Watershed health and ecological security modeling using anthropogenic, hydrologic, and climatic factors

Assessing the health and security of natural ecosystems is unavoidable to regulate and manage deterioration and prioritize management and conservation activities. As a result, the current study was conducted to assess the health and ecological security of the Galazchai Watershed in Oshnavieh Township, West Azerbaijan Province, Iran, utilizing criteria linked to various hydrologic, anthropogenic, and climatic aspects. In the conceptual pressure–state–response (PSR) method, 26 criteria with acceptable variance inflation factor and Durbin–Watson statistics were chosen. The Galazchai Watershed's ecology's health and security status and spatial differences were also chosen and analyzed at 17 subwatersheds. The contributions of the pressure (P), status (S), and reaction (R) indices to the watershed health index were 38.3, 30.1, and 31.6%, respectively, according to the findings. The health index for the watershed ranged from 0.388 to 0.688. Almost 64% of the watershed's area was in relatively healthy conditions. Additionally, the maximum and the minimum values of the ecological security index were 0.608 and 0.236, allocating 58% of the area to moderate ecological security. The northern part of the watershed had healthier and moderately desirable ecological security than the southern part, with unhealthy and relatively undesirable conditions. In this sense, the main problems of the study area were primarily related to anthropogenic factors followed by hydrologic ones. The procedure used in the current study based on health and ecological security status can be applied as a practical guideline by local managers for efficient and adaptive watershed management.

