Factors influencing the productivity of irrigated crops in Southern Peru, in relation to prediction by simulation models

The objective of the study was to determine the growth potential of alfalfa, potato, Rhodes grass and maize in the irrigated desert of S. Peru, as at that production level the highest utilization efficiency of irrigation water is usually obtained. Important growth- influencing factors were identified for each crop. In addition, measured results were compared to results from simulation models developed in Wageningen.Maximum annual alfalfa production was 27 t/ha dry forage. Experimental evidence suggested that assimilate partitioning favouring the root system, was a significant reason why yields were not higher. Other factors analysed were N 2 fixation, NPK fertilization, plant density decline, cultivar differences and weed infestation. Photosynthetis measurements on artificial swards are also presented.Potato yields of 80 t/ha were recorded. Very high growth rates of 270 kg/ha.d were obtained with the local andigena cultivar, but growth duration was shorter and harvest index lower than those obtained in temperate regions. Results of a line source sprinkler irrigation experiment indicated low transpiration coefficients, but also large irrigation losses due to a low uptake capacity of the roots. Further data are given on leaf area development, light interception, cultivar differences, split fertilizer application, NPK uptake and NO 3 -N concentration in the petioles.Both C 4 crops, maize and Rhodes grass, demonstrated high growth rates during summer, but dramatic declines in the other seasons, probably due to cold nights. For high production, ample N supply is of paramount importance. Data concerning NPK concentrations and leaf area index are given for both C 4 crops. Maize growth was also strongly hampered by mechanical resistance of sandy desert soils, but on clay loam very high growth rates of 500 kg/ha.d were reached. Hybrids from other sources and maize grown in a warmer environment at sea level, showed slightly higher initial growth but at the expense of final yield.The simulation model BACROS predicted the growth of maize and Rhodes grass during summer very well, but during the other seasons the simulations were too high as permanent damage to photosynthetic capacity due to low night temperatures was not taken into account. The very high growth rates of maize on clay loam could not be simulated by the model either. For the latter phenomenon and for the influence of mechanical soil resistance, model adaptations are presented and resulting simulations discussed.The model PHOTON both under- and over-estimated the photosynthesis measured on artificial swards of alfalfa.The simulation model ARID CROP was adapted to an alfalfa cutting management model that simulated reasonably accurately above and below- ground growth and reserve levels measured in S. Peru, at different cutting intervals from 21 to 53 days.Main factors influencing regional irrigated crop production and usefulness of simulation for agriculture in developing countries are discussed.