clustering

Introduction and Goal:

The decreasing trend of precipitation in many regions of the country, including the northwestern provinces, has made it more necessary to investigate new methodologies to complete the databases of the country's watersheds. In this regard, understanding the changes in precipitation barycenters, as an emerging concept, that indicates the long-term spatial distribution of regional precipitation is especially important for the administration of watershed management issues such as reducing drought effect, flood control, and water resource conservation. To this end, the present study analyzes the precipitation barycenter and its relationship with the spatial distribution of the rain gauge network in Ardabil Province.

To calculate the precipitation amount, the statistics of 49 meteorological stations during the statistical period of 45 years (1971-2016) were used. Based on the principle of station extraction and considering the rationality of the lowest density and uniformity, five stations were determined as the minimum possible number. Then, densities of 10, 20, 30, and 40 stations were extracted randomly to compare the distribution of precipitation centroids with the density of the station network. The statistics above were used to implement the gravity center model, calculate the standard deviation ellipse in the ArcMap 10.8 software environment, and conventional correlation analysis on monthly, seasonal, and annual time scales using XLSTAT.

The results showed that in spring (April to June), the direction of movement of the precipitation barycenter was significantly different and the movement distance was 18.17 km. In summer (July to September), the precipitation barycenter mainly moved 20.18 km to the north. In autumn (October to December), the precipitation barycenter mainly moves 20.49 km to the south and has the longest movement distance among the seasons. In winter (December to March), the precipitation barycenter moves in different directions and has the smallest movement distance (8.35). The annual precipitation barycenter migrated mostly in the southeast direction. The maximum migration of barycenter for 1979 was towards the northwest with 111.78 km. Besides, the annual precipitation barycenter in Ardabil Province in the three decades of 1971-1981 migrated mostly to the northwest and in the other decades to the southeast. The stational network density was positively correlated with changes in the precipitation barycenter (CCA = 0.65 at a significance level of 5%).

Conclusion and Suggestion:

The highest conventional correlation coefficient of 0.80 was obtained for the density of 40 stations among different densities of 5, 10, 20, 30, 40, and 50, and the lowest coefficient of 0.04 was obtained for the minimum density of 5 stations. Based on the results of this study, changes in the movement of precipitation patterns that can be effective in drought and wet periods were identified. It is suggested that the relationship between climate change and elevation factors with spatial-temporal changes in the precipitation center of gravity be investigated.

Considering the role and importance of environmental factors in the distribution and spread of vegetation, as well as the existence of wind of 120 days in the Sistan region, and the necessity and importance of vegetation and flood spreading in the region, this research was done to classification and ordination of plant communities in Niatak flood spreading area, Zabol. Vegetation types were determined by Braun-Blanquet's physiognomy-floristic method and cluster analysis based on species composition and soil characteristics were used to classify vegetation. In each plant type, soil and vegetation sampling was done. In the key area of each plant type, 3 transects of 30 m were established, and during each transect, all surface features were measured. In addition, in each transect, soil samples were taken and physical and chemical properties were determined. The results of the preliminary studies obtained from the initial typology by the physiognomy-floristic method in 17 investigated types led to the identification of five plant communities, such as Ta. stricta, Ta. aphylla, Ha. ammodendron, Ha. persicum and Pr. stephaniana. In addition, the classification of vegetation by cluster analysis method based on species composition and soil properties was identified as five and three plant communities, respectively. The ordination results showed that the first axis and the second axis together with 64% of the variance of the distribution of plant communities are the most important axes. Sand percentage has the most positive correlation and absorbable phosphorus (Pav) has the most negative correlation with axis 1. Sodium (Na) and exchange sodium percentage (ESP) have the most positive correlation with axis 2. Redundancy analysis (RDA) biplot shows that characteristics, such as sand and fine sand percentage have a positive correlation with Ta. aphylla and Ha.ammodendron communities, and factors such as Pav, absorbable potassium (Kav), and cation exchange capacity (CEC) cause the distribution of Ta. stricta community in the region. Electrical conductivity (EC) and sodium absorption ratio (SAR) cause the presence of Ta. aphylla and Ha. ammodendron communities. There was a special relationship between different plant communities and soil characteristics.

