Prediction of Rubber Yield Based on Soil Water Content

Abstract

Abstract Fluctuation of monthly rainfall resulted in fluctuation of monthly rubber yield. This study was aimed to determine a model to predict rubber yield based on soil water content. This study was conducted at Indonesian Rubber Research Institute, South Sumatera in May 2022. The plant materials used in this study were rubber clones PB 260 that were planted in 2001 in two blocks of rubber field. The soil texture in these two blocks was clay loam with good drainage conditions. The average of monthly soil water content and rubber yield data for 11 years from the first block was used to determine the model. Whereas the average of monthly soil water content and rubber yield data for 12 years from the second block was used to validate the model. The soil water content was generated by field water balance calculation. The regression and correlation analysis showed that the highest correlation and the most suitable model to predict rubber yield based on soil water content was the power model with formula Y = 0.0668X2.1423 (coefficient of determination (R2) = 0.80). Where, Y = rubber yield (kg/month) and X = soil water content (%). Furthermore, the validation of the model showed that the goodness of match (GOM), Pearson correlation coefficient (R), coefficient of determination (R2), and root mean square error (RMSE) were 0.94, 0.95, 0.90. and 10.3 respectively. These values showed that the model was accurate and strong enough to predict rubber yield based on soil water content.