Nothing Special   »   [go: up one dir, main page]
Next Article in Journal
Automatic Extraction and Cluster Analysis of Natural Disaster Metadata Based on the Unified Metadata Framework
Previous Article in Journal
Effectiveness of Adjacent and Bivariate Maps in Communicating Global Sensitivity Analysis for Geodiversity Assessment
You seem to have javascript disabled. Please note that many of the page functionalities won't work as expected without javascript enabled.
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
IJGI
Volume 13
Issue 6
10.3390/ijgi13060200
ijgi-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Exploring the Influence of Terrain Blockage on Spatiotemporal Variations in Land Surface Temperature from the Perspective of Heat Energy Redistribution

by
Hong Gao
1,2,
Yong Dong
1,
Liang Zhou
1,2,3,* and
Xi Wang
1
1
Faculty of Geomatics, Lanzhou Jiaotong University, Lanzhou 730070, China
2
National-Local Joint Engineering Research Center of Technologies and Applications for National Geographic State Monitoring, Lanzhou 730070, China
3
Department of Geography, University of California, Santa Barbara, CA 94607, USA
*
Author to whom correspondence should be addressed.
ISPRS Int. J. Geo-Inf. 2024, 13(6), 200; https://doi.org/10.3390/ijgi13060200
Submission received: 18 April 2024 / Revised: 4 June 2024 / Accepted: 5 June 2024 / Published: 14 June 2024
Browse Figures
Figure 1
<p>Location of study area.</p> "> Figure 2
<p>The flow diagram of this study.</p> "> Figure 3
<p>Illustration for extracting section elevation series in four directions. (<b>a</b>) Determination for the four main directions according to LST data. (<b>b</b>) Results of section elevation series for the four main directions in the spatial grid named G1000.</p> "> Figure 4
<p>Coefficients of correlation between different TBFs and LST.</p> "> Figure 5
<p>Evaluation indices for LST simulation in different months.</p> "> Figure 6
<p>Importance of different types of TBFs.</p> "> Figure 7
<p>Importance of different directions of TBFs.</p> "> Figure 8
<p>Scatter plot of simulation for single days and nights (the dotted line is the 1:1 line, where the simulated value equals the original value).</p> "> Figure 9
<p>Spatial distribution of the origin and simulation and the associated errors.</p> "> Figure 10
<p>Simulation accuracy for LST diurnal variation.</p> ">
Review Reports Versions Notes

Abstract

:
Land surface temperature (LST) is a critical indicator of the earth’s surface environment, which has significant implications for research on the ecological environment and climate change. The influence of terrain on LST is complex due to its rugged and varied surface topography. The relationship between traditional terrain features and LST has been comprehensively discussed in the literature; however, terrain blockage has received less attention and could influence LST by hindering the redistribution of heat energy in mountain regions. Here, we investigate the influence of terrain blockage on the spatiotemporal variation in LST in mountain regions. We first propose a terrain feature framework to characterize the effect of terrain blockage from the perspective of heat energy redistribution and then adopt a random forest model to analyze the relationship between terrain blockage features and LST over a whole year. The results show that terrain blockage significantly influences the spatial heterogeneity of LST, which can be effectively simulated based on terrain blockage features, with a mean deviation of less than 0.15 K. Terrain blockage has a more pronounced influence on LST during the four months from June to September. This influence is also more evident during nighttime than daytime. Regarding LST in mountain regions, local terrain blockage features have a greater influence than global terrain blockage features. In spatial terms, the influence of terrain blockage on LST is uniform. Moreover, the diurnal variation in LST can also be effectively simulated based on terrain blockage. The contribution of this study lies in the finding that terrain blockage can influence the spatiotemporal variation in LST through the process of heat energy redistribution. The terrain blockage features proposed in this study may be useful for other studies of the ecological environment in mountain regions.

1. Introduction

Land surface temperature (LST) is a critical indicator of the earth’s surface environment, which is of major concern for the ecological environment as well as global climate change, especially in terms of the Sustainable Development Goals of the United Nations [1,2]. LST has been linked to numerous problems related to extreme climate, environment and ecology, vegetation productivity, and urban heat exposure [3,4,5]. The spatial distribution of LST is influenced by many geographical factors, including solar radiation, vegetation conditions, terrain effect, soil moisture, and land use [6]. Therefore, it is essential to fully discuss the influence of different geographic elements on LST to explore the spatiotemporal heterogeneity in the earth’s surface thermal environment.
In the remote sensing of the surface thermal environment, an important task is the retrieval of LST, which can be performed using data-driven and algorithm-based approaches. The data used for the first approach include thermal infrared [7,8,9], passive microwave remote sensing [10,11,12], and other remote sensing data [13,14]. The algorithms for the LST retrieval include the split-window algorithm [15,16], mono-window algorithm [17], and regression model [18,19]. The other important task is exploring the relationship between LST and geographic factors. Various studies have analyzed the relationship between LST and vegetation cover [20,21,22], land use [23,24,25], and terrain effect [26,27]; the terrain effect influences LST in a complex manner [6,26,28,29,30]. LST retrieval and analysis are difficult in mountain regions because of their complex surface structure; therefore, the relationship between terrain and LST requires further study.
In mountain regions, two modes of heat energy distribution cause the spatial variation in LST; these are first distribution and redistribution. In first distribution, some places can be directly illuminated by the Sun and are exposed to solar radiation for varying times, whereas others are not due to terrain occlusion. The second mode involves heat exchange between the land surface and near-surface air [31,32], such that heat energy is redistributed along with the near-surface air flow. Hence, the redistribution of heat energy refers to the transfer of heat energy between the land surface and the air from other regions, because air movement disrupts the original heat balance. In this process, the raised mountain can block horizontal air movement, hindering the redistribution of heat energy and influencing the spatial distribution of LST. In this study, this phenomenon is referred to as terrain blockage.
Studies to date have paid much attention to traditional terrain features that are mainly characteristic of the first distribution of heat energy. However, the influence of terrain blockage on heat energy redistribution is often ignored in analyses of the spatial variation in LST in mountain regions. The relationship between LST and elevation has been comprehensively discussed, with many studies find that LST is significantly influenced by altitude [27,33]. Furthermore, slope and aspect are also analyzed with respect to LST in mountain regions, because surface illumination conditions change under different slopes and aspects [26,28]. Elevation, slope, and aspect are primary and common terrain features that have been used to estimate LST through the construction of regression models [6]. In summary, using these three primary terrain features, it is possible to characterize the influence of terrain on the first distribution of heat energy, with little consideration of terrain blockage. There are few studies in which terrain blockage has been considered, where an algorithm was proposed for retrieving LST with the correction of topographic effects [29]. This algorithm uses the sky-view factor (SVF) to characterize the influence of terrain blockage on LST. The SVF is able to reflect terrain occlusion in the vertical direction but is less capable of indicating terrain occlusion in the horizontal direction. This results in a lack of terrain features for characterizing terrain blockage in the horizontal direction; consequently, heat energy redistribution has not been considered in and introduced to the analysis of LST in the remote sensing of mountains.
Existing studies [6,26,27,28] have demonstrated that traditional terrain features can influence the spatial distribution of LST in mountain regions. However, analyses of LST spatiotemporal variation have paid less attention to terrain blockage, which hinders the redistribution of heat energy. To fill this gap, our study focused on horizontal terrain blockage and its influence on the spatiotemporal variation in LST in mountain regions in an attempt to answer the following important questions: (1) How can horizontal terrain blockage be characterized in the process of heat energy redistribution? (2) What are the spatiotemporal patterns of the influence of terrain blockage on LST in mountain regions?
This paper is organized as follows. Section 2 introduces the materials of this paper, including the study area and data. Section 3 describes the method used to explore the influence of terrain blockage on the spatiotemporal variation in land surface temperature. It consists of three main parts: characterizing horizontal terrain blockage by proposing terrain blockage features (Section 3.1), simulating the relationship between each different TBF and LST using a random forest model (Section 3.2), and evaluating the model results using different statistical indices (Section 3.3). In Section 4, the results of this study are presented in full to uncover the spatiotemporal patterns of the influence of terrain blockage on LST in mountain regions. Section 5 contains important discussion related to this study. The main conclusions are of this study are detailed in Section 6.