problem-oriented watershed health assessment (Gari et al., 2015). The P, S, and R (PSR, i.e., pressure, state, and response) conceptual framework has some advantages, such as the interaction of different dimensions of a watershed system, floating criteria and factors based on regional needs, and paying serious attention to the response (R) index based on economic and social dimensions as an essential element of watershed sustainability . In the PSR conceptual framework, P refers to the pressures on the environment imposed on a watershed through human activities and natural factors. S describes the current state of the natural environment and watershed performance. R also expresses the degree of community response or different watershed outcomes to the driving forces imposed on the watershed system (Hazbavi & Sadeghi, 2017;Liang et al., 2010). This method, which organizes the criteria into complete indices of pressure, status, and response, may be used to estimate the health of an ecosystem by taking into account all of the regulative circumstances. So, watershed health and ecological security assessments can be utilized to make sound, efficient, and successful decisions (Bai & Tang, 2010).
Several studies have been conducted with different goals and approaches regarding assessing ecosystem health, most of which are index-based and have been used qualitatively and quantitatively. For instance, Ding et al. (2008) assessed the regional health of China's Inner Mongolia Province at the watershed scale using the PSR framework. The results showed that the middle and southwestern part of the region was in danger due to significant anthropogenic pressures. Bai and Tang (2010) used the PSR conceptual framework to evaluate the ecological security of Tianjin, China. It showed that Tianjin's urban ecological security status had improved from 2001 to 2007. In another study, Hong et al. (2010) also developed an indicator system for assessing the ecological security of regional aquatic in Jinan, China, based on the PSR framework. The results presented that the ecological security of the aquatic in the Tianqiao Region was in the weakest possible condition because of improper exploitation of groundwater, reduction in landscape diversity, and lack of scientific and technical investment. Sun et al. (2019) also assessed the health of Jiaozhou Bay wetlands using the PSR framework in Shandong Province, China. The general health score was 0.539, which indicated that Jiaozhou Bay was in moderate health status. Also, Zhang et al. ( 2020a) assessed the health status of the water cycle in 13 Chinese cities. The results showed that the water cycle health status varied from unhealthy to healthy, and the quantity of water had the worst condition in the study cities. In another study, Li et al. (2021) assessed Alpine pastoral ecosystems' spatial and temporal health using the PSR framework in northwest China. The health of the Alpine pastoral ecosystem was assessed in this study using a novel PSR evaluation paradigm based on government response to pressure. The findings revealed that between 2000 and 2015, the health levels of Gannan pastoral ecosystems exhibited a regional distribution pattern that deteriorated from southwest to northeast. Ebrahimi Gatgash and Sadeghi (2022) examined the health of the Mikhsaz Watershed in Mazandaran Province, Iran. The findings revealed that the P, S, and R indices were 0.68, 0.61, and 0.75, respectively, indicating a dominantly generally healthy status for the watershed health and an overall value of 0.68. It was also revealed that hydrologic factors contributed 56.07% and 80.11% to control state and response levels. In the Pishkuh Watershed in Iran, Sadeghi et al. (2022) looked into a humaninduced watershed's security and ecological viability. According to the findings, 91% and 9% of the watershed were rated moderately healthy and relatively unhealthy, respectively. The ecological security index data also showed that 41% of the region was assigned a relatively poor status. Based on pasture management, Ashworth et al. (2022) examined the watershed health. The study demonstrated that pasture management and precipitation positively impacted water quality. It was also discovered that soil compaction harmed the health of the watershed. In the Chaohu Lake Watershed in China, Duan et al. (2022) recently assessed the health of the watershed and related motivating and impediment factors using an integrated multistatistic analysis-pressure-state-impact-response (PSIR) framework. The findings revealed that the watershed's export of nutrients and sediment grew, and the water quality of the study watershed's rivers decreased due to socioeconomic and natural forces. Growing the amount of forest and rangeland was a successful remedy for halting the deterioration of river water quality.
Ecological security, which goes beyond watershed health and sustainability, is the state of an ecosystem's stability in the face of external pressures (Huang et al., 2020). In the last decades, studies on ecological security have received the majority of attention Peng et al., 2018;Yu et al., 2009). In this regard, Chang and Zhang (2014) looked into the ecological security of the Shuleh River in China. They discovered the early warning signs of the landscape pattern impacting the ecological security of the study watershed. Zhang et al. (2020b) also examined the Yangtze River basin's ecological security projections for land-use/cover. Given an ecological security assumption, they provide essential scientific guidelines for spatial land development. Huang et al. (2020) looked at the degradation of ecological security worldwide due to climate change. They demonstrated a dramatic loss in ecological security in dry lands that had spread into neighboring areas during the previous 60 years. They also stated that ecological security is anticipated to respond more forcefully to global warming and human activity.
Multiple studies have been carried out with various aims about watershed health and ecological security assessment. Watershed health indices were offered as a novel technique to prioritize watersheds in a study (Salamat & Najafinejad, 2013). Bahraminejad et al. (2018) reported another study that proposed an early warning system to safeguard the ecological security of protected regions in Iran's east. In addition, an Index-based methodology for a thorough assessment of watershed health and sustainability was offered, emphasizing future research . In addition, other related studies have been done, such as evaluating watershed health indicators, comparative analysis of variability of health indices of P, S, and R in the Shazand Watershed , assessing the health of the Asiab Rud Watershed using the conceptual framework of PSR , predicting watershed health based on variables of surface water quality in the Taleghan Watershed , dynamic zoning of the Shazand Watershed health based on the characteristic of low flow and high flow discharges (Sadeghi et al., 2019), assessing and analyzing the health status of the Kuzah Topraqi Watershed, Ardabil Province (Hazbavi & Sadeghi, 2017), and assessing the health of the Shazand Watershed using PSR framework . In a different research, Salehpour Jam et al. (2021) used the driver-PSIR (DPSIR) framework to assess the management responses for the Chehel-Chay Watershed health improvement. The findings indicated that for the components of P, S, I, and R, the compliance rate of common priority among 40% of the top priorities was 73%, 60%, and 58%, respectively. Also, a study was conducted by Mosaffaie et al. (2021) to analyze the main environmental problems of the Gorgan Roud Watershed in Iran using the framework (DPSIR) in the period 2004-2018. The results showed that socioeconomic activities and pressures had deteriorated the watershed's health. Also, some practical measures have been taken to balance other DPSIR indices, but more was needed. Tehran's Kand Watershed thoroughly assessed soil conservation methods to lower the possibility of soil erosion by Salehpour . In comparison to the base year (i.e., 1997), the data indicated a decrease in soil erosion and the harmful effects that came with it in 2019. Based on the integrated index, soil erosion status, and associated consequences have decreased by 16% and 35%, respectively. However, despite this decline, the watershed still experiences significant soil erosion (26.27 t ha −1 year −1 ). Recently, Ebrahimi Gatgash and Sadeghi (2022), based on the results conducted for the Mikhsaz Watershed in northern Iran, disclosed that a watershed health assessment is a perfect basis for analyzing issues and impactful variables that lead to sustainable watershed management and that a health assessment is a valuable method for assessing and identifying effective human, ecological, and environmental resource management strategies at the watershed scale. It allows watershed managers to focus their efforts on priority subwatersheds to efficiently solve existing concerns by correctly classifying useful features and assessing degrees of controllability. According to Mirchooli et al. (2022), anthropogenic impacts and natural driving forces such as climate change may affect sustainability. As a result, local governments should investigate the geographical and temporal conditions of determinant elements, such as land-use changes, to address sustainability concerns and create watershed management policies.
The review of the research background presented that various critical criteria and indices have been introduced and developed using different approaches to evaluate the health and security status of different ecosystems. Also, studies showed that each approach used different criteria and factors, few of which reported the basis for the selection. Nonetheless, it has been approved that using one criterion cannot comprehensively indicate ecosystems' health and security status. Accordingly, combining the different criteria and indices was the most challenging issue in assessing watershed health and ecological security. Meanwhile, an insightful literature review further showed that no problem-oriented study had been reported assessing ecosystems' health and security in changing watersheds, particularly in developing countries. Therefore, in the present study, the interactive effect of anthropogenic, hydrologic, and climatic criteria on the health and ecological security of the Galazchai Watershed in Oshnavieh Township, northwestern Iran, was considered. Since being located in the country's border areas, the study area was exposed to various types of human interventions, destructive factors, and threatening triggers to the health of the watershed resources. Scrutinizing the governing conditions in the region showed that the failure to instantly analyze the situation of the study watershed leading to proper management of the soil and water resources might have irreparable consequences for local and regional policy/decision-makers and all other stakeholders. As a result, the study will combine key and influential components and indicators such as anthropogenic, hydrologic, and climatic aspects to assess the health and security of the Galazchai Watershed. Another goal is to identify the most effective indices for the entire watershed and a combination of current indices to develop a subwatershed-scale spatial map of the watershed's health and security, which will be presented to experts, stakeholders, and managers. In other words, watershed health assessments and ecological security can aid in identifying and protecting priority regions and suggest management strategies.

