Investigation of vegetation dynamics with a focus on agricultural land cover and its relation with meteorological parameters based on the remote sensing techniques: a case study of the Gavkhoni watershed**

. Background and Aims: This research investigates vegetation dynamics in the Gavkhouni catchment from 2001 to 2021, focusing on the spring season. The aim is to analyse the relationship between aridity, vegetation, and rainfall. Moreover, additional emphasis was placed on exploring the impact of these dynamics on agricultural land cover thereby contributing to our understanding of the environmental dynamics in the Gavkhouni catchment. Methods: The study made use of MODIS data, includ - ing the Enhanced Vegetation Index and Vegetation Condition Index, along with monthly rainfall statistics from Chirps. Analytical methods include time series analyses using correlation and regression analysis. Results: Throughout the study period, the average spring vegetation cover was 9276.33 km². The years 2001 and 2018 had the lowest degree of vegetation (15.53, and 17.3% of the watershed area). Conversely, 2013, 2019, and 2020 had the most coverage (27.4, 26.8, and 26.3%). The Enhanced Vegetation Index highlighted the arid years (2001, 2008, 2011, and 2018) and the years with the lowest drought prevalence (2006, 2007, 2010, 2013). Enhanced Vegetation Index correlated with spring rainfall. Cropland cover declined over the study period, and a close correlation was found between winter rainfall and spring agricultural coverage.


