Simulating water limiting effects on production of grain maize in lowland and upland environments.

Water-limiting effects on the production of maize, cv. Suwan 1, in rainfed rice or lowland environment and upland environment were simulated using CORNCROS, a maize crop growth model assembled from modules of the MACROS model. Validation against data for six growing seasons at Bertam Mardi station (lowland environment) showed simulated yield on a sandy clay loam soil overlying a clay loam with a water table at 1.5m depth, to differ from the observed by less than 14% in five out of six cases. The simulated yields that ranged from 3234 kg/ha to 5249 kg/ha were reduced by 0.1-1.2 t/ha as a result of waterlogging if a plough pan or a layer with impeded drainage were present at a depth of 20cm. In upland environment, represented by Serdang, maximum rooting depth has limited influence on yield when there is sample supply of water but plays an important role when there is insufficient rain. Yield on a heavy clay in the rainy season may exceed more than 3.5 t/ha providing water table can be maintained at 2-3 meters and there is minimal impedance to root growth. Mean simulated rainfed yields on sandy clay loam overlying a clay loam at Bertam over the period 1979-88 range between 5.0 t/ha for April sowing to 1.2 t/ha for December sowing. On an identical soil in Serdang, the yields vary between 4.6 t/ha for March sowing and 3.6 t/ha for June and November sowing. The yield variations at both sites are fair relections of the monthly rainfall patterns. Frequency distribution analyses for March 15 sowing at Bertam and March 1 sowing at Serdang based on 20 years of weather data indicate very high probability af attaining rainfed yields of 4.0 t/ha and above over wide range of soil conditions. Nevertheless, the best yields are obtained from deep, medium-textured soils