2. Materials

2.1. Study Area

The study area is a part of Qilian mountains in the western China (Figure 1), with a size of 214 km × 114 km, and its central coordinates are 102.1 E, 37.1 N. The elevation shows large variation, ranging from 1787 m to 4974 m. There is alternation of raised mountains with low valleys within the study area, resulting in the land surface having a complex three-dimensional structure. The west and north of the study area are characterized by continuous mountains, and the south is a valley area. There are a few plains in the northeast, and such landforms exhibit surface characteristics similar to those of mountain regions. Notably, the study area is located in the transitional zone between the Tibet Plateau and the Loess Plateau, which are two main plateaus in China. Therefore, this study area is a representative mountain region for inland regions, appropriate for exploring the relationship between terrain and LST.

2.2. Data

2.2.1. Digital Elevation Model

The digital elevation model (DEM) used in this study is from Shuttle Radar Topography Mission (SRTM) digital elevation model dataset and has a spatial resolution of 30 m × 30 m. This type of DEM dataset has been extensively applied in many studies to characterize terrain distribution [34] with excellent accuracy. In this study, DEM was used to represent the mountain region topography. The changes in elevation from the DEM were used to extract the digital features of terrain blockage in characterizing the heat energy redistribution.

2.2.2. LST Data

LST data with a spatial resolution of 1000 m × 1000 m were used in this study. They were derived from a dataset named TRIMS LST-TP; 2000-2021) V2 [35], which can be download from the National Tibetan Plateau Data Center in China (https://data.tpdc.ac.cn (accessed on 1 November 2023)). This dataset was produced using the Enhanced Reanalysis and Thermal infrared remote sensing Merging (E-RTM) method for the spatial scope of China mainland. It is indicated in the relevant literature [36,37,38] that this dataset has good image quality and seamless spatial accuracy. Moreover, this LST dataset demonstrates high consistency with the 1 km Terra/Aqua MODIS LST product in terms of amplitude and spatial distribution. Therefore, this dataset appropriately and accurately reflects the spatial distribution of the LST of the study area.
LST exhibits obvious changes between day and night as well as across different months. Therefore, in this study, the LST for each day and night over the course of a whole year was analyzed along with the relationship with terrain blockage. Specifically, LST data from this dataset for each single day and night across the whole year of 2021 were selected to explore temporal changes in an LST simulation by considering terrain blockage.

3. Methods

The methods used in this study involved two steps. The first step was determining how to characterize terrain blockage. For this, a series of terrain features was proposed for quantifying terrain blockage in raised mountains; these are called terrain blockage features (TBFs). In the second step, a random forest model was used to explore the influence of terrain blockage on LST in mountain regions and analyze its spatiotemporal patterns according to the data for one whole year. The flow diagram of this study is presented in Figure 2.

3.1. Characterization of Terrain Blockage in the Process of Heat Energy Redistribution

In this study, a series of TBFs for characterizing terrain blockage were proposed. To generate the TBFs, this study firstly extracted the elevation sequence in four main directions for each grid of LST. Then, according to the serial elevation values for each direction, different types of TBFs were proposed and calculated.

3.1.1. Extracting Section Elevation Series in Major Directions

To extract the section elevation in the major directions, the central location (geometric centroid) in each grid of LST was first determined. Then, straight lines from the central location outward in four main directions—northward, eastward, southward, and westward—were generated (Figure 3a). Finally, the elevation values of the DEM grids were extracted where four directional lines passed through to obtain a section elevation series for the four main directions. In each direction, severe and frequent elevation changes would incur obvious terrain blockage of near-surface air movements, while smaller and less frequent elevation changes would not. Figure 3b shows the section elevation series of the four main directions in the spatial grid named G1000.

3.1.2. Generation of TBFs Using Serial Values of Section Elevation

Based on the section elevation series for the four main directions, this study proposed five terrain features for measuring the blocking effects of raised mountains on near-surface air movements and the corresponding heat energy redistribution. These terrain blockage features are average elevation, average neighbor change, elevation changing range, maximum continuous climbing height, and local peak quantity. Explanations for the TBFs are shown in Table 1.
To summarize, the terrain feature framework is calculated according to four main directions, and five features are generated for each direction. Therefore, this study proposes 20 terrain blockage features for reflecting the redistribution of heat energy in mountain regions.

