فصلنامه محیط شناسی
سال چهل و ششم شماره 4 (پیاپی 96، زمستان 1399)

The Karun River is the biggest river basin in Iran, which supplies water demands of about 16 cities, several villages, thousands of hectares of agricultural. This river polluted because of domestic and urban sewerage, industrial sources, and irrigation of agricultural land, Hospital sewage and high tide level of Persian Gulf.Therefore, because of the importance of this river, the water salinity of this river is determined in this study. The traditional methods of determining water salinity are costly in comparison with remote sensing methods. In the present study, Landsat 8 (OLI) data was used to calculate the water salinity map for Karun River since not only it is free, but it also has an acceptable resolution.

Landsat 8 (OLI) images were used to calculate reflectance for a pixel and were attained from (US Geological Survey (USGS) 2019). First, radiometric correction was applied to normalize satellite images. This process convert Digital Number into radiance. Second, in order to attain the surface reflectance values, the process of atmospheric correction was applied using the Fast Line-of-sight Atmospheric Analysis of Spectral Hypercubes (FLAASH).Water salinity was calculate by Iran Water and Power Recourses Development Company. Eight stations are located in the crucial point for EC measuring ALIKALE, GOTVAND, SHOOSHTAR, SHOTEYT, GARGAR, DEZ, AHVAZ, and ABADAN. Iran Water and Power Recourses Development Company obtained 102 observed EC samples from June 2013 to July 2018 along the Karun River. The Support Vector Machine was classically used for classification, Support Vector Classification, but extended for using along with regression issue, namely Support Vector Regression.The results related to the quality of the SVR depend on some factors: the loss function Ɛ, the error penalty factor C and the kernel function parameters. Usually, radial basis kernel function (RBF), k(x, x΄) = k(x,x΄)=exp⁡〖( -||x-x΄〗 2/σ^2), has been used in remote sensing studies, so, it is implemented in this study. Finally, the Genetic Algorithm (GA) is employed to optimize some parameters including C, Ɛ and σ.GA is an optimization technique create by Holland (1975) and discussed the mechanism of GA in solving nonlinear optimization problems.Besides, the Genetic Algorithm (GA) was applied to determine the best performer bands combination. Furthermore, we employ GA to optimize SVR parameters and number of layers and neurons of MLP neural network in order to maximize model accuracy.

Salinity intrusion is a complex issue in coastal, hot, and dry areas. Currently, remote sensing techniques have been widely used to monitor water salinity changes, ranging from inland river networks to deep oceans. The Karun River basin, with a basin area of 67,000 km^2, is located in the southern part of Iran. The salinity of Karun River has been increasing due to some critical factors, e.g. severe climate condition and regional physiography, industrial sources, domestic and urban sewerage, irrigation of agricultural land, fish hatchery, hospital sewage, and high tide level of Persian Gulf .This study aimed at building Support Vector Regression (SVR) and Multilayer Perceptron (MLP) models to realize the salinity intrusion through the relationship between reflectance from the Landsat-8 Operational Land Imager images and salinity levels measured in situ. 102 observed samples were divided into 75% training and 25% test. Besides, the Genetic Algorithm (GA) was applied to determine the best performer bands combination. Furthermore, we employ GA to optimize SVR parameters and number of layers and neurons of MLP neural network in order to maximize model accuracy. The result showed that the MLP approach was the better model to estimate water salinity along the Karun River network in the study area, which coefficient of determination (R^2) and RMSE of test data is obtained as 0.73 and 390μs 〖cm〗^(-1).GA analysis proved that bands 1, 2 and 3 are the best for modeling water salinity. In this study, the GA is used to determine the SVR meta-parameters including the loss function Ɛ, the error penalty factor C and σ parameters, which are obtained to be〖1×10〗^(-9), 1099 and 0.96, respectively, and number of layers and neurons of MLP neural network, which are obtained to be 5 and 35, respectively.The result showed that the MLP approach was the better model to estimate water salinity along the Karun River network in the study area, which coefficient of determination (R^2) and RMSE of test data is obtained as 0.73 and 390μs 〖cm〗^(-1).

