Impacts of the Middle Route of the South-to-North Water Diversion Project on Land Surface Temperature and Fractional Vegetation Coverage in the Danjiang River Basin

Abstract

The Middle Route of the South-to-North Water Diversion Project is a critical infrastructure that ensures optimal water resource distribution across river basins and safeguards the livelihood of people in China. This study investigated its effects on the land surface temperature (LST) and fractional vegetation coverage (FVC) in the Danjiang River Basin. Moreover, it examined the spatial and temporal patterns of this project, providing a scientific basis for the safe supply of water and ecological preservation. We used the improved interpolation of mean anomaly (IMA) method based on the digital elevation model (DEM) to reconstruct LST while FVC was estimated using the image element dichotomous model. Our findings indicated a general increase in the average LST in the Danjiang River Basin post-project implementation. During both wet and dry seasons, the cooling effect was primarily observed in the south-central region during the daytime, with extreme values of 6.1 °C and 5.9 °C. Conversely, during the nighttime, the cooling effect was more prevalent in the northern region, with extreme values of 3.0 °C and 2.3 °C. In contrast, the warming effect during both seasons was predominantly located in the northern region during the daytime, with extreme values of 5.3 °C and 5.5 °C. At night, the warming effect was chiefly observed in the south-central region, with extreme values of 5.8 °C and 5.9 °C. FVC displayed a seasonal trend, with higher values in the wet season and overall improvement over time. Statistical analysis revealed a negative correlation between vegetation change and daytime temperature variations in both periods (r = −0.184, r = −0.195). Furthermore, a significant positive correlation existed between vegetation change and nighttime temperature changes (r = 0.315, r = 0.328). Overall, the project contributed to regulating LST, fostering FVC development, and enhancing ecological stability in the Danjiang River Basin.

1. Introduction

Large water conservation projects significantly impact the regional distribution of water resources, physical geography, and urban development processes. This leads to various ecological and environmental effects, with changes in land surface temperature (LST) and fractional vegetation coverage (FVC) being crucial for regional ecological security and sustainable resource use [1]. LST is a key indicator for studying the Earth’s physical change process and global resource and environmental monitoring, as well as climate change. Precise estimation of LST aids in studying global climate change, understanding long-term climate change patterns, and managing natural disasters induced by climate change [2,3,4]. Spatial and temporal changes in FVC exhibit regional differences. Long-term analysis of FVC evolution for specific regions provides a scientific basis for local ecological protection and land use planning [5,6,7,8].

The accurate reconstruction of LST data is crucial for generating final LST data. The accuracy of LST is greatly affected by cloudy images. Currently, there have been many research studies on methods to address the missing data in LST caused by cloudy conditions. For example, Duan et al. [9] used thermal infrared (TIR) and microwave (PMW) data to generate all-weather LST, and estimated LST using the temperature lapse rate. Researchers such as Ge et al. [10], Zhang et al. [11], Tang et al. [12], and Zhao et al. [13] often rely on Landsat data for radiative transfer modeling to analyze LST. Zhu et al. [14]. used MODIS (Moderate Resolution Imaging Spectroradiometer) atmospheric products (MOD06_L2) to estimate LST and MOD07_L2 to estimate near-LST. They developed a regression model without other auxiliary parameters to estimate LST under cloudy conditions during the day. Another commonly used method is to establish a model based on the temporal variation of the data to estimate the missing values between LST data and other data [15,16]. These methods must satisfy the temporal variation pattern, and commonly used approaches include time series harmonic analysis (HANTS) [17,18], the Savitzky–Golay filtering method [19,20], and TIMESAT [21], which have shown good performance in reconstructing 8-day LST products. Li et al. [22] reconstructed LST using Gressman’s objective analysis method and the asymmetric Gaussian fitting method based on the spatio-temporal characteristics of LST data. It is also common to use interpolation methods to fill in missing values from Landsat or MODIS data to achieve a reconstructed LST [23]. This effectively fills in the missing values of LST, but this method is only applicable to multi-scene LST data and only considers the impact of elevation and land cover [24,25,26,27]. However, there is a lack of an efficient spatio-temporal reconstruction method for LST that considers both terrain effects and the correlation between long time series images, which is necessary for addressing long-term monitoring of cloud cover-induced remote sensing images and missing data in complex terrains. For vegetation analysis, common methods include the use of the normalized differential Vegetation Index (NDVI) and vegetation coverage (FVC). For example, Feng et al. [28] and Mao. et al. [29] calculated the NDVI based on Landsat data and analyzed its changes, while Deng et al. [30] directly used MODIS NDVI data to analyze NDVI changes in the study area. Regarding FVC, it is often evaluated using pixel dichotomy [31,32] or linear hybrid spectral model [33] calculations. A notable limitation within existing studies is their dependency on existing data sets, often without adequate adjustments for missing or anomalous data points. This can potentially reduce the accuracy of the results, making methodological improvements crucial for enhancing the precision of analyses concerning LST and FVC. Therefore, improving the accuracy of these measurements is of great importance for advancing our understanding of these environmental variables.

In recent years, investigations into changes in LST and FVC have attracted significant attention within both domestic and international research communities [34,35,36,37]. While most existing studies, such as those by Zhang et al. [11], Wang et al. [38], and Liu et al. [39], have primarily focused on analyzing either LST or FVC in isolation, only a handful have examined the correlation between the two variables [40]. Moreover, the majority of research in this area focuses on long-term assessments spanning large geographical scales [41]. According to research by Mal et al. [42], the summer multi-year average LST in the Ganges River Basin shows an increasing trend, while the winter multi-year average LST shows a decreasing trend, especially in areas with reduced FVC. Liu et al.’s study [43] indicates that the interannual LST in the Yangtze River Basin fluctuates and shows an overall rising trend, with nighttime LST increasing at a faster rate compared to daytime LST. After the construction of the Three Gorges Dam, the LST shows an increasing trend, and areas near the dam experience an increase in FVC, resulting in a noticeable cooling effect [44,45]. Research conducted by He et al. [46] confirms that large-scale vegetation restoration projects on the Loess Plateau have an impact on LST while changing FVC. Studies specifically examining the effects of engineering projects or localized phenomena on LST and FVC are notably sparse.

