Soil moisture estimation from Landsat-8 imagery using triangle method in Imphal–Iril river catchment, Manipur, India

Abstract

Soil moisture is an essential indicator for long-term sustainable agriculture. Crop growth and production are highly dependent on cropland soil moisture conditions. The primary goal of this study is to assess surface soil moisture indirectly using satellite-derived temperature vegetation dryness index (TVDI) based on the triangle approach within the Imphal-Iril river catchment. TVDI was derived from the Landsat-8 imagery dataset for three different periods based on normalised difference vegetation index (NDVI) and land surface temperature (LST). Maximum and minimum LST values were calculated to compute TVDI using LST and NDVI. According to the results, NDVI and minimum LST showed a positive correlation, whereas NDVI and maximum LST showed a negative correlation. The TVDI was correlated with in situ soil moisture measurements using a regression analysis. Fifty in situ soil moisture measurement data were collected for three different periods over the catchment using time domain reflectometry (TDR-300). These ground observation data were used to compute regression parameters and validate the study. The computed soil moisture values were validated against in situ volumetric soil moisture measurements. The results were statistically significant, with R2 values of 0.83, 0.85 and 0.86 for 22 November 2018, 8 December 2018 and 9 January 2019, respectively. The results of this study revealed that TVDI can depict soil moisture variation under various land uses in a region. The study found that settlement areas and higher-elevated hilly regions have lower soil moisture content, whereas agricultural zones in the valley and densely vegetated areas show significantly higher soil moisture levels.