Estimasi Produksi Kelapa Sawit (Elaeis guineensis Jacq.) dengan Metode NDVI di Bukit Pinang Estate

Abstract

Estimasi produksi kelapa sawit diperlukan untuk mendukung perencanaan panen dan pengelolaan perkebunan secara efektif. Penelitian bertujuan menerapkan teknologi penginderaan jauh berbasis normalized difference vegetation index (NDVI), menganalisis pengaruh umur tanaman, NDVI, luasan, topografi, pemupukan ZA dan dolomit terhadap produksi kelapa sawit, serta membandingkan hasil estimasi dengan produksi aktual di Bukit Pinang Estate, Sumatera Selatan. Penelitian dilaksanakan pada Februari sampai Mei 2026 menggunakan citra Sentinel–2, ArcGIS, dan analisis regresi linear berganda. Model yang diperoleh adalah Y = 238 – 473X1 + 7,4X2 – 5,00 X3 – 0,057X4 + 53,2X5 – 60,5X6. Hasil penelitian menunjukkan bahwa umur tanaman, NDVI, luasan, topografi, dan pemupukan ZA berpengaruh positif terhadap produksi, sedangkan dolomit berpengaruh negatif. Model memiliki nilai R² sebesar 90,68%, akurasi 83,75%, dan RMSE 15,61 ton, sehingga mampu mengestimasi produksi kelapa sawit dengan tingkat ketelitian yang baik.

Oil palm production estimation is essential for supporting harvest planning and effective plantation management. This study aimed to apply remote sensing technology based on the normalized difference vegetation index (NDVI), analyze the effects of plant age, NDVI, plantation area, topography, ZA fertilizer application, and dolomite fertilizer application on oil palm production, and compare the estimated production with the actual production at Bukit Pinang Estate, South Sumatra. The study was conducted from February to May 2026 using Sentinel–2 satellite imagery, ArcGIS, and multiple linear regression analysis. The resulting regression model was Y = 238 – 473X1 + 7.4X2 – 5.00X3 – 0.057X4 + 53.2X5 – 60.5X6, where X1 represents NDVI, X2 represents plant age, X3 represents plantation area, X4 represents topography, X5 represents ZA fertilizer application, and X6 represents dolomite fertilizer application. The results showed that plant age, NDVI, plantation area, topography, and ZA fertilizer application had positive effects on oil palm production, whereas dolomite fertilizer application had a negative effect. The model achieved a coefficient of determination (R²) of 90.68%, with an accuracy of 83.75% and a root mean square error (RMSE) of 15.61 tons, indicating that the model was able to estimate oil palm production with a good level of accuracy.