EVALUATION OF CROPGRO-PEANUT MODEL FOR PREDICTING GROWTH AND YIELD OF GROUNDNUT (Arachis hypogaea L.) IN IBADAN AND OGBOMOSO, NIGERIA

Abstract

Groundnut is one of the most consumed oil crops in the tropics. Application of starter nitrogen and appropriate sowing dates are necessary for optimum growth and yield of groundnut. The use of crop growth models such as CROPGRO-Peanut can facilitate the evaluation of beneficial crop management practices with reduced undesirable impact on the environment. However, such models must be calibrated and evaluated for crop varieties in a given location before use. In this study, CROPGRO-Peanut model was calibrated and evaluated for groundnut performance in Ibadan and Ogbomoso, Nigeria. In a field experiment, four groundnut varieties (SAMNUT10, SAMNUT22, SAMNUT23 and Kampala-local check) were grown at four Sowing Dates (SD1, SD2, SD3 and SD4), spaced at seven days interval from 19 April to 10 May, at Ogbomoso. Data collected were used to calibrate the CROPGRO-Peanut model. Thereafter, SAMNUT23 and Kampala sown at SD2 and SD3 were evaluated with application of three starter nitrogen rates (0, 10 and 20 kg N/ha). Data were collected and used to evaluate the model. In Ibadan and Ogbomoso, three sources each of starter nitrogen (NPK 15:15:15, organomineral fertilizer and poultry manure) applied at 20 kg N/ha were evaluated on SAMNUT23 and Kampala at SD2. Data were collected and used to validate the model. All experiments were laid-out as split-plot in randomised complete block design replicated thrice. Data were collected on Days to Emergence (DE), Days to Flowering (DF), Pod Yield (PY) and weather parameters. The reliability of the model was evaluated using Percentage Error-PE (excellent: 0-10%, good: 11-20%; fair, 21-30%, and poor: >30%). Groundnut pod yield was predicted from 2010 to 2050 using daily weather data generated from GCM-DSSAT file generator. The simulated DE was excellent at SD1 (10%), but poor at SD2 (54%), SD3 (74%) and SD4 (358%). Simulated DF at SD1 was fair (30%) but excellent (2%)at SD2 while the simulated DF was good (20%) at SD3 and poor (39%) at SD4. Calibration of PY across the varieties was poor for all the sowing dates. The simulated DF with application of starter nitrogen at SD2 was fair for SAMNUT23 (25%) and poor for Kampala (35%), but fair for Kampala (28%) at SD3 and poor for SAMNUT23 (36%). The simulated PY was fair for both SAMNUT23 (26%) and Kampala (28%) at SD2, while it was fair for SAMNUT23 (22%) but poor for Kampala (36%) at SD3.For all sources of starter nitrogen, simulated DF was poor for both SAMNUT23 (43%), and Kampala (46%) while PY was excellent for SAMNUT23 (9%) and fair for Kampala (30%) (indicating the reliability of the model for the parameters). Model prediction for 2010 to 2050 showed decline in PY from 2,627 kg/ha to 700 kg/ha in Ibadan and from 4,568 kg/ha to 1,200 kg/ha at Ogbomoso. The CROPGRO-Peanut model adequately simulated days to emergence and was good for predicting pod yield of groundnuts regardless of starter nitrogen source and sowing dates. Ibadan and Ogbomoso lacked potential for long term production of groundnut.