3.2. Simulating the Relationship between TBFs and LST Using a Random Forest Model

To extract the relationship between LST and each TBF, a random forest (RF) model was used in this study. RF is an ensemble method for classification and regression tasks in machine learning [39]. It has been shown that RF demonstrates high and stable accuracy in geographic modeling, such as in soil mapping [40,41] and land cover mapping [42]. In particular, RF has also been widely used to estimate LST in mountainous regions [6,43]. In the random forest regression model of this study, five kinds of terrain blockage features in four different directions were taken as explanatory variables, and the LST in the daytime and nighttime were each taken as target variables. The random forest model was implemented using the randomForest package in R software (version 4.2.3) [44].
There are two important parameters when using a random forest model. One is the number of decision trees, which was set to 500 in this study, because the variation in training accuracies decreases when starting from this value in model building. The other parameter is the number of variables randomly sampled as candidates at each split, which is set as 6 and is calculated according to the number of explanatory variables as suggested by the random forest model.
In the random forest model, the dataset with 24,396 samples, which is the number of raster cells in the LST data, was randomly divided into training data (80%) and test data (20%). The training data were used to build a random forest model that could simulate the influence of TBFs on LST, and the test data were used for evaluating the generalizability of such influence.

3.3. Statistical Indices for Evaluating Model Results

The model results were evaluated based on the original and simulated LST values. This study calculated four statistical indices to evaluate the simulation results by using the original LST values and the corresponding simulated values. The indices are the coefficient of determination (R-squared), mean absolute error (MAE), root-mean-square error (RMSE), and mean deviation (MD). These indices were calculated according to the following equations:
R-square = 1 − ∑ i = 1 n ( y i − y ^ ) 2 ∑ i = 1 n ( y i − y ¯ ) 2
MAE = 1 n ∑ i = 1 n y i − y ^ i
RMSE = 1 n ∑ i = 1 n ( y i − y ^ i ) 2
MD = 1 n ∑ i = 1 n y ^ i − 1 n ∑ i = 1 n y i
where  y i  is the original LST value in cell  i ;  y ^ i  is the simulated LST value in cell  i ;  y ¯  is the mean of all the original values; and  n  is the number of LST cells. A higher R-squared and lower MAE, RMSE, and MD indicate more accurate results in the LST simulation. These indices were used to evaluate the results of the LST simulation.

4. Results

4.1. Correlation Coefficient Analysis for TBF and LST

In this study, the Pearson coefficient of the correlation between each TBF and LST was first calculated to analyze their linear relationship. The Pearson coefficients of the correlations between different TBFs and LST in daytime and nighttime are shown in Figure 4. AE, ANC, MCCH, and ECR were found to be negatively correlated with LST for both daytime and nighttime. This indicates that stronger terrain blockage results in a lower LST. The reason might be that terrain blockage hinders the movement of heated near-surface air to cold places, and these places would not consequently be warmed up by the redistribution of heat energy. AE exhibits the highest correlation with LST in both daytime and nighttime. LPQ and LST show a very low positive correlation, indicating that LPQ has limited predictive power for LST. This is because a few local mountain peaks are found in each direction for most of the spatial units. There is little variation in LPQ such that it could not possibly reflect the diversity of LST, and for which there is a correspondingly low correlation coefficient. In terms of the difference between daytime and nighttime, there is lower correlation between AE and LST in daytime than in nighttime; however, the correlations between NC, MCCH, ECR, and LST are higher in daytime than nighttime.

4.2. Temporal Patterns for LST Simulated Using TBF and Their Accuracy

All of the different TBFs in each direction were used to simulate LST in the random forest model for each day and night of one year. The values of four indices were calculated to evaluate the simulation results and then aggregated into a monthly average. The evaluation indices for the LST simulation for different months are shown in Figure 5, and the error bars indicate standard deviations within the indicated month.
According to the results for the training data (Figure 5a), the TBFs can effectively simulate LST for both daytime and nighttime, showing high accuracy. The average R-squared for all months exceeds 0.9, the average MAEs are all less than 1.5 K, and the average RMSEs are all less than 2.5 K. The MD values in different months are all near 0, and their absolute values are all less than 0.15 K. Furthermore, the simulation accuracy for nighttime is consistently higher than that for daytime for each month. The reason might be that there is no direct solar radiation input during nighttime. The influence of terrain blockage on heat energy redistribution and thus the spatial distribution of LST would be dominant in nighttime, such that, for LST simulated using TBFs, the accuracy would be higher in nighttime than in daytime. The accuracy was higher for the four months from June to September than the other months. This is because these four months correspond to summer in the study area, when the air temperature is highest and the heat exchange between the air and the land surface is the most intense; thus, there is the highest frequency of near-surface air flowing after being heated. Therefore, in terms of seasonal variation, the influence of terrain blockage on LST is most obvious in summer, and the results of the LST simulation using TBFs in summer are the most accurate.
Figure 5b shows the simulation results generated by applying the model to the test dataset, which could reflect the generalization ability of the model fitted using TBFs. It can also be observed that there is higher accuracy in nighttime than in daytime for the test dataset, as well as for the four months from June to September than the other months in the test dataset. These phenomena are also observed for the training dataset because the redistribution of heat energy is prominent during these time periods. Furthermore, the test dataset has lower accuracy than the training dataset, showing high variation across the different months. The limited generalization ability of TBFs is due to the fact that near-surface air movement is frequent and causes redistribution of heat energy. If near-surface air movement is frequent and causes obvious heat energy redistribution, TBFs could be applied to effectively simulate and predict LST for an unknown area.
According to the variable importance obtained from the random forest model, this study further analyzed the importance of different types of TBF in different directions on LST in mountainous regions. The variable importance was measured according to the mean decrease in accuracy from the random forest model. For a definition of the variable importance measurement, we refer the reader to [44]. For each tree, the prediction error for the out-of-bag portion of the data was recorded. This was reiterated after permuting each predictor variable. The difference between the two were then averaged over all the trees and normalized using the standard deviation of the differences. If the standard deviation of the differences was 0 for a variable, then division was not performed (but the average was almost always equal to 0 in that case).
Figure 6 shows the importance of the different types of TBFs in each month and their average importance. The importance ranking of TBFs shows similar patterns for the different months. ANC is a local terrain blockage characteristic and has the highest importance, indicating that local features of terrain blockage have a greater influence on the spatial variation in LST. MCCH and ECR, global features of terrain blockage, are lower in importance than ANC, which means that global features of terrain blockage have a smaller influence on LST than local features. AE and LPQ have relatively low feature importance. The correlation coefficients of ANC, MCCH, and ECR with LST are relatively low. However, in the random forest model, ANC, MCCH, and ECR have the highest importance for simulating LST, indicating that ANC, MCCH, and ECR have a complex nonlinear relationship with LST.
Figure 7 shows the importance of the different directions of TBFs in each month and their average importance. In the daytime, the different directions of TBFs have similar importance in different months. In the nighttime, the importance of different directions of TBFs is lower in May, August, and September than in the other months. Overall, the average importance of different directions is similar, indicating that LST is not affected by the specific direction of near-surface air flow. The reason may be that there is heat exchange between the surface and the near-surface air, and the flow of near ground air is mainly caused by thermal differences on the earth’s surface. Specific wind directions can cause changes in the upper air temperature in mountainous areas, thus having a relatively small impact on LST. Thus, TBFs have a similar importance for LST for the different directions.