| Study area
The Galazchai Watershed, with an area of 104 km 2 and a mean slope of 32%, is located in northwest Iran in Oshnavieh Township, West Azerbaijan Province. The mean elevation is 2390 m, ranging from 1492 to 3273 m above mean sea level. The mainstream length is 19.3 km, and the concentration time is about 2 h. The Gravilius and Horton shape coefficients are 1.3 and 0.92, respectively (Mostafazadeh et al., 2014). The mean annual precipitation and temperature of the Galazchai Watershed are 482 mm and 11.8°C. The dominant land-use is rangeland covering about 85% of the study area. According to the population and housing census, the watershed includes two main residential areas of Galaz and Zemmeh, with 1270 in 2006 and 1550 in 2016. In addition, nomadic stockbreeders partly live in the upper rangelands of the study watershed. A general view of the Galazchai Watershed, subwatershed delineation, and Galaz and Zemmeh villages are shown in Figure 1.

| Conceptualization of the PSR framework
Many studies have employed ecological ideas, methods, and frameworks to define and rate the health of watersheds. Nonetheless, it should be understood that biological processes are only one part of the watershed. Consequently, the watershed health assessment must consider the interconnected processes of the biophysical environment, human society, and relationships . Several conceptual frameworks have been offered to describe this goal's applicable criteria, indicators, and compilation methods (Sadeghi et al., 2019). In this context, the health and ecological security of the Galazchai Watershed in Iran were assessed using the conceptual framework of the P, S, and R index. The conceptual framework of PSR has determined health in terms of all the circumstances controlling a system by arranging the criteria as complete indicators of pressure, state, and response (Hazbavi & Sadeghi, 2017). To do this, the governing circumstances for the study watershed were theoretically modeled using current concerns and challenges at the subwatershed scale, considering intrinsic linkages between issues and watershed outcomes.

| Problem diagnosis in the study watershed
The study watershed is one of the border watersheds of the country, where smuggling is one of the main problems of its proximity to the Turkey and Iraq borders. Because of smuggling, nomads' settlement in summer pastures, and the high potential of pharmaceutical plant production in the region, the rate of human intervention, and road density are significantly high. The watershed problems were identified by several visits to subwatersheds in different periods. Also, the essential information to analyze the health and ecological security of the watershed was collected according to the problems and through reviewing various studies (Ding et al., 2008;Mao et al., 2014;Sadeghi et al., 2019;Wang et al., 2019). Then, the required indices and criteria were extracted to assess the watershed health. Some criteria were forcibly removed due to a need for more information and data. The determinant criteria were ultimately finalized, as described in Table 1. The long-term mean of the data was used for climatic and hydrological factors. The latest data, mainly from 2021, were used for human factors.

| Estimation of the pressure index
The pressure (P) index assesses watershed health as the direct and indirect driving forces of human and natural activities and causes environmental changes .
Due to smuggling, nomads' settlement, and pharmaceutical plant production in the region, the rate of human intervention and road density is significant. Therefore, a road network was drawn, and a field visit and Google Earth software in each subwatershed calculated road density. Since road construction caused disturbances in the drainage network and diversion of the flow, the intersection of roads and streams were extracted and calculated for each subwatershed. Also, 14 other criteria were used to calculate the pressure index. Climatic data for 12 stations around the watershed, such as rainfall, evapotranspiration, and temperature on an annual scale, was collected from Water Resources Management Company studies. Geostatistical methods in GIS prepared the distribution map. In the next step, the rate of human intervention, population density, and population growth rate were considered to achieve other adverse human effects. The human intervention was calculated from the ratio of nonnatural land area to each subwatershed area. Besides, population density and growth rates were estimated from 2006 to 2016 (Taghavi, 2004). The environmental sensitivity area index (ESAI) is a suitable criterion to indicate human activities' role in land degradation (Ali & El Baroudy, 2008;Basso et al., 2000;Ferrara et al., 2012;Marull et al., 2007). It expresses the effect of various variables affecting degradation and the role of human activities on watershed health. Different variables such as climatic, geological, soil, vegetation, and human and managerial factors were considered to estimate the P index. For instance, variables such as mean annual rainfall, drought index, slope direction related to climatic criteria and slope percentage, soil T A B L E 1 Initial indices, criteria, and variables in assessing the health and ecological security of the Galazchai Watershed in Oshnavieh Township, West Azerbaijan Province, Iran. texture, and soil depth related to soil and geologic criteria were selected. Also, other variables allied to management quality and anthropogenic factors were considered, like population density, population growth, land-use change, vegetation factor, erosion condition, and vegetation management. Afterward, the P index was calculated using the geometric mean for the study area (Hazbavi & Sadeghi, 2017;. In hydrological criteria, the mean slope and slope of contribution area in runoff generation were prepared using a digital elevation model (DEM) and combining slope and runoff maps, respectively. Besides, the area of agricultural land with a slope above 25% was made by integrating slope and agricultural landuse maps . The river length was extracted from the DEM with ArcHydro extension in ArcGIS 10.8 software. In addition, the standardized precipitation index (SPI) was considered because of using available precipitation data, was simple to calculate, and calculated for any time scale and weather (Bazrafshan et al., 2020;Miryaghoubzadeh et al., 2019). SPI is a nondimensional index established to monitor and track drought. SPI allows the analyst to determine drought and flood occurrences on a given time scale for any location (Kherad narooei et al., 2020;Sarvi Sadrabad et al., 2020). Moreover, the annual runoff flow was calculated using Justin's method for better performance in similar areas (Munna et al., 2021). Flow discharge was estimated using Dickon's method due to the lack of hydrometric stations in the study area, ease of calculation, and calibration in the country's northwestern regions (Sarvati et al., 2014).

