Mapping Land Surface Drought in Water-Scarce Arid Environments Using Satellite-Based TVDI Analysis

Abstract

Drought, a natural phenomenon intricately intertwined with the broader canvas of climate change, exacts a heavy toll by ushering in acute terrestrial water scarcity. Its ramifications reverberate most acutely within the agricultural heartlands, particularly those nestled in arid regions. To address this pressing issue, this study harnesses the temperature vegetation dryness index (TVDI) as a robust drought indicator, enabling a granular estimation of land water content trends. This endeavor unfolds through the sophisticated integration of geographic information systems (GISs) and remote sensing technologies (RSTs). The methodology bedrock lies in the judicious utilization of 72 high-resolution satellite images captured by the Landsat 7 and 8 platforms. These images serve as the foundational building blocks for computing TVDI values, a key metric that encapsulates the dynamic interplay between the normalized difference vegetation index (NDVI) and the land surface temperature (LST). The findings resonate with significance, unveiling a conspicuous and statistically significant uptick in the TVDI time series. This shift, observed at a confidence level of 0.05 (Z_S = 1.648), raises a crucial alarm. Remarkably, this notable surge in the TVDI exists in tandem with relatively insignificant upticks in short-term precipitation rates and LST, at statistically comparable significance levels. The implications are both pivotal and starkly clear: this profound upswing in the TVDI within agricultural domains harbors tangible environmental threats, particularly to groundwater resources, which form the lifeblood of these regions. The call to action resounds strongly, imploring judicious water management practices and a conscientious reduction in water withdrawal from reservoirs. These measures, embraced in unison, represent the imperative steps needed to defuse the looming crisis.

1. Introduction

Drought, a consequential natural hazard, exerts profound impacts on hydrological ecosystems. It manifests through an abnormal, sustained absence of precipitation, creating a protracted dry weather pattern. Recognizing drought as a multifaceted phenomenon, comprehensive studies have become imperative for effective water resource management, compelling diverse attempts to define and understand it across various sectors. Within the lexicon of drought classifications, distinct categories emerge, each addressing specific aspects of its impact. Agricultural drought unfolds as a critical facet, characterizing the overall deficit in terrestrial water content. This deficiency poses a substantial threat to crops, creating conditions ripe for agricultural challenges and potentially jeopardizing food security [1,2]. Meteorological drought, another pivotal classification, pivots on the assessment of precipitation and evapotranspiration rates. The interplay between these meteorological factors contributes significantly to the manifestation and severity of drought conditions [3,4]. Finally, hydrological drought extends its purview to the intricate dynamics of water flow in streams and runoff waters. This classification delves into the availability and distribution of water within the broader hydrological system, encapsulating the impact on water courses and aquatic ecosystems [5]. The varied nature of these drought classifications underscores the necessity of a holistic approach to comprehend the diverse dimensions of this natural hazard. By investigating drought through different perspectives, researchers gain a more nuanced understanding of its impacts on human societies, agriculture, meteorological patterns, and aquatic ecosystems.

Historically, the practice of monitoring drought predominantly relied on climatic data, employing indices such as the Palmer drought severity index (PDSI), standardized precipitation index (SPI), and standardized precipitation evapotranspiration index (SPEI). While proven effective in discerning drought patterns, these approaches incurred substantial costs attributed to the installation and maintenance of an extensive network of meteorological stations. The financial burden associated with sustaining these stations, particularly in vast geographic areas, prompted the exploration of alternative methodologies. In response to the financial constraints of traditional methods, the paradigm shifted towards the integration of remote sensing technologies (RSTs), offering a cost-effective and efficient means of capturing drought conditions, especially in regions where ground measurements are inadequate [6,7,8,9]. This transition represents a significant advancement in the field, capitalizing on technological innovations to overcome the limitations posed by the traditional approach. One of the widely adopted and versatile remotely sensed indices in this transition is the normalized difference vegetation index (NDVI). This index derives from the ratio of the difference between the maximum visible red band and the maximum reflectance of the near-infrared spectrum [10,11,12,13]. The NDVI’s sensitivity to water stress, affecting the growth and vitality of vegetation, positions it as a key indicator of drought conditions in various scientific investigations [14,15,16]. The utilization of the NDVI in drought assessment underscores its effectiveness in capturing variations in vegetation health and, by extension, water availability [13]. By leveraging the spectral reflectance characteristics of different bands, the NDVI facilitates a nuanced understanding of the impact of water stress on vegetation, offering a valuable tool for monitoring drought conditions over large spatial extents. This transition from traditional climatic data reliance to the integration of RSTs, particularly employing the NDVI, marks a pivotal shift in the field of drought monitoring.

A pivotal advancement in the realm of remotely sensed drought indices is the temperature vegetation dryness index (TVDI), representing a significant leap forward in understanding land water content conditions. The TVDI emerges from an empirical parameterization that establishes a relationship between two critical variables: the NDVI and the land surface temperature (LST). This innovative index integrates these parameters to provide a more comprehensive depiction of terrestrial water content dynamics [17,18,19]. The LST, acting as a crucial component of the TVDI, holds an indirect yet influential relationship with the land water content by regulating canopy temperature. This relationship serves as an indicator of the overall water stress experienced by the soil. As the LST interacts with the surrounding environment, the sensitivity it exhibits to variations in soil water content becomes a valuable asset in gauging the level of soil water stress [20,21,22]. The triangulation of information facilitated by the TVDI is instrumental in its effectiveness as a land surface drought assessment tool. This triangulation is based on the distribution of pixels in the characteristic space defined by the LST and NDVI. By employing a triangular method, the TVDI captures the intricate interplay between the land surface temperature and vegetation characteristics, offering nuanced insights into both water availability and canopy resistance. The TVDI, therefore, emerges as a powerful and efficient instrument in unraveling the complexities of soil drought conditions [23,24]. The TVDI’s capacity to consider the dynamic relationship between the land surface temperature and vegetation health ensures a comprehensive evaluation of soil water content conditions. The effectiveness of the TVDI in soil drought assessment extends beyond traditional methods, offering a more sophisticated and accurate representation of the intricate balance between water availability and vegetation response.

