Use of NDVI and NDWI to estimate soil moisture content in dry season

Abstract

Soil moisture content is an important parameter that indicates water contained in soil and plants. In dry season with limited water, there is a risk of damage to crop yields from lack of water. Monitoring of soil moisture content that covers a wide area of cultivated land is necessary to give an overview of water demand at the area level. Geo-informatics technology was applied in an attempt to estimate the soil moisture (SM) in dry season via two parameters: NDVI and NDWI parameters. Remote sensing data from LANDSAT 8 satellite recorded from January to April 2018 in the study area was analyzed. The field study was conducted to check soil moisture content at 15 cm depth of the study area by using gravimetric methods. Then, the prediction equation of soil moisture content for each month was developed by using multiple linear regression analysis. The equations to predict SM for January (SM_Jan), February (SM_Feb), March (SM_Mar), and April (SM_Apr) were SM_Jan = 11.909+(-32.244)NDVI+39.082NDWI, SM_Feb = (-2.322)+26.702NDVI+22.081NDWI, SM_Mar = 3.625+15.690NDVI+53.362NDWI and SM_Apr = (-4.312)+45.189NDVI+ (-3.212)NDWI, respectively. The coefficients of determination (R²) of these equations were 84.1, 65.9, 79.2, and 57.2, respectively. The developed equations can be used to predict SM content of cultivation area in the study site, and provide a soil moisture map for cultivation area. The predicted SM derived from the developed equation, integrated both spatial and temporal data, can be used for crop planning as well as water allocation plan with due consideration to actual cultivated area. This will lead to more accurate and precise water allocation.