Simulating forage yields and soil organic carbon under Brachiaria hybrid cv. Cayman in Tanzania with the CROPGRO perennial forage model

Land and soil degradation in cropping systems in sub-Saharan Africa has been exacerbated by inappropriate use of landscapes and poor management practices that result into environmental and subsequential social damages. Biophysical models are key to inform management activities that can restore degraded soils and ultimately improve yields and soil organic carbon (SOC) sequestration. Numerous modelling studies have been conducted on annual cropping systems, however there are no modelling studies on perennial forages. The goal of this study was to adjust and evaluate the ability of DSSAT CROPGRO-Perennial Forage model version 4.7.5.0, which was initially parameterised for Brachiaria cv. Marandu in Brazil, to simulate biomass yields and SOC under Brachiaria cv. hybrid Cayman (BHC) in three districts in the southern highlands of Tanzania. The key adjusted parameters were soil water (lower limit, drained upper limit, saturated water content) and stable soil organic carbon. After model calibration, the root means square error ranged from 638 to 2111 kg/ha for harvested biomass. The d-Statistic for harvested biomass ranged from 0.78 to 0.97. The RMSE for % SOC ranged from 0.26 to 1.01 % and 0.23 to 1.55 % at 0-20 cm and 20-40 cm depth respectively. The d-Statistic for SOC from ranged 0.19 to 0.35 and 0.40 to 0.53 for 0-20 cm and 20-40 cm respectively. The results indicate that the model can be used to simulate the growth of Brachiaria cv. Cayman under different soils and weather conditions with an acceptable adjustment of specific parameters including soil water (lower limit, drained upper limit, saturated water content) and stable soil organic carbon. Also, the model simulated SOC reasonably well despite the wide variability between observed and simulated values, which was attributed to short period for experimentation and other factors not captured by the model including residue return among others. The adapted parameterised model for Brachiaria cv. Marandu performed reasonably well in simulating biomass and SOC in a different region with different soils, climate and management. Hence, the parameterised model for Brachiaria cv. Marandu can also be used for Brachiaria cv. Cayman in a different region with different soils and climate conditions.

Korir, Jebet; Ordóñez, Leornardo; Hoogenboom, Gerrit; Boote, Ken; Birthe, Paul; Nyawira, Sarah Sylvia.