Applying vegetation indices to detect high water table zones in humid warm-temperate regions using satellite remote sensing

This study developed a geobotanical remote sensing method for detecting high water table zones using differences in the conditions of forest trees induced by groundwater supply in a humid warm-temperate region. A new vegetation index (VI) termed added green band NDVI (AgbNDVI) was proposed to discriminate the differences. The AgbNDVI proved to be more sensitive to water stress on green vegetation than existing VIs, such as SAVI and EVI2, and possessed a strong linear correlation with the vegetation fraction. To validate a proposed vegetation index method, a 23 km2 study area was selected in the Tono region of Gifu prefecture, central Japan. The AgbNDVI values were calculated from atmospheric corrected SPOT HRV data. To correctly extract high VI points, the influence factors on forest tree growth were identified using the AgbNDVI values, DEM and forest type data; the study area was then divided into 555 domains chosen from a combination of the influence factors and forest types. Thresholds for extracting high VI points were defined for each domain based on histograms of AgbNDVI values. By superimposing the high VI points on topographic and geologic maps, most high VI points are clearly located on either concave or convex slopes, and are found to be proximal to geologic boundaries—particularly the boundary between the Pliocene gravel layer and the Cretaceous granite, which should act as a groundwater flow path. In addition, field investigations support the correctness of the high VI points, because they are located around groundwater seeps and in high water table zones where the growth increments and biomass of trees are greater than at low VI points.

► We developed a geobotanical method for detecting high water table zones in forest areas, and proposed a new vegetation index, AgbNDVI, to discriminate differences in vegetation growth. ► High VI points were extracted correctly from the AgNDVI image using SPOT HRV data, which were correlated well with high water table zones. ► Our method can be useful for detecting high water table zones even in humid temperate forest areas.