Long-term assessment of nitrogen and variety technologies on attainable maize yields in Nigeria using CERES-maize

Maize simulation models are proposed as tools for assessment of response to nitrogen (N) fertilizer and varieties in order to explore potential target zones for improved maize varieties. The CERES-maize model in the Decision Support System for Agrotechnology Transfer version 2.1 was tested using international testing nurseries at Ibadan, Mokwa, and Kaduna situated, respectively, in the derived savanna, southern guinea savanna, and northern guinea savanna zones in Nigeria during 1992-95. Historical weather data spanning 20 years were used at the target production environments to generate probabilistic estimates of maize yields; nitrogen use efficiency (NUE) associated with fertilizer and variety technologies. Analysis shows with high probability that, under rainfed conditions and N fertilizer input, the 90-110 day varieties (MDV) would yield better than 120-150 day varieties (LDV) at Mokwa and Ibadan, with superior NUE. The risk of crop failure with no N input was, however, substantial. Although response to N varied dramatically from year to year in association with the rainfall, there appears to be no advantage in adjusting N-input strategy for a variety. NUE was predicted to be best at the 60 kg N/ha input strategy, indicating potentials of further yield increase if methods of enhancing NUE at the higher N input levels could be further investigated. The NUE was found to be always lowest at Ibadan, in the derived savanna transition zone where rainfall and cloud cover were higher. (C) 1999 Elsevier Science Ltd. All rights reserved.