The present study calculated the relationship between reflectance retrieved from Landsat-8 OLI and water salinity in the Karun River. SVR and MLP models had acceptable operation by considering the large size, geographic complexity of the study domain and the wide range of field data that change between 385 and 4310μs cm^(-1). Augmentation field data is the critical priority work for future study to probe the relationship between water salinity and satellite images.In addition, the contribution of thermal bands can help to increase accuracy of models. Salinity intrusion is a complex issue in coastal and hot and dry areas. Currently, remote sensing techniques have been widely used to monitor water salinity changes, ranging from inland river networks to deep oceans. The Karun River basin, with a basin area of 67,000 km2, is located in the southern part of Iran. The salinity of Karun River has been increasing due to some critical factors, e.g. severe climate condition and regional physiography, industrial sources, domestic and urban sewerage, irrigation of agricultural land, fish hatchery, hospital sewage, and high tide level of Persian Gulf .This study aimed at building Support Vector Regression (SVR) and Multilayer Perceptron (MLP) models to realize the salinity intrusion through the relationship between reflectance from the Landsat-8 Operational Land Imager images and salinity levels measured in situ. A total of 102 observed samples were divided into 75% training and 25% test. Besides, the Genetic Algorithm (GA) was applied to determine the best performer bands combination. Furthermore, we employ GA to optimize SVR parameters and number of layers and neurons of MLP neural network in order to maximize model accuracy. The result showed that the MLP approach was the better model to estimate water salinity along the Karun River network in the study area, which coefficient of determination (R2) and RMSE of test data is obtained as 0.73 and 390μscm-1

Permanent or temporary migration has always been one of the most important strategies adopted by human societies and individuals in the face of ecological or man-made disasters. However, our knowledge about complex relationship between ecological change as cause and migration as effect remains limited. Migration due to drought and climatic change is one of the examples of ecological migration. Drought in Iran is one of the most important climatic hazards as its consequences are evident today in various sectors such as water resources, environment, drying of wetlands and drying of lakes in different parts of country. Monitoring of drought in the past years, can increase our understanding and awareness about climate change and drought and would improve insights, predictions and future planning. Standard Precipitation Index (SPI) is one of the most well-known and widely used index to monitor drought in any scale.Investigating the relationship between ecological factors and migration would be effective in increasing the understanding of planners and decision makers. Iran internal migration statistics in the last three decades shows that the northwestern counties of the country almost have a negative net migration rate. Therefore, the aim of this paper is to monitor drought and investigate the relationship between drought and internal migration in the northwestern of Iran (Fig.1).

In order to achieve the goal of the research, the research process was followed in two separate sections. In the first part, migration studies, raw data of the internal migration matrix in county scale were received from the Statistics Center of Iran. By processing internal migration data, the net migration rate was prepared for each of the counties in the northwestern region of the country. In the drought studies section, initially, representative synoptic stations in the study area were selected. In selecting these stations, having 30 years of continuous statistics and appropriate spatial distribution in the study area was considered. Precipitation data of selected stations were obtained from the Meteorological Organization of Iran and were initially processed. Then SPI index was calculated using Meteorological Drought Monitor (MDM) software. After that, the shapefile of the administrative areas of the study area at the county level was taken from the Ministry of Interior of Iran. Then, the values of the standard precipitation index along with the values of the net migration rate were entered into the GIS environment and attributed to the relevant counties. By using SPI values, the study area was interpolated using thiessen polygons and Inverse Distance Weighting (IDW). Then thiessen polygons were converted to raster format and then the values of each county were calculated using the majority function of the zoning statistics tool.After this phase, spatial analysis techniques were used to investigate the effect of drought on internal migration in the study area. In this regard, the existence of autocorrelation in dependent and independent variables data were first investigated using the Moran’s I in GIS environment. Since the existence of autocorrelation and clustering pattern was evident in the data, Geographical Weight Regression in GIS environment was used to analyze the relationship between drought and internal migration. Fig 1. Location of the study area in Iran