The many investigations conducted by Gong, Zhao [25], De Keersmaecker, Lhermitte [26], Łabędzki [27], Riley, Calhoun [28], and Šebenik, Brilly [29] underscore the persistent reliance on and effectiveness of traditional climatic drought monitoring indices on a global scale. These indices, such as the PDSI and SPI, have proven instrumental in characterizing and understanding drought patterns. Despite their efficacy, these traditional approaches face a significant challenge in terms of their applicability in extensive areas, primarily due to the prohibitive costs associated with establishing and maintaining meteorological stations. In response to the limitations imposed by the high costs and logistical challenges of meteorological stations, recent studies by Tao, Ryu [21], Helali, Asaadi [30], Mehravar, Amani [31], and Shashikant, Mohamed Shariff [32] have accentuated the escalating prominence of remote sensing techniques as a viable and cost-effective alternative for monitoring drought conditions. The NDVI, initially introduced by Bhandari, Kumar [33], continues to stand out as a pivotal tool in drought assessment. The NDVI, based on the ratio of the difference between the maximum visible red band and the maximum reflectance of the near-infrared spectrum, proves highly sensitive to water stress affecting vegetation growth. This sensitivity allows the NDVI to serve as a cornerstone in various drought studies, as evidenced by its incorporation in the research conducted by Elkatoury, Alazba [10], Ding, He [15], Wang, Li [16], Shashikant, Mohamed Shariff [32], and Guo, Han [34]. The enduring relevance of traditional climatic indices emphasizes their historical significance and the wealth of knowledge they have provided in understanding and characterizing drought. However, the recognition of the limitations imposed by these methods, particularly in terms of their cost and spatial coverage, has spurred a paradigm shift towards embracing remote sensing techniques. This transition not only addresses the challenges associated with traditional methods but also opens avenues for more comprehensive and cost-effective drought monitoring, enabling researchers to explore and understand the intricacies of drought conditions in diverse and expansive geographic settings.

The primary objectives of this study are focused on a meticulous assessment of soil water content conditions within the arid region of At-Tawdihiya in the Kingdom of Saudi Arabia (KSA), utilizing the TVDI as an integrated measure of the LST and NDVI. This approach seeks to unravel the intricate dynamics of soil water content in a challenging arid environment. The overarching goal is to provide a comprehensive understanding of agricultural drought in At-Tawdihiya, with the anticipation that the findings can be extrapolated to serve as a microcosm for neighboring regions grappling with analogous challenges. The significance of this research becomes evident in its potential to yield valuable insights into the nuanced complexities of soil water content dynamics in arid climates, leveraging the sophistication of advanced remotely sensed indices like the TVDI. By focusing on At-Tawdihiya as a representative case study, this research aims to shed light on the specific challenges and opportunities inherent in managing agricultural drought in arid regions. The knowledge gleaned from this investigation holds the promise to inform and guide water resource management strategies not only in At-Tawdihiya but also in analogous arid areas within the Kingdom of Saudi Arabia and beyond. The broader implications of this research extend to regions where water resources are scarce, and climate change exerts a profound impact. By understanding the agricultural drought situation in At-Tawdihiya, this study aims to provide actionable insights that can be incorporated into adaptive water resource management strategies. The limited water resources in arid climates make effective management strategies crucial, and the findings of this research are poised to contribute to the development of sustainable and resilient approaches in the face of evolving climatic patterns. In conclusion, this study serves a pivotal role in bridging the existing gaps in the scientific literature by applying the TVDI as a specialized tool for monitoring soil drought in the unique context of arid climates. By doing so, it makes a novel contribution to the field of drought monitoring, offering a contextualized understanding of soil water content dynamics in regions characterized by aridity. The anticipated insights derived from this research endeavor are poised to advance the collective understanding of soil water content dynamics, laying the foundation for more effective drought monitoring strategies and informed water resource management practices in arid regions globally.

2. Methodology

2.1. Study Area Description

The investigation was meticulously conducted in the expansive region of At-Tawdihiya farm, Al-Kharj Governorate of the Kingdom of Saudi Arabia (KSA), covering approximately 205 km². Positioned strategically, it is situated within the geographical coordinates delineated by latitudes ranging from 24°09′35″ to 24°13′08″ N and longitudes spanning from 47°54′38″ to 48°05′40″ E (Figure 1). The topography of this region unfolds at an average elevation of around 391 m above sea level, contributing to its distinctive character. At-Tawdihiya experiences a quintessential desert climate, marked by scorching summers and cool winters. A notable feature of this climatic profile is the inverse relationship between humidity and temperature. As temperatures rise, relative humidity decreases, leading to arid atmospheric conditions. Precipitation in At-Tawdihiya is irregular, with sporadic occurrences scattered throughout the year. Significantly, a substantial portion of the annual precipitation tends to concentrate within a limited number of days, emphasizing the area’s arid climate profile. This climatic backdrop poses unique challenges and opportunities that are instrumental in shaping the focus and outcomes of this research study. At-Tawdihiya is strategically positioned within the mentioned coordinates, encapsulating diverse topographical features. The region’s elevation of 391 m influences local climate patterns and hydrological processes. Notably, the terrain exhibits variations that may affect the land water content distribution, making it an ideal locale for studying land drought dynamics. The study area comprises a mosaic of land cover types, including natural vegetation, agricultural fields, and potentially urban or infrastructure zones. Understanding the land cover is crucial for interpreting satellite-derived indices like the temperature vegetation dryness index (TVDI), as different surfaces exhibit distinct thermal and vegetation responses to drought conditions. At-Tawdihiya was selected for this investigation due to its representative nature of arid climates, data availability, and limited water resources in the Kingdom of Saudi Arabia. The unique climatic and topographic features provide an excellent case study for understanding soil drought dynamics, making the findings applicable to similar arid regions globally. The challenges posed by irregular precipitation and the distinct climatic characteristics of At-Tawdihiya underscore its significance in advancing our understanding of soil drought monitoring methodologies.

2.2. Data Manipulation and Drought Index Modeling

This study relied on three primary datasets consisting of remote sensing and monthly precipitation data. The remote sensing data obtained from the Landsat satellite images was sourced from the Earth Explorer service provided by the United States Geological Survey (USGS) “https://earthexplorer.usgs.gov/ (accessed on 1 November 2018)”. Two multi-sensor satellite missions, namely Landsat 7, with enhanced thematic mapper plus (ETM+), and Landsat 8, with operational land imager/thermal infrared sensor (OLI/TIRS), together captured around 144 satellite pictures. The Landsat 7 satellite images were acquired from January 2010 to June 2014, while the Landsat 8 satellite images were acquired from July 2014 to December 2015. The two satellite missions possess a recurring duration of 16 days and exhibit a pixel resolution capacity of 30 m. Both missions were prioritized based on the availability of digital photos with the necessary bands. The monthly mean of each of the two photographs was calculated, resulting in a total of 72 digital images studied over the research period. Additionally, a total of seven historical digital photos, namely those taken in December 1985, were acquired from Google Earth Pro in order to provide a comprehensive understanding of the conditions of agricultural areas within the study. The historical photos were used to compare land use changes over time and provide context for the current state of the agricultural areas. This combination of recent satellite data and historical imagery allowed for a thorough analysis of the study area’s land cover dynamics.

