Elham Mehrabi Gohari; Roghaye Shahriyaripour; Ahmad Tagabadipoor; Seyed Roohollah Mousavi
Abstract
This study aims to evaluate and compare the efficiency of Artificial Neural Network (ANN), Regression Tree (RT) and Neuro-Fuzzy (ANFIS) models using a digital soil mapping framework to predict soil texture in a part of Sirjan province. Sampling was carried out at 84 observation points with a regular ...
Read More
This study aims to evaluate and compare the efficiency of Artificial Neural Network (ANN), Regression Tree (RT) and Neuro-Fuzzy (ANFIS) models using a digital soil mapping framework to predict soil texture in a part of Sirjan province. Sampling was carried out at 84 observation points with a regular grid of 2x2 km, and soil texture components were determined from the soil surface depth of 0 to 30 cm. Auxiliary variables included primary and secondary derivatives of the digital elevation model (DEM), a geomorphological map and remote sensing (RS) spectral indices. The appropriate variables selected using the Principal Component Analysis (PCA) feature selection method. Based on PCA, eight topographic variables and six vegetation indices and spectra from RS selected to predict soil texture components (sand, silt and clay). The efficiency of the models was evaluated using coefficient of determination (R2), mean error (ME), root mean square error (RMSE) and normalised root mean square error (nRMSE). The RMSE values in the neuro-fuzzy model compared with the regression tree model. The results of the neuro-fuzzy model were 1.43% for clay, 1.98% for sand and 2.1% for silt, which were 4.32%, 5% and 4.54% lower respectively compared to the regression tree model. The results of this study showed that the ANFIS model was more accurate in predicting clay, silt and sand compared to ANN and RT. Also, the geomorphology map, topographic wetness index, multi-resolution valley bottumn flatness index and Landsat 8 bands 5 and 6 had the highest relative importance in predicting soil texture components.
Hannaneh Sadat Sadat Mousavi; Afshin Danehkar; Ali Jahani; Vahid Etemad; Farnoush Attar Sahragard
Abstract
Different forms of land use development and human activities in protected areas are considered to be the main drivers of change, which have many effects on habitats, habitats, diversity and richness of species. The purpose of this research is to model the effect of human activities on the diversity of ...
Read More
Different forms of land use development and human activities in protected areas are considered to be the main drivers of change, which have many effects on habitats, habitats, diversity and richness of species. The purpose of this research is to model the effect of human activities on the diversity of vegetation using the artificial neural network method and determine the impact of ecological and human variables on them. This research was done in the Central Alborz protected area under the management of Alborz Province. To achieve the mentioned purpose, firstly, 101 plots and 101 soil samples were collected and, soil and vegetation analysis were performed on the samples. Finally, using the multilayer perceptron neural network method and using 18 input variables including physical and chemical variables of the soil, physiographic variables, and human factors variables , the effect of human activities on the diversity of vegetation in the study area modeled. According to the results, the vegetation diversity model with the structure of 1-5-18 according to the highest value of the coefficients of determination in the three categories of training, validation, and test data is equal to 0.82. 0.81 and 0.68 show the best structure optimization performance, distance from roads, electrical conductivity, and percentage of organic matter in the soil show the greatest effect on the diversity of vegetation in the study area. The model presented in this research is used as a decision support system in evaluating the effects of human activities on the diversity of vegetation in protected areas and provides the possibility of predicting the extent of these effects on the diversity of vegetation in these areas.
Zahra Khosravani; Mohammad Akhavan Ghalibaf; Maryam Dehghani; Vali Derhami; Mustafa Bolca
Abstract
The aim of this study was to model the subsidence of Abarkouh plain using inSAR and artificial intelligence techniques. At first, the subsidence map was prepared using the 46 Sentinel - 1 radar images (2014 – 2018) and radar interferometry techniques. Then, the Feedforward artificial neural network ...
