Reflectance spectroscopic approach for estimation of soil properties in hot arid western Rajasthan, India
2015
Santra, Priyabrata | Singh, Ravindra | Sarathjith, M. C. | Panwar, N. R. | Varghese, Preeti | Das, B. S.
Periodic and regular assessment of land degradation in arid regions of the world is essential for implementing suitable corrective measures in time. Assessment of soil properties based on soil sampling from hot arid tracts followed by laboratory analysis is a formidable task. Reflectance spectroscopy appears to be an emerging technology for the assessment of soils in extreme environment. In this study, soil spectral library of 138 soil samples from hot arid western Rajasthan have been created in visible, near-infrared and short wave infrared (350–2500 nm) region of the electromagnetic spectrum along with the measurements of basic soil properties. Further, spectral reflectance-based algorithms have been developed for rapid assessment of soil resources of arid regions. Results showed that sand and clay content may be satisfactorily estimated from linear models involving principal components (PCs) or derived band reflectance as the input variables (R ² = 0.41–0.43). Organic carbon (OC) content of soil was also found satisfactorily correlated with spectral data (R ² = 0.27). Derived band reflectance corresponding to Operational Land Imager bands of Landsat-8 has been found best to predict soil properties. Soil OC content has been found to be best estimated by derived spectral band data corresponding to spectral bands of IRS-P6 satellite. Partial least square regression-based models were found even better than the PCs-based and band reflectance-based multiple regression models for estimating soil properties. Thus, the present study indicates that soil spectral reflectance data captured by remote sensing satellites may have a great potential for rapid assessment of soil resources in arid regions.
显示更多 [+] 显示较少 [-]