| Sumario: | Early maturing rice varieties are crucial for climate-resilient agriculture, yet nitrogen optimization in these varieties remains under-explored. Most existing studies focus on conventional varieties and lack an integrated approach combining agronomic traits, remote sensing, and statistical modeling. The objective of this study was to determine evaluation criteria and develop a model to predict the productivity of short-season rice varieties. Experiments were conducted in different seasons at two locations in Sidenreng Rappang and Maros, South Sulawesi, using a nested split-plot design with three replicates. The main plots consisted of five nitrogen levels, while the subplots included five early maturing rice varieties and two moderate age as control. Key findings of this study is that the stepwise regression model combining NDVI and yield per clump showed strong performance, with R2 = 0.65/0.73, RMSE = 0.65/0.61, and MAPE = 9.72%/10.81% for training/testing, respectively. This regression model effectively evaluates how rice growth responds to varying nitrogen fertilizer doses, particularly in early-maturing varieties. Therefore, it can be reliably used to predict the future yield of these varieties.
|
|---|