Assessing precipitation alterations in a large area like Iran is required for the identification of those areas that are more vulnerable to changes in precipitation patterns, considering the fact that such changes may significantly influence water availability, agriculture, and other sectors that are dependent on water resources.  On the other hand, understanding the spatial variability of precipitation patterns can help develop purposive strategies, including drought or flood management in specific regions. Moreover, as severe weather events such as floods and droughts can devastate communities and their infrastructure, such an understanding can inform decisions made concerning disaster risk reduction efforts. Therefore, assessing precipitation variations is essential for the effective management of water resources and the reduction of disaster risks.
 

This study suggests a new method for analyzing precipitation properties in Iran, using a mixture of Maximal Overlap Discrete Wavelet Transform (MODWT) and Multiscale Entropy (MDE) techniques. This approach allows for a more detailed and accurate assessment of the spatial and temporal characteristics of precipitation properties, preparing the ground for the development of appropriate strategies for different regions in Iran. To this end, annual precipitation data collected from fifty Iranian synoptic stations for 1980-2020 were analyzed. Then, after classifying the precipitation data into different subseries, the concept of entropy was used to measure precipitation variability. Moreover, MDE values were used as input data for clustering purposes, followed by the calculation of internal evaluation criteria to be used for the determination of the optimal number of clusters and the most suitable clustering method calculated.
 

The variations and trends of the precipitation data can be identified through the analysis of partial coefficients D1-D4 and the approximation coefficient. Accordingly, while the smaller partial sub-categories indicate more rapid variations at higher frequencies, the greater partial coefficients show more moderate variations at lower frequencies. Moreover, the approximation coefficient reveals the slightest variations at low frequencies in annual time series. The study’s results suggested that northern and northwestern Iranian regions that are primarily characterized by rainy, cold, and in some cases semi-arid climates experienced the greatest variations in terms of annual precipitation. On the other hand, the eastern and southern parts of Iran, which are mostly dry areas, experienced more moderate variations in annual precipitation rates. Therefore, according to the results found in this study, it could be argued that northern and northwestern Iran enjoy more precipitation variability than other parts of the country. Furthermore, the D3 sub-category (eight years) was found to have the greatest variations in terms of MWE. On the other hand, based on the values of Sci, DBi, and CH index, the k-means clustering method performed better than SOM (Sci=0.41, DBi= 1.22, and CH= 14.58). Finally, fifty Iranian synoptic stations were categorized into four clusters based on the MWE index, with Chabahar, Shahrud, Abadan, and Zanjan being selected as core clusters.
 

The current study proposed a methodology for analyzing and zoning Iran’s annual precipitation based on the multiscale entropy method, considering the fact that the period and trend of the annual time series could be identified via the analysis of the precipitation series. Following the analysis of the collected data, this study used the multiscale entropy method to record precipitation variability in each synoptic station. On the other hand, the determination of SOM and k-means input data based on MWE values helped reduce the input data, leading to an increase in the accuracy of the zoning method and the selection of the homogenous clusters based on the proposed methodology. The results of the study indicated that generally, the K-means method offered more homogeneous areas than the SOM method. Moreover, the homogeneous distribution of annual precipitation variations obtained based on the K-means-derived clusters confirmed the positive performance of the methodology proposed in this study, which involves the identification of hydrologic uncertainty and temporal-spatial variations of precipitation in those stations related to a specific cluster. The methodology seeks to create distinct clusters of stations that share common features. This approach can provide insights into the spatial variability of hydrological processes and help improve water resource management by identifying areas susceptible to hydrological extremes such as floods or droughts. Some studies have already been conducted on the spatial clustering of precipitation stations in Iran using different methods, including the application of geographical proximity or precipitation rate as criteria for clustering. However, the method used in the current study involved clustering based on similarity in hydrological uncertainty and temporal-spatial complexity, which is consistent with the findings of Roshangar and Alizadeh (2019), and Roshangar et al. (2019). Identifying hydrologically homogeneous regions and their associated precipitation characteristics can significantly enhance the effective management of water resources in terms of adapting to climate change, preventing damage to water environments, and mitigating the impact of climate-related disasters. Therefore, the proposed methodology for spatial clustering of synoptic stations could be useful in managing water resources and all precipitation-related sectors and variables such as runoff and soil moisture. However, this study faced some limitations such as the small number of synoptic stations and the short length of the statistical period. Therefore, it is recommended that the methodology be applied to more stations at more varied time scales.