4.3. Comparison of Spatial Distribution between Original and Simulated LST

Figure 8 shows a scatter plot of the simulation results for single days and nights. From all of the days and nights in the year of 2021, simulation results for two days and two nights were selected for illustration. These are the day with the highest simulation accuracy (Figure 8a, Day 111), the day with the lowest simulation accuracy (Figure 8b, Day 142), the night with the highest simulation accuracy (Figure 8c, Night 187), and the night with the lowest simulation accuracy (Figure 8d, Night 284). For the results with the highest simulation accuracy (Figure 8a,c), the scatter points formed by the simulated and original values are closely distributed near the 1:1 line (black dashed line in the picture), indicating that they are very close to each other and the high performance of the simulation.
For the results with the lowest simulation accuracy (Figure 8b,d), the scatter points formed by the simulated and original values are only found closely distributed near the 1:1 line if the original LST is within the range of high values (>270 K). However, if the original LST is within the range of low values (<270 K), the scatter points formed by the simulated and original values are above the 1:1 line, indicating that the simulated value is higher than the original value. This indicates that the TBF-based simulation model is likely to overestimate the LST when the original LST is in the range of low values.
This study further analyzed the spatial distribution of errors, namely the simulated LST minus the original LST. The errors are small, indicating that the LST value is effectively simulated using TBFs and that terrain blockage has a strong influence on LST in these locations. Therefore, this study mainly explored the spatial patterns of the influence of terrain blockage through the spatial distribution of errors. Figure 9 shows the spatial distribution of the simulation results for the single day (Day 111) and night (Night 187) with the highest accuracy.
It can be found that the simulated and original LST have similar patterns of spatial distribution. The simulated and original image are similar in terms of the low values at the ridge, the high values at the valley, and the transition from the ridge to valley. This demonstrates the similar spatial distribution of the simulation original results, indicating that TBFs can be used for the high-accuracy simulation of LST. The error from the simulation is larger in daytime than in nighttime because LST is often higher in daytime. The negative and positive error values in the daytime images are in relative equilibrium. However, for the nighttime images, there are more grids with positive than negative error values, which indicates that the simulated LST is higher than the original LST, and that LST in nighttime may be overestimated when using TBFs.
To summarize, the simulation errors are relatively small and their spatial distribution is homogeneous, indicating that terrain blockage has a universal influence on LST over the entire mountain region.

5. Discussion

5.1. Comparing First Distribution with Redistribution of Heat Energy in Mountain Regions

This study summarizes geographic processes that affect LST in mountain regions into the first distribution and redistribution of heat energy. The first distribution of heat energy mainly refers to the temperature changes caused by direct sunlight exposure, while redistribution refers to the temperature changes caused by air from other regions. The first distribution and redistribution of heat energy in mountain regions can be influenced by numerous terrain features. Traditional terrain features, such as latitude, slope, aspect, and hill shade, can influence the first distribution of heat energy by inhibiting sunlight exposure or affecting the direction and duration of such exposure. In heat energy redistribution, terrain blockage features are the most important influencing factors. The raised mountains hinder the horizontal flow of near-surface air, which weakens the process of heat energy redistribution. Thus, from the perspective of the redistribution of heat energy, this is another way in which the terrain influences LST.
In reality, the first distribution and redistribution of heat energy co-exist, and these co-determine the spatial distribution of LST. The first distribution of heat energy is obvious and dominant in determinations of LST in mountain regions, with many examples of studies focusing on first distribution to discuss the effect of terrain on LST. However, our results demonstrate that the redistribution of heat energy also influences the spatial distribution of LST in mountain regions. We find that TBFs can better simulate LST in nighttime than in daytime, and the simulation accuracy is higher in the months from June to September than other months. When there is obvious and strong heat energy redistribution in mountain regions, considering terrain blockage would result in higher accuracy for LST derived from simulations.

5.2. Directions in Characterizing Terrain Blockage Caused by Raised Mountains

To reflect near-surface air movements, the calculation of the proposed TBFs was based on four main directions, which are northward, southward, eastward, and westward. These four directions represent the basic and universal directions of near-surface air movements. TBFs were calculated according to these four directions and can illustrate the overall difficulty for near-surface air to move in the spatial unit. The results show that the TBFs proposed in this study can effectively simulate the spatial distribution of LST in mountain regions. However, these four directions might not represent the actual directions of near-surface air movements in mountain regions.
In this study, a framework that uses four cardinal directions to generate terrain blockage features was proposed. Based on this framework, other approaches to select directions could be implemented. In related future work, there are two kinds of approaches that require further discussion regarding the selection of directions of near-surface air movements. On one hand, determinations could be based on more than four directions, such as 8 or 16, to fully explore the abundant and possible directions of near-surface air movement. TBFs could be calculated using more precise directions closer to the actual direction of near-surface air movement. On the other hand, near-surface air movements accompany certain geographical processes. We can determine the direction for characterizing terrain blockage according to geographic process of near-surface air.
The importance of direction lies in the fact that terrain blockage is directed against specific directions. Terrain blockage might therefore vary in intensity for different directions. In addition, five kinds of TBFs are proposed in this study that could be applicable for other directions. For other directions, we could also extract serial values of section elevation, which can be used to calculate the proposed TBFs. Therefore, the TBFs proposed in this study can be used in other studies characterizing terrain blockage.

