Search

In the present study, the potential map of groundwater springs was prepared using the famous and efficient Maxent model as a benchmark in Namroud catchment, Tehran province. In this study, the effect of three different scenarios on the number of presence points (100,500,800 springs), raster resolution (100,50,30) and sample size (10/90, 20/80, 30/70, 50/50) on the predicted map was evaluated. First, 18 factors affecting the emergence of springs including lithological formations, the distance from the fault, fault density, elevation classes, slope percentage, slope direction, slope length factor, curvature maps, the distance from the waterway, waterway density, the distance from the road, topographic moisture index, relative slope position, flow power index, soil texture, surface roughness index and land use cover were selected and their maps were prepared in the ArcGIS10.5 and SAGA systems. After the correlation test and classification of the effective layers, the percentage and frequency of groundwater potential in each class were obtained using the frequency ratio method, which calculates the exact weight of each class. The relative operating characteristic curve (ROC) was used to evaluate the performance of these models. The results showed that the combination of raster resolution scenarios: 30-number of points: 100-sample size: 90/20 with the highest AUC (0.953 in training phase and 0.927 in validation phase) has goodness-of-fit and prediction accuracy compared to the combinations of other scenarios. In addition, about 9.14% of the study area had high and very high potential for groundwater. Based on the combination of the best scenario, the factors of distance from the waterway, relative slope position, drainage density and topographic moisture index with values of 17.3, 14.2, 13.8 and 10.2%, respectively, are the main factors of spatial control in the study area influencing the spring occurrence. [ABSTRACT FROM AUTHOR]

The objective of the current study was to zone flood probability in the Marzdaran watershed. Since the allocated budget for management work is limited and it is not possible to carry out operations in the whole area, having a map that has prioritized different areas in terms of the probability of flood occurrence will be very useful and necessary. A well-known data mining model namely MaxEnt (ME) is applied due to its robust computational algorithm. Flood inventories are gathered through several field surveys using local information and available organizational resources, and the corresponding map is created in the geographic information system. The twelve predisposing variables are selected and the corresponding maps are generated in the geographic information system by reviewing several studies. The area under the curve (ROC) is used to evaluate the modeling results. Then, the most prone areas of flood occurrence which are prioritized for management operations are identified based on the prepared map. Based on the results, about 100 km2 of the study area is identified as the most prone area for management operations. The results showed that the accuracy of the maximum entropy model is 98% in the training phase and 95% in the validation phase. The distance from the river, drainage density, and topographic wetness index are identified as the most effective factors in the occurrence of floods, respectively. [ABSTRACT FROM AUTHOR]

Due to drought and climate change, estimation and prediction of rainfall is quite important in various areas all over the world. In this study, a novel artificial intelligence (AI) technique (WGEP) was developed to model long-term rainfall (67 years period) in Anzali city for the first time. This model was combined using Wavelet Transform (WT) and Gene Expression Programming (GEP) model. Firstly, the most optimized member of wavelet families was chosen. Then, by analyzing the numerical models, the most accurate linking function and fitness function were selected for the GEP model. Next, using the autocorrelation function (ACF), the partial autocorrelation function (PACF) and different lags, 15 WGEP models were introduced. The GEP models were trained, tested and validated in 37, 20- and 10-years periods, respectively. Also, using sensitivity analysis, the superior model and the most effective lags for estimating long-term rainfall were identified. The superior model estimated the target function with high accuracy. For instance, correlation coefficient and scatter index for this model were 0.946 and 0.310, respectively. Additionally, lags 1, 2, 4 and 12 were proposed as the most effective lags for simulating rainfall using hybrid model. Furthermore, results of the superior hybrid model were compared with GEP model that the hybrid model had more accuracy. [ABSTRACT FROM AUTHOR]