Drought Monthly SPI showed that in the past 30 years, the frequency of dry months has been the lowest in the southern areas of Lake Urmia and increased to the central areas, so that the northern part of the region has the highest frequency of dry months. Based on six-month SPI, different pattern of the drought frequencies was seen in the study area. In this regard, Urmia and Miyaneh synoptic stations had the most drought frequencies.The one-year SPI also showed that Mahabad station and surrounding areas had the lowest drought frequency (2 years). This station had a similar pattern in the monthly and six-month SPI. Among the remaining eleven stations, Ardabil, Khoy, Jolfa, and Tabriz have experienced 6 years of drought and Miyaneh, Mako, Parsabad and Maragheh have experienced 5 years and Khalkhal, Ahar and Urmia have experienced 4 years. One-year SPI revealed that the pattern of drought distribution extends from the southwest, i.e. Mahabad station, to other geographical directions. Migration In the last three decades, only 26 percent of the region's counties have had a positive net migration rate. In fact, less than one-third and 74 percent of study area counties have a negative net migration rate. The spatial pattern of distribution of counties with a positive net migration rate is mainly concentrated around Lake Urmia, and this spatial pattern is well visible in the last three decades. Relationship The overall results revealed that there is a relatively low coefficient of determination (average 21.5 percent) between the independent and dependent variables. However, this amount of explanatory is not far from expectation because many variables are influential in decisions leading to migration and this amount of explanatory seems to be significant for the drought variable.

Ecological migration is one of the issues that has attracted the attention of various researchers due to the extensive changes in the ecological context that made by humans and sometimes by natural processes. Drought is one of the ecological factors that can cause population movements, especially in arid and semi-arid regions. Although drought is not an unfamiliar phenomenon for arid and semi-arid climates, climate change and excessive use of surface and groundwater resources have intensified its impact. Migration is affected by many factors and understanding drought as one of the migration causes is very complex. The Relationship between drought and migration in the study area indicates a significant but relatively weak relationship. This is primarily due to the nature of migration, which is influenced by various economic, social, cultural, political and environmental factors. Also, drought has several dimensions that in the present study only its climatic dimension has been studied. The SPI revealed that little drought has occurred in the study period. So weak relationship can also be due to the timely distribution of precipitation, as the region has received average precipitation, but this has not happened in the growing season. This adds to the complexity of this relationship. However, the amount obtained R2 is significant given the nature of the migration.

Landscape fragmentation reduces the patch area of internal habitat, hinders the operating and regulating ability of normal landscape ecological processes, and damages ecological corridors. Therefore, connecting isolated broken ecological patches and stepping stones through potential corridors within the borders can improve the impact of fragmented landscapes on biodiversity and the connectivity of landscape and promote the exchanges of genetic material and species between patches, which would effectively improve the service functions of natural ecosystems and have an important ecological significance. Basically correct landscape pattern requires ecological network and ecological system. Ecological network helps planners to increase the landscape connectivity between habitat patches. Network optimization is mainly based on the improvement in network connectivity, including the optimization of corridors and nodes. The optimization of corridors mainly refers to the increase in the number of corridors and the repair of ecological breakpoints in the corridors based on the degree of connectivity. Corridor connectivity should be increased in areas with low landscape connectivity. In recent years, the morphological spatial pattern analysis (MSPA) approach, which mainly focuses on structural connectivity, has been increasingly applied in ecological network analysis. This model is mainly used for the analysis of structural connectivity and can be used to accurately distinguish between landscape types and structures. The MSPA method applies four parameters, namely “connectivity”, “edge width”, “transition” and “intext” to classify landscape. Landscape connectivity can be used as a quantitative indicator of how facilitating a source landscape patch is for species migration, as a high degree of connectivity facilitates biodiversity protection and the maintenance of landscape ecological functions. The connectivity of the landscape and the importance of the various landscape patches to landscape connectivity can be reflected under graph.In northern Iranian provinces like Gilan province, cities have experienced irregular and horizontal urban sprawls during recent decades due to the existence of Hyrcanian Forests, special climatic setting, presence of green areas and adjacency to Caspian Sea, high population density, and the development of economic activities across the region. As a result of land-use change, urban growth and land degradation, the distributions of some terrestrial species have changed in recent years. Phasianus colchicus is one of the focal species in this region. Dispersal distance, which is species specific, is a critical process determining the distance threshold. The maximal dispersal distance of the Phasianus colchicus 3.2. The species prefers forests with canopy cover of 5–25% because these forests are largely covered by shrubs and bushes, which common pheasant use as a refuge. Pheasants live out their lives within a home range of about one square mile (640 acres), requiring all habitat components (nesting cover, brood habitat, winter cover and food plots) to be in close proximity. Ideally, a minimum of 30-60 acres (about 5-10 percent) of this range should be nesting cover. Larger blocks of cover are preferable to narrow linear strips. In this study, seeking to make a more comprehensive assessment of landscape connectivity, the core habitats and corridors will be identified according to the habitat type and dispersal distance of the focal species.

