Open AccessDigital Mapping Algorithms to Estimate Soil Salinity in Indira Gandhi Nahar Pariyojana (IGNP) Command area of India
Author(s) -
P. C. Moharana,
S. Dharumarajan,
Nirmal Tej Kumar,
U. K. Pradhan,
R. K. Jena,
R. K. Naitam,
Sunil Kumar,
R. S. Singh,
R. S. Meena,
M. Nogiya,
R. L. Meena,
B. L. Tailor
Publication year - 2019
Publication title -
agropedology
Language(s) - English
Resource type - Journals
ISSN - 0971-1570
DOI - 10.47114/j.agroped.2021.dec2
Subject(s) - algorithm , salinity , geography , computer science , mathematics , geology , oceanography
In the present study, the distribution of salinity was investigated using digital soil mapping (DSM) algorithmsin 5 km buffer zone of both side of Indira Gandhi Nahar Pariyojana (IGNP) canal system of Suratgarh tehshil in Rajasthan. To achieve this goal, 64 soil samples were used with 21 environmental covariates and 3 DSM algorithms. Result from the study showed that the difference between the minimum and maximum ECe is very high (35.55 dS m-1) in the different irrigation zone of IGNP canal system. The ECe ranged from 0.50 to 36.05 dSm-1. Results indicated that the most important environmental covariates were annual precipitation, elevation and valley depth. Among the DSM algorithms, RF model showed the best performance in predicting ECe at the regional level. Results showed that the RF algorithm could predict ECe with an R2, RMSE and MAE of 0.701, 3.367 and 1.722, respectively. RF and QRF showed similar performance in predicting ECe, while SVM showed lower efficiency than the other models in terms of R2 and prediction errors. Salinity prediction map shows that the vulnerability to soil salinity is high in Anupgarh branch of canal, and low in the IGNP main and Bikaner canal area. Furthermore, the model developed in this study provides comprehensive guidance for the land planners and decision makers to develop amicable strategies for management of IGNP canal system.