommended as a management tool.