| Estimation of the state index
Nine different criteria were used to calculate the state (S) index, as summarized in Table 1. Therefore, the mean normalized difference vegetation index (NDVI) was considered for the time series of May due to the lack of grazing and harvesting of pharmaceutical plants. The Google Earth Engine was used to estimate NDVI in the range from −1 to +1 because of the power to process a large amount of data in the short term (Soltani & Mohammadnejad, 2021). Also, the criteria of organic matter, electrical conductivity, and soil acidity were extracted to estimate the physical condition of the study area (Zabihi Seilabi et al., 2020). Moreover, the state of the landscape and the spatial analysis of various land-use measures were performed using the Fragstats software. Shannon's diversity index is calculated based on the number of distinct spots of different land-use types and land covers and their distributions in the study area. These two characteristics are Richness and Evenness, with a range equal to zero or higher. Equation (1) was used to calculate Shannon's diversity index in each subwatershed.
where k is the number of land-use types and land cover or spot richness in each subwatershed, pi is the percentage of each land-use category and unconstructed land cover, and pu is the percentage of urban land-use and development to reduce habitat capacity (Darvishi et al., 2021). The S index examines environmental conditions, including physical, chemical, biological, natural, and human well-being . Additionally, the indices of Shannon's diversity and the appropriateness of Simpson's diversity were estimated. Two other essential biologic density and naturalness index criteria were also used in the S index, determined using Equations (2) and (3), respectively, where Bd is the biological density index, p is the number of population, Cn is the area of nonnatural land, NI is the naturalness index, C 1 is the area of agricultural land, C 2 is the area of seminatural areas, C 3 is the area of entirely natural areas. Thus, poor rangelands resulting from overgrazing and trampled areas resulted from nomadic settlements in various subwatersheds that were supposed to be seminatural. Other criteria, such as the ratio of contribution area to runoff generation to the total area and the ratio of precipitation to evapotranspiration, were studied among the hydrological factors.

| Estimation of the R index
The R index was also employed to show any response or change due to the pressures . In this part, hydrologic and anthropogenic criteria were considered to calculate the R index. Four variables of forest land percentage, rangeland area, grazing season vegetation cover, and NDVI were related to anthropogenic criteria. Also, two other variables of grazing season vegetation cover and NDVI were selected to recognize the effects of harvesting pharmaceutical plants and grazing on the watershed health (Haq et al., 2020). Therefore, the mean NDVI index and the monthly NDVI changes for the grazing season were extracted using the Google Earth Engine. The land-use map also estimated the area of rangelands and forest land. In addition, drainage density, mean storm-wise erosion, storm-wise erosion difference (Katebikord et al., 2017;Khademalrasoul & Amerikhah, 2021), time of concentration (Mostafazadeh et al., 2014), and erosion status based on BLM form (Sadeghi et al., 2021) were used to assess the hydrological response.

