Utilising Landsat-8 OLI in determining Soil pH using single and combination band method for paddy fields area in Malaysia.

Abstract

The pH of a soil is a measure of its acidity or alkalinity. The pH of the soil is important in agricultural activities because it has an impact on crop yield. Remote sensing, Geographic Information Systems (GIS), and digital soil maps are becoming more appreciated in soil science studies. The research thus focuses on creating a predicted soil pH map of the study area using Landsat-8 satellite imagery and GIS. The single and combination of spectral bands were used to generate three models using simple linear regression. The results suggested that the approach is not sensitive enough for prediction of soil pH in the study area. R2 value obtained are 0.049 from Model 1, 0.016 from Model 3, and 0.0003 for Model 2. All the models indicate that the soil pH is in the acid situation but the full range of observed pH is not matched by any predicted model. In the validation process, Model 1 has an RMSE value of 0.397, whereas both Model 2 and 3 have RMSE value of 0.405. To obtain a more promising pH result, it is suggested to use indices such as vegetable indices (VI), salinity index (SI), and a combination of band ratios.