Gully erosion is one of the main sources of soil loss in watersheds and also one of the main causes of sediment storage in dam reservoirs and river alluvium. In addition, human activities can exacerbate this erosion, which has very worrying consequences. Therefore, the present study was conducted to determine the environmental thresholds of gully erosion and also the morphoclimatic classification of these areas in Kermanshah province.

To do this, first by collecting information from relevant authorities and also using aerial photographs 1: 20,000 and 1: 50,000, the main gully of the province that were visible on the pictures and then this information using field operations and specifying them on corrected on topographic map. Climatic zones of gully in the province were identified using existing maps based on the modified Domarten method. Then, in each climate, one to three gully were identified as targets, and in each of them, a representative gully was identified and two replications were identified, and the field identification of these gully was completed by field operations and field visits. These characteristics included location, land type, dimensions at the head of the gully and at intervals of 25, 50 and 75% of the top, length, profile and general plan of the gully. In addition, soil sampling was performed in the above sections and some physical and chemical properties of soil were measured in the laboratory. The classical statistical method of cluster analysis was also used to classify gully.

The results of this study showed that there are differences in the important morphometric characteristics and topographic threshold of the studied gully in different gully in the province and in different climates, which are mainly due to the climate, type of formation, topography and vegetation of this area. Classification of gully based on the classification method showed that with a similarity level of 97.5%, three general classes can be identified for gully.

gully areas in arid desert climates (warm regions) of the province in one class, gully zones located in the middle altitudes and cold semi-arid climate in the second class and gully zones in cold semi-arid climates to the Mediterranean colds are in the third class similar to gully. The active gully of the province are mainly located in the hill and plain lands. The results of this study show a picture of effective and influential factors in creating gully, based on which it is possible to apply how to control them in future planning and management of officials.

Sediment and erosion are two natural phenomena in watersheds. Due to irregular recording and sampling difficulties, daily data are not available for sediment records. Therefore, decision makers and researchers have to apply interpolating methods to estimated sediment yields. In this study, 30 watershed characteristics including physiography, geomorphology, vegetation, climate conditions in 69 watersheds located in the Karoon and Karkheh basins were used to statistical analysis. Based on the principle component analysis, eight characteristics including area, perimeter, river length, relief, mean of elevation at 85% upstream and 15% point of longest flow path and the number of landslide events were selected. Then using Cluster Analysis, six homogenous regions were identified and multiple regression models were applied. Due to constriction of large dames on the studied watersheds, access to the reliable data is a challenges for sediment yield analysis. Based on the sediment-precipitation double-mass curves 29 out of 35 stations were influenced by upstream dam. Results indicated that the effects of large reservoir dams can influence the downstream sediment yield along 98 Km of river length. The results show that in each group a particular combination of variables influence the sediment yields of the watersheds. According to the validation indices (NS and R2) the obtained models have the high performance (R2 = 0.71 and NS=0.72). In general, the physiographic characteristics of the watershed such as length, area, main flow path and relief are more important than other climatic, vegetation and geological factors. The total explain variance by the mentioned variables is 87.3%.

Analysis of suspended sediment load data in rivers is the basis for understanding the trend of erosion and sediment in the management and planning of soil and water resources. Due to lack of access to daily suspended sediment loading data with direct measurement, it is important to use methods for modeling and estimating it in watersheds. One of the best methods used in this field is the use of artificial neural networks. To evaluate daily suspended sediment load, Sira hydrometric station was studied in Karaj River watershed. The number of data used in this study included 624 information records of 31 years (1981–2011) statistical period .Input data to the artificial neural network models included instantaneous flow discharge, average daily flow discharge, average daily flow discharge with a delay of three days, average daily precipitation and average daily precipitation with a delay of three days. Output data to models was daily suspended sediment load. In this research, gamma test and genetic algorithm were used to obtain optimal variables and best combination of variables for entering the model. Then, these combinations with some combination of test and error variables were entered to artificial neural network models. The self-organizing map neural network was used for data clustering and all data were divided into three homogeneous groups: 70 percentage training data, 15 percentage validation data and 15 percentage test data. Then, the combination of variables entered to neural network models with activation functions log sigmoid and tangent sigmoid. The results showed that the neural networks using the optimal variable combinations in comparison with manual combinations have a more accurate estimate for suspended sediment load. In all combinations of inputs to neural network models, a model with tangent sigmoid activation function, with input variables combination including, instantaneous flow discharge (Q), average daily flow discharge (Qi), average daily flow discharge for two day ago (Qi-2), average daily flow discharge for three day ago (Qi-3), average daily precipitation (Pi), average daily precipitation for two day ago (Pi-2) and average daily precipitation for three day ago (Pi-3), was the best model for estimating daily suspended sediment load. This model has the lowest of error (MAE=500.05 (ton/day), RMSE=1995.33(ton/day) and Erel=7%), the highest accuracy (R2=0.96), the highest performance model (NSE=0.96) and has the lowest general standard deviation (GSD=0.97) compared to other models. Also, this model is the best combination with the most influential input variables derived from gamma test and genetic algorithm for estimating SSL.