The monthly precipitation data’s time series average was acquired from the tropical rainfall measurement mission’s TRMM data, which was accessed via the website “https://giovanni.gsfc.nasa.gov/giovanni/ (accessed on 1 January 2019). The TRMM data was assigned the same time period as the Landsat satellite images. The use of this particular data has been effectively employed in several study investigations [35,36,37]. This integration of various data sources provided a comprehensive understanding of the changes in land cover over time. By combining satellite data with precipitation data, researchers were able to analyze the relationship between land cover dynamics and climate patterns in the study area. The modeling of the drought index used an approach including three independent stages. The process of generating land surface temperature (LST) maps involves the conversion of raster maps. The subsequent step involves the generation of the normalized difference vegetation index (NDVI). The third and final step of this study includes merging the LST and NDVI maps to generate the temperature vegetation dryness index (TVDI), which serves as an indication of drought and soil water content in the research area. Furthermore, the following section offers a thorough explanation of each of the aforementioned techniques. The TVDI is a valuable tool for monitoring and assessing drought conditions, as it combines information from both the LST and NDVI. By merging these two datasets, researchers can gain a comprehensive understanding of the environmental conditions in the study area.

The satellite data employed in this study have been previously validated in our earlier publications [9,10]. In that study, we conducted a comprehensive validation of the Landsat-derived LST, NDVI, and TVDI against ground measurements and the MODIS data within the same arid environment. The results indicated strong correlations and acceptable accuracy levels, thus confirming the reliability and robustness of the satellite-based indices for drought and land surface monitoring in such contexts.

2.3. Land Surface Temperature (LST)

The use of the RSTs has seen a recent surge in popularity owing to their capacity to provide immediate and all-encompassing perspectives for a particular region. The LST may be obtained by using RST methods to analyze the thermal infrared bands of Landsat 7 and 8. Thermal bands are often available in the form of digital numbers (DNs). Consequently, the satellite images acquired were transformed from absolute spectral radiance to integer spectral radiance. The aforementioned procedure was accomplished by using picture elements that were acquired with the satellite images in the form of a metadata file. The decision to convert was made due to the fact that the DNs have dimensionless values. The collected pictures from both the Landsat flights were transformed into spectral radiance using Equations (1) and (2).

For Landsat 7:

$${\mathrm{L}}_{\mathsf{\lambda }}={\displaystyle \frac{\left({\mathrm{L}}_{{\mathsf{\lambda }}_{\mathrm{m}\mathrm{a}\mathrm{x}}}-{\mathrm{L}}_{{\mathsf{\lambda }}_{\mathrm{m}\mathrm{i}\mathrm{n}}}\right)}{({\mathrm{Q}\mathrm{c}\mathrm{a}\mathrm{l}}_{\mathrm{m}\mathrm{a}\mathrm{x}}-{\mathrm{Q}\mathrm{c}\mathrm{a}\mathrm{l}}_{\mathrm{m}\mathrm{i}\mathrm{n}})}}\times \left(\mathrm{D}\mathrm{N}-{\mathrm{Q}\mathrm{c}\mathrm{a}\mathrm{l}}_{\mathrm{m}\mathrm{i}\mathrm{n}}\right)+{\mathrm{L}}_{{\mathsf{\lambda }}_{\mathrm{m}\mathrm{i}\mathrm{n}}}$$

(1)

For Landsat 8:

$${\mathrm{L}}_{\mathsf{\lambda }}={\mathrm{M}}_{\mathrm{l}}\times \mathrm{Q}\mathrm{c}\mathrm{a}\mathrm{l}+{\mathrm{A}}_{\mathrm{L}}$$

(2)

The spectral band radiance at a specific wavelength (W∙m⁻² sr μm) is denoted as ${\mathrm{L}}_{\mathsf{\lambda }}$. The digital number is represented as DN. The minimum and maximum spectral radiances for each band, measured in W∙m⁻² sr μm, are denoted as ${\mathrm{L}}_{{\mathsf{\lambda }}_{\mathrm{m}\mathrm{i}\mathrm{n}}}$ and ${\mathrm{L}}_{{\mathsf{\lambda }}_{\mathrm{m}\mathrm{a}\mathrm{x}}}$, respectively. The minimum and maximum quantized calibrated pixel values at ${\mathrm{L}}_{{\mathsf{\lambda }}_{\mathrm{m}\mathrm{i}\mathrm{n}}}$ and ${\mathrm{L}}_{{\mathsf{\lambda }}_{\mathrm{m}\mathrm{a}\mathrm{x}}}$ are denoted as ${\mathrm{Q}\mathrm{c}\mathrm{a}\mathrm{l}}_{\mathrm{m}\mathrm{i}\mathrm{n}}$ and ${\mathrm{Q}\mathrm{c}\mathrm{a}\mathrm{l}}_{\mathrm{m}\mathrm{a}\mathrm{x}}$, with DN values of 1 and 225, respectively. ${\mathrm{M}}_{\mathrm{l}}$ is the factor used for multiplying the brightness of the band; ${\mathrm{A}}_{\mathrm{L}}$ represents the factor used for adding radiance to the band; and $\mathrm{Q}\mathrm{c}\mathrm{a}\mathrm{l}$ represents the pixel value in DNs.

The surface temperature was determined by using the calculation provided in the Landsat handbook.

$$\mathrm{L}\mathrm{S}\mathrm{T}={\displaystyle \frac{{\mathrm{K}}_2}{\mathrm{l}\mathrm{n}(\frac{{\mathrm{K}}_1}{{\mathrm{L}}_{\mathsf{\lambda }}}+1)}}$$

(3)

The LST is measured in Kelvin, whereas the coefficients ${\mathrm{K}}_1$ and ${\mathrm{K}}_2$, as shown in

Table 1. Conversion constants (K₁ and K₂) for the calibration of the thermal spectral bands.

