Identification of paddy fields using temporal NDVI changes in Minamiaso Village, Kumamoto Prefecture, Japan to estimate groundwater recharge

Abstract

In Minamiaso Village (Kumamoto Prefecture, Japan), land covers of forests, grasslands, and paddy fields play an important role in groundwater recharge. The land covers have been changed in a past few decades by agricultural systems and natural disasters: landslides, slope failures, and damage of irrigation canals caused by the 2016 Kumamoto earthquake in Japan. In this study, we extracted training data for classifying the land covers, and detected the temporal change of NDVI (Normalized Difference Vegetation Index) for each land cover. To quantify the impact of past land cover change on the groundwater recharge, we need to grasp the area of each land cover with the different hydrological parameters such as infiltration capacities. In the case of paddy fields, its distribution also should be known because the infiltration rates that we investigated vary depending on the paddy fields. We attempted to identify the paddy field distribution by the obtained NDVI values, and confirmed that the paddy field areas near the earthquake faults damaged by the earthquake where are still not usable were well distinguished. The paddy field distribution based on NDVI was almost matched with that based on SAR.