The present study is aimed at estimating the suspended load discharged from watersheds in the Province of Hormozgan. The mean annual suspended loads were considered as the dependent variable and 21 parameters of morphological and climatic factors were selected as the independent parameters. The data were collected from 19 hydrometry stations for a 21-years period (1991–2011). This analysis is presented in two steps; a principal component analysis (PCA) was further applied to extract the underlying factors (principal component) and use the hierarchical clustering analysis in order to identify homogeneous groups. Multi regression was performed to identify the contribution of each variable. Different clustering scenarios were considered for selecting the optimal number of homogeneity groups, and their overall impacts on cluster validation indices were assessed. These three scenarios include dependent variable, independent parameters and the combined variables. The performance of the models was evaluated using the coefficient of determination (R2) and the Nash-Sutcliffe model efficiency. The watersheds were divided into two clusters and multiple regression models were derived for each homogenous region. Eventually, the models with coefficients of determination of 0.78 and 0.84 were used to estimate the mean annual suspended load. It is recommended to study more watersheds with different condition to reach the plurality and improve the quality of these models for estimation of the suspended load in the ungauged watersheds.

Correlation between vegetation and environmental variables is the most important issue related to the structure and function of rangeland ecosystems. Environmental factors such as climate, topography, and soil can alter the dynamics of plant communities. In this study, the effects of environmental factors on vegetation of some rangeland sites in Semirom region of Isfahan province were evaluated. Sampling was carried out at 52 rangeland sites in a random-systematic way. At each location, 40 plots of one square meter were placed along four transects, and the percentage of vegetation cover in each plot was measured using ocular-estimate method. The grid layers of bioclimatic and topographic variables were prepared in ArcGIS 10.1 and values for each range site were extracted. Soil samples were collected from range sites and some physical and chemical properties of the samples were measured. Redundancy Analysis (RDA) was used to evaluate the relationships between environmental factors and the plant species cover. Cluster analysis was also used to group the studied range sites based on environmental factors. According to the results, annual mean temperature, isothermality, mean temperature of wettest quarter, soil texture, pH, organic matter, and elevation were identified as the most important ecological factors affecting plant species cover.

The estimation of the correct amount of suspended sediment has an important role in the optimal design of water structures, erosion studies and water quality studies. The sediment rating curve (SRC) is a conventional and well-known regression model. However, due to logarithmic transformations in calibrating this model, its estimated values ​​are often less than actual values. In the present study, using the instantaneous flow discharge and suspended sediment load of Beytas hydrometric station in the Mahabad-Chai River, the SRC model was calibrated, and then using Non-dominated Sorting Genetic Algorithm II (NSGA-II), the coefficients of this model optimized again. This algorithm is an automatic procedure and can use different objective functions in the calibration process simultaneously. In this regard, in the calibration process of the model, four objective functions RMSE, MAE, NSE, and LOGE were used as pairwise combinations. According to the results of the model evaluation, the NSE and LOGE objective functions were selected as the best objective functions for optimization of the model. In order to increase the power of the model's generalization, the self-organizing map (SOM) neural network was used to cluster data and form two homogeneous data sets (calibration and evaluation sets) of 70% and 30% respectively. The results showed that the use of the NSGA II algorithm resulted in improved model efficiency so that the results are much more favorable than the other results of conventional SRC models (such as the rating curve of mean load within discharge classes, SRC models corrected by correction factors). In this regard, the error value (RMSE) of the test data set in the best model of the conventional SRC models was 383.65 tons/day, which was reduced by using the NSGA II algorithm to 102.94 tons/day. In sum, using the NSGA-II algorithm, we can optimize the coefficients of the SRC model, which is more efficient than the other conventional models.

