The CROPGRO-Peanut and DSSAT-Maize models (Boote et al. 1998) were used to simulate the long-term yields of peanut and maize at Akatsi. The peanut model was previously calibrated and validated for the coastal savanna zones by Adiku et al. (2001). The maize model was validated in 2003 using data from an ongoing study on maize-based cropping system at Kpeve located in the southeastern ecological zone of Ghana. Soil input data for the crop models were obtained for Akatsi from previous studies at Akatsi (Adiku et al. 2001). Weather input data were obtained from the Meteorological Services Department, Accra as noted above. As the data did not include solar radiation, sunshine hours were converted to solar radiation according to the Angstrom formula (FAO 1998):
Rs = K + h where: Rs = solar or shortwave radiation (MJ m-2 d-1); n = actual duration of sunshine (hours); N = maximum possible duration of sunshine or daylight hours (hours); n/N = relative sunshine duration; Ra = extraterrestrial radiation (MJm-2 d-1); as, bs = Angstrom values; as + bs = fraction of extraterrestrial radiation reaching the earth on clear days.
The values of as and bs were set at 0.25 and 0.50 respectively according to FAO (1988). The yields of peanut and maize were simulated for each of the 24 years at Akatsi (1976-2000) using three planting dates (PD): Julian day 73 (i.e. 14 March, as early planting date), 97 (i.e. 7 April, as intermediate planting date) and 120 (i.e. 30 April, as late planting date). At the beginning of each simulation, moisture content of the top 30 cm of the soil profile was set to field capacity. The planting density for peanut was set at 13 plants m-2 and that for maize was 6 plants m-2. These conform to the densities commonly found in farmers' fields. For maize, nitrogen application was varied between 0 kgNha-1 (control), 60 kgNha-1 (recommended rate) and 120 kgNha-1 (high application rate).
The simulated peanut and maize yields were sorted according to ENSO phases, ranked from smallest to the largest and the cumulative relative frequencies determined. Probability or cumulative distribution functions (CDFs) were constructed for each planting date for the various ENSO phases. Preference for a given planting date was based on the stochastic dominance concept (Anderson et al. 1977). From a pair-wise comparison of the CDFs of any two cropping strategies, the strategy whose CDF lies to the right is considered preferred (more is preferred to less). Maize yield increases with respect to increased applied nitrogen formed the basis for the choice of nitrogen fertilizer rate.
Was this article helpful?