INTRODUCTION
Investigations dealing with vegetation coverage fluctuations constitute an important area of study for vegetation management and control that aligns with the principles of sustainable development.The pivotal role of vegetation within the ecosystem is a well-established fact acknowledged by researchers (Gupta et al., 2020;Peng et al., 2012) as it facilitates energy exchange, water circulation, and biogeochemical cycles on the Earth's surface.Indeed, the capacity of vegetation to influence habitat conditions and species composition represents a significant area of interest within the biological sciences (Rannow and Neubert, 2014).
Remote sensing has emerged as an efficacious approach for quantifying the influence of climate and topography on spatial vegetation patterns (Couteron et al., 2014).By discerning vegetation-specific radiation, this technology facilitates the discernment of changes in plant reflection (Manesh et al., 2018).Employing satellite data to monitor extensive plant activities has distinct advantages over conventional methods (Pei et al., 2018).Over recent decades, the significance of indicators such as Normalized Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI) with reference to the areas of vegetation dynamics, drought monitoring, and evaluation have been repeatedly described (Olafsson and Rousta, 2021;Wan et al., 2004).The EVI, an enhanced NDVI, is calculated using MODIS sensor data and constitutes valuable time series data, andis indispensable for monitoring global vegetation dynamics (Ghafarian Malamiri et al., 2018).Relevant stu-dies have demonstrated that climate plays a dominant and controlling role as an environmental factor.In particular, its two primary parameters, temperature and precipitation, significantly influence the distribution of vegetation types within terrestrial ecosystems and contribute to their temporal and spatial diversity.These effects occur on both global and local scales (Chang et al., 2014;Luan et al., 2018;Raynolds et al., 2008;Zhong et al., 2010).Precipitation is a particularly important climate parameter, having a significant influence over vegetation dynamics within arid and semi-arid ecosystems (Martiny et al., 2006;Rousta et al., 2014).Moreover, precipitation directly impacts the water balance and serves as a significant factor in altering soil moisture and influencing plant growth (Farajzadeh et al., 2011).Ground-truth data for validating and calibrating satellite-derived estimates are provided through the use of stationary measurements of soil moisture.These measurements involve the use of in-situ sensors such as time domain reflectometry (TDR) probes, capacitance sensors, or neutron probes to monitor moisture changes within the soil profile at specific locations over time (Majcher et al., 2021).By capturing moisture dynamics at different depths, stationary measurements offer insights into vertical water movement, root zone moisture distribution, and soil moisture variability across various landscapes (Kafarski et al., 2019).However, at the present time drought and flood monitoring is often based on remote sensing rainfall images provided by satellites as a viable substitute for conventional measurements and meteorological data (Gupta et al., 2020;Rousta et al., 2018).
In recent years, drought has emerged as a prevalent and critical natural disaster and climatic phenomenon.It exerts profound impacts on communities, water resources, and ecosystems (Bhuiyan et al., 2006;Dorjsuren et al., 2016;Rousta et al., 2017;Zhang et al., 2013).Numerous researchers have emphasized that the most crucial factor driving changes in vegetation is drought and the variations resulting from shifts in rainfall patterns (Farahani et al., 2022).The consequences of drought are especially severe for agricultural communities that depend on water resources, by contrast non-agricultural communities aren't so severely affected.Since the rural economy in many developing countries relies heavily on agriculture, and this agriculture is solely dependent on rainfall, the impact of drought is even more pronounced in such cases (Ashraf and Routray, 2013).In meteorological terms, drought manifests as a period marked by below-average rainfall, this results in a temporary shift in weather patterns.Subsequently, contingent upon soil moisture levels, agricultural drought ensues and also, if prolonged, hydrological drought that diminishes river flow and other water reservoirs.The ultimate result of a prolonged lack of rainfall encompasses economic drought, which imposes detrimental consequences upon both society and the environment (Mansourmoghaddam et al., 2022).As a consequence, based on the influence of distinct environmental factors, drought can be classified into meteorological, agricultural, hydrological, and socioeconomic subtypes (Peters, 2003;Tate and Gustard, 2000).
The Gavkhouni watershed is a part of the watershed of the Central Plateau of Iran, it exhibits a diverse array of climate patterns, ranging from exceedingly dry to substantially humid owing to its altitude range between 1466 to 3974 m.This basin is classified as being exceptionally important due to its size and location in central Iran and also the availability of suitable soils for agriculture.The existence of about 70 cities and 1720 villages in the region, the presence of the metropolis of Isfahan, the existence of the Gavkhouni wetland and the Zayandeh Roud watershed further accentuate its significance.In recent years, climate change and diminished precipitation have resulted in the permanent depletion of Zayandeh Roud's flow and in the desiccation of the ecologically vital Gavkhouni wetland (Mirahsani et al., 2018).To date, these areas have been preserving biodiversity, controlling fine dust and air pollution, impacting local, provincial, regional, national, and transnational climate/temperature equilibrium, and influencing the living conditions in Central Iran.As a consequence, such topics as climate and weather regulation, maintaining the stable flow of the Zayandeh Roud from Sarab to Payab, hydrological control and protection, water resource quality and production, groundwater replenishment, soil erosion prevention, as well as various economic and social dimensions, including tourism, agriculture (Jafari and Bakhshandehmehr, 2016), animal husbandry, food production, employment, green economy development, the livelihood of local communities, disease prevention, health maintenance and meeting consumption needs are emerging as focal areas of concern.This applies especially in the central and eastern regions, the negative impact on the regional economy in general and agriculture in particular, and also the well-being of numerous animals and birds, cannot be overlooked.Thus, a comprehensive investigation into droughts and climate change in this region is indispensable.
Given the dry and semi-arid climate prevalent in Central Iran and the varying rainfall distribution across different areas of the basin (northern and western sections exhibit increased rainfall due to higher altitudes), assessing the processes leading to changes in vegetation relative to basin rainfall acquires significance from the viewpoint of ecological sustainability.Increased vegetation cover yields benefits such as carbon dioxide reduction, enhanced air quality, soil moisture preservation, water and soil conservation, and flood mitigation, these are all integral to ecological stability.By contrast, environmental instability may lead to desertification, soil erosion (Weishou et al., 2011), climate change, and other issues, inflicting adverse economic, social, and climatic consequences.As drought is a prime factor contributing to vegetation reduction, this further underscores the importance of the subject under investigation, with climate changes and atmospheric circulation shifts serving as underlying causative agents (Khosravi et al., 2023).
Also, changes in climate, including the precipitation occurring over the last few decades, are important in terms of being one of the factors affecting the reduction in soil resources and in the low levels of water found in rivers.The dependence of the agricultural sector on water is undeniable.The rural employment rate in the agriculture sector in the Gavkhouni watershed is 89%, so this basin is known as one of the agricultural poles in the basins of Iran and is very important (Saraei et al., 2017).
In this study, the Enhanced Vegetation Index (EVI), Vegetation Condition Index (VCI), and Standardized Precipitation Index (SPI) were used to reveal drought occurrences and vegetation area changes across different zones within the basin throughout the period spanning from 2001 to 2021.Special attention was focused on agricultural land cover in order to assess its dynamics and the degree of interdependence with climatic variables.

Study area
The Gavkhouni watershed area spans 41 552.3 km 2 and is located at an elevation of 1 450 m above sea level.From the viewpoint of the general division of Iran's hydrology, the Gavkhouni watershed is an integral component of the Central Plateau of Iran's broader watershed system, it is bordered by the Namak Lake watershed to the north, the Degh Sohrh and Siahkouh desert basins to the east, the Abarqo desert basin to the south, and the Karun watershed to the west and south.It is notable that a significant portion, constituting 83% of the Gavkhouni International Lagoon catchment area, falls within the Isfahan province, while 9% extends into Fars province, 4.5% into Chaharmahal and Bakhtiari province, and 3.5% into the Yazd province.More than two-thirds of the Gavkhouni watershed area belongs to the Zayandeh Roud basin, which encompasses an area of 26 917 km 2 .This region is of major importance due to the establishment of major population centres there, industrial developments, and the primary concentration of agricultural activities within the Gavkhouni basin.The geographic coordinates of the basin fall between 50°2' to 53°24' east and 31°12' to 33°42' north and the elevation ranges from 1 450 to 3 900 m above sea level (Fig. 1).