The Middle Route of the South-to-North Water Diversion Project is a large-scale inter-basin water transfer project implemented to solve the serious water shortage problem in northern China. It was constructed in December 2003 and officially commenced operation in December 2014, drawing water from the Danjiangkou Reservoir on the Han River, a tributary of the Yangtze River, with the help of the terrain difference. The water source is self-flowing, and the water will be sent to northern China, finally arriving at Tuanjiehu Lake; the total length of the project is about 1267 km. This project, key for optimal water resource allocation across river basins and the protection of people’s livelihoods, significantly influences the spatial distribution of water resources in China. The alterations in water levels triggered by this project inevitably change the spatial and temporal distribution of water resources, thereby influencing LST and FVC in the Danjiang River Basin. The observed changes in LST and FVC significantly affect the native climate and regional ecology of the Danjiang River Basin [47]. Given these implications, it is crucial to assess the impact of the Middle Route of the South-to-North Water Diversion Project on these environmental variables within the basin.

In summary, this study focuses on the Danjiang River Basin as the research area and utilizes MODIS, Landsat, digital elevation model (DEM), and meteorological station temperature data. An improved interpolation of mean anomaly (IMA) method based on DEM is constructed to complete the missing temperature values, improve the accuracy of the reconstructed LST, and analyze the evolution of LST in the Danjiang River basin before and after the implementation of the South-to-North Water Diversion Project. To achieve a more precise analysis, this study employs a pixel-based binary model for estimating FVC in the Danjiang River Basin. The research aims to quantitatively establish the correlation between shifts in LST and alterations in FVC within the basin. The results can provide scientific evidence for water supply security and ecological protection of the water source area in the Middle Route of the South-to-North Water Diversion Project.

2. Data and Methods

2.1. Study Area

Danjiangkou Reservoir is the largest artificial freshwater lake in Asia and the water source for the Middle route of China’s South-to-North Water Diversion Project, which plays a key role in China’s allocation of water resources. After the completion of the implementation of the Middle Route of the South-to-North Water Diversion Project, the watershed area of the study area was increased by about 169 km², and the water level was raised to 170 m [47]. The Middle Route of the South-to-North Water Diversion Project has improved water shortage and quality issues in the Danjiang River Basin, alleviated river sedimentation, and promoted local ecological restoration. It is located in the southwest of China’s Henan Province (Figure 1). It is high in the northwest and low in the southeast, with complex terrain and large fluctuations. Covering a total area of 8411.57 km², it is located at 110°35′–111°58′E and 33°30′–34°01′N. The study area experiences a monsoonal climate, transitioning from subtropical to warm temperate, with an average annual temperature of about 16.9 °C. The average annual rainfall is abundant, but the interannual precipitation distribution is uneven and varies significantly, mainly concentrated in August–October, with sparse rainfall in December–February, and the annual precipitation ranges between 850 and 1000 mm, with the same period of rain and heat. The region supports various forest types and FVC, including deciduous broad-leaved forests, mixed coniferous forests, and rock-deserted barren hills.

2.2. Data Sources

For the temporal scope of the study, two intervals are selected: the wet season, spanning from July to September, and the dry season, which extends from December to February of the following year. The data sets involve the years 2002–2006 and 2018–2022, providing a comprehensive view of environmental changes over time. The period from 2002–2006 represents the pre-construction phase of the Middle Route of the South-to-North Water Diversion Project, while the period from 2018–2022 represents the post-construction phase of the Middle Route of the South-to-North Water Diversion Project. The LST data used are from the MYD11A2 dataset, obtained from the official website of the National Aeronautics and Space Administration (NASA) of the United States (https://www.nasa.gov/), accessed on 1 September 2022, with a spatial resolution of 1 km and a temporal resolution of 8 days (

Table 1. Data sources.

Data Type	Data Name	Data Source
Raster data	MODIS remote sensing image data (MYD11A2)	https://www.nasa.gov/
	MODIS remote sensing image data (MYD13Q1)	https://www.nasa.gov/
	DEM data	https://www.gscloud.cn/
	Landsat 7 remote sensing image data	https://www.gscloud.cn/
Meteorological data	Daily average temperature data	https://data.cma.cn/

). FVC data are from the MYD13Q1 dataset, also obtained from NASA (https://www.nasa.gov/), with a spatial resolution of 250 m and a temporal resolution of 16 days. The meteorological station data used for validation consist of daily average temperature data from four meteorological stations within the study area (2003–2005, 2019–2021, and the months of January–March and July–December each year), sourced from the China Meteorological Administration Scientific Data Center (https://data.cma.cn/), accessed on 1 September 2022. The DEM data used are from the Shuttle Radar Topography Mission (SRTM), retrieved from the China Geospatial Cloud (https://www.gscloud.cn/), accessed on 1 September 2022, with a resolution of 90 m. Landsat 7 data are also sourced from the China Geospatial Cloud, with a selection criteria of cloud cover below 5%.

