Estimation Model of Robusta Coffee (Coffea canephora) Productivity Based on Soil, Plant, and Remote Sensing Data

Coffee is an important global commodity, and understanding the relationships among factors influencing its productivity is essential for improving production efficiency. This study aimed to evaluate the effects of soil, plant, and remote sensing variables on Robusta coffee productivity. The production estimation model included soil variables (potassium, pH, and electrical conductivity), crop variables (plant height, crown diameter, and chlorophyll content), and remote sensing data (NDVI). Data were collected directly from field plots measuring 10 m × 10 m. Multiple linear regression models were developed to improve prediction performance. Model accuracy was evaluated using paired t-tests, RMSE, and RRMSE. The results showed that the model based on soil and crop data (R² = 0.85) performed slightly better than the model based on soil, plant, and NDVI data (R² = 0.88). Furthermore, the soil and crop data-based model produced lower error values (RMSE = 2659.44; RRMSE = 11%) than the model incorporating NDVI (RMSE = 2737.10; RRMSE = 12%). These findings indicate that soil and plant variables play a dominant role in predicting coffee productivity, while remote sensing data provide complementary information. This study provides a comprehensive understanding of the integrated influence of soil, plant, and remote sensing variables in estimating and improving Robusta coffee productivity.