Estimation of Vertisols Soil Nutrients by Hyperion Satellite Data: Case Study in Deccan Plateau of India

Abstract

Soil nutrients are essential for agricultural purpose. There are efforts for estimation of topsoil properties using visible and near infrared reflectance (VIS–NIR) of Hyperion satellite. Notwithstanding this, there should be more research on variety of soils and fields of practicable size especially in the Indian context, necessary for using this on a practical basis. To bridge this gap, estimation of selected soil nutrients (nitrogen (N), potassium (K), copper (Cu) and iron (Fe)) from small sized and randomly scattered vertisols fields taken up in Deccan plateau of India using Hyperion satellite data. The nutrient index (NI) for Fe was estimated to be higher (NI = 2.76) than other nutrients in the study area, which influences the spectral behavior. The pretreatment of Hyperion reflectance data by Savitzky-Golay filter (window size 15, second-order derivative) and partial least square regression (PLSR) analysis resulted in low to moderate estimations of soil nutrients. The variable importance projection (VIP) for each soil nutrient has been estimated. The important wavelengths were identified in the mid infrared region for nitrogen. For Potassium, the wavelengths were identified in the visible, near infrared and the mid infrared regions. The near infrared and midinfrared region for iron. Lastly for Cu, in the green region and mid infrared region were identified. The prediction accuracy for N, K, Fe, and Cu were estimated to be medium, with coefficients of determination values as 54%, 45%, 40%, and 41%, respectively. The vertisols in the study region demonstrated low reflectance that is deficient in humus, due to low permeability and moisture stress throughout the drought. Hence the presence of soluble nutrients concentration is low compared to other soils. In this study considering the results of R2, the iron has good prediction, then other soil properties. Thus, the present research implied that, Hyperion satellite data provides moderate potential to estimate the Indian vertisols soil nutrients. © 2022, Indian Society of Remote Sensing.