Comparison of Two Electrical Conductivity Measurement Methods for Soil Salinity Assessment in the Ouedrigh Region (Northern Sahara, Algeria)

Soil salinity poses a significant challenge to sustainable agriculture in numerous irrigated regions of Algeria.The measurement of electrical conductivity (EC) provides a method for assessing soil salinity.Most soil laboratories in Algeria employ electrical conductivity tests of 1:2.5 and 1:5 soil-to-water suspensions (EC1/2.5 and EC1/5) to assess soil salinity due to their simplicity.The electrical conductivity of the saturated paste extract (ECe) is the primary determinant of the impact of soil salinity on plant growth.Therefore, EC1/2.5 and EC1/5 must be modified to ECe.The objective of this study was to develop regression models to predict ECe utilizing EC1/2.5 and EC1/5 data.The operation was conducted upstream of the OuedRigh region in northern Algeria, within the Blidet Amar palm grove located inside boundary 108.Fifty-one soil samples for EC1/2.5 and EC1/5 were collected and analyzed to develop the models.The results indicated significant variability in the electrical conductivity of the soil samples, with values spanning from 2.63 to 21.34 mS/cm for 1:5 soil-water extracts, 3.52 to 32.4 mS/cm for 1:2.5 extracts, and 5.92 to 142 mS/cm for saturated paste extracts.The proposed linear regression equations demonstrated robust linear correlations between ECe and EC1/2.5, as well as EC1/5.Upon validation, the equations ECe = 4.84* EC1/2.5 -5.48 and ECe = 7.99* EC1/5 -6.31 were selected by the study to forecast ECe based on EC1/2.5 and EC1/5, achieving an R2 of 0.95.Twelve equations derived from eight supplementary EC conversion tests were utilized to evaluate the applicability of these models.Relative to the 12 models employed in this study, ECe prediction errors were reduced by a factor of 7.42 to 4.71.This method obviates the necessity for saturated paste extraction, which is costlier and more time-intensive, hence facilitating precise assessment of soil salinity.Finally, soil salinity maps were generated with ordinary kriging.The mean error (ME) and root mean square error (RMSE) for each map were computed to evaluate the model's efficacy.