2.3. Study Methods

2.3.1. IMA Algorithm Based on DEM Improvement (DEM-IMA)

Studies on LST estimation using MODIS LST have proven that there is a linear relationship between LST and elevation, and LST decreases when elevation increases [48]. In traditional spatial temperature interpolation, when the observed temperature data from meteorological stations are used to spatially interpolate the air temperature in a certain region, the vertical rate of temperature decrease has an important influence on interpolation accuracy. One crucial aspect that enhances the robustness of this study is the inclusion of elevation changes in the reconstruction of LST. The need to consider elevation is highlighted by its crucial role in enhancing spatial interpolation accuracy and in providing a correct interpretation of spatial variation in LST. As supported by research such as the study by Tian et al. [49], the effect of elevation on LST cannot be ignored. Elevational differences can yield significant disparities in LST between areas situated at varying altitudes. Recognizing a linear relationship between LST and elevation, this study normalized the LST at the value point to geoid LST based on DEM elevation data and the vertical rate of temperature decrease. The accuracy of the LST reconstruction is improved by utilizing the target image neighborhood to capture the temporal dependence between proximity images while performing correction of the vertical decreasing rate of temperature. The IMA was used to perform spatial interpolation of the geodetic LST to obtain complete LST data results after interpolation. Then, the reconstructed LST results of the normal elevation were calculated backward using the vertical decreasing rate of the elevation data and the temperature. To validate the reconstructed LST, this study utilized data from weather stations situated within the research area. According to the geographical location of the weather station, the LST at different time points was extracted, and the reconstructed LST was verified by comparison to evaluate the effectiveness of the reconstructed LST.

The correction formula for the positive and negative calculation of elevation and LST is expressed in Equation (1).

$$\left\{\begin{array}{c}{T}_{sealevel}=T_{MODLST}-a\times H_1/100\\ T_{result}=T_{Interpolation}+a\times H_2/100\end{array}\right.$$

(1)

where $T_{sealevel}$ represents the LST (°C) of the MYD11A2 temperature product with value pixels converted to the geodetic level, $T_{MODLST}$ denotes the temperature value of the MYD11A2 temperature product, a is the vertical decreasing rate of temperature (a = 0.65 is selected by referring to the research of Ojha [50] and Yue et al. [51] in this paper), $H_1$ symbolizes the elevation (m) of the MYD11A2 temperature product with value points, $H_2$ signifies the elevation of each pixel point within the study area data (m), $T_{Interpolation}$ represents the spatial interpolation result of $T_{sealevel}$ after IMA, and $T_{result}$ denotes the result based on the conversion of the $T_{Interpolation}$ spatial interpolation result from the geodetic level to actual elevation.

2.3.2. Temperature Effect Change Intensity Index

To investigate the influence of temperature effect changes due to the Middle Route of the South-to-North Water Diversion Project, the Euclidean allocation principle was employed. This principle assigned the temperature of the river water surface to other locations in the study area based on the minimum Euclidean distance principle. Subsequently, the Euclidean allocation temperature ($LS{T}_{EA}$) for each point was obtained. The temperature difference between the river surface and the surrounding environment was calculated by subtracting the Euclidean allocation temperature from the LST. The equation for this is as follows:

$$TD=LST-LS{T}_{EA}$$

(2)

where $TD$ denotes the LST difference (°C), $LST$ represents the LST extracted from MODIS remote sensing images, and $LS{T}_{EA}$ is the Euclidean distance of the nearest river water LST (°C) assigned by the Euclidean allocation principle.

The temperature effect change intensity index (RECI) is defined as the difference between the LST difference before and after the construction of the Middle Route of the South-to-North Water Diversion Project. This index is used to represent the intensity of temperature effect changes due to the project. The equation for the RECI is as follows:

$$RECI=T{D}_L-T{D}_F$$

(3)

where $RECI$ represents the difference between the LST difference before and after the construction of the Middle Route of the South-to-North Water Diversion Project, $T{D}_F$ represents the LST difference before the construction of the Middle Route of the South-to-North Water Diversion Project, and $T{D}_L$ represents the LST difference after the construction of the project.

To provide an accurate assessment of the grade and a precise representation and evaluation of the spatial and temporal distribution of the LST difference, the intensity of the temperature effect change caused by the project is standardized. The normalized temperature effect change intensity is calculated using Equation (4).

$$NRECI={\displaystyle \frac{{RECI-RECI}_{MIN}}{{RECI}_{MAX}-{RECI}_{MIN}}}$$

(4)

where $NRECI$ represents the normalized temperature effect change intensity index of the Middle Route of the South-to-North Water Diversion Project and $RECI$ represents the temperature effect change intensity index of the project. Finally, the normalized LST difference results were categorized into five levels based on equal intervals, as presented in

Table 2. Classification levels based on NRECI ranges.

Level	NRECI
Extremely weak zone of variation	0.0–0.2
Relatively weak zone of variation	0.2–0.4
Medium zone of variation	0.4–0.6
Relatively strong zone of variation	0.6–0.8
Extremely strong zone of variation	0.8–1.0

. This reflects the spatial distribution of the LST difference in a more objective manner.

2.3.3. Thermoregulation Effect

In this paper, different combinations of TDF water (before the construction of the project), TDL (after the construction of the project), and RECI are considered. A comprehensive analysis is conducted based on the positive and negative combinations of three indices: the LST difference before and after the project’s construction and the temperature regulation effect of the project. This led to the establishment of the impact assessment of the temperature regulation effect of the project, summarized in

Table 3. Impact of the Middle Route of the South-to-North Water Diversion Project on temperature regulation.

No.	Land Surface Temperature Difference before the Construction of the Middle Route of the South-to-North Water Diversion Project	Land Surface Temperature Difference after the Construction of the Middle Route of the South-to-North Water Diversion Project	RECI	Cooling Effect	Warming Effect	No Effect
1	>0	>0	≥0			√
2	>0	≥0	<0	√
3	>0	<0	≥0			√
4	>0	≤0	<0	√
5	<0	≥0	>0		√
6	<0	>0	≤0			√
7	<0	≤0	>0		√
8	<0	<0	≤0			√

. Eight combinations of scenarios were included, capturing all possible scenarios. These were used to determine whether the temperature effect of the project’s construction belonged to cooling, warming, or no effect categories.