Satellite Mission	Band Number	K1, W∙m−2∙sr−1∙μm−1	K2, Kelvin
Landsat 7	6	666.09	1282.71
Landsat 8	10	774.89	1321.08
	11	480.89	1201.14

W∙m⁻²∙sr⁻¹∙μm⁻¹ = Watt per square meter per steradian per micrometer.

, are determined by the effective wavelength of the satellite sensor, as stated in the metadata.

The data accuracy derived from the Landsat missions was assessed by comparing it with the time series data acquired from the moderate-resolution imaging spectroradiometer (MODIS). The MODIS data time series was obtained from the website https://giovanni.gsfc.nasa.gov/giovanni/ (accessed on 1 January 2019). In a similar vein, Chen, Son [38] conducted a verification of the LST data by means of a comparative analysis using the MODIS data. Figure 2 and Figure 3 are used to assess the appropriateness of the LST data for this research. Based on the data shown in Figure 2, it can be seen that the frequency distribution of both the Landsat and MODIS LST adheres to a normal distribution. According to the data shown in Figure 3, there are no atypical data points within the dataset that possess the potential to significantly impact the ultimate outcomes. A statistically significant association has been observed between the LST time series derived from the Landsat satellite images, with a significance threshold of 0.05 (p-value = 0.005). Approximately a 69.8% disparity has been observed between the LST derived from the Landsat pictures and the LST derived from the MODIS data. The equation derived from the regression model is as follows: LST_Landsat = 0.9189LST_MODIS + 24.4.

2.4. Normalized Difference Vegetation Index (NDVI)

The Normalized Difference Vegetation Index (NDVI) is a widely used metric for evaluating the condition of vegetation, particularly in relation to vegetation attributes like the leaf area index (LAI). The NDVI values range from −1.0 to 1.0, with values close to 0 indicating the absence of green patches. According to Elkatoury, Alazba [39], greater NDVI values are indicative of dense vegetation, while lower values suggest the existence of stressed or non-green vegetation regions.

The NDVI is a metric used to assess the state of vegetation by comparing the absorption of red and near-infrared bands by vegetation, as measured by Landsat 7 and Landsat 8. The equation for calculating the NDVI is as follows:

$$\mathrm{N}\mathrm{D}\mathrm{V}\mathrm{I}={\displaystyle \frac{({\mathsf{\lambda }}_{\mathrm{n}\mathrm{i}\mathrm{r}}-{\mathsf{\lambda }}_{\mathrm{r}\mathrm{e}\mathrm{d}})}{({\mathsf{\lambda }}_{\mathrm{n}\mathrm{i}\mathrm{r}}+{\mathsf{\lambda }}_{\mathrm{r}\mathrm{e}\mathrm{d}})}}$$

(4)

The near infrared band, denoted as ${\mathsf{\lambda }}_{\mathrm{n}\mathrm{i}\mathrm{r}}$, is represented by bands 4 and 5 in Landsat 7 and 8, respectively. Similarly, the red band, denoted as ${\mathsf{\lambda }}_{\mathrm{r}\mathrm{e}\mathrm{d}}$, is represented by bands 3 and 4 in Landsat 7 and 8, respectively.

The research region exhibits a range of NDVI values, spanning from a minimum of −0.14 to a high of 0.50, with a standard deviation (±SD) of 0.093. Figure 4 depicts the geographical distribution of the mean NDVI across the duration of this research. The figure illustrates that a significant portion of the examined area exhibits low NDVI values, with an average of 0.087 and a ±SD of 0.090. The regions that use center pivot irrigation systems or plantation lands have the highest NDVI values. Based on the histogram shown in Figure 5, it can be seen that the distribution of the study area pixels adheres to a normal distribution. Specifically, the highest number of pixels, amounting to 9683, is observed at an NDVI value of 0.074. This finding substantiates the notion that a significant section of the land within the research region remains unoccupied, devoid of any agricultural endeavors.

2.5. Temperature Vegetation Dryness Index (TVDI)

RSTs are crucial for comprehending and estimating drought conditions. The TVDI is a water content indicator that measures the severity of drought conditions. Given the premise that there exists an inverse correlation between the temperature of plants and the process of transpiration, the link between temperature and plant density is significantly influenced by the highest LST. The NDVI and LST yield a more complete dataset pertaining to the surface soil water content [17,18,19]. Hence, the TVDI serves as a drought-monitoring indicator to assess the drought pattern over the designated research region. The fundamental principle behind the TVDI is derived from the empirical correlation seen in the NDVI-LST triangle, as shown in Figure 6. The TVDI is clearly correlated with the level of vegetation in a given location, often serving as an indicator of drought conditions. Additionally, the maximum temperature seen in the NDVI-LST triangle corresponds to an arid region characterized by the lowest soil water content, assuming equivalent plant cover. The wet edge with the greatest surface water content is represented by the lowest temperature value in the triangle. The calculation of the TVDI is represented by Equation (6).

$$\mathrm{T}\mathrm{V}\mathrm{D}\mathrm{I}={\displaystyle \frac{(\mathrm{L}\mathrm{S}\mathrm{T}-{\mathrm{L}\mathrm{S}\mathrm{T}}_{\mathrm{m}\mathrm{i}\mathrm{n}})}{(\mathrm{a}+\mathrm{b}\mathrm{N}\mathrm{D}\mathrm{V}\mathrm{I})-{\mathrm{L}\mathrm{S}\mathrm{T}}_{\mathrm{m}\mathrm{i}\mathrm{n}}}}$$

(5)

The determination of the position of each point (pixel) inside the NDVI-LST triangle is influenced by several parameters, including evapotranspiration, fractional plant cover, net radiation, surface roughness, and thermal characteristics of the surface, as stated by Sandholt, Rasmussen [40]. The LST is a thermal characteristic that is highly influenced by the status of vegetation and soil water content. The link between land surfaces and the LST has been investigated in several studies using the NDVI and satellite imagery [41,42,43].

The constants (a, b) in the triangle represent the intercept and slope of the dry edge. They are derived from the correlation between the NDVI and LST (${LST}_{max}$ = a + bNDVI), where ${LST}_{max}$ is the highest temperature on the Earth’s surface for a certain NDVI. The triangular region is enclosed by a horizontal line at a lower position, denoting the wet edge (${LST}_{min}$). In order to determine the constants (a, b) for each pixel, the curve fit for a pixel level raster regression tool was used. The average of the raster maps for the (a, b) constants during the research time limitations is shown in Figure 7 and Figure 8. Based on the zonal statistics of raster maps, the constant (a) in the research region has a range of values from 0.011 to −0.0094. The average value of the constant is 0.0012, with a standard deviation of 0.0018. The constant b exhibits a maximum value of 0.604 and a lowest value of −0.343. It has an average of 0.052, with a standard deviation of 0.127.