Zayandeh Rood Basin has a vital role in Iran's poetry, biomass, agriculture, industry and tourism, faced with drought problems. Clustering approach can be a management approach to reduce drought risk impacts which groups the members with regard to the division based on the Euclidean distance of stations. In this research, the approach of determining the spatial-temporal distribution of drought clusters in watersheds is used to express variations based on precipitation precipitation index (SPI) parameters of stations, which depends on the probability of precipitation for any time scale. Since the maximum spatial distribution of the meteorological stations in the region and the maximum time period of the long-term and possible long-term statistical period were considered the same, the data of 26 stations from 12 years (2003 to 2014) was used as reference data. In this regard, the12 months SPI index was first calculated. Then, the 12-month SPI index, which ended in December, was used for cluster analysis of the SPI, and then 144 data were clustered into four groups. Further, zoning analysis was performed on data clusters. Then, the relationship between elevations as an effective landform factor in drought with SPI drought index cluster was investigated using correlation of variables. SPI correlation with mean height of each cluster stations was studied and the results were compared and analyzed. The results of the SPI drought fluctuation chart showed a very severe drought in 2008 and 2009 and 2010, and severe drought in 2010 and mild drought in 2003, 2005 and 2013. Also, 12-month SPI drought data showed a high and negative correlation with height data. Consequently, spatial-temporal monitoring of drought indicators clusters is recommended as a way to manage the impacts of drought risk.

Peak flow estimation is one of the major issues in water resources and flood management that have basic role in the design of hydraulic structures and biomechanics activities in basins. So that a proper assessment has a basic role in the success of administrative works. In this paper, using artificial intelligence methods (MLP Neural Network, the mixture of SOFM with MLP, the mixture of FCM with ANFIS) to estimate Yalfan River’s peak discharge in hydrometer local station. For these models, eight variables have been considered as the inputs that includes rainfall amount in the occurrence time of flood, rainfall of five days ago from occurrence of flood, curve number of the basin (CN), basic discharge and finally peak discharge are considered as the output. In the artificial intelligences after preprocessing of the data, the optimal structure of the models are determined with input and output data, evaluation criteria and trial and error. At the end, the MLP model had better performance compared to ANFIS+FCM, MLP+SOFM, GRNN models.

Estimating the runoff coefficient that is influenced by morphometric, geologic and hydro-climatologically factors are the most important issues in hydrology and information of its role in the planning and management of water resources is more important. Clustering catchments is the best method for the analysis of hydrological parameters in the absence of full coverage of hydrological data. In this research, twenty two hydrometric stations with common period from 1974 to 1999 were selected. Physiographic parameters of the catchments were extracted. Runoff coefficient was calculated and then base flow was extracted from using one parameter recursive digital filters. Lithological units using digital geological map, with the scale of 1: 250,000, based on expert opinion divided on two classes and area covered by each unit in each catchment were calculated. Factor analysis using 15 parameters were conducted. Catchments using independent factors in different hierarchy methods includes: nearest neighbor, furthest neighbor, median clustering, centroid clustering and Ward method were classified. Then, the regional equations using linear regression at 1% significant level were determined. To compare and evaluate the accuracy and efficiency of the models, independence errors, colinerity and normal distribution of error were tested. The results of factor analysis showed that all variables are to be classified in terms of five factors which 85.9% of the variance was included. Results of homogeneity showed that the basins in homogeneous methods of nearest neighbor, furthest neighbor, centroid clustering and median clustering, were all the same and classified in two groups with the similar components. The results of accuracy assessment showed that the accuracy of nearest neighbor methods was more accurate, and because of low relative error (25.4%) and MAE of 7.85 and RMSE of 9.62 was diagnosed as the best method for regional analyzing of runoff coefficient in the study area.