2.3.4. Inversion of Surface FVC

To estimate FVC, the image element dichotomy model was used in this study. Image element dichotomy is one of the most commonly used image element linear decomposition models, which assumes that surfaces covered by vegetation and bare land without vegetation cover are one image element, and determines the degree of vegetation coverage in the image element by calculating the proportion of the area each occupies in the image element, in which the percentage of the image element occupied by the vegetation-covered surface is the degree of vegetation coverage of the image element [52,53,54]. This empirical model uses NDVI as a remote sensing variable and solves the limitation of NDVI, which is less effective in the study of areas with very high or low vegetation cover and is the ratio of the difference between the values of the near-infrared and visible red bands to the sum of the two bands, which partially eliminates the effects of solar altitude angle, satellite observation angle, terrain, clouds, shadows, and irradiance related to the atmospheric conditions and is a good solution for vegetation cover. It is the best indicator of vegetation growth status and FVC, and the conditions of NDVI conforming to the image element dichotomous model have a good correlation with the vegetation summer cover, and a large number of research studies [55,56] have demonstrated the ability and advantage of substituting NDVI into the dichotomous model for the estimation of vegetation coverage. Therefore, through this process, synthetic images describing the vegetation cover in different years in the Danjiang River Basin were finally generated.

The equation for the image element dichotomous model is as follows:

$$FVC=(NDVI-NDV{I}_{min})/(NDV{I}_{max}-NDV{I}_{min})$$

(5)

where $NDV{I}_{min}$ represents the minimum value of NDVI in the image, and $NDV{I}_{max}$ represents the maximum value of NDVI. As noise is inevitable, $NDV{I}_{min}$ and $NDV{I}_{max}$ are taken as the minimum and maximum values within a certain confidence level, and the emissivity (an indicator of how well vegetation-covered surfaces can reflect and absorb radiation energy) is calculated based on this method. The vegetation index NDVI is considered to represent full FVC when its value exceeds 0.7, and bare land when NDVI is less than 0.05. Using the NDVI thresholds of 0.7 and 0.05, FVC is assigned a value of 1 when NDVI exceeds 0.7, and 0 when NDVI is less than 0.05 [34].

$$\left\{\begin{array}{c}{F}_{Cwater}=0\hfill \\ F_{Cvegetation}=1\hfill \\ F_{Cother}=(NDVI-NDV{I}_{min})/(NDV{I}_{max}-NDV{I}_{min})\hfill \end{array}\right.$$

(6)

Considering the actual ecological environment of the study area, corresponding classes were established to accurately describe the distribution of FVC. This classification was based on existing studies [57,58]. Detailed descriptions are provided in

Table 4. Classification of FVC.

Level	FVC
Extremely high coverage	>70%
Relatively high coverage	50–70%
Medium coverage	30–50%
Relatively low coverage	10–30%
Extremely low coverage	<10%

.

2.3.5. FVC Shift Direction Level

In this study, ArcGIS tools were employed to assign different values to different levels of FVC data. In this scale, higher numbers correspond to higher FVC levels. The data assigned to FVC post-construction of the Middle Route of the South-to-North Water Diversion Project were subtracted from the pre-construction data. The resulting difference represents the direction and intensity of FVC changes. Positive values denote benign FVC development, while negative values signify malignant development. A larger absolute value indicates a more significant change, with the direction and intensity of this change displayed in

Table 5. Level and intensity of FVC shift direction.

Benign Development of FVC		Malignant Development of FVC
Level	Intensity	Level	Intensity
0	Not	−1	weak
1	Weak	−2	strong
2	strong	−3	relatively strong
3	relatively strong	−4	extremely strong
4	extremely strong

.

3. Results

3.1. Analysis of the Impact of the Middle Route of the South-to-North Water Diversion Project on LST Changes in the Danjiang River Basin

3.1.1. Analysis of Temporal Changes in LST and Characteristics

Temperature is greatly influenced by the water body, and this paper analyzes the LST changes in the Danjiang River Basin from the two periods of the wet season and the dry season before and after the middle route of the South-to-North Water Diversion project, respectively (Figure 2 and Figure 3). The temperature distribution is similar during dry and wet seasons. Remarkably, the study also observed notable temperature variations in the river during the day and night. Specifically, the river manifests lower temperatures than its surrounding areas during the daytime. Conversely, the nighttime temperatures of the river are notably higher than the adjacent regions. In terms of the wet and dry seasons before and after the implementation of the Middle Route of the South-to-North Water Diversion Project, the lowest daily average LSTs are recorded as 0.8 °C, −6.7 °C, 16.6 °C, and 8.4 °C. These temperatures are primarily located in the high-altitude, mountainous areas in the northern region. On the other hand, the highest values of 15.3 °C, 8.1 °C, 30.7 °C, and 21.5 °C are predominantly found in the valleys and plains of the central and southern regions, which feature lower terrain. However, there are differences in the comparison of specific time periods. In the northern mountainous areas, there are more scattered high-temperature areas in the wet season than in the dry season, and the northern high-temperature areas in the wet season are significantly increased at night, while the northern high-temperature areas in the dry season are significantly decreased.

As shown in Figure 4, NRECI was used to analyze the intensity of LST changes, and the intensity of temperature changes in the Middle Route of the South-to-North Water Diversion Project showed different rules in different periods. The temperature variation effect of the Middle Route of the South-to-North Water Diversion Project is more obvious in the dry season than in the wet season, and more obvious at night than in daytime.

During the wet season, the daytime temperature changes in the Danjiang River Basin predominantly occur within the medium zone of change, which represents 73.39% of the total area. Both the relatively strong and weak zones of change are less frequent, accounting for 11.19% and 14.78%, respectively. The relatively weak zones are mostly located in the central and northern regions, while the relatively strong zones are primarily in the southern region. The extremely weak and extremely strong zones have limited coverage, constituting only 0.46% and 0.18% of the total area, respectively. At nighttime in the wet season, the medium zone of change continues to dominate, covering 61.56% of the area. Unlike the daytime, the relatively strong zone of change is more significant at night, making up 31.36% of the area. This zone is more scattered in the northern region and more concentrated in the southern region. The relatively weak and extremely weak zones account for a smaller 4.41% and 0.26%, mainly dispersed in the southern region. The extremely strong zone, although minimal, represents 2.40% and is found in the central and southern areas.