The TVDI readings range from 0 to 1, with a value of 1.0 at the dry edge indicating the absence of surface evaporation or a restricted water content supply. In contrast, when the TVDI value is equal to 0, it signifies the wet edge, characterized by significant surface evaporation or a restricted water content supply. The TVDI is categorized into five groups that describe dryness conditions, as shown in

Table 2. Drought condition classes according to temperature vegetation dryness index (TVDI).

Drought Class	Range
Severe Drought	0.8−1
Drought	0.6−0.8
Normal	0.4−0.6
Severe Wet	0.2−0.4
Wet	0−0.2

[34]. The current investigation revealed that the relative cumulative frequencies were determined to arrange the types of drought circumstances in a certain manner. The first stage in determining the relative cumulative frequencies was the calculation of the cumulative frequencies, subsequent to the establishment of a frequency distribution table for the pixels within the designated research region. Furthermore, the cumulative frequencies that were computed were divided by the total values of the cumulative frequencies.

2.6. Mann–Kendall (MK) Test for the TVDI

The Mann–Kendall (MK) test is a nonparametric statistical test used to find and detect the presence of monotonic properties in a time series dataset, specifically in relation to persistent trending changes over time, such as growing or decreasing trends. The MK test encompasses several characteristics pertaining to time series patterns, which may be enumerated as follows: (1) the application of this test is not limited to any specific distribution, as long as it satisfies the assumptions of normality. (2) The time series data utilized in the test must be devoid of autocorrelation, indicating the absence of serial correlation. (3) The test remains applicable even in cases where there is a seasonal component present in the time series. (4) All-time series data are treated as independent variables in the test. Hamed [44] stated that the MK test relies on the correlation between the observed rankings of time series and their corresponding chronological order. The MK trend analysis statistic used in this work for the time series of the TVDI is as follows: $\mathrm{T}\mathrm{V}\mathrm{D}\mathrm{I}=\left\{\begin{array}{ccc}{\mathrm{T}\mathrm{V}\mathrm{D}\mathrm{I}}_1,& {\mathrm{T}\mathrm{V}\mathrm{D}\mathrm{I}}_2,& \dots ,{\mathrm{T}\mathrm{V}\mathrm{D}\mathrm{I}}_{\mathrm{n}}\end{array}\right\}$

$$\mathrm{S}={\sum }_{\mathrm{i}<\mathrm{j}}\mathrm{s}\mathrm{g}\mathrm{n}\left({\mathrm{T}\mathrm{V}\mathrm{D}\mathrm{I}}_{\mathrm{j}}-{\mathrm{T}\mathrm{V}\mathrm{D}\mathrm{I}}_{\mathrm{i}}\right)=\mathrm{s}\mathrm{g}\mathrm{n}\left({\mathrm{R}}_{\mathrm{j}}-{\mathrm{R}}_{\mathrm{i}}\right)=\left\{\begin{array}{c}+\begin{array}{cc}1& {\mathrm{T}\mathrm{V}\mathrm{D}\mathrm{I}}_{\mathrm{i}}<{\mathrm{T}\mathrm{V}\mathrm{D}\mathrm{I}}_{\mathrm{j}}\end{array}\\ \begin{array}{cc}0& {\mathrm{T}\mathrm{V}\mathrm{D}\mathrm{I}}_{\mathrm{i}}=\end{array}{\mathrm{T}\mathrm{V}\mathrm{D}\mathrm{I}}_{\mathrm{j}}\\ \begin{array}{cc}-1& {\mathrm{T}\mathrm{V}\mathrm{D}\mathrm{I}}_{\mathrm{i}}\end{array}>{\mathrm{T}\mathrm{V}\mathrm{D}\mathrm{I}}_{\mathrm{j}}\end{array}\right.$$

(6)

$$\mathrm{V}\mathrm{a}\mathrm{r}(\mathrm{S})=\left\{\mathrm{n}\left(\mathrm{n}-1\right)\left(2\mathrm{n}+5\right)-{\sum }_{\mathrm{j}=1}^{\mathrm{p}}{\mathrm{t}}_{\mathrm{j}}\left({\mathrm{t}}_{\mathrm{j}}-1\right)\left(2{\mathrm{t}}_{\mathrm{j}}+5\right)\right\}/18$$

(7)

The rankings of the observations of ${TVDI}_j$ and ${TVDI}_i$ are denoted as R_j and R_i, respectively. The variable Var(S) represents the variance of the MK statistic (S). The variable n denotes the sample size used for doing the trend analysis. The variable p represents the number of tied groups in the time series. Lastly, the variable $t_j$ represents the number of data points in the jth tied group.

Hence, the substantial positive values of S indicate a robust positive trend in the time series, characterized by an upward trajectory. Concurrently, the negative values of S indicate a downward tendency in the time series. If the Z-transformation is used, the statistics S may be approximated to follow a normal distribution.

$$Z=\left\{\begin{array}{ccc}{\displaystyle \frac{S-1}{\sqrt{Var\left(S\right)}}}& if& S>0\\ 0& if& S=0\\ {\displaystyle \frac{S+1}{\sqrt{Var\left(S\right)}}}& if& S>0\end{array}\right.$$

(8)

3. Results and Discussion

3.1. Soil Water Content Based on TVDI

This study estimated the TVDI value for each pixel using the NDVI-LST triangle. Figure 9 divides the drought condition into five classes based on the TVDI average from 2010 to 2015. The zonal statistics of the TVDI average classes during 2010–2015 for the study area classify 39.4% and 38.0% of the total study area as normal and drought conditions, respectively, while the rest of the area is classified as wet, severe wet, and severe drought by 10.1%, 5.1%, and 7.4% of the total area, respectively. The map reveals that most areas experiencing an increase in humidity are undergoing agricultural expansion, while the remaining areas are deserts devoid of urban activity.

This study used the cumulative relative frequency to determine which portion of the total observed pixels falls below the upper limit of the frequency period. Figure 10 shows the cumulative relative frequency of the pixel ratio of the drought categories in the region. It is noticeable from this figure that the highest cumulative relative frequency of these groups has appeared in the case of drought. Then, the rest of the categories came after the drought conditions, and they were in ascending order as follows: severe drought, normal, wet, and severe wet.