Data and methodology
The land cover map utilized in this study (Fig. 2) is the MCD12Q1 annual product (IGBP) acquired from the MODIS system, it was obtained from the United States Geological Survey website (https://Ipdaacsvc.cr.usgs.gov/appeears/task/area).This map, embracing 17 distinct classes, meticulously characterizes the surface cover across the Earth (Belward et al., 1999;Loveland and Belward, 1997).For the period spanning from 2001 to 2021, the land cover within the Gavkhouni watershed is delineated into 7 discrete classes, designated as classes 7, 9, 10, 12, 13, 16, and 17, in accordance with the classification scheme devised by the University of Maryland (Hansen et al., 2000).These classes correspond, respectively, to shrubland, tree cover (canopy exceeding 2 m in height), grassland, agricultural land, urban and man-made areas, bare land, and wetland areas.As illustrated in Fig. 2, the vegetation within the Gavkhouni catchment basin is predominantly represented by shrubland and grassland (pasture) formations, occupying central, northwestern, and western sectors of the basin, where the extent of the vegetation coverage corresponds to the elevated regions.The Digital Elevation Map (DEM) data was made using USGS EarthExplorer data (https:// earthexplorer.usgs.gov/).
For this study, a dataset consisting of 484 EVI images, derived from Moderate Resolution Imaging Spectroradiometer (MODIS), was compiled for the Gavkhouni watershed, the dataset spanned a period of 21 years .The MOD13Q1 MODIS sensor offers a spatial resolution of 250 m, a radiometric resolution of 12 bits, and a spectral range of 0.4 to 14.4 μm.Moreover, it provides observations at various temporal resolutions, including daily, 4-day, 8-day, 16-day, quarterly, and annual intervals.The EVI images were extracted from MODIS sensor observations at a 250 m spatial resolution, they were acquired every 16 days using the Google Earth Engine system.The periods were partitioned as follows: the winter vegetation data was derived from an average of 6 images ranging from image number 353 to 065, corresponding to a period ranging from December 19 to March 6; the spring vegetation data was derived from an average of 6 images spanning from image 081 to 161, this corresponds to the period of March 22 to June 10; the summer vegetation data was derived from an average of 6 images spanning from image 177 to 257, it encompassed June 26 to September 14; the autumn vegetation data was calculated from an average of 5 images ranging from 273 to 337, it corresponded to September 30 to December 3; and the annual vegetation data was obtained from an average of 23 images ranging from 017 to 001, representing January 17 of the preceding year to January 1 of the subsequent year.In the next step, 251 monthly images of precipitation were downloaded and processed.The total annual rainfall was calculated as the sum of precipitation across all 12 months, while seasonal rainfall data was determined for winter (December, January, and February), spring (March, April, and May), summer (June, July, and August), and autumn (September, October, and November).The seasonal and annual EVI, VCI, and precipitation images were processed using the ArcGIS environment.Subsequently, all pixel values for daily, seasonal, and annual datasets were imported into EXCEL software, where extensive analysis, classification, graph plotting, inter-indicator correlations, and assessments of changes in precipitation and vegetation were conducted.Additionally, vegetation area calculations were performed using this software.Then, the cropland class was extracted separately from the landcover maps of each year and checked according to the average spring season maps of the basin.
Also, the Chirps rainfall dataset was employed to assess precipitation patterns.The Chirps data encompasses the region between 50°N and 50°S, it spanned all longitudes, with temporal coverage extending from the commencement of January 1981 to the present day, with daily, monthly, and seasonal values.The CHIRPS product is a compilation of over thirty years of global rainfall data, with a spatial resolution of approximately 0.05° on a global scale.This product utilizes not only satellite information but is also validated by ground stations, thereby making it a widely employed resource for drought monitoring.These images were obtained monthly during the period ranging from 2001 to 2021 via the website https://www.chc.ucsb.edu/data/chirps.Subsequently, they were extracted on a quarterly and annual basis within the ArcGIS environment.

Enhanced Vegetation Index (EVI)
The EVI has been developed as an advancement over the NDVI, the aim is to optimize vegetation signals within the leaf surface index range.By incorporating the blue band reflection to mitigate soil background signals and account for atmospheric effects, including the dispersion of suspended particles, the EVI demonstrates superior performance (Huete et al., 1994).The mathematical formulation of the EVI is represented as: (1) Here, B NIR , B RED , and B BLUE denote the near-infrared band, red band, and blue band, respectively.L serves as the land and soil adjustment factor, it is set to 1.The coefficients C 1 and C 2 are utilized for correcting aerosol dispersion.Notably, in the red bands, the values of C 1 and C 2 are equal to 6 and 7.5, respectively.The weight factor G, which was set at 2.5, further enhances the performance of the EVI (Huete et al., 1994(Huete et al., , 1997(Huete et al., , 1999(Huete et al., , 2002)).The annual and seasonal improved EVI was calculated as follows: (2)

Standardized Precipitation Index (SPI)
The SPI constitutes a prominent method utilized for drought assessment and is formulated as follows: (7) In the relationship above, x i denotes the amount of precipitation for each month, season, or year, x m represents the average precipitation observed during the statistically relevant period, and SD denotes the standard deviation observed during that period.A comprehensive interpretation of the SPI results is presented in Table 1, where positive values indicate rainfall which surpasses the average, whereas negative values suggest the opposite scenario (Cancelliere et al., 2007;Shah et al., 2015).