Read More
The aim of this study was to model the subsidence of Abarkouh plain using inSAR and artificial intelligence techniques. At first, the subsidence map was prepared using the 46 Sentinel - 1 radar images (2014 – 2018) and radar interferometry techniques. Then, the Feedforward artificial neural network (ANN) algorithm was used to model the subsidence. In this algorithm, groundwater level changes (2014-2018), groundwater level, aquifer thickness, clay thickness in the aquifer and the clay thickness in the range of groundwater level changes (2014 - 2018) were introduced as input layers and the subsidence layer obtained from the radar interferometry method was introduced as an output layer to model training. These five parameters were obtained from the measured data set of 34 piezometer wells and 77 logs available in the archive of Regional Water Company of Yazd province. After initial checking of the data accuracy, the Kriging interpolation method was used to extend the five parameters to the whole region and the raster layers were prepared. The results of inSAR showed that maximum subsidence in parts of the studied area, i.e. in east, north - east and north, were 6, 2.7 and 1.6 cm/year respectively. Also, in order to verify the accuracy of the map resulting from using a neural network model, it was compared with the map with the radar imaging method. For this purpose, model evaluation criteria such as Nash-Sutcliffe (NS), RMSE,,MAE)and MARE were used, which 0.9524, 0.0018, 0.0012 and 0.1545 were obtained respectively.
Alireza Sepahvand; Hasan Ahmadi; Aliakbar Nazari Samani; Sebastiano Trevisani
Abstract
The geomorphometric indexes have been widely used for separation of surface landform features in the geomorphology science over the past decades. In this study, Multilayer Perceptron Neural Network (MPNN) was used to provide karstic landform classification. To that regard, initially, geomorphometric ...
Read More
The geomorphometric indexes have been widely used for separation of surface landform features in the geomorphology science over the past decades. In this study, Multilayer Perceptron Neural Network (MPNN) was used to provide karstic landform classification. To that regard, initially, geomorphometric indicators were extracted from Digital Elevation Model (DEM), and then these indexes were used as neurons of input layer in artificial neural network. Furthermore, the box plots were applied to analyze the relationship between karstic landforms (such as dolines, hills, karstic plains, karstic valley and headland) and geomorphometric indexes. The results showed that 34, 6.9, 1.07, 48.5, 9.51 percent of the studying area are spatially covered by valleys, plains, dolines, highlands and hills respectively. It has also been found that the optimal structure of artificial neural networks for classification of landform is model No. 12-9-1 by having the learning rate 0.1 and 87.18 percent of determination coefficient. Also, it should be noted that the accuracy of the innovative method for classification of karstic landform is 90.58 percent. The analysis revealed that variations in geomorphometric indexes are very visible in the landform of hills, highlands and karstic valleys, whereas there are slightly overlapping in the plains and dolines.
hamide afkhami; azam habibi pour; mohammad reza ekhtesasi
Abstract
Evaporation is considered one of the key climatic variables, especially in arid regions and evaporation losses is one of the important issues in irrigation and water resources management in these areas. Therefore, it is important being aware of the amount of evaporation and its modeling, as one of the ...
Read More
Evaporation is considered one of the key climatic variables, especially in arid regions and evaporation losses is one of the important issues in irrigation and water resources management in these areas. Therefore, it is important being aware of the amount of evaporation and its modeling, as one of the most important hydrological variables in agricultural research and water and soil conservation. In recent decades, artificial intelligence techniques have proven high capability and flexibility to estimate and predict nonlinear phenomena. In this study, three important data mining techniques including Artificial Neural Network, Active Neuro-Fuzzy Inference System and Regression Decision Tree were used for predicting evaporation. For this purpose, 8 climatic variables (Minimum average temperature, average maximum temperature, average temperature, sunshine hours, wind speed, wind direction, relative humidity and evaporation averages) were employed in this study. The results showed three models are able to predict evaporation for 12 months after. Finally among the used models, ANN showed better performance with coefficient efficiency of 0.97 and RMSE of 5.1and ME of 0.48. Also, The results showed that there is not significant difference in simulation results to predict the evaporation between two scenario, original data and normalized data.
elahe zafarian; Ataollah Ebrahimi; Reza Omidipour
Abstract
Land cover mapping is essential for natural resource management. Satellite imagery can be used for mapping land cover. Several methods are available for land cover mapping whilst choosing the best method is one of the most important issue in this context. To compare pros and cons of three methods of ...