3.2. Trends of TVDI, Precipitation, and LST

Figure 11 illustrates the suitability of the created time series comprising the average TVDI drought over the study area. The normal probability plot of the residuals versus the percent of their expected value shows that most of the points approximately follow a straight line. The residuals versus fitted values indicate that the values are distributed randomly on both sides of the zero line, without any clear patterns in the values. The histogram of the residual plot indicates that the time series of the average TVDI is almost symmetric around the zero value. As shown in the residuals versus order plot, there is no pattern indicating that the trend line does not fit the TVDI data. Therefore, the trend analysis appropriately uses the TVDI time series.

In the present study, we used precipitation data to compare the sensitivity between the TVDI and precipitation. Figure 12 shows that the TVDI’s average value for the study area fluctuates over time. Several cases link the decline in the TVDI value to severe rainfall, but we cannot generalize this observation because some precipitation events may occur, offset by a decrease in the rain index value, indicating a drought. For example, the rainy conditions at observation orders of 49 to 53 (corresponding to January to May 2014) have severe drought conditions, although there is precipitation. In another way, this comparison showed that there was an unclear influence of the precipitation rate on the TVDI in the event of high monthly average precipitation, which indicated a rise in soil water content.

It is also noticeable from the same figure that there are no significant rainfall rates in the region, where the highest precipitation value in April 2013 was 52.6 mm/month. The drought associated with this amount of rain was due to dry conditions. Linking rainfall to drought in the region is very difficult, since it is difficult to understand the causes of drought. However, Ref. [45] concluded that the trend in rainfall occurrence is decreasing on a decadal basis (2001–2010). The researcher also noted that Saudi Arabia stands out due to its significant annual rainfall variability.

The El Niño Southern Oscillation (ENSO) is the natural reversal climatic cycle in the Pacific Ocean’s temperatures, winds, and clouds. Many scientists attribute the world’s uncertain climate to this phenomenon [46,47,48]. This phenomenon can shift climate drivers, such as wind patterns, clouds, perceptions, and temperatures [49]. This climatic shifting can lead to higher risks of extreme climatic conditions, e.g., severe droughts and devastating floods [50]. Consequently, the ENSO affects all human activities and ecosystems on Earth, which could be amplified due to global warming.

The Arabian Peninsula’s climate may differ due to its connection to broader climate change. Abid, Almazroui [51] conducted a study on the Arabian Peninsula, using a model capable of simulating the ENSO and related climate phenomena to predict the rainfall state in the region. This study concluded that the ENSO has an impact on the sustainability of agricultural activities in the region. Huang, Tian [52] discovered a correlation between the TVDI variation and the increased frequencies and intensities of the ENSO phenomenon.

According to many scientific studies in Saudi Arabia, e.g., [53], precipitation occurs primarily during two seasons: the dry season and the wet season. The dry period is between June and September, while the wet months are between November and April. During the dry season, about 12.94% of the total annual rainfall occurs, whereas the rest of the annual rainfall occurs in the wet season [54].

The general statistics of the average TVDI time series show a maximum value of 0.986 and a minimum value of 0.480, with an average of 0.739 and an ±SD of 0.104. The coefficient of variation (CV) is equal to 0.14. The MK test has been applied to understand the trend behavior of the average TVDI time series. According to this test, there is statistically significant evidence of an increasing trend at a significance level of 0.05 (p-value = 0.0497), with a standardized value of the MK test (Z_S) of 1.648. At the same time, when looking at the results of the MK trend test for the monthly rainfall during the study period, we find that there is no significant evidence of any general trend, either increasing or decreasing, at a level of significance equal to 0.05 (p-value = 0.396) with a Z_S of −0.264.

Figure 13 shows the time series of the average monthly temperature in the study area. This figure displays that the average temperature of the study area is almost constant throughout the study period. There were no significant variations in temperatures within the months of the same season from January 2010 until December 2015. In more detail, the lower temperatures were in the winters of 2011 and 2014, with recordings of 23.8 °C and 23.6 °C in December and January, respectively. The investigation recorded about sixteen temperature records exceeding 50 °C in June, July, August, and September. The present work observed high temperatures of 50.6, 52.2, 52.3, and 50.1 °C in July, August 2010, and September 2011, respectively. On the contrary, instead of rising drought conditions, the region was experiencing drought levels of severe wetness.

By analyzing the trend of the LST, we notice that there are no trends (either increasing or decreasing) in the temperature of the study area, which may indicate a lack of clear correlation between drought conditions and the LST in the short term. This contradicts global conditions, as ever-increasing drought frequencies have negatively impacted the environment due to the rise in temperatures [55,56]. However, given that previous studies that have taken into account the long-term trend analysis of temperatures, the LST changed dramatically over the period. Accordingly, regional temperatures increased in the context of climate change and global warming [57,58]. According to Shi, Wen [59], there is an obvious increase in consecutive hot conditions with a decrease in other weather types like snowfall, thunderstorms, and fog. The changes in these augmented temperatures can be either variability or extremes [60,61,62].

Many studies have discussed the general trends of drought and its variability around the world. Somorowska [63] pointed out that the spatial extent of droughts has demonstrated broad variability using the SPEI as a measurement of drought conditions related to agriculture and hydrology. Shiru, Shahid [64] studied the recent trends for droughts in Nigeria and found that climate trends will increase drought frequency in the future. Spinoni, Naumann [65] showed an increasing linear trend in drought variables in the period of 1950–2012 in southwestern Europe, with diminishing precipitation and rising potential evapotranspiration (PET). Coll, Aguilar [66] expected that there would be increasing drought conditions over the Iberian Peninsula due to future climate change. Spinoni, Naumann [67] have highlighted that the frequency and severity of droughts have increased in the summer and autumn over southern and eastern Europe. In addition, some of these studies have tried to find the linkage between the TVDI and other drought indices. Chen, Zhang [68] have proven that there is a strong correlation between the standardized precipitation and evapotranspiration index (SPEI) and TVDI values. As well, they are suggesting that the smaller values of the TVDI are correlated with the SPEI, which in turn depends on several factors, including precipitation and PET. The aforementioned discussions show that the general trend in droughts in many regions of the world is toward a significant increase, which is consistent with this study. The intricate phenomenon of drought cannot be simplified to a broad pattern of precipitation or specific regional temperatures.

