Impacts of the dry season on rice production using Landsat 8 in West Sumatra

This paper examines the impact of the dry season on rice production using the Normalized Difference Vegetation Index (NDVI) of Landsat 8 Operational Land Imager (OLI) in Ampek Angkek sub-district, West Sumatra, Indonesia. The objectives of this study are (i) to develop a monitoring system for rice production in the dry season by utilizing Landsat 8 OLI and (ii) to evaluate the impact of the dry season on rice production at the sub-district scales. A supervised classification method was applied to detect rice cultivation during the dry season from Landsat 8 OLI imageries. The reference of NDVI was obtained by a UAV equipped with a multispectral camera. This study shows that the NDVI of Landsat 8 Lv2 had a good correlation with the NDVI of UAV (R = 0.826) compared with Landsat 8 Lv1 (R = 0.783). The result of the error matrix and accuracy of NDVI classification of Landsat 8 Lv2 found that the overall accuracy of NDVI Landsat 8 Lv2 was 87.1%, with Kappa accuracy of 71.8%. This study predicted rice production in the dry season by a linear function of NDVI from Landsat 8 Lv2 with a significant correlation (R = 0.658). The total rainfall during the dry season positively correlated with rice production with a high correlation (R = 0.923). In addition, this study could explain that the dry season parameters (i.e., the amount of rainfall, the onset, the end, and the duration) had an impact on the predicted rice production in Ampek Angkek sub-district.