The study area in this study is located in the two watersheds of Lahijan Chabaksar (49 12 to 5005 E, 37 07 to 37 25 N) and Astaneh-Kuchesfahan (5021 to 50 26 E, 37 02 to 37 06 N), in the east and center of Gilan province, respectively. In the first step to classify the land cover in this study, the total Landsat 8 images in the period 01/01/2019 to 31/12/2019, which had a cloud cover below 10%, were used. Then, using Google Earth Engine and the products and instructions of vegetation index (NDVI) which related to the four seasons in 2019, urban lands, tree canopy cover to identify forest areas with trees height above 30 meters and finally the data removed from the ground and entered into the system by the user Land cover was classified into eight categories: forest land, rangeland, farmland, water, residential area, and tea farmland, garden and open space. According to the classified map of NDVI and land cover index and finally the identification of rangelands, gardens, forest lands with canopy cover less than 30%, agricultural lands and tea cultivation on the one hand and on the other hand considering the minimum area, elevation (Less than 1200 m above sea level) and slope (low to medium) required for the habitat of this species, the habitats of pheasant species in the region were identified. Then, MSPA analysis was used to form the ecological network and obtain core area. So forest land is extracted to be the foreground, and other land as the background, a series of image processing methods are used to divide the foreground into seven non-overlapping categories (namely, core, bridge, edge, branch, loop, islet and preformation), and then categories that are important for maintaining connectivity are identified, which increases the scientific nature of the ecological source and ecological corridor selection. The level of landscape connectivity in a region can quantitatively characterize whether a certain landscape type is suitable for species exchange and migration, which is of great significance for biodiversity protection and ecosystem balance. In this study, in the aspect of landscape connectivity evaluation, the integral index of connectivity (IIC), the probability of connectivity (PC), the delta of PC (dPC) and the delta of IIC (dIIC) are commonly used as the important indicators of landscape pattern and function, which can reflect well the degree of connection between core patches in the regional level and are calculated by Conefor 2.6 software. As the dispersal ability of different species varies, we assigned the dispersal distance 3.2 km and ring-necked pheasant, respectively. Finally, the top 8 patches with value of dPC above 4 were chosen as the most important habitats. The using least-cost path the corridors between them were determined. The least-cost path is often used to optimize a grid module. The resistance value of a grid describes its facilitating or impeding influences on dispersal processes of species. The resistance value is attached to each land cover unit to calculate the connectivity between two habitats (Table 1). The least-cost path model makes it possible to calculate the minimum cumulative link (corridors) between the target patch and the nearest source patch (habitat). We calculated the path of least resistance for the organism to migrate along and obtained the potential corridors between source patches using the “cost path” analysis in ArcGIS. The different resistance values of each land cover class were the key factors affecting the result.

Migration to cities and urban development have led to the irregular growth of cities and the uncontrolled transformation of natural land cover into artificial and impenetrable cover. As a result, it has created numerous environmental consequences for cities, including the phenomenon of heat islands, as a result of which the temperature of urban areas has increased compared to the surrounding areas, causing changes in ambient temperature, air pollution and harmful effects such as greenhouse emissions. Therefore, measuring and controlling the effects of urban heat islands, based on scientific and justifiable principles, helps decision-makers to overcome the resulting problems. Today, one of the most effective ways to control the effects of heat islands in developed countries is to use less heat-absorbing coverings, such as green infrastructure, high-albedo materials and flagstone, to cover the roofs of buildings. Therefore, in this study, an optimal planning based on spatial analysis, using the remote sensing and computational intelligence in the form of a spatial decision support system that can determine the effects of changing the roof covering of buildings in the study area.