Vegetation Condition Index (VCI)
The concept of the VCI was initially introduced by Kogan in 1995 (Kogan, 1995).The VCI is computed and standardized against a range of long-term NDVI values.The NDVI is derived from the contrast between the red and near-infrared bands within an image, thereby providing valuable insights into the vegetation percentage, plant photosynthetic activity, surface water presence, leaf area index, and biomass quantity.This index is assigned values within the range of -1 to 1.As the EVI represents an enhancement of the NDVI, it has also been used for VCI computation.The numerical output of the VCI index is expressed as a percentage within the range of 0 to 100.Values nearing zero signify the presence of stress and severe drought in the region, whereas values approaching 100 indicate favourable vegetation conditions with no water stress (Kogan, 1997).It is calculated as follows: (8) In the relationship above, EVI denotes the improved seasonal or annual vegetation cover index, whereas EVI min and EVI max denote the minimum and maximum long-term EVI values observed across the entire study area during the period spanning from 2001 to 2021, respectively.

Temporal and spatial changes in vegetation
From the results presented in Fig. 3, it is evident that the Gavkhouni catchment area exhibits the most extensive vegetation area during the spring season.The period of robust vegetation growth commences on late March, this extends Ta b l e 1. Classification of wet and dry periods according to the Standard Precipitation Index (SPI)

Description Class
Extremely wet ≥ 2 Very wet 1.5 -2 Moderately wet 1 -1.5 Near normal -1 -1 Moderate drought -1.5 --1 Severe drought -2 --1.5 Extreme drought ≤ -2 throughout the entirety of the summer, with peak vegetation cover in the region occurring on the beginning of May, accounting for 27.35% of the total area (11 363.8 km 2 ).Subsequently, from late June onward, the vegetation undergoes a gradual decline, but it persists until the conclusion of winter.The winter season in general, but culminating on the middle of January, records the lowest vegetation area at 5.1% (2 130 km 2 ).It is evident that the observed decline in vegetation area during the summer season, which includes the period from late June to the middle of September, is concurrent with the drying of vegetation experienced in this season.This phenomenon results in diminished vegetation coverage as compared to the flourishing conditions observed during the spring season.Additionally, the relative humidity of the air plays a significant role, as it exhibits a positive correlation with the rate of water evaporation from both the leaf surfaces and the ground.Consequently, not only does this favour an augmented water supply to the plants in the soil, but it also reduces the daily water requirement of the plants (Rousta et al., 2020a).In the summer season, only green trees and parks in urban areas and at higher altitudes manage to preserve their vegetation due to reduced temperatures, this leads to decreased evaporation and transpiration rates in these plants within the studied area.
Based on the findings presented in Fig. 4, it is evident that the spring season demonstrates the most substantial vegetation in terms of its extent, encompassing an average area of 9 276.3 km 2 .Subsequently, the summer season follows suit with the largest vegetation area after spring, covering 6986.36 km 2 it is the largest area of vegetation among the seasons after the spring season.Conversely, during the autumn season, the vegetation area undergoes a reduction due to the complete evaporation of water from the leaves of dry trees, with an average area of 3 710.03km 2 observed, it is mainly poor vegetation (EVI from 0.1 to 0.2), with limited occurrences confined primarily to urban areas and the northwest and west elevations.In the winter season, as the catchment area experiences increased rainfall, particularly in the northwest and west elevations, and reduced air temperatures, the vegetation area in these regions diminishes to zero.Moreover, the vegetation cover during this season remains scant in the centre of the basin, but relatively more concentrated in urban areas, occupying an area of 2 485.64 km 2 .Figure 4 offers a comprehensive depiction of the average vegetation cover fluctuations across the four distinct seasons.
Figure 5 illustrates the temporal dynamics of vegetation cover as measured by the EVI throughout the period in the Gavkhouni catchment area.Within this figure, positive values displayed in green signify increments in vegetation cover, whereas negative values displayed in red indicate reductions in vegetation during the specified period.Areas characterized by intermediate values and represented in white denote regions with no significant vegetation changes.A notable observation from the figure reveals green areas, they signify an upward trend in vegetation cover, predominantly situated in the northwest, west, in certain central areas, and to a lesser extent in the northeast and southwest regions, extending southwards within the Gavkhouni catchment area.Additionally, scattered instances of increasing vegetation cover are evident within   the central basin, they are particularly concentrated around urban areas.Conversely, the most substantial decrease in vegetation cover is observed at the centre of the Gavkhouni catchment area, further extending towards the southwest and southeast regions.Notably, the white areas indicate a trend of unchanged vegetation cover, attesting to the achievement of stability during the study period.