During the dry season’s daytime, the relatively weak zone of change is the most prevalent, accounting for 59.55% of the area, followed by the medium zone at 35.52%. The extremely weak and strong zones are relatively rare, constituting 2.07% and 2.77%, respectively, mainly located in the central and southern regions. The extremely strong zone is the least common, at just 0.09%, and is scattered in the south-central region. At nighttime in the dry season, the medium zone of change again dominates, constituting 61.51% of the total area. The relatively weak zone accounts for 26.13% and is primarily found in the southern and northern regions. The relatively strong zone represents 9.00% and is mainly situated in the south-central region. Both the extremely strong and extremely weak zones are less common, covering 1.78% and 1.59% of the area, respectively, and are dispersed primarily in the central and southern regions.

3.1.2. Analysis of the Temperature Regulation Effect of the Middle Route of the South-to-North Water Diversion Project

(1)

Cooling effect of the Middle Route of the South-to-North Water Diversion Project

The cooling effect of the Middle Route of the South-to-North Water Diversion Project on the local climate was assessed by integrating data on LST differences before and after the project’s implementation, as well as the intensity of temperature effect changes induced by the project. A segmentation method was used for this analysis, and the corresponding results are tabulated in Figure 5.

Based on the analysis of Figure 5, this study demonstrates that the construction and operation of the Middle Route of the South-to-North Water Diversion Project exert a cooling influence on the LST in the Danjiang River Basin. During the wet season, the cooling effect reaches an extreme of −6.1 °C over an area of 657.12 km² in the daytime, while at nighttime, the cooling extreme is −3.0 °C, affecting an area of 709.97 km². During the dry season, the daytime cooling extreme is −5.9 °C over an area of 827.83 km², and at nighttime, the extreme is −2.3 °C, impacting an area of 1036.42 km². The most significant cooling effects are more pronounced during the day than at night and are marginally more noticeable during the wet season than the dry season. In addition, the extent of the cooled area during the daytime in the dry season surpasses that of the wet season, as does the area experiencing cooling at night. The daytime cooled area is primarily located in the south-central region, near river water surfaces. In contrast, the cooled area at night is mainly found in the northern region, where elevation changes are more distinct and distance from the river water surface is greater. This results in a more dispersed distribution and a lessened cooling effect on the river itself during the wet season’s nighttime.

(2)

Warming effect of the Middle Route of the South-to-North Water Diversion Project

The warming effect of the Middle Route of the South-to-North Water Diversion Project on the surrounding area was also obtained using the segmentation method, and the results are depicted in Figure 6.

Conversely, the analysis of Figure 6 reveals a warming impact on the basin’s LST due to the same project. Specifically, during the wet season, the warming effect peaks at 5.3 °C in the daytime, affecting an area of 540.19 km². At nighttime, the extreme value is 5.8 °C, with a warming area of 743.64 km². In the dry season, the daytime warming extreme is 5.5 °C over an area of 994.33 km², while at night, the extreme is 5.9 °C, impacting a smaller area of 461.62 km². The warming effect tends to be more pronounced at night compared to daytime, and it is also more apparent during the dry season as opposed to the wet season. The area experiencing warming during the daytime of the dry season is more expansive than that during the wet season, and the cooled area during the nighttime of the wet season is larger than that of the dry season. The daytime warming region is mostly located in the north, away from the river water surface, and exhibits a relatively scattered pattern. Conversely, the nighttime warming area is primarily concentrated in the south-central region, in close proximity to the river water surface.

In summary, the diurnal distribution of LST in the Danjiang River Basin is similar in the dry and wet seasons, with the low mean surface temperature zones distributed in the high-altitude mountainous areas in the north, and the high zones distributed in the higher terrain areas in the central and southern parts of the basin. The implementation of the South-to-North Water Diversion Project affected the changes of LST in the Danjiang River Basin. In the wet season, the main temperature changes in the Danjiang River Basin were in the medium zone during the daytime, and the southern part of the basin was more obvious at night; in the dry season, the relatively weak zone dominated during the daytime, and the medium zone was still dominant at night. The construction of the South-to-North Water Diversion Project simultaneously exhibits both warming and cooling effects on the surrounding environment. The intensity and distribution of the warming and cooling effects before and after the construction of the South-to-North Water Diversion Project vary at different periods, with both effects coexisting. In the same period, the cooling effects in the northern mountainous areas are opposite to those in the central and southern plains. During the wet or dry season, the cooling effects in the northern mountainous areas or the central and southern plains have opposite directions for both daytime and nighttime temperatures.

3.2. Analysis of the Impact of the Middle Route of the South-to-North Water Diversion Project on FVC Changes in the Danjiang River Basin

3.2.1. Analysis of Temporal Changes in FVC

The distribution of overall FVC, as demonstrated in Figure 7 and

Table 6. Distribution of FVC during different periods.

Time Period/FVC Level			Extremely High Coverage	Relatively High Coverage	Medium Coverage	Relatively Low Coverage	Extremely Low Coverage
Before the implementation of the project	wet season	area (km2)	6085.62	1957.13	189.65	38.00	141.18
		proportion (%)	72.35	23.27	2.25	0.45	1.68
	dry season	area (km2)	378.61	1094.40	5675.17	1064.89	198.51
		proportion (%)	4.50	13.01	67.47	12.66	2.36
After the implementation of the project	wet season	area (km2)	6481.92	1354.79	226.76	79.60	268.50
		proportion (%)	77.06	16.11	2.70	0.95	3.19
	dry season	area (km2)	487.85	1515.87	5768.67	334.95	304.24
		proportion (%)	5.80	18.02	68.58	3.98	3.62

, follows a trend of being “high during the wet season and low during the dry season”. Prior to the construction of the Middle Route of the South-to-North Water Diversion Project, FVC was high. However, it decreased after the project’s implementation, especially in the plain and river valley areas. Extremely low FVC was observed in the watershed and northern river valley regions.