The MK trend test was utilized for analyzing the changing drought conditions using the TVDI during the study period at a significance level of 0.05 (p-value).

Table 3. The Mann–Kendall test statistics for the percent of areas exposed to drought with an ordinary least squares (OLS) regression trend line and Theil–Sen trend line parameters (slope and intercept) for drought conditions throughout the four seasons and annual time series.

Season	Drought Condition	Mann–Kendall Test			OLS Regression Line		Theil–Sen Trend Line
		S	Zs	p-Value	Slope	Intercept	Slope	Intercept
Spring	Severe Drought	19	0.68	0.2480	0.10	9.48	0.11	9.16
	Drought	51	1.89	0.0291 **	0.35	77.26	0.24	79.41
	Normal	−65	−2.42	0.0077 *	−0.22	8.90	−0.19	8.12
	Severe Wet	−20	−0.72	0.2360	−0.01	0.54	−0.01	0.53
	Wet	−11	−0.38	0.3520	−0.08	3.34	−0.01	2.24
Summer	Severe Drought	25	0.91	0.1820	0.06	10.36	0.14	9.25
	Drought	47	1.74	0.0407 **	0.40	73.73	0.37	74.22
	Normal	−57	−2.12	0.0170 *	−0.22	9.57	−0.19	9.34
	Severe Wet	−4	−0.11	0.4550	0.00	0.73	0.00	0.53
	Wet	−25	−0.91	0.1820	−0.15	4.69	−0.06	3.27
Autumn	Severe Drought	−18	−0.64	0.2600	−0.08	13.17	−0.08	13.57
	Drought	41	1.52	0.0649	0.46	70.20	0.37	70.74
	Normal	−37	−1.36	0.0863 *	−0.21	10.24	−0.20	10.11
	Severe Wet	−3	−0.08	0.4700	0.00	0.91	−0.01	0.68
	Wet	−37	−1.36	0.0863 *	−0.22	6.04	−0.15	5.00
Winter	Severe Drought	−6	−0.19	0.4250	−0.01	11.89	−0.02	11.84
	Drought	−3	−0.08	0.4700	−0.06	78.00	−0.01	77.54
	Normal	−5	−0.15	0.4400	−0.07	7.71	−0.03	7.14
	Severe Wet	47	1.75	0.0403 **	0.05	0.39	0.04	0.35
	Wet	32	1.18	0.1200	0.09	2.01	0.11	1.57
	Severe Drought	126	0.61	0.2720	0.004	11.240	0.007	11.036
	Drought	423	2.05	0.0201 **	0.063	75.246	0.053	76.199
Annual	Normal	−584	−2.83	0.0023 *	−0.041	8.880	−0.040	8.537
	Severe Wet	16	0.07	0.4710	0.003	0.648	0.000	0.515
	Wet	−160	−0.77	0.2200	−0.020	3.894	−0.009	2.951

Z_s = standardized value of Mann–Kendall test value (S); * and ** indicate significant trend evidence of decreasing and increasing, respectively, at a significance level of 0.05.

reveals the results of the MK test throughout the four seasons of the year, furthering the annual values. The results showed that there were some differences in the seasonal and annual rates for land areas exposed to drought conditions. In the spring, the MK analysis shows that there is a significant increase in the general trend of land percentage under drought conditions at a significance level of 0.05 (p-value = 0.0291), while the overall trend of drought conditions below the normal level decreased significantly at a significance level of 0.05 (p-value = 0.0077). The land ratios for the rest of the drought conditions remained significantly unchanged in the proportion of land exposed to the other drought conditions.

In the summer, the general trend of drought conditions did not differ from the spring. The percentage area of the TVDI in drought conditions increased significantly (p-value = 0.0407). The general trend of land exposed to the TVDI with a normal condition was significantly decreased at a significance level of 0.05 (p-value = 0.0170).

In the autumn, the trend of the areas exposed to normal and wet conditions significantly decreased at a significance level of 0.05 and a p-value equal to 0.0863. While the winter showed a different trend from the rest of the three seasons mentioned earlier, the proportion of the land area exposed to the TVDI with extreme wetness decreased significantly. The overall trends of these lands remained unchanged.

Considering drought-prone land on an annualized basis, we note that drought-prone areas have increased significantly, while land with dominant normal conditions has declined significantly. The rest of the annual TVDI and the other drought conditions remained unchanged.

Thus, the MK trend analysis result shows a significant increase and decrease during the spring and summer seasons under drought and normal conditions, respectively (p-value < 0.05). The increase in drought conditions during the spring was Z_S = 1.89, and the decrease in normal conditions was Z_S = −2.42. Also, during the summer months, the Z_S of drought and normal conditions were 1.47 and −2.12, respectively. Also, Z_S in normal and wet conditions was −1.36 during the autumn season. On the other hand, severe drought conditions showed a significant increase during the winter season, with a Z_S of 1.75.

On an annual basis, the proportion of the area exposed to drought conditions revealed a significant increase of Z_S = 2.05 and a significant decrease in the normal condition, with Z_S = −2.83. Moreover, the trend line of both methods (ordinary least squares (OLSs) and Theil–Sen) shows the same trend slope in cases of significant decreases or increases during the spring, summer, autumn, and winter seasons. The approximate slope values of the OLSs and Theil–Sen trend lines are shown in Table 3. The positive slope appeared during the significant increase in drought conditions that occurred throughout the four seasons. In addition, the intercept values of the two regression trend lines were convergent.

In general, the observed TVDI trend is an increasing trend and may threaten a larger proportion of drought-prone land in the future. The current drought situation, coupled with the global lack of significant increases in precipitation and temperatures, exacerbates the conditions and accelerates the process of desertification. Therefore, care should be taken by officials to mimic nature carefully to prevent further problems associated with this phenomenon (e.g., groundwater depletion).

Figure 14 depicts the raster map of the average trend slope of the TVDI. Accordingly, the TVDI range value of the trend slope is about 0.071, with an average and an ±SD of 0.013 and 0.006, respectively. Areas with positive change values of the slope, from 0.017 to 0.052, cover most of the pixel count (909,708) of this raster map. We can also observe that the areas with a high positive slope around 0.052 are relatively few and primarily concentrated in agricultural areas. We can attribute this significant change in the TVDI to the increasing tendency of drought conditions in these regions, which result in seasonal variations in their microclimates due to cultivated crop cycles.