Results of vegetation change evaluation
The variations in vegetation area across different seasons during the statistical study period were also investigated.In the present study, the investigations focused on the vegetation area of the spring season because in this season the highest level of vegetation coverage was recorded.The results revealed that throughout the study period, the spring seasons of 2001, 2002, 2008, and 2018 produced the lowest vegetation cover, accounting for 15. 53, 17.8, 18.39, and 17.3% of the Gavkhouni catchment area, respectively (Table 2).Conversely, the region also encountered wetter years, such as 2006, 2007, 2010, 2013, 2019, and 2020, where vegetation coverage reached 25.19, 25.9, 28.2, 27.4, 26.8, and 26.3% of the total study area, respectively.Table 2 provides insights into the variations within the different layers of the EVI coverage in the Gavkhouni catchment basin during the spring season throughout the period from 2001 to 2021.Based on distinct vegetation classes, an EVI range of 0.1 to 0.2 exhibited a significant increasing trend (R = 0.5), an EVI range of 0.2 to 0.4 displayed no significant trend, while an EVI range of 0.4 to 0.6 demonstrated a significant decreasing trend during the study period, EVI values exceeding 0.6 also showed a decreasing trend.In Fig. 6  In Figure 7, the relationship between spring vegetation and drought severity, which was computed using the EVI and the VCI in the Gavkhouni catchment area during the period spanning from 2001 to 2021, is presented.Drought areas, as delineated in the research conducted by Dikici (2022) are characterized by values ranging from 0 to 37.5, whereas areas without drought are identified by values exceeding 37.5.Within the entirety of the Gavkhouni catchment area, the spring season produces fluctuations each year, with an average vegetation area of 9276.3 km 2 (accounting for 22.32% of the total study area) during the entire period under examination.As indicated by the VCI index, the years 2001, 2008, 2011, and 2018 exhibit the highest levels of drought severity, representing 83.4,78.9, 68.3, and 79.3% of the total catchment area, respectively.Conversely, the lowest percentages of drought areas were In contrast to the VCI, the EVI exhibits its lowest extent during the years 2001, 2002, 2008, 2011, and 2018, constituting 15.53, 17.85, 18.39, 19.42, and 17.39% of the total area, respectively.This noteworthy finding underscores a notably strong negative correlation (-0.667) between the two aforementioned indices, with the statistical significance at the 0.01 level.Table 3 offers insights into the correlation of various EVI classes with the area affected by dry conditions.The areas covered by poor vegetation (0.1 to 0.2) and moderate vegetation (0.2 to 0.4) demonstrate a significant negative correlation with the dry area at p = 0.01.On the other hand, good vegetation (0.4 to 0.6) exhibits a negative trend without statistical significance, while extensive vegetation (exceeding 0.6) displays a negative trend with statistical significance at the 0.05 level.
Figure 8 illustrates the spatial distribution of spring vegetation density in two drought years, namely 2001 and 2018, as well as two wet years, namely 2007 and 2010.The central region of the watershed and the relatively flat terrain surrounding the cities of Varzaneh, Isfahan, and Zarinshahr, primarily comprising agricultural zones, exhibit the highest concentration of moderate (0.2-0.4) and good (0.4-0.6) vegetation class.Gradually, as one moves towards the mountains and elevations in the western and northwestern areas, the density of the vegetation diminishes.In these elevated regions, the most prevalent vegetation type is that of the poor category (0.1-0.2), which is predominantly found in bushes and grasslands.Conversely, the eastern, northeastern, southern, and southeastern segments of the basin are characterized by desert and arid regions, thereby showing a considerably lower vegetation density.One intriguing observation pertains to the springtime periods of 2011 and 2018, these were previously classified as dry years, with dry area percentages of 68.3 and 79.3, respectively.However, Fig. 10 reveals that during these two spring seasons, rainfall remained within the normal range.An analysis of the obtained data indicates that the average rainfall in the winter season of 2011 was 83.6 mm, which stood below the overall period's average of 99.1 mm.Similarly, the spring season of 2018 produced an average rainfall of 63.7 mm, which was lower than the statistical average of 70.5 mm for the period.Moreover, during the winter season of 2018, the recorded rainfall not only fell below the average for the period but also deviated below the normal threshold.Additionally, it was previously noted   that the years 2010 and 2013 featured low dry percentages (21.4 and 12.6, respectively).However, upon closer examination of Figs 9 and 10, it is evident that the spring season of 2010 produced normal rainfall, while the winter season had rainfall levels that were almost normal and surpassed the normal threshold, which could have influenced the spring vegetation.Similarly, in 2013, both seasons produced a level of rainfall within the normal range, the substantial vegetation cover of 2 277.9 km 2 which occurred during the spring season could have been influenced by other factors such as the temperature of the study area.

