Simulating potential growth and yield of oil palm (Elaeis guineensis) with PALMSIM: Model description, evaluation and application

摘要

Reducing the gap between water-limited potential yield and actual yield in oil palm production systems through intensification is seen as an important option for sustainably increasing palm oil production. Simulation models can play an important role in quantifying water-limited potential yield, and therefore the scope for intensification, but no oil palm model exists that is both simple enough and at the same time incorporates sufficient plant physiological knowledge to be generally applicable across sites with different growing conditions. The objectives of this study therefore were to develop a model (PALMSIM) that simulates, on a monthly time step, the potential growth of oil palm as determined by solar radiation and to evaluate model performance against measured oil palm yields under optimal water and nutrient management for a range of sites across Indonesia and Malaysia. The maximum observed yield in the field matches the corresponding simulated yield for dry bunch weight with a RMSE of 1.7 Mg ha(-1) year(-1) against an observed yield of 18.8 Mg ha(-1). Sensitivity analysis showed that PALMSIM is robust: simulated changes in yield caused by modifying the parameters by 10% are comparable to other tree crop model evaluations. While we acknowledge that, depending on the soils and climatic environment, yields may be often water limited, we suggest a relatively simple physiological approach to simulate potential yield, which can be usefully applied to high rainfall environments and is considered as a first step in developing an oil palm model that also simulates water-limited potential yield. To illustrate the application possibilities of the model, PALMSIM was used to create a potential yield map for Indonesia and Malaysia by simulating the growth and yield at a resolution of 0.1 degrees. This map of potential yield is considered as a first step towards a decision support tool that can identify potentially productive, but at the moment degraded sites in Indonesia and Malaysia