During the wet season, the landscape is dominated by areas of extremely high FVC, while medium coverage prevails in the dry season. After the middle route of the South-to-North Water diversion project was completed, the government adopted afforestation, restoration of degraded forest, and other measures to increase the greening around the reservoir. The construction of the project has influenced notable shifts in FVC within the study area, with an overall increase observed. During the wet season, the average FVC increased by 1.10% following the project’s construction. Moreover, the area characterized by extremely high coverage expanded by 5.01%, primarily in the central and southern regions. The areas showing substantial changes in water imaged by the project have primarily evolved from the original areas with relatively high coverage, resulting in a reduction in these areas. The percentage of extremely low coverage rose from 1.68% to 3.19%, while medium coverage increased from 67.47% to 68.58%. Concurrently, the relatively low coverage area decreased from 12.66% to 3.98%, which can be considered a typical consequence of the reduction in green areas caused by the rise in water surface.

During the dry season, a noticeable enhancement was witnessed in average FVC within the study area. Prior to the construction of the Middle Route of the South-to-North Water Diversion Project, this value was 40.01%, and it rose to 43.68% post-construction—an improvement more significant than that observed in the wet season. A marked change was seen in the area of relatively high coverage, which expanded from 13.01% to 18.02%, aligning with the shift in abundance. There was a slight increase in extremely high coverage and medium coverage areas, at 1.30% and 1.11%, respectively. In contrast, areas of relatively low coverage were notably reduced, with this reduction primarily contributing to the expansion of relatively high coverage classes. Consequently, the percentage of relatively low coverage areas decreased from 12.66% to 3.98%. The ecological improvement is evident, primarily in the upper and central-western regions. As evidenced by the above results, overall FVC in the wet season shows an increasing trend. Areas with extremely low coverage expanded from 2.36% to 3.62%, an expected result given the reduction in green areas due to the rising water level.

The overall distribution of FVC shows a pattern of high FVC during the wet season and low FVC during the dry season. Before the construction of the South-to-North Water Diversion Project, FVC was high, whereas after the construction, it decreased. The plains and river valleys have low FVC, while other regions have high FVC. Extremely low FVC is mainly found in water bodies and the northern river valley area. During the wet season, FVC is primarily in the extremely high FVC zone, while during the dry season, it is mainly in the medium FVC zone.

3.2.2. Spatial Variation Analysis of FVC

After the completion of the construction of the Middle Route of the South-to-North Water Diversion Project, there was an overall significant increase in FVC during the wet season and dry season, mainly in the northern river valley area and the central and southern plains (Figure 8). The expansion of the water surface area due to the project led to a reduction in FVC in certain areas during both wet and dry seasons, as previously existing vegetation near the waterfront was partially submerged. The growth of vegetation was mostly observed in the northern river valleys and the plain areas in the central and southern regions. There was a greater increase in the dry season compared to the wet season, with the differences primarily concentrated around the northern river valleys.

Upon integrating the DEM analysis, it becomes clear that the lower elevation central area was impacted more than the higher elevation northern area due to more prominent river surface changes and closer proximity. The northern higher elevation area was less affected due to topographic limitations. Areas with benign FVC development accounted for 15.27% of the total study area. Reductions in FVC class amounted to 11.08%. The expansion and impoundment of the reservoir in the Middle Route of the South-to-North Water Diversion Project led to elevated water levels, resulting in the inundation of certain terrestrial and vegetative areas. This alteration in water levels poses challenges to vegetation, which may struggle to adapt and will consequently decline, thereby reducing FVC. Furthermore, issues may arise during the process of water allocation. For instance, inadequate or unequal water distribution could lead to water scarcity in certain parts of the Danjiang River Basin, adversely affecting plant growth and FVC. The dry season displayed a clear increasing trend, with areas showing benign FVC development constituting 24.89% of the total study area. This was an increase of 9.62% compared to the wet season.

After the construction of the Middle Route of the South-to-North Water Diversion Project, a clear development towards benign FVC was observed in the study area. This change varied in intensity across different periods, with the dry season exhibiting more noticeable improvement. Moreover, the spatial distribution of benign development differed during different periods: it was more concentrated during the wet season and more widespread and scattered during the dry season.

3.3. Correlation Analysis of LST Change and FVC Change in the Danjiang River Basin

3.3.1. Correlation Analysis of Temperature Change and FVC Change during the Wet Season

The change in temperature intensity, both during the day and at night, during the wet season was correlated with the shift in FVC level in those same periods, and the results are presented in Figure 9 and

Table 7. Correlation coefficients between daytime/nighttime LST and FVC during the wet season.

Correlation
		Shifts in FVC	LST variations during the daytime	LST variations during the nighttime
Shifts in FVC	Pearson Correlation	1	−0.195 **	0.315 **
	Significance (two-tailed)		<0.001	<0.001
	Number of cases	651	651	651

** The correlation is significant at the 0.01 level (two-tailed).

.

Analysis of the Pearson correlation coefficient table yields two conclusions:

(1)

A negative correlation (r = −0.195, p < 0.001) was observed between the change in FVC and the change in daytime LST during the wet season. This implies that the greater the change in daytime temperature during these periods, the lesser the change in vegetation transfer.

(2)

A positive correlation (r = 0.315, p < 0.001) was seen between the change in FVC and the change in nighttime LST during the wet season. This suggests that the greater the change in nighttime temperature during these periods, the greater the shift in vegetation.