Comparison of annual and seasonal changes of vegetation with precipitation
Table 4 presents the correlation matrix between the EVI index and precipitation over a 21-year period.Remarkably, the spring vegetation area demonstrates a statistically significant correlation with precipitation at a significance level of 0.05.By contrast, the vegetation areas for summer, autumn, winter, and the entire annual period show no significant correlation with precipitation. Figure 11   areas exhibited no significant correlation with precipitation.Moreover, Table 5 provides insights into the seasonal and annual variations in vegetation cover within the study area, with particular consideration being focused on the delineation of wet and dry years during the spring season.These findings corroborate the observed percentage fluctuations in annual vegetation cover.

Results of the evaluation of cropland cover changes
Given the significance of agriculture within the study area, an independent investigation was conducted to assess the changes in agricultural land cover between the years 2001 and 2021.As the data presented in Table 6 shows, the years 2018, 2019, 2020, and 2021 stand out as having the smallest total crop cover area.This observation aligns with the outcomes presented in Table 2 and depicted in Fig. 7, where these years correspond to periods characterized by dry climatic conditions and a reduction in the extent of the vegetation.Conversely, the most extensive agricultural land coverage was observed during the years 2006, 2007, and 2010, this finding coincided with periods of the least pronounced drought within the studied statistical timeframe.
In Fig. 12, the changes in cropland cover during the spring season within the studied region, for the distinct EVI classes, are depicted.Across all classes, a negative trend is observed throughout the period spanning 2001 to 2021, with this decline becoming particularly pronounced from 2018 onwards.The findings presented in Table 7 reveal noteworthy insights.Specifically, EVI classes 0.1-0.2 and 0.2-0.4showed a statistically significant and negative trend (at a significance level of 0.05), and the total annual area had a statistically significant and negative trend (at a significance level of 0.01).Conversely, for EVI classes 0.4-0.6 and >0.6 insignificant negative trends were observed.

Relationship between cropland cover changes and the SPI
In Table 7 the correlation coefficients between changes in distinct EVI classes for agricultural land cover and precipitation are presented.The findings indicate a significant correlation (at a significance level of 0.05) between winter precipitation and spring agricultural coverage.This implies that an augmentation in rainfall is associated with an augmentation in vegetation cover.As illustrated in Fig. 13, a discernible pattern emerges wherein the reduction in precipitation during the winter season corresponds to a negative trend in agricultural land coverage within the Gavkhouni catchment region in the years from 2012 to 2021.During the spring season, although the trend may lack significance, a decrease in rainfall during this period aligns with reduced cropland coverage within the studied area.