| The health index zonation
After determining the final criteria for watershed health assessment, owing to the differences in units and magnitudes, all input data were normalized in the range of 0-1 according to their contributions to PSR conceptual framework using Equations (4) and (5) for the criteria with positive and negative contributions, respectively D. Liu & Hao, 2017), where X s , X i , X min , and X max are the normalized, actual, minimum, and maximum values of the desired criterion, respectively.
Moreover, statistical analyses, including the Durbin-Watson test and the variance inflation factor (VIF) assessment, were performed to test the autocorrelation and multicollinearity among the study variables. The favorable range of 1.5-2.5 for Durbin-Watson and VIF < 10 was considered for finalizing the study criteria for the next steps . Afterward, to determine the P, S, and R indices, the arithmetic means of the standardized criteria values were calculated, and then the health status of the study area was determined using the geometric mean (Hazbavi & Sadeghi, 2017;Hazbavi et al., 2018). The health status was eventually assessed at the subwatershed scale using Equation (6) and also classified into one of the five categories healthy (0.81-1), relatively healthy (0.61-0.80), moderately healthy (0.41-0.60), relatively unhealthy (0.21-0.40), and unhealthy (0.00-0.20). The spatial distribution of the health status of the entire watershed was then mapped using the Arc-GIS software package In Equation (6),  X n k n =1 and k are the product and the number of the indices, respectively.

| Ecological security index zonation
To clarify the practical applicability of the health of the subwatersheds, the ecological security Xia et al., 2014) was evaluated by integrating the PSR indices using Equation (7) Security index S R P = × .
Ecological security includes the ratio of the two indices of S and R to pressure Equation (7), and it can complete the health results and supplement the assessment. The fraction equation emphasizes the role of indices and provides a comprehensive view of watershed health (Ma et al., 2019).

| RESULTS AND DISCUSSION
The present study studied the health and ecological security assessment of the Galazchai Watershed in Oshnavieh Township, West Azerbaijan Province. Accordingly, the study area was zoned based on the health condition and ecological security index after selecting influential factors and criteria. The statistical analyses showed that 26 out of the total variables for P, S, and R indices were acceptable based on the VIF and Durbin-Watson test, as shown in Table 2. Thus, annual evapotranspiration, mean temperature, population density, mean slope, stream length, SPI, flow discharge, Simpson's diversity index, and vegetation cover of the growing season were excluded from the calculation process (Table 1).

| P index
The results of pressure-related criteria showed that the maximum and the minimum effects of the slope of contribution area in runoff generation were in subwatershed 17 with 32% and subwatershed 14 with 5.74%, respectively (Figure 2). The criterion values are very high except in some northern subwatersheds for their mountainous nature and elevation difference. This criterion shows that the watershed's western and southern parts (mountainous areas) are sensitive to erosion. Also, the maximum area of agricultural lands with a slope >25% is in T A B L E 2 Statistical analyses of the criteria affecting the PSR conceptual framework of the Galazchai Watershed in Oshnavieh Township, West Azerbaijan Province, Iran. subwatershed 13 with an area of 0.138 km 2 , and subwatersheds 2, 3, 4, 5, 7, 9, 10, 11, and 12 had no problem. It should be noted that there was no residential area in these subwatersheds. The soil depth and topography were not suitable for farming too. Road density was one of the most critical man-made pressures on the natural condition of the study watershed. So, the results showed that the highest and the lowest road density were allied to subwatersheds 15 and 11 with values of 4.82 and 0.49, respectively. It also displayed that human activities and settlements were the origins of most pressures and destructions in natural ecosystems. So, subwatersheds with residential areas had the highest road density. Agricultural and human-induced land-uses, and consequently road construction for easier access, destroyed the natural ecosystems of the region. Another criterion was the degree of human intervention, which had the highest value in subwatershed 14. Also, the minimum amount of human intervention belonged to subwatersheds 9 and 12 with a zero value, indicating 100% of the natural land-use coverage. Subwatersheds 5, 7, and 15, with 2.34%, had the highest population growth, and subwatersheds 11, 10, 9, 8, 6, 4, 3, 2, 1, and 12 were zero. In addition, the ESAI, as an integrated variable for assessing the pressure on the ecosystem health, showed that subwatersheds 13 and 16 had the highest and lowest ESAIs, respectively. The results presented that land degradation and sensitivity to pressure in watersheds with the human population were higher than those without residents. The essential criteria of the stream and road intersection reflecting the rate of human intervention in the natural flow of ecosystems and the high degradation of natural lands to road constructions were also evaluated. It is also the starting point of accelerated erosion in nature. The results presented that the highest and lowest intersections were in subwatersheds 15, with 28 intersections, and subwatersheds 10 and 11, with no intersections. Among the most impressive factors in increasing human interventions was the smuggling of goods in this border watershed. The rural and western subwatersheds had many intersections that clearly showed the impact of human pressures. The results of road density, human intervention, population growth, and road intersection showed human impacts' role in reducing ecosystems' health and sustainability.
Meanwhile, the minimum and the maximum annual runoff flow were assigned to subwatersheds 4 (197 mm) and 16 (334 mm). Likewise, the maximum and the minimum annual precipitation were allied to subwatersheds 3 with a mean of 455.69 mm and 17 with 347.3 mm, respectively. In general, the outcomes of the P index showed that the watersheds with unfavorable conditions were directly related to population distribution. So, human settlements and other needs increased road density, intersections, agricultural land with slopes above 25%, human intervention, and other possible pressures. Therefore, the ecological sensitivity of areas where the pressures exceeded the ecosystem capacity caused an unsecured condition, consistent with .