5.3. Predictive Power of Terrain Blockage for Other Properties of Thermal Environment

Diurnal temperature variation in mountain regions is meaningful for vegetative growth and biomass accumulation, which is also an important geographic characteristic of mountain regions. Therefore, the simulation results for LST were also discussed in terms of diurnal variation (LST in daytime minus LST in nighttime for the same day) to demonstrate the predictive power of TBFs for other properties, besides LST, of the thermal environment in mountain regions.
The simulation results of LST for single days and nights show that the accuracy is similar when comparing two adjacent days. This is similar to the simulation results for LST diurnal variation. To ensure obvious temporal regularity while maintaining a low calculation cost, it was determined that a time interval of five days should be used in the simulation of LST diurnal variation. Figure 10 shows the simulation accuracy of LST diurnal variation every five days in a year. It is found that the TBF method has higher accuracy in simulating LST diurnal variation.
Diurnal variation in LST is mainly controlled by heat exchange between the land surface and near-surface air. TBFs can be used in characterizing the difficulty for horizontal movement of near-surface air. A higher TBF value indicates that the raised mountains make it more difficult for near-surface air to move horizontally. As a result, terrain blockage would keep near-surface air thermostatic during the transition from daytime to nighttime. Therefore, TBFs also have a strong relationship with the changes in LST from daytime to nighttime. It was found that TBFs also demonstrate predictive ability for other properties of the thermal environment.

5.4. Limitations and Usefulness

In existing studies, traditional terrain features and other geographic elements have been simultaneously used as predictors for estimating LST [6,26]. However, there is insufficient consideration of terrain blockage in these studies, whose contribution has not been separately evaluated. This study extracts terrain blockage features from the DEM data to analyze their influence on LST. Therefore, other geographic elements are not used in the random forest model. The results of this study mainly reflect the influence of terrain blockage on LST in mountain regions, with the influence of other geographic elements excluded. Other geographic elements, such as NDVI, are also important and prevalent regarding LST in mountain regions. In future work, we would combine the TBFs proposed in this study with other geographic elements to estimate LST, so as to discuss the differences in their influence on LST.
The TBFs proposed in this study are compared with traditional terrain features for the simulation of LST using a random forest model. Traditional terrain features that are widely used in existing studies [6,26,28,45,46] to estimate LST are elevation, slope, and aspect. These three traditional features were used to simulate LST based on a random forest model, and the comparison of the results is shown in Table 2.
According to the R-squared, RMSE, and MAE, the average accuracy of the TBF-based method is higher than that of traditional terrain features, whereas the standard deviation of the accuracy is lower. Based on the MD, the accuracy of the TBF-based method is very close to that of traditional terrain features. Compared with traditional terrain features, there is a 2% improvement in the simulation accuracy of the TBF-based method and a 24% reduction in simulation variation in terms of R-squared. Thus, in the simulation of LST, the TBF-based method allows more stable and precise results than the method based on traditional terrain features. The reason for the high accuracy of TBFs is the richness in the types and directions of terrain features. Using the TBF framework, many more refined features for simulating LST can be obtained, so as to ensure good simulation results with increased computation.
The usefulness of TBFs mainly relies on their application in characterizing terrain blockage in raised mountains, such that they can be used as indicators of terrain blockage in other studies of LST. Many other studies estimating LST in mountain regions ignore the redistribution of heat energy, which is mainly controlled by terrain blockage. For example, TBFs could be set as additional terrain features that are considered together with traditional terrain features and other geographic factors in estimating or retrieval of LST in mountain regions. Moreover, TBFs could also be used to estimate rainfall in mountainous regions and toward discussions of spatial distribution of soil and vegetation. Therefore, TBFs could have wide application in remote sensing studies of mountain regions when considering terrain blockage.

6. Conclusions

In this study, a method for characterizing and analyzing terrain blockage and its influence on LST in mountain regions was proposed. By using terrain blockage features to simulate LST, we found that terrain blockage has a significant influence on the spatiotemporal variation in LST. The mean deviation of the simulation results ranges from −0.11 K to 0.003 K in the daytime and from −0.04 K to 0.003 K in the nighttime. The influence of terrain blockage on LST is more pronounced during the four-month period from June to September and more evident during nighttime than daytime. Local features of terrain blockage have a greater influence than global features on the LST in mountain regions. In spatial terms, the influence of terrain blockage on LST is universal and uniform. Moreover, terrain blockage can also be used to reveal the LST diurnal variation in mountain regions.
The contribution of this study lies in the introduction and characterization of terrain blockage from the perspective of heat energy redistribution, with exploration of the influence of terrain blockage features on the spatial heterogeneity of LST. The terrain blockage features proposed in this study may be useful for other studies involving remote sensing of mountains.

Author Contributions

Conceptualization, Hong Gao; data curation, Yong Dong and Liang Zhou; formal analysis, Hong Gao and Liang Zhou; funding acquisition, Hong Gao and Liang Zhou; methodology, Hong Gao and Xi Wang; supervision, Liang Zhou; writing—original draft, Hong Gao; writing—review and editing, Yong Dong, Liang Zhou and Xi Wang. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by the Science-Technology Foundation for Young Scientists of Gansu Province (No. 22JR11RA149), National Natural Science Foundation of China (No. 42271214), Young Scholars Science Foundation of Lanzhou Jiaotong University (No. 2022005), Natural Science Foundation of Gansu Province (No. 21JR7RA281, 21JR7RA278), the CAS ‘Light of West China’ Program, the Basic Research Top Talent Plan of Lanzhou Jiaotong University (No. 2022JC01), and the Talent Innovation and Entrepreneurship Project of Lanzhou (No. 2023-RC-31).

Data Availability Statement

The LST data that support the findings of this study are available from the National Tibetan Plateau Data Center in China (https://data.tpdc.ac.cn (accessed on 1 November 2023)). The DEM data that support the findings of this study are available from the NASA EOSDIS Land Processes Distributed Active Archive Center.