During the wet season, due to the implementation of the South-to-North Water Diversion Project, FVC in most areas of the Danjiang River Basin remained relatively unchanged, with the more significant areas of change concentrated in the northern valley areas and the south-central plains. Daytime LST changes in the Danjiang River Basin after the implementation of the project were mainly concentrated in areas where FVC changes were not significant, and the rest of the area was mainly retarded by the increase in the water body to slow down the degree of LST changes. Nighttime LST changes are mainly concentrated in the south-central sub-region where FVC changes are more significant. There was a balanced distribution of positive and negative temperature differences. At night, during the wet season, the variance was greater than during the day. The most extreme decreases in FVC were primarily due to watershed expansion. The resulting temperature difference remained relatively stable at plus or minus 3 °C. The daytime exhibited a significant negative relationship between temperature and vegetation change, with a greater increase in temperature when vegetation was reduced. Conversely, a significant positive relationship was observed at night between temperature and vegetation change. Although the correlation was statistically significant, the strength of the association was weak.

FVC changes and the temperature regulation effect of the Middle Route of the South-to-North Water Diversion Project during periods of water abundance were predominantly centralized in their respective low-change areas. Correlations were more apparent at points where changes were more pronounced. This analysis proves the relationship between the temperature regulation effect of the project and the degree of FVC change during the wet season. The higher the degree of FVC change, the stronger the temperature regulation effect of the project, with opposite regulation directions observed during the daytime and nighttime. The correlation between the nighttime temperature regulation effect and FVC change during periods of water abundance is stronger.

3.3.2. Correlation Analysis of Temperature Change and FVC Change during the Dry Season

The intensity of temperature change during both the daytime and nighttime of the dry season, linked to the Middle Route of the South-to-North Water Diversion Project, was found to correlate with shifts in FVC level during the wet season. The results are listed in Figure 10 and

Table 8. Correlation coefficients between daytime/nighttime LST and FVC during the dry season.

Correlation
		Shifts in FVC	LST variations during the daytime	LST variations during the nighttime
Shifts in FVC	Pearson Correlation	1	−0.184 **	0.328 **
	Significance (two-tailed)		<0.001	<0.001
	Number of cases	651	651	651

** The correlation is significant at the 0.01 level (two-tailed).

.

The Pearson correlation coefficient table allowed us to infer two key findings:

(1)

There was a significant but relatively weak negative correlation (r = −0.184, p < 0.001) between changes in FVC and temperature during the daytime dry season. In other words, a greater temperature change during the daytime dry season corresponds with a smaller shift in FVC during the same period.

(2)

Conversely, a significant positive correlation (r = 0.305, p < 0.001) was found between changes in FVC and nighttime temperature during the dry season. This suggests that a greater temperature change during the nighttime dry season correlates with a greater shift in FVC during the same period.

During the dry season, the implementation of the South-to-North Water Diversion Project promotes the increase of the overall FVC in the Danjiang River Basin, but there are more areas with basically unchanged FVC, and the areas with more significant changes in FVC are concentrated in the south-central subplain area. The areas with basically unchanged FVC account for most of the areas. After the implementation of the South-to-North Water Diversion Project, the distribution of areas with significant LST changes in the Danjiang River Basin during the day was scattered, and the areas with more significant LST changes at night were mainly concentrated in the south-central sub-plain areas with more significant FVC changes, slightly favoring the areas with increased vegetation. The distribution of positive and negative temperature differences remained largely balanced, with greater dispersion during the dry season. Extreme reductions in FVC primarily occurred due to watershed area expansion, resulting in a relatively stable temperature difference of plus or minus 2 °C. While the correlation was significant, the overall significance was low.

The temperature regulation effect of the Middle Route of the South-to-North Water Diversion Project, and shifts in FVC during the dry season, mostly occurred in areas with minimal changes. There was a slight bias towards vegetation growth when compared to the wet season. More evident correlations were observed at a few points where changes were more pronounced. This correlation analysis verifies a relationship between the temperature regulation effect of the project and shifts in FVC during the dry season. The higher the change in FVC, the stronger the temperature regulation effect of the project, with opposing regulation directions in the daytime and nighttime. This correlation was stronger at night.

4. Discussion

4.1. Impact of the Middle Route of the South-to-North Water Diversion Project on LST in the Danjiang River Basin

In this study, the IMA algorithm based on DEM improvement is used to reconstruct the surface temperature of the Danjiang River Basin before and after the construction of the South-to-North Water Diversion Project by introducing the concepts of average anomaly and the vertical rate of temperature decrease. According to the existing temperature interpolation methods, it is difficult to interpolate the low-quality long time series images while ensuring the accuracy of the data, and the reconstruction of the long time series data for complex terrains is complicated and inefficient. The method proposed in this study solves the problem of missing long time series images under long time series climatic events, improves the reliability of the missing long time series image interpolation, improves the efficiency of the inversion, and reduces the probability of anomalies appearing.

The implementation of the Middle Route of the South-to-North Water Diversion Project had a significant impact on the LST in the Danjiang River Basin. The data showed that after the implementation of the Middle Route of the South-to-North Water Diversion Project, the LST in the Danjiang River Basin showed an overall upward trend, accompanied by significant diurnal temperature variations. This study suggests that water storage around the reservoir area in the Danjiang River Basin has a moderating effect on the rising trend of LST and can effectively slow down the rising trend of LST. These findings are consistent with those of Mao et al. [59] on the spatial and temporal variations of LST across China and Liu et al. [39] on the diurnal heat island intensity in Chinese cities. The increase in water bodies due to project construction leads to a decrease in LST. This is consistent with the study of LST changes after the construction of the Three Gorges Reservoir [44] and the Yangtze River Basin [43], which showed an overall increasing trend in terms of LST, but the warming and cooling effects were more pronounced in areas close to the watershed. Obviously, water transfer leads to the redistribution of surface water in the basin, which affects the spatial temperature distribution.