To survey the research, a neighborhood from a central region of Tehran, the 7th region, was chosen to develop a software package for green roof planning. It is expected that the UHI effect has a significant role in this neighborhood since the region that the neighborhood belongs to, is one of the central regions in Tehran. Moreover, for developing the software package, map of parcels with attributes related to the area and land use and Images of Landsat 8 over the neighborhood are employed. Two main groups have a pivotal role in calculating UHI indices including vegetation and urban groups. When the indices are developed, the relationship between UHI and indices is investigated using the linear regression method (LRM) to obtain indices’ coefficients. Afterward, the software package tries to find some parcels, which constitute a certain and predefined percentage of area, that have a significant impact on UHI’s standard deviation by changing their roofs’ covers into three types cover including green, high albedo materials, and flagstones as the novelty of the research. Since there are a lot of feasible solutions, it is necessary to use a metaheuristic algorithm for finding the optimal solution. Therefore, in the second step of the proposed method, the optimal solution is conveyed by the Genetic algorithm (GA), as the most common algorithm in metaheuristic algorithms. After finding the optimum parcels for changing roofs’ cover, the UHI effect is computed once again to show the improvements.

As mentioned, the Genetic Algorithm is used to select the optimum subset of parcels for changing their roofs’ infrastructure with three covering classes including vegetation, high-albedo materials, and flagstones. This subset is assumed as 10 percent of all parcels in the area. However, some parameters should be set before using the algorithm such as the number of population and generation, the ratio of selection, crossover, and mutation. Besides, minimizing the standard deviation of SHI values was assumed as a fitness function for GA. As a result of the algorithm, the selected parcels and their appropriate roofs’ infrastructure for minimizing the standard deviation of SHI in the area are presented. This optimal solution was obtained through 252 generations that its convergence trend is presented. Additionally, based on the changes of selected parcels for roofs’ cover, the SHI values for the study area are computed again. These new values for the SHI and UHI effects are presented. However, the obtained standard deviation of SHI values for the changed roofs’ cover is 10.781°C while this value before changes is 13.222°C.By examining the selected parcels obtained from the GA results with green spaces in and around the study area, it is found that the GA selects parcels for changing the roof covering with vegetation that is not contiguous with the green spaces in their surrounding area. whereas, according to results, the GA did not choose any parcel in these areas to change their roofs’ infrastructure to vegetation cover. However, highly efficient covering in SHI values such as vegetation and high albedo materials circumscribed the study area. This fact shows that in order to control the variation of UHI in the center of the area, it is necessary to curb the SHI values in the border of the study area. However, the less efficient cover compared to vegetation and high-albedo materials, which is flagstones, are located dominantly in the center of the study area since their influence is more limited and local than the other types. It is also can be perceived that all changes in roofs’ infrastructure are not in line with changing to the vegetation cover, although this type of covering has the best effect to reduce SHI value. This consequence is because of the fitness function of GA, which is based on the standard deviation and not the mean value. The type of vegetation for covering decreases the SHI value, and thereby leads to decreasing mean value, while the objective of the software is to minimize the variation of SHI values. Therefore, vegetation cover is used in a location where the study area confronts with hotspot SHI value at that location. To verify this claim, the vegetation cover is utilized for all parcels selected by the GA to compute the SHI value for this scenario