Read More
Land cover mapping is essential for natural resource management. Satellite imagery can be used for mapping land cover. Several methods are available for land cover mapping whilst choosing the best method is one of the most important issue in this context. To compare pros and cons of three methods of classification including maximum likelihood, object-based segmentation and artificial neural network, first, the efficiency of these three methods were evaluated. Then the trend of land cover changes in Shahrekord basin was investigated for 1999, 2009 and 2015 using Landsat images of TM, ETM+ and OLI sensors, respectively. After geometric and radiometric corrections, the land cover map of 2015 was prepared based on the three methods. The results of the validation mapping methods revealed that object-based method was more promising than the others with 93 and 90% for total accuracy and Kappa coefficients of agreement, respectively. So, the object-based segmentation method is recommended for monitoring of land cover changes. The results of the land cover change indicated that residential areas increased from 1.727% in 1999 to 2.98% in 2015 and agricultural lands increased from 5.73% to 12.60% but rangelands were decreased by 9.05 in total. Moreover, bare-lands were increased from 1999 to 2009 by 6.19% but decreased from 2009 to 2015 by 5.27%. The result of this study showed that the object-based method is superior to pixel based method of Maximum-liklihood and neural network. So, object-based segmentation is recommended for estimating land cover changes.
Hamide Afkhami; mohammad dastorani; farzaneh fotouhi firuzabadi
Abstract
Due to the nature of the sediment data, selection of appropriate methods for processing the data before entering them to the artificial intelligence models can enhance the reliability of simulations results. In this study, the effects of sediment data processing procedures on ANN and ANFIS models outputs ...
Read More
Due to the nature of the sediment data, selection of appropriate methods for processing the data before entering them to the artificial intelligence models can enhance the reliability of simulations results. In this study, the effects of sediment data processing procedures on ANN and ANFIS models outputs in 7 Dez Basin stations were evaluated. Accordingly, three scenarios were considered: In the first scenario, original data was used without exerting any processing technique; in the second scenario, the data was normalized; and in the third scenario, logarithm of data were used according to logarithmic distribution governing. The simulation results showed that using data logarithm leads to higher performance and lower error, especially in stations where the best fit probability distribution is one of the log family distributions. Finally, among applied models, ANFIS showed the best performance with coefficient efficiency of 0.95 and RMSE of 5.4, MSE of 1.4 and ME of 0.42 in Biatoon gauging station and using the third scenario.
Amin Zoratipour
Abstract
Abstract
Estimation of fine suspended load rivers is important in designing reserves, transition volume ofsediment, and estimating lake pollution. Thus, some methods are needed for determining damagescaused by sedimentations in environment and determining its effects on the watersheds. There aremany ...
Read More
Abstract
Estimation of fine suspended load rivers is important in designing reserves, transition volume ofsediment, and estimating lake pollution. Thus, some methods are needed for determining damagescaused by sedimentations in environment and determining its effects on the watersheds. There aremany methods for estimating suspended load, one of these methods that solves the problems ofsediment discharge and can predict it is using Neuro fuzzy or ANFIS (Adaptive Network FuzzyInference System), and ANN (Artificial Neural Network) methods. These make a function betweensediment and simultaneous discharge by use of different algorithms. The goal of this research iscomparing the effectiveness of Neuro fuzzy, neural network artificial and statistical methods forestimating suspended load river in Glinak station of Taleghan Basin. It was found out thatsuspended load estimations of Nero fuzzy method with MAE 1006 ton/day, and correlationefficiency (R) 77%, RMSE 2621 ton/day and Nash-Sutcliff error (NS) 0.51 is better than NeuralNetwork Artificial and Statistical methods and Artificial Neural Network method rather thanStatistical Method are more proper. Also, contracting both neural networks artificial to fuzzy lawscan be illustrated better than other methods, variation of sediment Load River. One more merit ofthis method is that it is not sensitive to few errors in early statistical data and this fact enables betterestimation of neural network model in comparison with statistical model. Finally, Neuro fuzzymethod works better as the percent of train data to test data increases.