Acknowledgments

The authors would like to express their gratitude to the editors and anonymous reviewers for their time and effort in processing this manuscript, which have greatly improved its quality.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Fuldauer, L.I.; Thacker, S.; Haggis, R.A.; Fuso-Nerini, F.; Nicholls, R.J.; Hall, J.W. Targeting climate adaptation to safeguard and advance the Sustainable Development Goals. Nat. Commun. 2022, 13, 15. [Google Scholar] [CrossRef] [PubMed]
  2. Liu, J.L.; Tang, H.Y.; Yan, F.Q.; Liu, S.W.; Tang, X.G.; Ding, Z.; Yu, P.J. Future variation of land surface temperature in the Yangtze River Basin based on CMIP6 model. Int. J. Digit. Earth 2023, 16, 2776–2796. [Google Scholar] [CrossRef]
  3. Massaro, E.; Schifanella, R.; Piccardo, M.; Caporaso, L.; Taubenböck, H.; Cescatti, A.; Duveiller, G. Spatially-optimized urban greening for reduction of population exposure to land surface temperature extremes. Nat. Commun. 2023, 14, 10. [Google Scholar] [CrossRef] [PubMed]
  4. Xie, X.Y.; Tian, J.; Wu, C.L.; Li, A.N.; Jin, H.A.; Bian, J.H.; Zhang, Z.J.; Nan, X.; Jin, Y. Long-term topographic effect on remotely sensed vegetation index-based gross primary productivity (GPP) estimation at the watershed scale. Int. J. Appl. Earth Obs. Geoinf. 2022, 108, 12. [Google Scholar] [CrossRef]
  5. Mushore, T.D.; Mutanga, O.; Odindi, J. Understanding Growth-Induced Trends in Local Climate Zones, Land Surface Temperature, and Extreme Temperature Events in a Rapidly Growing City: A Case of Bulawayo Metropolitan City in Zimbabwe. Front. Environ. Sci. 2022, 10, 14. [Google Scholar] [CrossRef]
  6. Zhao, W.; Duan, S.B.; Li, A.N.; Yin, G.F. A practical method for reducing terrain effect on land surface temperature using random forest regression. Remote Sens. Environ. 2019, 221, 635–649. [Google Scholar] [CrossRef]
  7. Zhu, J.; Ren, H.; Ye, X.; Teng, Y.; Zeng, H.; Liu, Y.; Fan, W. PKULAST-An Extendable Model for Land Surface Temperature Retrieval From Thermal Infrared Remote Sensing Data. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 2022, 15, 9278–9292. [Google Scholar] [CrossRef]
  8. Zhao, H.; Ren, H.; Fu, G. An Analytic Solution Method for Retrieving Land Surface Temperature from Remotely Sensed Thermal Infrared Imagery. J. Indian Soc. Remote Sens. 2015, 43, 279–286. [Google Scholar] [CrossRef]
  9. Duan, S.-B.; Li, Z.-L.; Leng, P. A framework for the retrieval of all-weather land surface temperature at a high spatial resolution from polar-orbiting thermal infrared and passive microwave data. Remote Sens. Environ. 2017, 195, 107–117. [Google Scholar] [CrossRef]
  10. Duan, S.-B.; Han, X.-J.; Huang, C.; Li, Z.-L.; Wu, H.; Qian, Y.; Gao, M.; Leng, P. Land Surface Temperature Retrieval from Passive Microwave Satellite Observations: State-of-the-Art and Future Directions. Remote Sens. 2020, 12, 2573. [Google Scholar] [CrossRef]
  11. Mao, K.; Wang, H.; Shi, J.; Heggy, E.; Wu, S.; Bateni, S.M.M.; Du, G. A General Paradigm for Retrieving Soil Moisture and Surface Temperature from Passive Microwave Remote Sensing Data Based on Artificial Intelligence. Remote Sens. 2023, 15, 1793. [Google Scholar] [CrossRef]
  12. Wang, X.; Zhong, L.; Ma, Y. Estimation of 30 m land surface temperatures over the entire Tibetan Plateau based on Landsat-7 ETM+ data and machine learning methods. Int. J. Digit. Earth 2022, 15, 1038–1055. [Google Scholar] [CrossRef]
  13. Ermida, S.L.; Soares, P.; Mantas, V.; Goettsche, F.-M.; Trigo, I.E. Google Earth Engine Open-Source Code for Land Surface Temperature Estimation from the Landsat Series. Remote Sens. 2020, 12, 1471. [Google Scholar] [CrossRef]
  14. Liu, W.; Shi, J.; Liang, S.; Zhou, S.; Cheng, J. Simultaneous retrieval of land surface temperature and emissivity from the FengYun-4A advanced geosynchronous radiation imager. Int. J. Digit. Earth 2022, 15, 198–225. [Google Scholar] [CrossRef]
  15. Ye, X.; Ren, H.; Zhu, J.; Fan, W.; Qin, Q. Split-Window Algorithm for Land Surface Temperature Retrieval From Landsat-9 Remote Sensing Images. IEEE Geosci. Remote Sens. Lett. 2022, 19, 7507205. [Google Scholar] [CrossRef]
  16. Rongali, G.; Keshari, A.K.; Gosain, A.K.; Khosa, R. Split-Window Algorithm for Retrieval of Land Surface Temperature Using Landsat 8 Thermal Infrared Data. J. Geovis. Spat. Anal. 2018, 2, 14. [Google Scholar] [CrossRef]
  17. Wang, F.; Qin, Z.; Song, C.; Tu, L.; Karnieli, A.; Zhao, S. An Improved Mono-Window Algorithm for Land Surface Temperature Retrieval from Landsat 8 Thermal Infrared Sensor Data. Remote Sens. 2015, 7, 4268–4289. [Google Scholar] [CrossRef]
  18. Saradjian, M.R.; Jouybari-Moghaddam, Y. Land Surface Emissivity and temperature retrieval from Landsat-8 satellite data using Support Vector Regression and weighted least squares approach. Remote Sens. Lett. 2019, 10, 439–448. [Google Scholar] [CrossRef]
  19. Kiavarz, M.; Firozjaei, M.K.; Alavipanah, S.K.; Hassan, Q.K.; Malbeteau, Y.; Duan, S.-B. A new approach to LST modeling and normalization under clear-sky conditions based on a local optimization strategy. Int. J. Digit. Earth 2022, 15, 1833–1854. [Google Scholar] [CrossRef]
  20. Guha, S.; Govil, H. Annual assessment on the relationship between land surface temperature and six remote sensing indices using landsat data from 1988 to 2019. Geocarto Int. 2022, 37, 4292–4311. [Google Scholar] [CrossRef]
  21. Song, Z.; Yang, H.; Huang, X.; Yu, W.; Huang, J.; Ma, M. The spatiotemporal pattern and influencing factors of land surface temperature change in China from 2003 to 2019. Int. J. Appl. Earth Obs. Geoinf. 2021, 104, 102537. [Google Scholar] [CrossRef]
  22. Liu, J.; Zhou, X.; Tang, H.; Yan, F.; Liu, S.; Tang, X.; Ding, Z.; Jiang, K.; Yu, P. Feedback and contribution of vegetation, air temperature and precipitation to land surface temperature in the Yangtze River Basin considering statistical analysis. Int. J. Digit. Earth 2023, 16, 2941–2961. [Google Scholar] [CrossRef]