With the widespread growth of cities and the increase in population, natural covers have been changed to artificial and impenetrable land cover, which lead to several environmental problems for cities, including UHI effects. Due to these changes, which are caused by the UHI effects, the temperature of urban areas becomes higher than the surrounding areas. One of the most practical and efficient methods for controlling the effects of urban heat islands is utilizing the green infrastructure and high albedo materials for roofs’ infrastructures; however, previous studies did not model this subject in quantitative practice. Based on this shortcoming, the present study proposed a software package to investigate quantitatively the changes of UHI effects based on the substitution of present roofs’ infrastructures to three selected types of covering class including vegetation, high-albedo materials, and flagstones. Additionally, the software used GA as a sub-model of the software to select the best set of parcels in the study area for changing their roofs’ infrastructure according to a specified fitness function. The fitness function assumed for this research is the standard deviation of the SHI values in the study area. This fitness function controls the variation of the SHI values and prevents the creation of UHI hotspots in the study area. . Examining the selected parcels obtained from the results of the GA with the surrounding green space and the study area, it was found that the genetic algorithm selects parcels to change the roof covering with vegetation that is not adjacent to the green space. This fact shows that in order to control the change of UHI in the center of the region, it is necessary to limit these values at the border of the region. However, less efficient cover compared to vegetation and high albedo materials, which is flagstone, is predominantly studied in the center of the area. It can be also seen that not all changes in roof infrastructure are consistent with changes in vegetation, although this type of cover has the best effect on reducing the amount of urban heat islands. This result is due to the fitness function of the genetic algorithm, which is based on minimizing the standard deviation of SHI in the area. Therefore, the vegetation is used in a place where the study area is exposed to a high amount of urban heat islands. Additionally, this cover type is more effective in the range of 100 and 150 meters of green areas.

Landscape fragmentation is one of the consequences of the increasing socio-economic pressures that many parts of the world are facing today. This situation is more likely to occur in areas where socio-economic development leads to increase communication networks. Qazvin province has become more sensitive due to its proximity to the city of Tehran and the role of the province as a socio-economic passage through the northwestern connection of the country to the center passes, as well as its role as a bio-corridor on a national scale. Also, the existence of 5 areas under the management of Department of Environment in Qazvin province has made Qazvin as interesting landscape for researchers. Various approaches and methods have been developed to quantify fragmentation in landscape ecology studies as Landscape division, splitting index, and effective mesh size. These methods are based on the ability of two randomly chosen animals, which placed in the different areas in the landscape, to find each other. in other word, the chance that two randomly chosen places in a landscape will be found in the same patch type. Effective Mesh Size (EMS) is one of the most widely used landscape metrics in the worldwide and was first developed by Geager (2000). The aim of the present research was to investigate the new Cross-Boundary Connectivity (CBC) methodology using GIS software to measure the Effective Mesh Size (EMS) that presented by Moser et al. (2007). this method eliminates marginal effects, which has been evaluated in Qazvin province with emphasizing on protected areas that is being implemented for the first time in Iran. On the other hand, in order to make an optimal comparison, the results of the CBC map and the traditional methods were compared. Also, we evaluated protected areas by overlaying them with the results of EMS. Finally, the typological analysis of the province has been performed. This measurement can give more accurate results than similar indices due to the elimination of the marginal effects in the calculation.

In this study, in order to calculate the Effective Mesh Size (EMS) index, Qazvin province was divided into 1354 study sample units. The size of each unit is 1100 hectares. In addition to the sample units, Implementing the EMS calculation model requires selecting the type of landscape elements that have been disconnected. To distinguish landscape elements, Land Use and Land Cover (LULC) map have been prepared. LULC types, which provide the resource for biodiversity needs, such as agriculture, rangelands, and etc., have been identified as potential habitats. LULC types, which have led to destruction of habitats or limit wildlife moves, have been identified as material and energy flow barriers. To apply the model, the Cross-Boundary Connections (CBC) method has been used, which presented in Moser et al. (2007), to consider the area of all patches located wholly or partially in the reporting unit (hexagon), as well as, the area of some patches spread beyond the reporting unit borders. The result of multiplying patches within the reporting unit and the total area of the same patches have been divided by reporting unit area to calculate EMS.Finally, by overlaying the results of the EMS with the Qazvin protected areas, the qualitative status of the protected areas has been evaluated.