Figure 15 shows the spatial distribution of the TVDI in two selected months (December 2011 and December 2014) to illustrate the state of soil water content in the area under study. This figure reflects the spatial distribution of the TVDI values that occurred at the average of the maximum and minimum temperature spikes for December, which were recorded during the study period. The lowest average temperature recorded for December was in 2011 at 23.8 °C, while the highest temperature for the same month was in 2014 at 27.8 °C. The difference in the increase was 14.38%.

It is noticeable by following the map colors (Figure 15) that there has been an evident change in drought levels in these two months. The percentage of land exposed to the wet and severe wet categories increased from 5% and 6% to 8% and 9% in December 2011 and December 2014, respectively. The proportion of land subjected to the normal and severe drought categories increased from 17% to 15% to 28% and 19% in the same two months, respectively. Meanwhile, the percentage of land with drought conditions dropped from 57% in December 2011 to 36% in December 2014.

Similarly, it is possible to observe that the areas with an increasing percentage of land exposed to wet and severe wet are irrigated agricultural lands. This indicates that the source of this soil water is groundwater, in light of the low precipitation rates, as well as the lack of surface water supply in the region. This may indicate pressure to withdraw water from the groundwater reservoir through sustained pumping, which prefigures the imaginable depletion of the groundwater reservoir in the future [69,70].

Figure 16 shows the agricultural time evolution of the study area for December since 1985 at intervals of five years. This figure exhibits the extent of the scarcity that occurred in the agricultural sector. Around 64.6% of the area was under cultivation in 1985. Thereafter, a significant deterioration began, and in 1990 and 1995, this percentage of the cultivated land dropped to about 47.4% and 53.9%, respectively. After that, at the beginning of 2000, the percentage of agricultural land fell sharply to about 28.6%. This percentage continued to fluctuate up and down around this value, reaching 28.8% in December 2015. Figure 12’s trend analysis of the TVDI, which recorded a positive general trend during the study period, aligns with this figure. This situation foresees the chance of increasing drought in the region in the coming years.

The findings from our TVDI analysis offer valuable insights that can directly inform and enhance local land and water resources management in dry regions. By providing spatially explicit maps and temporal trends of drought severity, our methodology enables stakeholders to identify emerging drought conditions early, facilitating proactive water conservation and allocation measures. This can help mitigate groundwater depletion by directing resources to high-risk areas and optimizing irrigation schedules. Additionally, the trend analysis offers evidence to support the development of sustainable water management policies, emphasizing the necessity of adaptive strategies, such as crop diversification, improved irrigation practices, and groundwater recharge initiatives. The cost-effective and repeatable nature of satellite-based monitoring allows authorities to regularly assess drought conditions and evaluate the effectiveness of implemented mitigation measures over time. Ultimately, our results serve as a practical decision-support tool, empowering local agencies to make informed, timely decisions that promote resilient and sustainable land and water use in water-scarce arid environments.

4. Conclusions

This research aimed to calculate the temperature vegetation dryness index (TVDI) as an indicator of soil water drought. By combining the data from remote sensing technologies (RSTs) and geographic information systems (GISs), this was possible. Following this, the TVDI was used in a rigorous examination of drought patterns, helped by the Mann–Kendall (MK) technique. The results of this extensive undertaking have brought attention to several significant patterns and occurrences. The notable upward trajectory seen in the TVDI, which signifies a state of soil water drought, provides evidence of an escalating issue of water shortage. This tendency, substantiated by rigorous statistical analysis, has ramifications that go beyond simply numerical data. This suggests an increasing strain on the nearby ecological system, particularly in a region that largely depends on groundwater for intensive farming methods. The use of groundwater for agricultural purposes, while crucial for food production, gives rise to apprehensions over the exhaustion of these vital water sources.

At a confidence level of 0.05, the findings initially revealed a very noteworthy positive pattern in the TVDI time series. A Z_S statistic of 1.648 supports this conclusion. This study, on the other hand, looked at how well two surface energy balance (SEB) models could estimate evapotranspiration (ET) using satellite imagery in dry areas. It found a small increase in the rate of precipitation and the land surface temperature (LST), which matches the trend analysis of the TVDI at a significance level. Further examination in this study revealed a significant increase in the total areas susceptible to drought conditions, as indicated by a Z_S value of 2.05. Conversely, regions with TVDI values that suggest typical drought conditions had a significant decline, as shown by a Z_S value of −2.83. The trends showed changes at the seasonal level. The regions that experienced drought and normal circumstances throughout the spring and summer periods had similar patterns, as shown by the Z_S values of “1.89 and 1.74” and “2.42 and −2.12”, respectively. On the other hand, the winter season had a notable rise in regions characterized by intense precipitation, as shown by a Z_S value of 1.75.

The notable rise in the proportion of land regions susceptible to drought conditions is a matter of concern. The observed phenomenon indicates an increasing ecological disparity and emphasizes the pressing need to adopt sustainable approaches for managing water resources. On the other hand, the regions that are seeing a decline in TVDI levels, which are suggestive of typical drought conditions, highlight the dynamic character of the regional climate and the difficulties it presents for conventional agricultural methods. The variations in TVDI trends throughout the year shed light on the ever-changing nature of drought patterns at a seasonal level. During the warmer months, such as spring and summer, there is a consistent occurrence of drought and normal circumstances, which emphasizes the ongoing nature of these difficulties. Conversely, the significant rise in regions experiencing intense precipitation throughout the winter season may seem contradictory. Nevertheless, the phenomenon might be ascribed to the intricate interaction of many climatic elements, emphasizing the need for meticulous examination while tackling matters pertaining to drought.

The observable patterns have notable ecological implications for the fragile biological system of the region, especially in an area marked by extensive agricultural practices that significantly depend on the extraction of groundwater. The persistent and excessive use of groundwater resources inevitably results in the depletion of reservoirs, a problem that is further intensified by the significant increase in drought patterns seen. There is a growing body of evidence suggesting that the ecological system may encounter difficulties in adapting, which might possibly lead to a significant environmental disaster. Therefore, it is imperative for policymakers to promptly implement strategies that guarantee the sustainable governance of the region’s irreplaceable natural resources, thereby preventing imminent ecological disruption. To summarize, these results emphasize the urgent need for preemptive actions. It is essential for decision-makers to undertake a course of action that guarantees the sustainable management of the valuable natural resources within the area. This necessitates a nuanced equilibrium between addressing the requirements of agriculture and preserving the ecological system. This research acts as a compelling appeal, promoting the need to maintain environmental balance in response to changing climate conditions.