  23. Nega, W.; Balew, A. The relationship between land use land cover and land surface temperature using remote sensing: Systematic reviews of studies globally over the past 5 years. Environ. Sci. Pollut. Res. 2022, 29, 42493–42508. [Google Scholar] [CrossRef] [PubMed]
  24. Muro, J.; Strauch, A.; Heinemann, S.; Steinbach, S.; Thonfeld, F.; Waske, B.; Diekkrueger, B. Land surface temperature trends as indicator of land use changes in wetlands. Int. J. Appl. Earth Obs. Geoinf. 2018, 70, 62–71. [Google Scholar] [CrossRef]
  25. Hasan, M.A.; Mia, M.B.; Khan, M.R.; Alam, M.J.; Chowdury, T.; Al Amin, M.; Ahmed, K.M.U. Temporal Changes in Land Cover, Land Surface Temperature, Soil Moisture, and Evapotranspiration Using Remote Sensing Techniques—A Case Study of Kutupalong Rohingya Refugee Camp in Bangladesh. J. Geovis.Spat. Anal. 2023, 7, 11. [Google Scholar] [CrossRef]
  26. He, J.L.; Zhao, W.; Li, A.N.; Wen, F.P.; Yu, D.J. The impact of the terrain effect on land surface temperature variation based on Landsat-8 observations in mountainous areas. Int. J. Remote Sens. 2019, 40, 1808–1827. [Google Scholar] [CrossRef]
  27. Bindajam, A.A.; Mallick, J.; AlQadhi, S.; Singh, C.K.; Hang, H.T. Impacts of Vegetation and Topography on Land Surface Temperature Variability over the Semi-Arid Mountain Cities of Saudi Arabia. Atmosphere 2020, 11, 28. [Google Scholar] [CrossRef]
  28. Peng, W.F.; Zhou, J.M.; Wen, L.J.; Xue, S.; Dong, L.J. Land surface temperature and its impact factors in Western Sichuan Plateau, China. Geocarto Int. 2017, 32, 919–934. [Google Scholar] [CrossRef]
  29. Zhu, X.L.; Duan, S.B.; Li, Z.L.; Zhao, W.; Wu, H.; Leng, P.; Gao, M.F.; Zhou, X.M. Retrieval of Land Surface Temperature With Topographic Effect Correction From Landsat 8 Thermal Infrared Data in Mountainous Areas. IEEE Trans. Geosci. Remote Sens. 2021, 59, 6674–6687. [Google Scholar] [CrossRef]
  30. Bento, V.A.; DaCamara, C.C.; Trigo, I.F.; Martins, J.P.A.; Duguay-Tetzlaff, A. Improving Land Surface Temperature Retrievals over Mountainous Regions. Remote Sens. 2017, 9, 13. [Google Scholar] [CrossRef]
  31. Mutiibwa, D.; Strachan, S.; Albright, T. Land Surface Temperature and Surface Air Temperature in Complex Terrain. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 2015, 8, 4762–4774. [Google Scholar] [CrossRef]
  32. Oyler, J.W.; Dobrowski, S.Z.; Holden, Z.A.; Running, S.W. Remotely Sensed Land Skin Temperature as a Spatial Predictor of Air Temperature across the Conterminous United States. J. Appl. Meteorol. Climatol. 2016, 55, 1441–1457. [Google Scholar] [CrossRef]
  33. Aguilar-Lome, J.; Espinoza-Villar, R.; Espinoza, J.-C.; Rojas-Acuna, J.; Leo Willems, B.; Leyva-Molina, W.-M. Elevation-dependent warming of land surface temperatures in the Andes assessed using MODIS LST time series (2000–2017). Int. J. Appl. Earth Obs. Geoinf. 2019, 77, 119–128. [Google Scholar] [CrossRef]
  34. Rana, V.K.; Suryanarayana, T.M.V. Visual and statistical comparison of ASTER, SRTM, and Cartosat digital elevation models for watershed. J. Geovis. Spat. Anal. 2019, 3, 12. [Google Scholar] [CrossRef]
  35. Zhou, J.; Zhang, X.; Tang, W.; Ding, L.; Ma, J.; Zhang, X. Daily 1-km All-Weather Land Surface Temperature Dataset for the Chinese Landmass and Its Surrounding Areas (TRIMS LST; 2000–2021); National Tibetan Plateau Data Center, Ed.; National Tibetan Plateau Data Center: Beijing, China, 2021. [Google Scholar] [CrossRef]
  36. Zhang, X.; Zhou, J.; Göttsche, F.M.; Zhan, W.; Liu, S.; Cao, R. A Method Based on Temporal Component Decomposition for Estimating 1-km All-Weather Land Surface Temperature by Merging Satellite Thermal Infrared and Passive Microwave Observations. IEEE Trans. Geosci. Remote Sens. 2019, 57, 4670–4691. [Google Scholar] [CrossRef]
  37. Zhou, J.; Zhang, X.; Zhan, W.; Göttsche, F.M.; Liu, S.; Olesen, F.S.; Hu, W.; Dai, F. A Thermal Sampling Depth Correction Method for Land Surface Temperature Estimation From Satellite Passive Microwave Observation Over Barren Land. IEEE Trans. Geosci. Remote Sens. 2017, 55, 4743–4756. [Google Scholar] [CrossRef]
  38. Zhang, X.; Zhou, J.; Liang, S.; Wang, D. A practical reanalysis data and thermal infrared remote sensing data merging (RTM) method for reconstruction of a 1-km all-weather land surface temperature. Remote Sens. Environ. 2021, 260, 112437. [Google Scholar] [CrossRef]
  39. Breiman, L. Random Forests. Mach. Learn. 2001, 45, 5–32. [Google Scholar] [CrossRef]
  40. Makungwe, M.; Chabala, L.M.; Chishala, B.H.; Lark, R.M. Performance of linear mixed models and random forests for spatial prediction of soil pH. Geoderma 2021, 397, 115079. [Google Scholar] [CrossRef]
  41. Gao, H.; Zhang, X.; Wang, L.; He, X.; Shen, F.; Yang, L. Selection of training samples for updating conventional soil map based on spatial neighborhood analysis of environmental covariates. Geoderma 2020, 366, 114244. [Google Scholar] [CrossRef]
  42. Agarwal, S.; Nagendra, H. Classification of Indian cities using Google Earth Engine. J. Land Use Sci. 2019, 14, 425–439. [Google Scholar] [CrossRef]
  43. Phan Thanh, N.; Degener, J.; Kappas, M. Comparison of Multiple Linear Regression, Cubist Regression, and Random Forest Algorithms to Estimate Daily Air Surface Temperature from Dynamic Combinations of MODIS LST Data. Remote Sens. 2017, 9, 398. [Google Scholar] [CrossRef]
  44. Liaw, A.; Wiener, M. Classification and Regression by randomForest. R News 2002, 2, 18–22. [Google Scholar]
  45. Taripanah, F.; Ranjbar, A. Quantitative analysis of spatial distribution of land surface temperature (LST) in relation Ecohydrological, terrain and socio- economic factors based on Landsat data in mountainous area. Adv. Space Res. 2021, 68, 3622–3640. [Google Scholar] [CrossRef]
