---

Using remote sensing data due to providing updated information, cover repetitive, low-cost assessment of natural resources have a special place. Also change detection in the management and evaluation of natural resources is one of the basic needs. Thus the value of change of land use /land cover (LULC) is the result of the change detection process can obtain on multi-temporal remote sensing images. Therefore, in this study, both of the Landsat satellite images 8 (OLI&TIRS) the year 2013 and 7(ETM⁺) the year 2000 were used as input data for land cover/ use mapping level 1 and 2. In the meantime, because of the new images OLI, radiometric corrections was formulation with existing equation with using in Erdas software model maker.also from Normalize Difference Vegetation Index (NDVI), Bare Soil Index (BI) and three main components from Principal Component Analyze (PCA) as input alongside other bands were used to increase the accuracy of classification. The polynomial 5 degree from SVM method compared with artificial neural network (ANN) and maximum likelihood classification (MLC). Results showed that support vector machine method using Polynomial kernel and degree 5 (accuracy 92%) gives overall accuracy higher than artificial neural network method (accuracy 89% ) and maximum likelihood method (accuracy 91.8%) . Also SVM method shows better performance where classes exhibit similar spectral behavior. Post classification method used for detect changes in the timeframe of 13 years. The results show large changes in (LULC) was occurred thus need monitoring and proper management is needed for this watershed.

---

Evaporation is one of most important parameters which are affected by many variables such as rainfall, wind velocity, sunny hours, and relative humidity etc. Evaporation estimation is important for any area with surface water resources because of its effect on dam lakes, precipitation-runoff modelling, river area performance, water management – calculating amount of water that plants need and planning for watering and so on. Evaporation can have significant effect on water balance of a river or a reservoir and it may be cause water level to decrease.

Due to hydraulic system complications caused by statistical information imperfection and determining all parameters involved, complete hydraulic system modelling is impossible. At such circumstances using al mathematical modelling system will be considered. 

Matherials & Method

In this study we tried to estimate pan evaporation using two models including Artificial Neural Network (ANN) and Support Vector Machine (SVM) with data preprocessing (gamma test and principal component analysis) to determine affective inputs into two models. For this matter data gatherd from three synoptic stations at Astara, Kiashahr and Talesh at Guilan province has been used. Synoptic stations data includes evaporation, wind velocity at two meter altitude, temperature (minimum, average and maximum), humidity (minimum, average and maximum), sunny and rainy hours. Statistical period of data for Astara and Talesh synoptic stations were 1384 to 1393 and for Kiashahr were 1385 to 1393. 80 percent of meteorology data were used for calibration and other 20 percent were used for model validation. In this study we used multilayer perceptron artificial neural network with sigmoid tangent function and 1 to 20 neurons for hidden layer and support vector machine with radial based kernel function.

Calculations has been made in to section with two data preprocess methods. At first section input variable has been selected by gamma test and pan evaporation estimation was made by both models. At second section modelling has been pulled out by input variables selected by principal component analysis.

Discussion of results

At gamma test section pan evaporation estimation parameters were as follows: minimum temperature, maximum humidity, minimum humidity, rainfall and sunny hours for Astara station; maximum temperature, minimum temperature, minimum humidity, rainfall and sunny hours for Kiashahr station and maximum temperature, minimum temperature, maximum humidity, average humidity, rainfall and sunny hours for Talesh station. According to principal component analysis results on Astara, Kiashahr and Talesh stations, five, five and four principal component were used in modeling these stations respectively. At first section input compound determined by gamma test to estimate Pan evaporation of the selected stations were used. Pan Evaporation estimation results shows that at Astara station GT-ANN model has less root mean square error than GT-SVM model and beter performance. Pan Evaporation estimation at Kiashahr station was done suitably with both models. At this station GT-SVM had a better performance with root mean square error of 1.295 compared to GT-ANN model with 1.356.

At Talesh station both models had close results but results for GT-SVM were more accurate compared to GT-ANN. Nash Sutcliffe coefficient attained for Astara and talesh stations acknowledges their  excellent results and for Kiashahr station shows the satisfactory results.

At second section modelling were done by using selected inputs by PCA preprocess method. Accordint to results, PCA-ANN model had better performance estimating pan evaporation at Astara and talesh stations than PCA-SVM model as its root mean square error was lower. Value of Nash Sutcliffe coefficient shows the suitable performance of both models at both stations. PCA-SVM model had better performance estimation pan evaporation than PCA-ANN with lower root mean square error at Kiashahr station. Nash Sutcliffe coefficient of PCA-SVM model was 0.666 and for PCA-ANN model was 0.634 which shows the satisfactory performance of both models.

Conclusions

Results shows the good performance of preprocessing methods (principal component analysis and gamma test). Actually performance of GT-ANN, PCA-ANN, GT-SVM and PCA-SVM models performance estimating pan evaporation of each one of the stations are very close to each other. This similarity is caused by performance of gamma test and principal component analysis preprocessing methods. Principal component analysis converts input variables to independent principal component using linear relation between input variables. Actually this method reduces the effect of the variables with similar information by giving them lower factor. But in gamma test method consider to gamma factor attained from various input compounds, variable that has a negative effect on output will be determined and eliminated from final input compound. As we said before, nature of none linear Gamma and linear PCA methods are different but when PCA method decreases the factor that is eliminated in gamma test to a small value, inputs determined by both methods will be close to each other. This can be one of the reasons that close the estimating models results to each other. So we cannot recommend one preprocessing methods better than the other. We can conclude that for estimating pan evaporation at these stations both preprocessing methods are suitable.

According to results PCA-ANN for Astara and Talesh and GT-SVM model for Kiashahr station had better performance than others.

