JWSS - Journal of Water and Soil Science

Knowledge about the spatial distribution of soil organic carbon (SOC) is one of the practical tools in determining sustainable land management strategies. During the last two decades, the utilization of data mining approaches in spatial modeling of SOC using machine learning algorithms have been widely taken into consideration. The essential step in applying these methods is to determine the environmental predictors of SOC optimally. This research was carried out for modeling and digital mapping of surface SOC aided by soil properties ie., silt, clay, sand, calcium carbonate equivalent percentage, mean weight diameter (MWD) of aggregate, and pH by machine learning methods. In order to evaluate the accuracy of random forest (RF), cubist, partial least squares regression, multivariate linear regression, and ordinary kriging models for predicting surface SOC in 141 selected samples from 0-30 cm in 680 hectares of agricultural land in Khorramabad plain. The sensitivity analysis showed that silt (%), calcium carbonate equivalent, and MWD are the most important driving factors on spatial variability of SOC, respectively. Also, the comparison of different SOC prediction models, demonstrated that the RF model with a coefficient of determination (R2) and root mean square error (RMSE) of 0.75 and 0.25%, respectively, had the best performance rather than other models in the study area. Generally, nonlinear models rather than linear ones showed higher accuracy in modeling the spatial variability of SOC.

In the present study, a model was developed using a system dynamics approach to simulate and optimize the profitability of crops of the Jofeyr (Isargaran) Irrigation and Drainage Network located in Khuzestan Province. To validate the results, the statistical indicators of root mean square error (RMSE), standard error (SE), mean biased error (MBE), and determination coefficient (R²) were used. To validate the simulation results of the benefit-cost ratio, the values of these indicators were obtained 0.25, 0.19, 0.005, and 0.96, respectively. Then, to determine the optimal cultivated area of the network and increase the profitability, the cropping pattern was determined both non-stepwise and stepwise in 2013 to 2017 cropping years. In the non-stepwise, the cultivated area of each crop changed from zero to 2 times of current situation. In stepwise, due to social and cultural conditions of inhabitants, this change was slow and 10% of the current situation every year. The analysis of the results showed the success of the model in optimizing and achieving the desired goals and the total benefit-cost ratio increased in all years both non-stepwise and stepwise. For example, in 2017 compared to 2016, production costs decreased by 7.1 percent and sales prices increased by 5.8 percent, and increased the benefit-cost in 2017 compared to the previous year. The results showed that the present model has good accuracy in simulating and optimizing the irrigation network, its cropping pattern, and defining other scenarios.

The impact of different management options on the region and the existing conditions can be evaluated with minimal cost and time to select the most practical case using various tools including mathematical models. In this study, the SWAT hydrological model was performed from 2009 to 2019 using climatic, hydrological, and hydrometric data in the Malayer catchment, and the final model was validated by SWAT-CUP. To reduce the amount of uncertainty in the input parameters to the MODFLOW model, using the values of surface recharge from the implementation of the SWAT hydrological model, quantitative modeling of Malayer aquifer was performed more reliably in GMS software by using MODFLOW model. After modeling the study area in the 2009-2018 period and calibrating the model in the years from 2018 to 2019, the mean values of absolute error (MAE) were 0.35-0.65 m, and root means square error (RMSE) was 0.62-0.94 m, which seems acceptable considering computational and observational heads equal to 1650 m. Results of water level changes in observation wells located in the Malayer region indicate that the groundwater level in the aquifer has decreased by an average value of 9.7 m in the 10-year study period.

The land use change as well as changes in climatic parameters such as temperature increase affect many natural processes such as soil erosion and sediment production, floods, and degradation of physical and chemical properties of soil. Therefore, it is necessary to pay attention to different aspects of the effect of these changes in studies and macro decisions of the country. In the present study, the SWAT conceptual model was used to test and analyze the existing scenarios in the Marvast basin. After calibrating the model, the two scenarios were tested. The first scenario is in the field of agricultural management and conversion of gardens to agricultural lands and the second scenario is a 0.5-degree increase in temperature by assuming other conditions are constant. The calibration and validation results of the model with the Nash-Sutcliffe test showed 0.66 and 0.68 respectively, which indicate the acceptable performance of the model in the study area. Then, the results of using two scenarios of land use change and heating, especially in recent years showed the effect of 30 percent of the climate scenario on the increase of flooding in the basin. The scenario of changing the use of garden lands to agriculture in two cases of 20% and 50% change of use of 10% and 12% was added to the flooding of the basin. The results indicate that in similar areas of the study area which is located in a dry climate zone, a possible increase in temperature can have a significant effect on flooding in the basin. However, the indirect impact of the human factor in increasing greenhouse gases and flooding in the basin should not be ignored.

The agricultural sector as one of the most important sectors of water consumption has great importance for the sustainability of the country's water resources systems. The objective of this study was to estimate the river water abstraction (RWA) for agricultural consumption in the study area of Nobaran in the Namak Lake basin. The RWA was estimated using variables related to morphological, hydrological, and land use factors, as well as a combination of their variables collected through field sampling. Data mining methods such as adaptive-network-based fuzzy inference systems (ANFIS), group method of data handling (GMDH), radial basis function (RBF), and regression trees (Rtree) were also used to estimate the RWA variables. In the current study, the GMDH₂₄ model with a combined scenario including the variables of river width, river depth, minimum flow, maximum flow, average flow, crop, and the garden cultivated area was adopted as the best model to estimate the RWA variable. The RMSE value for the combined scenario of the GMDH₂₄ model was found to be 0.046 for estimating RWA in the Nobaran study area. The results showed that the performance of the GMDH₂₄ model for estimating RWA for maximum values is very acceptable and promising. Therefore, modeling and identifying various variables that affect the optimal RWA rate for agricultural purposes fulfills the objectives of integrated water resources management (IWRM).

Introducing reliable regional models to predict the maximum discharge of floods using characteristics of sub-basins has special importance in terms of flood management and designing hydraulic structures in basins that have no hydrometric station. The present study has tried to provide appropriate regional flood models using generalized linear models (GLMs) to estimate 2-, 10-, 50-, and 100-year maximum daily discharges of 62 sub-basins in Great-Karoon and Karkhe basins. According to the results, the sub-basins were categorized into four sub-regions based on some physiographic and climatic characteristics of the study sub-basins. The results showed that regional flood modeling was successful in all sub-regions except sub-region II, which includes very large basins (A̅≈17300 km2). The adjusted R² of the best models in sub-regions I, III, and IV were estimated at around 82.4, 91.3, and 90.6 percent, and these models have a relative error (RRMSE) of around 9.5, 9.23, and 6.7 percent, respectively. Also, it was found that more frequent floods with 2- and 10-year return periods are influenced by properties such as basin’s length, perimeter, and area, while rare floods with 50- and 100-year return periods are mostly influenced by the river systems characteristics such as the main river length, total lengths of the river system, and slope of the main river. According to the research, it can be stated that the behavior of maximum daily discharges in the study area is extremely influenced by the different climatic and physiographic characteristics of the watersheds. Therefore, the maximum daily discharges can be estimated accurately at ungauged sites by appropriate modeling in gauged catchments.