  46. Xie, X.Y.; Li, A.N. Development of a topographic-corrected temperature and greenness model (TG) for improving GPP estimation over mountainous areas. Agric. For. Meteorol. 2020, 295, 13. [Google Scholar] [CrossRef]
Ijgi 13 00200 g001
Figure 1. Location of study area.
Figure 1. Location of study area.
Ijgi 13 00200 g001
Ijgi 13 00200 g002
Figure 2. The flow diagram of this study.
Figure 2. The flow diagram of this study.
Ijgi 13 00200 g002
Ijgi 13 00200 g003
Figure 3. Illustration for extracting section elevation series in four directions. (a) Determination for the four main directions according to LST data. (b) Results of section elevation series for the four main directions in the spatial grid named G1000.
Figure 3. Illustration for extracting section elevation series in four directions. (a) Determination for the four main directions according to LST data. (b) Results of section elevation series for the four main directions in the spatial grid named G1000.
Ijgi 13 00200 g003
Ijgi 13 00200 g004
Figure 4. Coefficients of correlation between different TBFs and LST.
Figure 4. Coefficients of correlation between different TBFs and LST.
Ijgi 13 00200 g004
Ijgi 13 00200 g005
Figure 5. Evaluation indices for LST simulation in different months.
Figure 5. Evaluation indices for LST simulation in different months.
Ijgi 13 00200 g005
Ijgi 13 00200 g006
Figure 6. Importance of different types of TBFs.
Figure 6. Importance of different types of TBFs.
Ijgi 13 00200 g006
Ijgi 13 00200 g007
Figure 7. Importance of different directions of TBFs.
Figure 7. Importance of different directions of TBFs.
Ijgi 13 00200 g007
Ijgi 13 00200 g008
Figure 8. Scatter plot of simulation for single days and nights (the dotted line is the 1:1 line, where the simulated value equals the original value).
Figure 8. Scatter plot of simulation for single days and nights (the dotted line is the 1:1 line, where the simulated value equals the original value).
Ijgi 13 00200 g008
Ijgi 13 00200 g009
Figure 9. Spatial distribution of the origin and simulation and the associated errors.
Figure 9. Spatial distribution of the origin and simulation and the associated errors.
Ijgi 13 00200 g009
Ijgi 13 00200 g010
Figure 10. Simulation accuracy for LST diurnal variation.
Figure 10. Simulation accuracy for LST diurnal variation.
Ijgi 13 00200 g010
Table 1. The framework of terrain blockage features.
Table 1. The framework of terrain blockage features.
NameExplanations
Average Elevation (AE)AE is the average elevation in the section elevation series of a single direction. It is mainly used to illustrate the elevation where terrain blockage occurs. It is calculated as the statistical mean according to the values of section elevation series and is a global feature of terrain blockage.
Average Neighbor Change (ANC)ANC is the average value of all local changes in the section elevation series of a single direction and characterizes the difficulty of near-surface air in moving from one location to a neighboring location. A larger ANC indicates higher difficulty of short-distance movement for near-surface air. ANC is a local feature of terrain blockage.
Elevation Changing Range (ECR)ECR is the global range of elevation in the section elevation series of a single direction. ECR equals the difference between the maximum and the minimum elevation in a single direction. A larger ECR indicates that it is more difficult for near-surface air to realize long-distance movement and is a global feature of terrain blockage.
Maximum Continuous Climbing Height (MCCH)MCCH is the height difference in a part of section elevation series and represents the longest distance when elevation is continuously increasing. The continuous increase in elevation means the continuous climbing of near-surface air, which needs to overcome the barrier of terrain. Thus, MCCH is the largest barrier of terrain in the process of near-surface air movement and is a local feature of terrain blockage.
Local Peak Quantity (LPQ)LPQ is the quantity of local mountain peak in the section elevation series of a single direction. LPQ reflect the number of times near-surface air needs to overcome the barrier of terrain to climb and is a global feature of terrain blockage. The minimum value of LPQ is zero.
Table 2. Comparison between TBFs and traditional terrain features.
Table 2. Comparison between TBFs and traditional terrain features.
Terrain Blockage Features (TBFs)Traditional Terrain Features
AverageStandard DeviationAverageStandard Deviation
R-squared0.88340.01610.84740.0215
RMSE1.64180.23501.87710.2654
MAE1.22570.15111.41400.1753
MD−0.0070.015−0.0020.005
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Gao, H.; Dong, Y.; Zhou, L.; Wang, X. Exploring the Influence of Terrain Blockage on Spatiotemporal Variations in Land Surface Temperature from the Perspective of Heat Energy Redistribution. ISPRS Int. J. Geo-Inf. 2024, 13, 200. https://doi.org/10.3390/ijgi13060200

AMA Style

Gao H, Dong Y, Zhou L, Wang X. Exploring the Influence of Terrain Blockage on Spatiotemporal Variations in Land Surface Temperature from the Perspective of Heat Energy Redistribution. ISPRS International Journal of Geo-Information. 2024; 13(6):200. https://doi.org/10.3390/ijgi13060200

Chicago/Turabian Style

Gao, Hong, Yong Dong, Liang Zhou, and Xi Wang. 2024. "Exploring the Influence of Terrain Blockage on Spatiotemporal Variations in Land Surface Temperature from the Perspective of Heat Energy Redistribution" ISPRS International Journal of Geo-Information 13, no. 6: 200. https://doi.org/10.3390/ijgi13060200

APA Style

Gao, H., Dong, Y., Zhou, L., & Wang, X. (2024). Exploring the Influence of Terrain Blockage on Spatiotemporal Variations in Land Surface Temperature from the Perspective of Heat Energy Redistribution. ISPRS International Journal of Geo-Information, 13(6), 200. https://doi.org/10.3390/ijgi13060200

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop