5. Conclusion
The study was conducted in the Indo-Gangetic plain region to evaluate
the performance of reflectance spectroscopy in the Vis-Nir and MIR
regions for estimation of electrical conductivity which is known to
affect crop productivity and soil quality. The procedure adopted for
testing EC was one that is commonly used in the soil testing
laboratories of India and results suggested wide variation of electrical
conductivity in the study region. Spectral regions in 460-500 and
1890-1906 nm in the Vis-NIR region and 4200-4310, 5275-5280, 6660-6670,
7305-7310 and 8290-8300 nm in the MIR region were identified as
sensitive for estimation of EC. The PLSR model outperformed other models
in calibration and validation in the Vis-NIR range and the PLSR model in
calibration and RF model in validation was better in the MIR range
(followed by PLSR). Among the Vis-NIR and MIR regions, on the basis of
higher RPD values, the data fitted better in the Vis-NIR region and
would be the region of choice for predicting EC in the study area. Thus
reflectance spectroscopy in the Vis-NIR range with PLSR algorithm is
very well suited to replace conventional method of estimating electrical
conductivity in the intensively cultivated Indo-Gangetic plain regions
of India.