Morphometric analysis is considered as quantitative evaluation of geometric features landforms and landscape. In the study of basin tectonic features, the use of some morphometric parameters can provide very substantial information. Gavkoshak Basin with an area of 46.73 km2 is located in Zagros simply folded belt. The objective of this research is to use the morphometric indices such as hypsometric integral, basin shape, stream length-gradient, asymmetric factor, and the valley width/height ratio. As a result, these indices are converted to the tectonic activity index. This index can be used to assess the overall performance of the region's tectonic activity.Morphometric indices of area are studied by dividing this area into 43 sub-basins, using the Digital-Elevation Model (DEM) and Geographic Information System) GIS). Morphotectonics index method with the use of geographic information systems provide procedures and a powerful tool for estimating tectonic activity in the region. It should be borne in mind that the results of these indices can show different values in different sectors. By examining the relative index of tectonic activity, the basin in this research is divided into two parts namely active (25.6% of the watershed) and semi-active (74.4 % of the watershed). Through statistical analysis, the area under investigation includes four clusters: cluster one with 91. 83% similarity, cluster two with 95.19% similarity, cluster three with 96.18% similarity, and cluster four with 91.09% similarity. In this way, homogeneous regions were determined based on clustering algorithm ward. The active tectonic basin in this research can be studied in another research project, using other morphometric parameters.

The composition and amount of the soil seed bank will change by Management factor ecosystems, Understanding these changes can help explain vegetation change is affect resuscitation. For this purpose, this study examines the changes resulting from corrective action of planet non-native Atriplex canescens plant. Sampling in the three masses planting of Atriplex two, ten and twenty years in terms of environmental factors, and the homogeneity of similarities in places under the shrubs, between the shrubs were taken. In addition, a control area adjacent to each site where the plant had not been done to compare the sampling was done. In total 54 soil samples from the soil surface depth with approach Random – systematic before the start of the growing season Samples were taken and then soil samples were transferred to the greenhouse for seed germination. Data were collected species of grown and rank-abundance plot diagram For each of the sites studied, Traced and also Frequency distribution models Included Geometric, Log-series, Broken-stick and Log-normal On the data's Were fitted and Was selected The best model distribution. Also, Jaccard similarity index and divider clustering for different sites were calculated and plotted. The results of rank-abundance plot diagram Studied Places shows the Curve from control plots all sites have steeper from plot curve under and between shrubs has so species diversity of seed bank control Site is lower. Broken-stick model and Log-Normal conform to the treatment site 20, which represents a stable community with high diversity and evenness. The results also showed Jaccard similarity index the 10-year-old site under shrub and between shrubs are Maximum Likelihood (0.75) Soil seed banks and also in 20-years-old sites this similarity reaches its maximum. The overall results indicate that restoration of rangeland by planting Atriplex cause Increase the species diversity of soil seed bank in study area the of course the amount this changes depends on Duration of its implementation.

Self-Organizing Feature Maps (SOFM) are a variety of artificial neural networks that their applications in the areas of pattern recognition and data clustering makes them noticeable tools to perform regional flood frequency analysis (RFFA). In this study, ability of Self-Organizing Feature Maps for regionalization of Sefidrood watershed in order to perform regional flood frequency analysis using L-moment algorithm is assessed. Results of this study show that SOFMs may be used as an acceptable method for data clustering and regionalization of watersheds. Evaluation of values of cluster validity measures showed that they can’t be a determining factor to identify suitable number of regions for regional flood frequency analysis, but homogeneity of regions is main factor to determine desirable number of regions. According to homogeneity of regions and sizes of formed regions, regionalizations including two regions that formed by Ward’s algorithm and SOFM were chosen as optimum choices to regional flood frequency analysis on Sefidrood watershed. Furthermore, based on results of flood estimation by at-site FFA and two RFFA, regional estimates are very close to each other and their average relative difference is equal to 1% nearly. Also relative difference between regional and at-site estimates doesn’t exceed 17% in any station and its mean value is about 8%.

Droughts event can be effect on available groundwater and water resources in the many components that it cause deficit. This purpose, for the deal with droughts hazard for make decisions it is necessary to monitoring and assessing droughts feature in point of view concluding severity, geographic extending. In this study the hydrological droughts analyzed based on the clustering, surface water index and moving average methods in the 6 selected hydrometric stations in Ilam province over the 1982-2011. The results of the SDI showed that only occurred the moderate and slightly droughts states for the 3 and 6 reference periods. In the 9 and 12 reference periods, additionally we observed the severity droughts states in the three past decade. All of the stations experienced one wet year severity droughts in over the 1991 to 1996 period in the three past three decade. In contrast, all of the stations had been in the hydrological droughts event in over the 1996 to 2006. Hydrometric stations classified in three clusters based on surface water drought statuses. The results of clustering showed that Ivan, Holeylan and Tangsazin stations labeled in cluster 1 and Nazarabad station in cluser 2 and finally the Sarjooi and Dartoot stations classified in cluster 3. Ultimately, The Mann-Kendall and Linear Regression were applied to explore the relationship between hydrological droughts and river flow trends. The results of the trends test showed that all of the stations had decreasing trends.