4.2. Impact of the Middle Route of the South-to-North Water Diversion Project on FVC in the Danjiang River Basin

The Middle Route of the South-to-North Water Diversion Project has notably affected FVC in the Danjiang River Basin. FVC in the Danjiang River Basin exhibited an overall increasing trend both before and after the implementation of the Middle Route of the South-to-North Water Diversion Project. This observation aligns with the findings of Lu et al. [60]. However, in certain regions, FVC declined following the project’s construction. This decrease is attributable to the expansion of the reservoir’s water surface and the resultant inundation of adjacent areas, which created conditions unsuitable for some plant species. Conversely, FVC in other areas experienced a notable increase. This increase is largely due to various greening measures implemented subsequent to the construction of the Middle Route of the South-to-North Water Diversion Project. These initiatives effectively enhanced FVC, leading to the observed anomaly in vegetation cover trends [61,62,63,64,65,66]. Moreover, the project has altered the hydrological cycle in the Danjiang River Basin, promoting vegetation growth even further. Since the construction of the project, FVC in the Danjiang River Basin has seen a rise, particularly along the riverbanks and in wetland areas. However, the project also presents potential downsides, such as overuse of water sources, environmental degradation, and possible inundation of reservoir-surrounding areas. These factors may lead to vegetation destruction and reduction.

4.3. Interaction of LST and FVC

There is a close correlation between LST and FVC. Plants can reduce the energy input to the land surface by absorbing solar radiation and lower the LST by releasing water through evapotranspiration. Larger leaf areas can absorb more solar radiation, thus reducing LST. Greener and denser vegetation cover provides more shade and evaporative surfaces, reducing LST increases [67]. The leaves of plants absorb radiant energy while releasing some of the heat and emit the water absorbed by the plant roots to the atmosphere through the stomata via transpiration, a process that consumes a large amount of heat, thus reducing LST in areas with vegetation cover. As a result, LST is lower in areas with high FVC [68].

The project’s impact has instigated a correlation between LST and FVC in the Danjiang River Basin. The increase in water supply to the basin, brought about by the project, can lower the LST. This allows for increased soil moisture and water utilization, fostering a favorable environment for plant growth. Transpiration from the thriving vegetation can further decrease the LST, making areas with abundant vegetation cooler. Hence, the project’s influence on LST and FVC requires a comprehensive understanding. Additionally, there is a need for measures to protect the ecological environment and to ensure the sustainable use of water resources. LST and FVC changes in the Danjiang River Basin have significant environmental implications. A reduced LST can mitigate the urban heat island effect and enhance the city’s climate [69]. Increased FVC improves the soil’s water retention capacity, reducing the risks of soil erosion and flooding. Further, vegetation growth can bolster air quality and lessen air pollutant concentrations. Therefore, the Middle Route of the South-to-North Water Diversion Project’s impact on LST and FVC in the Danjiang River Basin is largely positive for the environment.

4.4. Limitations and Perspectives of the Study

This study reveals that the Middle Route of the South-to-North Water Diversion Project has favorably influenced the regional ecological environment. However, limitations exist due to data and technical constraints that require future improvements and exploration. These shortcomings include:

(1)

The resolution of MODIS satellite data is limited. The accuracy and resolution of the two MODIS satellites are insufficient for the data requirements of future studies due to scientific and technological constraints. Therefore, it is essential to investigate how to utilize higher-accuracy data in future studies. For example, Landsat, GOES (Geostationary Operational Environmental Satellite), ASTER (Advanced Spaceborne Thermal Emission and Reflection Radiometer), and other series of high-resolution satellite data [70,71,72] can be used to supplement experimental analysis to improve the robustness of the results. This will help to provide a more precise characterization of the study area, thereby enabling a more accurate assessment of the Middle Route of the South-to-North Water Diversion Project’s impact.

(2)

This paper primarily investigates the correlation and influence of elevation on LST and FVC, focusing on regions with significant topographic relief and apparent elevation differences. The correlation may vary slightly, contingent upon the slope direction and other factors in different regions. Therefore, the influence of elevation on LST and FVC should be assessed in conjunction with topographic factors.

(3)

The water diversion project carries a substantial impact, with intricate interactions between regional precipitation, FVC, and LST. Precipitation quantity also influences LST and FVC, with the flow of surface runoff in the reservoir area being closely connected. Hence, there is a need for further comprehensive investigation into the temperature regulation effect of the Middle Route of the South-to-North Water Diversion Project. This should incorporate an in-depth analysis of precipitation and various environmental and climatic factors.

5. Conclusions

Leveraging MODIS, Landsat, DEM, and meteorological station temperature data, this paper establishes an LST reconstruction method based on DEM-IMA interpolation and improves the accuracy of LST reconstruction. This method analyzes the LST variation pattern in the Danjiang River Basin, inverts FVC using an image dichotomous improvement model, and quantitatively examines the correlation between LST variations and FVC variations. The key conclusions are as follows:

(1)

The Middle Route of the South-to-North Water Diversion Project has a regulatory effect on the LST in the Danjiang River Basin. Average LST has risen overall before and after project implementation, with a more significant increase during the wet season. The construction of the project has a more noticeable dampening effect on warming during the dry season. During both daytime and nighttime, the cooling effect is more pronounced in the wet season compared to the dry season, with differences of 3.6 °C and 3.1 °C, respectively. Conversely, the warming effect shows a decrease during the wet season in both daytime and nighttime, with differences of −0.6 °C and −0.4 °C, respectively.

(2)

The Middle Route of the South-to-North Water Diversion Project has positively influenced FVC in the Danjiang River Basin. FVC distribution shows higher levels in the wet season and pre-project, and lower levels in the dry season and post-project, particularly in plains and river valleys. The proportion of the area where FVC showed benign development during the dry season reached 24.89% of the total study area, a 9.62% increase compared to the wet season. Thus, the construction of the project has significantly contributed to the benign development of FVC and ecological environment stability.

(3)

A correlation exists between LST changes and FVC changes in the Danjiang River Basin. The change in FVC is negatively correlated with the change in daytime temperature and positively correlated with the change in nighttime temperature during both the wet and dry seasons. The higher the intensity of FVC variation, the stronger the temperature regulation effect of the project. An inverse regulation direction exists between daytime and nighttime, and the correlation between the intensity of FVC variation and the temperature regulation effect is more pronounced during the dry season.