DISCUSSION
This study focused on investigating the trends in vegetation areas and their correlation with precipitation in the Gavkhouni catchment area over a period of 21 years  using remote sensing data.The study region exhibited diverse vegetation patterns, with the northwest, west, parts of the southwest, and northeast being primarily characterized by pastures and lacking forest areas, which resulted in the prevalence of weak vegetation cover (0.1-0.2) in these regions.The maps of vegetation changes during the 21-year period and the Land Cover map of the Gavkhouni catchment area indicated that most of the area remained unchanged, thus signifying the absence of vegetation in these regions.However, certain parts, such as the northwest, certain areas in the centre, and limited regions in the northeast and southwest of the basin, experienced an upward trend in vegetation, while the central basin area and its extension towards the southwest and southeast exhibited a declining trend over the studied period which is consistent with the studies of Mirahsani et al. (2019).The investigation specifically focused on changes in vegetation areas during the spring seasons.Findings showed  that the years 2001, 2002, 2008, and 2018 had the lowest total spring vegetation area based on the EVI, with percentages of 15.53, 17.8, 18.39, and 17.3, respectively. By contrast, the years 200615.53, 17.8, 18.39, and 17.3, respectively. By contrast, the years , 200715.53, 17.8, 18.39, and 17.3, respectively. By contrast, the years , 201015.53, 17.8, 18.39, and 17.3, respectively. By contrast, the years , 201315.53, 17.8, 18.39, and 17.3, respectively. By contrast, the years , 201915.53, 17.8, 18.39, and 17.3, respectively. By contrast, the years , and 2020 exhibited the highest level of vegetation cover, with percentages of 25.19, 25.9, 28.2, 27.4, 26.8, and 26.3, respectively, in the total catchment area (41 552.3 km 2 ).In the studies of Mahmoud et al. (2022) and Razipour (2019), the years 2001, 2008 and 2018 had the lowest level of spring vegetation and the years 2010 and 2019 had the highest.Furthermore, the VCI indicated that the highest dry area occurred in the years 2001, 2008, 2011, and 2018, with percentages of 83.4, 78.9, 68.3, and 79.3, respectively, of the total area.Conversely, the lowest percentage of drought was recorded in the years 2006, 2007, 2010, and 2013, with percentages of 23.1, 16.2, 21.4, and 12.6, respectively.The regions with the largest dry areas were identified in the north, northeast, east, southeast, and in certain parts of the centre, corresponding to cities such as Kuhpayeh, Antkhort, Mimeh, Izdkhasht, Varzaneh, Alavijeh, and Dehgh.During dry conditions, the central and northwestern regions of the basin experienced almost normal conditions.The results produced by the central basin areas were primarily reliant on urban vegetation, forest parks in Isfahan, Zarinshahr, and Fouladshahr, and irrigated agricultural lands.Also, the northwest of the basin, benefiting from the presence of rivers and the Zayandeh Dam, showed a relatively higher vegetation cover as compared to other areas.The primary vegetation in this region was pasture and grassland, this was concentrated mainly in the highlands.
Further analysis revealed a significant correlation (R = 0.5) between EVI and precipitation, indicating the existence of a significant relationship between the two parameters.Notably, the years 2001, 2008, and 2021 produced the lowest precipitation levels with 49.5, 32.26, and 37.4 mm, respectively, whereas the years 2007 and 2019 recorded the highest average rainfall with 133.9 and 116.96 mm, respectively, during the period.The observed vegetation cover aligned with these precipitation patterns, with the low rainfall years producing a decreased vegetation cover due to reduced soil moisture, leading to stress conditions for the vegetation.On the other hand, years with abovenormal rainfall produced increased vegetation cover.
A statistical analysis revealed varying trends among the different vegetation classes based on the EVI achieved during the studied period.Specifically, an EVI range of 0.1 to 0.2 demonstrated a significantly increasing trend (R = 0.5) during the spring season, with a confidence level of 95%.Conversely, poor (0.1 to 0.2) and moderate vegetation (0.2 to 0.4) exhibited a negative correlation with dry areas at a significance level of 0.01.The good vegetation (0.4 to 0.6) displayed a negative trend without significance, while the extensive vegetation class (above 0.6) exhibited a significant negative trend at a level of 0.05 which is consistent with the studies of Syed et al. (2021).
The findings of the study indicated a significant correlation (R = 0.5) between EVI and precipitation, thereby highlighting the substantial relationship between these two parameters which is consistent with the studies of Rousta et al. (2020b).The years 2001The years , 2008The years , and 2021 produced the lowest rainfall with 49.5, 32.26, and 37.4 mm, respectively, while the years 2007 and 2019 recorded the highest average rainfall with 133.9 and 116.96 mm, respectively, throughout the period.As a consequence, the years characterized by a lower rainfall exhibited a decrease in vegetation cover, while the years with above-average precipitation witnessed higher vegetation levels.These results underscore the pivotal role of precipitation and soil moisture in influencing EVI variations in the studied region.The years 2001The years , 2008The years , and 2021 were marked by decreased levels of soil moisture due to insufficient rainfall, and exhibited the lowest coverage, which resulted in stressful conditions for vegetation across an area of 1 290.51, 1 528.93, and 1 783.42 km 2 , respectively.Also, the decrease in precipitation in the winter and spring seasons from 2012 to 2021 caused a sharp decrease in crop area coverage, which is similar to the results achieved by Jafari and Hasheminasab (2017).

CONCLUSIONS
An important and noteworthy concern arises from the observed negative trend in vegetation coverage within the central region of the Gavkhouni catchment basin.This decline is particularly pronounced in densely populated urban areas, such as the metropolis of Isfahan, it is being driven by urbanization-induced population growth and construction activities.Also, a reduction in vegetation is evident in irrigated agricultural lands located in Zarinshahr and in the vicinity of Isfahan.The convergence of these factors has led to the erosion of ecological stability in these locations.A significant contributing factor is the diminishing water storage capacity in the Zayandeh-Roud dam, this was caused by reduced rainfall patterns and inefficiencies in water resource management.The root cause of the reduced vegetation is complex, it involves a combination of climatic, social, and cultural factors.Alongside urbanization and agricultural expansion, factors such as shifting climatic patterns, alterations in land use, and human activities have contributed to the decline in vegetation in the region.The repercussions of environmental instability are being manifested in various ways, including the onset of desertification, soil erosion, and climate change-related disruptions.The compounding effects of these factors will have profound economic, social, and climatic consequences, thereby underscoring the urgency of addressing these issues.The development of comprehensive strategies aimed at conserving and restoring vegetation cover in the Gavkhouni catchment basin are a necessity.In order to achieve this, integrated approaches that encompass sustainable land management, the optimization of water resources, and urban planning measures must be implemented to mitigate any further degradation and safeguard the ecological equilibrium of the area.It is imperative to recognize the cascading effects of vegetation loss and also to adopt evidence-based policies to foster ecosystem resilience and sustainable development.Collaborative efforts among scientific communities, governmental bodies, and local stakeholders have the potential to create harmonious landscapes that balance human well-being with ecological vitality, thereby paving the way for a more prosperous and resilient future in the Gavkhouni catchment area.