Although both models had acceptable performance estimating pan evaporation of stations but SVM model results were better than ANN model.

---

Classification of high-dimensional hyperspectral data with many spectral bands for the derivation of good accuracy is an important problem in hyperspectral remote sensing .The most of classification algorithms are based on spectral information .Here, in order to achieve an high classification accuracy, we can use the spatial information of data. Integration of hidden morkov random field that optimize spatial information by minimizing energy functions, with support vector machine that is an powerful method for classification of hyperspectral data, can improve classification accuracy in final classified map  properly. The purpose of this study is to improve the classification accuracy with a limited of training samples by combination of support vector machine algorithm and hidden morkov random field. In this study, tow hyperspectral dataset from Hyperion and AVIRIS sensors has been used. After the applying radiometric corrections like correcting embedded lines and remove bad bands , atmospheric correction Hyperion dataset done by FLAASH method and AVIRIS dataset by IAR algorithm. MNF transformation was used in order to dimensionally reduction and the endmembers were extracted from PPI band and then in order to spectral classification, used from SVM method. Finally, to improve classification accuracy in the final classified map, hidden Markov random field (HMRF) was used. So that after the extracting of Components from PCA and MNF Transformations, computing of some statistic parameters of classes in SVM classified map in order to use in inputs model and so configuration of iterations, SVM-HMRF model was applied.

The results show that the proposed model (SVM-HMRF) has improved overall classification accuracy in both of data sets. For example, the improved classification accuracy on some of land uses, were around 25 percent. Also regions of final classified map is much more homogeneous and salt and pepper nose drastically reduced.

---

 
Introduction
groundwater forms part of the water cycle and is a reliable source of water for human consumption, as well as in Iran, most of the water used in the drinking, agricultural and industrial sectors is supplied from groundwater. Due to the condition of Iran, due to the deficit surface water resources, the use of groundwater resources for water supply has been considered.
 
Materials and methods
1) Use of the Radial basis function of the neural network
If a generalized regression function of the neural network, PNN / GRNN, is selected, all of the network weights can be calculated as probable. In RBF, a Gaussian transmission function is used which is similar to a ring (GRNN) One of the benefits of these networks is its rapid learning of other networks, including the multi-layered perceptron network of MLPs. The Gaussian networks of the transfer function network are of an unidentified learning type, but the output is a controlled learning type. The network is very practical in simulating hydrological and hydrological issues, due to its rapid training, generalizability and ease of use.
 
2) Use a support vector machine
  A support vector machine is proposed based on the principle of minimizing structural error. A support vector machine can be used both for categorization issues and for the estimation of functions. used a new error function called ε-insensitive for machine application in regression problems, so that this function ignores errors that are at a given distance from actual values. This function is defined as How to design a network.
In this study, the data used were 95 piezo metric wells in the Amol-Babol Plain. Data were used with a mean average of 30 years. In order to simulate the depth of the groundwater table, effective factors such transmissivity of aquifer formations, precipitation values and distance from water resources. For the design of the network, for both models, there are two classes of training and testing data. One important criterion for training a network is the number of repetitions or epoch during training. The higher the number of replays, the error decreases so that training data can be converted, which will increase the number of unsuccessful repetitions at that time. for network optimization purposes, the goal of network training is to reduce network error, which can improve the relationship between the input and output of the model. Due to the lack of specific rules for designing artificial neural networks (ANNs), various structures have been investigated to optimize the design. Select the number and type of input parameters for the model is important. For this reason, seven design input patterns are given. Which was carried out in the software of the NeuroSolutions.
 
Discussion of Results
for the optimal simulation model based on all parameters and the provision of all its input data will require a great deal of time and cost, a method based on the main parameters of input (optimal inputs) is modeled and validated. it was observed that the predicted aquifer level for both models is about to its actual value.
 
Conclusions
The results obtained with the Radial basis function and the support vector machine represent this point where the support vector machine and the radial function have the ability to have approximately the same ability to predict and modeling, although generally the results of the Radial basis function are more acceptable. The results of the model test are shown. As the results of the survey are presented, among the methods implemented in the model, using effective factors such as transmissivity of aquifer formations, precipitation values and distance from water resources to predict level of level The aquifer was used. The results of the test showed that the Radial basis function of the support vector machine with a correlation coefficient of 0.82 and a mean absolute error of 1.94 is an appropriate tool for prediction of water resource management.
 

---

The simulation of the rainfall-runoff process is a crucial step in water resources management, watershed management, water scarcity crisis, and flood control. The intrinsic complexity of the rainfall-runoff process, spatiotemporal variability, and the factors affecting it make the simulation with physical or hydrological models difficult. Therefore, metaheuristic approaches, such as support vector machines, gene expression programming, and artificial neural networks, have been widely used in hydrological studies, and generally, in the phenomena without definite relationships. Due to the provision of drinking, agricultural, and industrial water, the drainage basin of Aharchay, located in the northwest of Iran, has an influential role in the development of the region. This paper has evaluated the models of support vector machine, gene expression programming, and artificial neural networks for the simulation of the rainfall-runoff process in the drainage basin of Aharchay at the hydrometric stations of Tazeh Kand, Ravasjan, Oushdilaq, Barmis, Owrang, and Kasin. In order to determine the input combination of the models, a list of independent variables associated with the runoff of each station was prepared. Then the appropriate inputs were chosen using the two criteria of the Pearson correlation coefficient and partial mutual information. The input combinations obtained from each criterion were evaluated in the simulation of the rainfall-runoff of the Aharchay drainage basin in the hydrometer stations. The results indicated the reasonable accuracy of the models of support vector machine and gene expression programming, and the relative superiority of the artificial neural network. Moreover, in determining the input variables, the Pearson correlation coefficient provided the best results or was close to them.