The size of habitat patches within the sample of study, as well as the size of the same patches, regardless of the boundary of the sample of study, are the two main indicators used to calculate the Effective Mesh Size (EMS) index. Therefore, size of the habitat patch inside the sample of study shows the amount of human impact and change in the landscape structure. The result shows that 26 sample areas are in the very high level and have suffered the most damage and structural change by humans. in other words, those areas that have been severely fragmented by communication network, urban and industrial development. 15 samples are in the high level and 99 samples have been changed on average (medium level). Low and very low levels have occurred in 179 and 620 units, respectively.According to the results of the LULC map, 96% of the case study is covered by potential habitats, some of which are natural, such as forests, pastures, semi-deserts and riverbeds, and some man-made ones such as irrigated agriculture, dry farming and groves. The northern and the southern strip include most natural and the central plain includes most man-made habitats. The results of the EMS showed that 310 and 425 samples are in the very low and low range of EMS respectively. The reason of these values is man-made barriers as communication networks that are mostly located in the center of Qazvin plain. Many studies including Zebardast et al. (2011), Girvetz et al. (2007) and Pătru-Stupariu et al. (2015), mentioned that communication networks have high effect in fragmentation and the most important reason for the low EMS. 352 samples have a moderate value, most of which have been covered by traditional agriculture and areas under average human influence. 153 and 115 samples have high and very high values of EMS respectively. These samples have been covered by continuous rangeland and semi-desert that are far from centralized human developments.Meanwhile, Alllah-Abad hunting prohibited region, with only 4% fragmentation, has the smallest fragmentation among the protected areas in the Qazvin province. Alamoot Protected Area with 22% of medium and high fragmentation, is in relatively good condition. Also, the Abgarm and Avaj hunting prohibited region has more than 50% of the units with high fragmentation. The Tarom and Bashgol protected areas with 92% and 96% of the fragmented units respectively, are in poor condition. The last three zones are fragmented more than others. Thus, to compare the protected areas in Qazvin province, Allah-Abad hunting prohibited region has the most favorable and Bashgol protected area has the most unfavorable situation in terms of EMS index.

The results of this study showed Qazvin province’s landscape has been fragmented in the center, where the communication network has developed. We highlighted the need to pay attention to ecological process and the matter and energy flows in the center of the study area. Protected areas should have a high EMS index due to their nature.This study also considered the role of communication network on habitat fragmentation, which is emphasized the attention should be paid to the issue of habitat fragmentation before roads and railroads projects are implemented. Also, by ranking the protected areas based on the EMS index, it was founded that some protected areas are severely fragmented, and special attention needs to be paid to these areas in management programs.The overall results of this study can be used for planning and protecting the biodiversity and identifying the new protected areas or also changing the protection levels, in addition, it can be used for land use planning and regional planning at the upper province level.

In 1958, UNESCO announced the first definition of literacy. This definition evolved in 1978 and 2005, particularly when the “plurality of literacy” could expand its perspectives. Environmental literacy is an example in plural literacy. It is the main objective of environmental education based on sustainable development goals (SDGs). The objective of environmental education is to improve the awareness, incentives, commitment, and skills of citizens to rationally deal with environmental challenges. In a nutshell, environmental literacy is made of three main pillars of knowledge, attitude and behavior that should be learnt continuously and in long-term. Knowledge represents basic information learnt by education or experience. Attitude points to the sensitive or sensible perspectives about a subject. Hopes, frustrations and values can be included as attitude. Behavior is an index for rational actions carried out in specific conditions.Water literacy can similarly include these three pillars (Figure 1). Water management and saving methods in urban societies should be educated for citizens, particularly in arid areas. Isfahan is a city with dramatic water scarcity. However, there is a lack of knowledge about the level of water literacy in this society. Therefore, this research calculates the water literacy of citizens living in Isfahan, with population more than 2 millions, based on field surveys and statistical analysis.
Figure 1. The three pillars of water literacy