Fig. 1 .
Fig. 1.Geographical location of the Gavkhouni watershed between the sub-basins of Iran (right), and a digital elevation model (DEM) map of the Gavkhouni watershed.

Fig. 4 .
Fig. 4. Average vegetation cover changes from spring to winter during the period of 2001-2021.

Fig. 5 .
Fig. 5. Trends in EVI changes in the Gavkhouni catchment during the period of 2001-2021.

Figure 9
Figure9depicts the relationship between the EVI and the SPI during the spring season, this reflects the amount of rainfall in the Gavkhouni catchment area.The results reveal the statistically significant relationship between EVI and precipitation, as denoted by the significant correlation coefficient of R = 0.5.Notably, the years 2001, 2008, and 2021 exhibit the lowest precipitation levels, with recorded values of 49.5, 32.26, and 37.4 mm, respectively, while the years 2007 and 2019 manifested the highest average rainfall during the period, reaching 133.9 and 116.96 mm, respectively.As can be seen in Fig.9, during the years of scarce rainfall, vegetation cover experienced a decline, whereas the years with above-normal precipitation witnessed a rise in vegetation cover.As a consequence, it may be inferred that precipitation and soil moisture play a decisive role in modulating the fluctuations of EVI in the studied area.Specifically, the years 2001, 2008, and 2021, with their diminished vegetation coverage and areas spanning 1 290.51, 1 528.93, and 1 783.42 km 2 , respectively, experienced water stress-induced conditions that impacted vegetation.Conversely, the years 2007 and 2019 were characterized by heightened precipitation during the period, this fostered increases in vegetation, encompassing areas of 2 158.58 and 2 230.5 km 2 , respectively.One intriguing observation pertains to the springtime periods of 2011 and 2018, these were previously classified as dry years, with dry area percentages of 68.3 and 79.3, respectively.However, Fig.10reveals that during these two spring seasons, rainfall remained within the normal range.An analysis of the obtained data indicates that the average rainfall in the winter season of 2011 was 83.6 mm, which stood below the overall period's average of 99.1 mm.Similarly, the spring season of 2018 produced an average rainfall of 63.7 mm, which was lower than the statistical average of 70.5 mm for the period.Moreover, during the winter season of 2018, the recorded rainfall not only fell below the average for the period but also deviated below the normal threshold.Additionally, it was previously noted

Ta b l e 3 .
Correlation of the area of different EVI classes with the dry area in the Gavkhoni catchment during the statistical period of 2001Correlation significant at: *p = 0.05 and **p = 0.01.

Fig. 9 .Fig. 10 .
Fig. 9. Time series of the vegetation anomaly with the SPI anomaly in the spring season in the Gavkhouni catchment area 2001-2021.
highlights the distribution of precipitation levels concerning the corresponding vegetation areas.Apart from the spring vegetation area, all other seasonal and annual vegetation Ta b l e 4. Correlation matrix of EVI with precipitation in Gavkhoni catchment during the statistical period of 2001

Fig. 11 .
Fig. 11.Distribution of the total vegetation area for the four seasons and the annual rainfall in the Gavkhouni catchment area.

Fig. 12 .
Fig. 12.Time series of different EVI classes for agricultural lands during the period of 2001-2021.

Fig. 13 .
Fig. 13.Time series of the agricultural land cover anomaly and precipitation in spring and winter during the period 2001-2021.
, and the lowest in 2001, with 1 363.43 km 2 .Good vegetation (0.4-0.6) reached its peak in 2006, it encompassed 584.65 km 2 , while the lowest extent was found in 2018, with 52.37 km 2 .Lastly, the extensive vegetation (EVI more than 0.6) peaked in 2004, covering 7.13 km 2 , while the smallest areas were recorded in 2001 and 2011, at 0.77 km 2 .The 21-year average area for each vegetation class is 7 050.7, 2 025.7, 197.39, and 2.42 km 2 , respectively.Area of coverage (km 2 ) of different EVI classes in the spring season in the Gavkhoni catchment area during the period of 2001-2021 Changes in the percentage of the total vegetation area of the four seasons and also annual changes in the Gavkhoni watershed 2001-2021 Area of coverage of different EVI classes in the spring season of agricultural lands (km 2 ) during the period of 2001-2021