A spatial site yield potential model of oil palm / Haslinda Musa

The site yield potential (SYP) has been used as a realistic target for the plantation management to achieve and maximize the oil palm production. This yield potential is currently computed using an empirical and numerical model, which is non-spatial. However, the oil palm yields are known to vary spatially across the plantation. To represent this, Geographical Information System (GIS) and Global Positioning System (GPS) technologies have been incorporated into the model. This has resulted in a spatial SYP model of oil palm that allows the management to quickly compare the target and actual yields over the whole plantation. The SYP model gave poor relative root mean square error of 18.12% and efficiency of -1.60, indicated that the model tended to over-estimate the site yield potential. These poor results were mainly caused by replanting area, the attacks of Rhinoceros beetles and wrong identification of Rengam soil series. When the above fields were removed from the analysis, the model gave low relative root mean square error of 10.9%. However, it still over-estimated the site yield potential by about 5% on average. Apart from this bias, the model performed satisfactorily based on the standard model parameters for precision. Results also show that the mean difference between slopes measured using clinometer and the readily available topographical map of Malaysia with a 20 m contour intervals using triangulated irregular network (TIN) model, were only 0.43 degree, which was below the accuracy of the clinometer of 1 degree, although they were statistically different. Therefore, an empirical model using GIS can be used as a tool to help the management and agronomists to focus on the main problems especially fields with large yield gap. This site yield potential model will require extensive validation before it can be adopted by the oil palm industry.