| S index
The results of the S index showed that the minimum and the maximum NIs were in subwatersheds 14 and 9 with the value of 0.4 and 0.95, sequentially (Figure 3). The NI values showed a better economic usage of natural resources than other economic sources such as mining, logging, and agriculture. The high coverage of the rangelands area and lack of human intervention in the subwatershed 9 were the most important reasons for high NI. While most areas in subwatershed 14 were related to agriculture, seminatural and natural lands had a small portion. Also, human interventions in this subwatershed reached >80%. Another reason for the higher NI in the western part of the watershed was the mountainous topography and steep slopes, which limited the extension of agriculture. The maximum and the minimum value of bio-density were for subwatershed 15 with a value of 3.38 and subwatersheds with no population. Ebrahimi Gatgash and Sadeghi (2022), Mirchooli et al. (2022), Mosaffaei et al. (2021), and Yasuri et al. (2012) reported that increasing biodensity causes pressure on soil and water resources and, in addition to destruction and pollution, leads to migration and other socioeconomic consequences. The high concentration of population in the subwatersheds 13 and 15 caused more pressure to reach needs. In other words, the higher biointensity index describes that more food should be produced from a piece of land, so that the amount of pressure applied to the land is more than its biological capacity. So Shannon's index related to the number of spots and land-use diversity was applied. The higher Shannon's index, the higher the ecological vulnerability. Thus, subwatersheds 12 and 15 had the lowest and highest Shannon's index due to high population density, which increased land degradation in the study area.
The results of the soil factor presented that the minimum organic matter with a value of 1.1% was in subwatershed 3, and the maximum with a value of 2.9% was in subwatershed 9. The high coverage of rangeland land-use in subwatershed 9 increased organic matter. Whereas the limited rangeland area and the considerable area of agricultural lands in subwatershed 3 F I G U R E 3 Spatial variation of the criteria affecting the state index of the Galazchai Watershed in Oshnavieh Township, West Azerbaijan Province, Iran.
were the most important reasons to reduce organic matter. The reduction in soil organic matter due to tillage operations has been reported by Hajabbasi (2001). The lowest soil salinity was estimated in subwatersheds 1, 2, and 10, and the highest was in subwatershed 9. Though there was no considerable difference in soil acidity across the study watershed. A similar finding in connection with soil limitation effects on watershed health was reported by Ebrahimi Gatgash and Sadeghi (2022) for the Mikhsaz Watershed in Iran.
The minimum and the maximum ratio of contribution area in runoff generation to the total area were calculated in subwatershed 4 with a value of 0.54 and in subwatershed 7 with a value of 0.613. The criterion is determined using Justin's runoff calculation method, in which slope is one of the parameters directly affecting runoff generation (Bahrami & Imeni, 2019). It had an inverse relationship with the watershed health, so the western and southern subwatersheds, with higher values than the northern subwatersheds, harmed the status index due to mountainous topography and high slope . Also, the minimum and the maximum ratio of precipitation to evapotranspiration were estimated in subwatersheds 17 and 3 with respective values of 7.76 and 8.12. Accordingly, the southern subwatersheds were in better condition than the northern subwatersheds, which agreed with the results of the precipitation criterion. However, it had insignificant differences due to the small number of climatic stations in the study area. In general, the results of the S index showed that the western subwatersheds were in a more desirable condition from the viewpoint of anthropogenic criteria such as NI, biodensity, and Shannon's diversity index. Because the S index assesses the status of various dimensions related to pressures in ecosystems, its outcome describes the state of ecosystems under pressure. It is consistent with observations in field visits.

| R index
The statistical analyses showed that all nine factors considered to calculate the R index had equal Durbin-Watson test with an acceptable VIF ( Figure 4).
As shown in Figure 4, the maximum and minimum drainage densities were sequentially in subwatersheds 15 and 3, about 2.2 and 0.09 km km −2 . In addition, the maximum and the minimum time of concentration were estimated for subwatersheds 2 and 17, respectively. Also, subwatershed 7 had a maximum forest cover of 26.23%, while subwatersheds 1, 2, 3, and 10 had no forest. Subwatersheds 9 and 14 had the maximum and the minimum coverage of rangelands, respectively. Therefore, the more distance from population settlements, the less degradation of rangelands.
Moreover, the highest value of mean storm-wise water erosion was related to subwatershed 13 with a rate of 0.0333 t, while the lowest soil erosion was 0.00065 t assigned to subwatershed 14. As the primary cause of storm-wise erosion, subwatersheds 13 and 14 had the maximum and the minimum topographic slopes of 34.27% and 10.46%, respectively. Also, the maximum storm-wise erosion difference was in subwatershed 13 with 0.279 t. The minimum storm-wise erosion difference was in subwatershed 14 with a 0.00065 t. Accordingly, most of the subwatersheds (i.e., 78%) had moderate erosion, and four subwatersheds just had low erosion rates, which is consistent with the studies by Sadeghi et al. (2021).
The maximum and the minimum amounts of grazing season vegetation cover were about 0.49 and 0.213 for subwatersheds 14 and 12, respectively. Therefore, it could be implied that the subwatersheds adjacent to Galaz Village were less grazed due to dominant orchard land-use; however, subwatershed 2 was more influenced by grazing because of nomadic settlements.
Meanwhile, vegetation cover change was positive in subwatershed 7 and negative in subwatershed 16, owing to several springs in subwatershed 7. Generally, the grazing factor's positive and negative trends described the dynamics of nature and vegetation degradation in the subwatersheds.