This research used questionnaires and field surveys to collect the required data from citizens in Isfahan City. The population of the study consisted of all citizens of Isfahan having more than 20 years old. Therefore, the sample size based on Kukran formula was 384 citizens. Accordingly, the whole questionnaires were totally collected on 2020, in two forms of virtual (37%) and physical (63%). Here, a self-made questionnaire with 36 questions was used. These questions enquire about the knowledge, attitude, and behavior (each 12 questions) of citizens about water in urban areas. These questions consist of different topics in urban water management such as: water supplies, water scarcity, leakage and bursts, water cycle, water saving methods, water tariffs, virtual water, the treatment process, water quality and sanitation, personal hygiene or obsessions, actions during water switch-offs, dependability to tap water, their responses to water misuse, and etc.It should be noted that questionnaires were developed by a two-step pre-test. The validity of the instrument was controlled by 4 experts of National Water and Wastewater Company (NWWC) and 3 professors in University of Isfahan. The reliability of questionnaire was measured by Cronbach's alpha as 0.73, 0.8 and 0.87 for water knowledge, attitude, and behavior, respectively.Data analysis was carried out by the application of descriptive and inferential statistics by SPSS (version 23) and Minitab (version 19). Kolmogorov–Smirnov (KS) and Anderson-Darling (AD) were used to control the normality of results about water literacy. One-way analysis of variance (ANOVA) was used to compare water literacy in different variables. The impacts of gender, age, education, employment, living property, and income status of respondents on water literacy were examined. In this test, the reliability of 95% confidence interval (P-value < 0.05) was set as the criterion. In addition, Pearson test was used for correlation, while path analysis used partial least squares (PLS) regression with standardized coefficient.

The normality tests verified that water literacy follows a normal probability distribution function. The skewness and kurtosis of these data were 0.18 and 0.27, respectively and KS and AD were 0.074 (p < 0.05) and 1.5 (p < 0.05), respectively. The mean of water literacy was calculated as 43.5 (out of 100) with standard deviation of 9.5. Here, the first and third quartiles were 37.1 and 50, respectively. It verified that the majority of citizens (about 89%) in Isfahan had moderate water literacy having score between 33 and 66.Among the three pillars of water literacy, surveys revealed that water knowledge could gain the lowest score (34.1), while attitude received the highest (47.6). Here the average score of behavior calculated as 43.9. It means that the citizens have little information about the basics of water but they have better understanding about its value and risks. This conclusion contradicts the results of previous studies in which knowledge was claimed as a prerequisite for good attitude or behavior in environmental literacy. It can be due to the fact that the life of people living in arid or semi arid area, like Isfahan, is very reliant on water. Therefore, the value, risks, and water saving methods may not necessarily dependent on the basic water knowledge. Path analysis with regression modeling also revealed that water literacy is mostly dependant on water behavior (β= 0.77) and attitude (β= 0.66) than basic knowledge (β= 0.57) as illustrated in Figure 2.The comparative statistical analysis also demonstrated that variables such as gender, age, education, employment and income were effective on water literacy in the study area. Table 1 outlines the mean, standard deviation, and P-value of water literacy for each variable based on one-way T-test. It can be concluded that having a college education, a stable job, higher income, or age more than 40 may give citizens some personality or a character that presents responsibility to the citizens for water and can be reflected in attitude or behavior. In a traditional masculine society of Isfahan, being a man may also bring this kind of responsibility. Therefore, the intrinsic motivation for life and the responsibility of water saving can improve water literacy in a society. However, higher education seems to be the most influential variable according to the path analysis. In addition, it is realized that higher water literacy may reduce the satisfaction of citizens about the performance of water supply companies. It roots in higher education and consequently increases the expectations from water companies. Figure 2: path analysis of water literacy and its variables Table 1: comparative statistical analysis of water literacy in different variables Variable Group Count Mean Std. deviation P-value (T-test) Gender Man 204 44.72 8.98 0.004 Woman 194 41.91 9.45 Age < 40 286 42.82 9.52 0.042 > 40 112 45.19 8.30 Education College 222 42.56 9.89 0.000 School 176 39.99 7.16 Employment Employed 236 44.62 9.29 0.000 Unemployed 162 41.09 8.92 Income High 179 45.15 10.05 0.002 Low 219 42.13 8.57 Living property Owner 209 43.51 9.69 0.716 On rent 189 43.15 8.79

This research accounts the water literacy of citizens in a mega city in Iran based on basic knowledge, attitude and behavior. According to the results, it can be concluded that:• Water literacy is moderate for the majority of citizens in Isfahan.• Basic knowledge, attitude and behavior can independently enhance water literacy in which behavior and attitude are mostly influential. Having higher education, a stable job, higher income as well as being in the middle age (> 40 years old) can provide an opportunity for citizens to enhance their water literacy. People with higher water literacy showed less satisfaction about the performance of water companies. It implies that water literacy can be introduced as 1) an index for public water education programs, and 2) a motivation for upgrading the performance of water companies.