The physiographical and climatological attributes of basins and maximum annual flood statics were obtained from 38 hydrometric stations of Mazandaran Province. The first, physiographical space based interpolation method was used for regional flood analysis with 10, 20, 50 and 100 year return periods. The Canonical Correlation Analysis was used to design physiographic space and regional analysis was obtained application of three interpolation methods include, ordinary Kriging, simple Kriging and IDW in the physiographic space. The results showed at-site flood quantiles have maximum correlation spatial in the physiographic space and their spatial structures following from Gaussian semi variogram. In order to evaluated the accuracy results using L-moments method for regional flood frequency analysis. In this method Ward clustering has been used to the determination of homogeneous regions. Homogeneity areas were adjusted by using homogenous and discordancy tests. Using goodness-of-fit Z DIST the Generalized Logistic distribution was selected for all clusters as the best regional distribution and regional estimates obtained by parameters selected distribution. Finally, performance methods evaluated using Jack Knife procedure and the five statistical indexes BIAS, BIASr, RMSE, RMSEr and NASH. Both methods give the same and desirable performance based on NASH; however results of other indexes showed that physiographical space based interpolation method provide better performance than linear moments and improving estimated by increasing the return periods, while the L-moments offers better performance for short return periods. The results of interpolation indicated that geostatistics estimates have more accuracy than deterministic estimates and ordinary Kriging has best performance.

Nowadays, the accurate estimation of rivers suspended sediment load (SSL), from various aspects, such as water resources engineering, environmental issues, water quality and so on is important. In this regard, because of various roles of fixed and dynamic variables of watersheds, the watershed hydrological models have not showen a proper efficiency in statimation of SSL. Also, the most SSL studies are based on only flow discharge variable whereas the results of the present study have proved that the efficiency of these modeles is very poor. On the other hands, the parameters such as rainfall type, year seasons and flow hydrograph shape have important role in watershed sediment yield that were ignored in the most SSL simulations. In the present study, multi layers perceptron neural network and hydro-meteorological data (daily flow discharge, suspended sediment concentration, daily rainfall and temperature) of Karaj dam watershed in a 30-year period (1981 to 2011) were used to estimate daily suspended sediment concentration of Sierra station. Due to the role of seasonal changes and flow conditions in sediment yield and sediment transport of the watershed, based on rainfall regime, hydrograph condition and runoff type, the data used in this study were first seperated into 5 groups and then for each group, a separate model was designed. In order to increase the generalization ability of the neural network models, self-organizing map (SOM) and Silhouette coefficient were used for data clustering and determination of the optimal number of clusters respectively. The research results showed that the use of daily precipitation and temperature variables along with flow discharge and data separating based on watershed time and hydro climatic conditions has had an important role in increasing the accurate estimation of the river sediment. In this regard, among the models, the maximum calculated error is when only a single model is used for all year seasons. The research results and its used structure can be used as a pattern to estimate and to forecast many environmental variables of watersheds.

The physiographical and climatological attributes of basins and maximum annual flood statistics were obtained from 38 hydrometric stations of Mazandaran province with minimum and maximum statistical periods of 13 and 56 years. The first، homogenous regions based on neighborhood approach were determined using canonical kriging method. The Canonical Correlation Analysis was used to design physiographic space using the six physiographic variables affecting the flood and two hydrological variables. The Gaussian model had the best fit to semi-variogram model in 10، 20، 50 and 100 year return periods، and regional estimates obtained using ordinary Kriging technique based on physiography space. The accuracy of results was evaluated using L-moments method for regional flood frequency analysis. In this method، Ward hierarchical clustering has been used to the determination of homogeneous regions. Homogeneity areas were adjusted by using homogenous and discordancy tests. Using goodness-of-fit Z DIST the Generalized Logistic distribution was selected for all clusters as the best regional distribution and regional estimates obtained by parameters selected distribution. Finally، performance methods evaluated using Jack Knife procedure and the five statistical indexes BIAS، BIASr، RMSE، RMSEr and NASH. Both methods give the same and desirable performance based on NASH، however results of other indexes showed that canonical kriging method provide better performance than linear moments. Increasing the return periods will improve the estimation، but the L-moments offers better performance for short return periods.