Evaluation and application of the CROPGRO – Cowpea model for simulating appropriate sowing window and planting density of cowpea varieties across contrasting environments

Kamara, A., Solomon, R., Tofa, A.I., Garba, I.I., Eseigbe, O.B., Jibrin, J.M., ... & Peter-Jerome, H. (2025). Evaluation and application of the CROPGRO–Cowpea model for simulating appropriate sowing window and planting density of cowpea varieties across contrasting environments. Field Crops Research, 331: 109988, 1-14.

Context Cowpea [Vigna unguiculata (L.) Walp.] is an important legume crop in sub-Saharan Africa where its grain and fodder are valued for food and feed. Grain yields are, however, low due to several biotic and abiotic constraints. Several improved stress-tolerant varieties and complementary agronomic management technologies have been developed to enhance its productivity and sustainability. Cropping simulation models are useful tools for evaluating the deployment of crop varieties and management options for target locations. While the CSM-CROPGRO model in DSSAT has been used to simulate the performance of several legume crops, only a few studies have evaluated and used the relatively new CSM-CROPGRO-cowpea model for use in West Africa. Objectives The objectives of this study were to (i) evaluate the performance of the CSM-CROPGRO-Cowpea model in simulating the cowpea growth and yields in contrasting environments (ii) use the model to assess the optimal sowing window and planting density of cowpea varieties across contrasting environments in the savannas of Nigeria. Methods Here, we used comprehensive savanna-wide datasets to calibrate and validate the CSM-CROPGRO-cowpea model for savannah environments. The evaluated model was then applied to assess the yield performance of cowpea varieties with varying plant densities and six sowing windows across four sites considering 36 growing seasons. Results The model accurately simulated cowpea phenology (RMSE 0.58–0.67 day; nRMSE 1.36–1.46 %; d-index > 0.90 for days to flowering, RMSE 0.82–1.73 days; nRMSE 1.09–2.29 %; d-index 0.88–0.99 for days to physiological maturity), grain yield (RMSE 86–121 kg ha−1; nRMSE 3.66–6.14 %; d-index > 0.90) and total dry matter (RMSE 260–295 kg ha−1; nRMSE 4.79–10.73 %; d-index = 0.87–0.95). The long-term simulation results indicate that SAMPEA 9 showed no response to sowing density beyond 13.3 plants m–2 across all locations, likely due to interplant competition at higher densities. In contrast, the simulated yield of SAMPEA 14 and FUAMPEA 1 increased as plant density increases from 13.3 to 40 plants m⁻². In northern Guinea savanna, sowing could be delayed until July 14 at Demsa and July 29 at Zaria for all tested varieties. In the Sudan savanna AEZ (SS), sowing should be done between July 1 and 14 for all varieties, beyond which there will be a significant reduction in yield. Conclusion Except for SAMPEA 9, the simulated optimum planting density for all the varieties is 40 plants m–2 in all AEZ, while the sowing window was dependent on location and AEZ. The variety SAMPEA 9 was the most yield-stable variety across the tested environments and did not require planting density above the current industry recommendations of 13.3 plants m−2. This study could fill the knowledge gap in understanding optimal cowpea management opportunities needed to maximize productivity and strengthen cropping resilience.