| Zonation of the health index in the Galazchai Watershed
The study presented that the mean and standard deviation of the P index in the Galazchai Watershed were 0.72 and 0.15, respectively ( Figure 5 and Table 3). Also, the lowest and the highest value of the P index were 0.31 and 0.87, related to the subwatersheds 15 and 12, sequentially. Accordingly, the P index was four low, relatively low, moderate, and relatively high, with a relative coverage of 39.74%, 30.16%, 25.9%, and 4% of the total watershed. Meanwhile, anthropogenic criteria (i.e., 65.6%), hydrologic criteria (i.e., 24.86%), and climatic criteria (i.e., 9.54%) have participated in determining the P index, which was consistent with the related studies of Ding et al. (2008), Ebrahimi Gatgash and Sadeghi (2022), Hazbavi et al. (2018), Mao et al. (2014), Mirchooli et al. (2022), and Mosaffaie et al. (2021) on the role of anthropogenic factors in determining the health of ecosystems. Government-related variables were also listed by Mao et al. (2014) as the most crucial elements affecting the health of Lake Ulansuhai in China. Consequently, the contributions of runoff and human intervention to the P index were calculated to be 14% and 13%, respectively ( Figure 6).
The maximum P index was related to the subwatershed 15, which had the highest level of nonnatural lands (i.e., 55%) compared to other subwatersheds. Also, population density in this subwatershed was significantly higher than in other subwatersheds. It had the highest road density, with a value of 4.82 km −2 , which agreed with the studies by  and Sadeghi et al. (2017). However, no high P index was observed in the Galazchai Watershed, mirroring a high restoration potential in the region.  The results confirmed ( Figure 6 and Table 3) that the mean and standard deviation of the S index were 0.56 and 0.09, respectively. The lowest value of the S index was 0.36 in subwatershed 15, and the maximum value was about 0.68 in subwatershed 5. Accordingly, the S index was classified into three categories of relatively undesirable (4%), moderately desirable (32%), and relatively desirable (64%) in the study area. Relatively undesirable and moderately desirable subwatersheds were directly connected to population centers with the lowest elevation. In contrast, the relatively desirable subwatersheds were often distributed in high-elevation subwatersheds, and the result is consistent with the studies of . Therefore, soil characteristics such as organic matter and soil salinity had better conditions in high-elevation areas. At the same time, a high amount of precipitation and low evapotranspiration have promoted the S index in upland subwatersheds. Also, due to minimum nonnatural lands (i.e., agriculture, garden, and residential land-uses), road density, and population density, an increase in NI and a decrease in Shannon's diversity index resulted in an improvement in the S index. The statistical results (Figure 7) further presented that soil physical, anthropogenic, and hydrologic criteria contributed to the S index at 34.34%, 37.76%, and 26.9%, respectively. However, some 10% variation was found in the contribution of the criteria to the S index in the study watershed. Moreover, analysis of the R index presented that the mean and standard deviation of the R index were 0.51 and 0.09, respectively. The lowest value of this index in the Galazchai Watershed was about 0.29, related to subwatershed 13, and the highest value was about 0.65, related to subwatershed 3. The R index was classified into relatively high, moderate, and relatively low, with 6.49%, 76.46%, and 17.05% shares in the study area. Thus, subwatersheds 1, 3, and 4 were relatively high, 11 and 13 subwatersheds were relatively low, and the rest were in a moderate status regarding the R index. The three northern subwatersheds have relatively high responses due to suitable drainage density, vegetation and erosion conditions, slope, and minimum human intervention. Among the criteria used to calculate the R index, rangeland area and mean storm-wise erosion, with a rate of 13.84% and 12.02%, and forest land percentage, with a rate of 9.63%, had the highest and the lowest contributions, respectively (Figure 8). The mean watershed health index for the Galazchai Watershed, calculated from the geometric mean of the P, S, and R indices, was about 0.58 (Table 3). Accordingly, about 63% of the study area had a relatively healthy status with a negative tendency, 26% had a moderate status with a negative tendency, 4% had a moderate status with a positive tendency, and about 7% had a relatively unhealthy status with a positive tendency. So, no healthy or unhealthy extreme status was observed in the study area. Subwatersheds 15 and 4, with health indices of 0.39 and 0.69, were classified as relatively unhealthy and relatively healthy, respectively. Based on Figure 9, the contribution of P, S, and R indices to health status was about 38.22%, 31%, and 30.78%, respectively, indicating the more considerable role of the P index and, and consequently, the role of anthropogenic criteria in the health status of the watershed. Regardless of the participation ratio of different criteria in the status of indices, variables such as the slope of contribution area in runoff generation, NI, and mean storm-wise erosion had the worst conditions related to the three indices of P, S, and R, respectively. Analyzing the trends in the health index of the study watershed declared that it should not be assured about relatively healthy subwatersheds and/or desperate about relatively unhealthy subwatersheds. Since the areas with relatively healthy status and a negative tendency can lead to a lower level of health without supervision, the areas with relatively unhealthy status would be promoted with a bit of attempt and care.
Based on the results, subwatersheds with human activities and residential areas were troubling, but the rest had a better condition and can promote a healthier condition through proper management. Anthropogenic criteria had the most significant impact on the health status of the Galazchai Watershed, consistent with  in the KoozehTopraghi Watershed in Ardebil Province with almost similar geographical conditions. The results also align with those of Ahn and Kim (2019) on declining health status downstream, attributed to excessive human intervention, land-use change, and agriculture. Ebrahimi Gatgash and Sadeghi (2022) found that biological density and the ratio of approved to illegal livestock contributed 36.54% to the pressure indicator. Hydrologic variables accounted for 56.07% of the state and 80.11% of the response statuses, respectively.

