The Least Limiting Water Range to Estimate Soil Water Content Using Random Forest Integrated With GIS and Geostatistical Approaches

Abstract

Algorithms that exist in every area today have become the center of our lives with technological developments. The uses of machine learning algorithms are being researched with the new developments in the agricultural field. The present study determined the least limiting water range (LLWR) contents of alluvial lands with different soils distributed in the Bafra Plain, where intensive agricultural activities are carried out, and revealed the compression and aeration problems in the area with distribution maps. Also, the predictability of LLWR was evaluated with the random forest (RF) algorithm, one of the machine learning algorithms, and the usability of the prediction values distribution maps was revealed. The LLWR contents of the soils varied in the range of 0.049-0.273 cm3 cm -3 for surface soils. There were aeration problems in 6.72%, compaction problems in 20.16%, and aeration and compaction problems in 0.8% of the surface soils examined in the study area. Furthermore, 72.32% of the soil was under optimal conditions. For the 20-40 cm depth, an aeration problem in 5.88%, a compaction problem in 28.57%, and both an aeration and a compaction problem in 2.52% of the points were detected. In estimating LLWR with the RF algorithm, the root mean square error (RMSE) value obtained for 0-20 cm depth was determined to be 0.0218 cm3 cm -3, and for 20-40 cm depth, it was 0.0247 cm3 cm -3. In the distribution maps of the observed and predicted values obtained, the lowest RMSE value was determined by the SK interpolation methods for 0-20 cm depth and the OK interpolation methods for 20-40 cm. The distribution of obtained and predicted values in surface soils was similar. However, variations were found in the distribution of areas with low LLWR below the surface. As a result of the study, it was determined that LLWR can be obtained with a low error rate with the RF algorithm, and distribution maps can be created with lower error in surface soils.