| Zonation of the ecological security index in the Galazchai Watershed
The ecological security index results from the interactive effects of three P, S, and R indices in the PSR conceptual framework. The index was derived from analyzing various anthropogenic, hydrologic, and climatic criteria. According to Figure 10, some 60% of the study area had moderate ecological security, 21% had low, and just 9% of the region had a high ecological security index. The lowest and the highest ecological security index belonged to subwatersheds 13 and 15, respectively. Although subwatershed 15 was relatively unhealthy, it is in a better status on the ecological security index than other subwatersheds. So, the watershed is better than the S and R than the P. In addition, subwatersheds 2 and 9 were relatively healthy. However, they were classified in a relatively low condition of ecological security, indicating these subwatersheds' sensitive and fragile nature. Despite the low P index, this characteristic provided a moderate R index, leading to a relatively low ecological security situation ( Figure 10). Also, subwatersheds 16, 10, 8 and 7, and 5 were relatively and moderately healthy, respectively. By integrating health and ecological security results, some subwatersheds, such as 9, had high health conditions but very low-security status due to unfavorable S and R reactions to reasonable P leading to reduced ecological security. Moreover, other subwatersheds, such as 15, had low health but relatively high-security status, so P had tried to reduce health, and the S and R indices attempted to compensate.

| CONCLUSIONS
This study evaluated the health and ecological security of the Galazchai Watershed in Oshnavieh. The influence of nomads and goods smuggling were among the 26 issues considered. Consequently, the Galazchai Watershed in Iran's Oshnavieh Township was evaluated using the PSR conceptual model based on the pressure, status, and reaction framework. The PSR conceptual framework was fundamental due to applying cause and effect relations to connect different dimensions of the watershed system. One of the model's advantages was emphasizing the R index based on economic and social dimensions as a fundamental element of watershed sustainability. It also determined the interactions between human and natural systems and combined various indices and variables. The problem-oriented PSR framework is recommended to natural resources managers and policy-makers as a comprehensive, systematic model for identifying variables and a reliable and appropriate management tool. The main problems in the Galazchai Watershed were related to anthropogenic and hydrologic factors. Furthermore, population growth and increasing human activities, land-use changes, and agricultural lands with steep slopes will affect watershed health and security, leading to further problems soon. Also, nomads' settlement, overuse of natural pharmaceutical plants, and traffic of border smugglers into this region have exacerbated the condition. As a result, despite the relatively good state of the watershed's health, ecological security is in a more precarious position. In this regard, it is projected that if these requirements are followed without consideration for boosting health and ecological security, the watershed's sustainability and resilience will plummet, causing instability under strain. In conclusion, it is suggested that management activities aimed at improving the health and ecological security of the watershed be prioritized. Due to the distribution of health status, key variables, and critical criteria in the examined watershed, comprehensive management policies in the context of adaptive management methods by responsible organizations will be feasible to offer. However, one of the study's weaknesses is the absence of detailed data on nomads and smuggling in the region and the lack of data on water quality and the quantity of chemical fertilizers and pesticides used in agricultural lands. As a result, conducting periodic monitoring with more comprehensive and timely data, particularly on the condition of nomads and smuggling in the study area, as well as analyzing the trend of relative changes in indices and criteria, is critical for implementing wise and adaptive watershed management and evaluating the effects of various management measures on promoting the watershed's health and ecological security. Local managers can also apply the procedure used in the current study and corresponding findings to designate watershed management plans' geospatial distributions appropriately.