Urban heat environment regulation method and system coupled with spatial structure

Abstract

The application discloses a city heat environment regulation method and system coupled with LCZ and spatial structure, which comprises the following steps: generating a local climate zone LCZ classification map and inverting a land surface temperature LST based on multi-source remote sensing and geographic data; identifying a high-temperature dominant LCZ type through a high-temperature area contribution quantification method; constructing a heat pattern comprehensive index HPI by using an entropy weight method; embedding LCZ as a key driving variable into a city form-heat environment coupling simulation framework; predicting a future LCZ spatial pattern by using an ECOBRIDGE expert system combined with a transfer matrix and a neighborhood competition mechanism; simulating a future city heat environment distribution under different scenarios by combining a Markov chain and a PLUS model; identifying a heat environment regulation area under different scenarios, determining an LCZ optimization priority and outputting a spatial regulation atlas based on a prediction result. The method can output a future heat environment evolution prediction result under a multi-city scale, and provide an intervention scheme for city high-temperature area relief, heat pattern improvement and climate adaptability planning.

Description

Technical Field

[0001] This invention belongs to the field of urban climate regulation and spatial optimization technology, specifically involving a method and system for urban thermal environment regulation coupled with LCZ and spatial structure. Background Technology

[0002] With the acceleration of global climate change and rapid urbanization, the urban heat island effect has become a significant problem affecting the urban ecological environment and residents' health. Numerous studies have shown that the spatial distribution of surface temperature is not only influenced by regional climate conditions but also closely related to urban land use structure and urban morphology. The large-scale expansion of high-density built-up areas, industrial land, and impermeable surfaces significantly enhances the urban thermal environment, while ecological spaces such as green spaces and water bodies effectively mitigate the urban heat island effect through evapotranspiration and shading. Therefore, revealing the formation mechanism of the urban thermal environment from the perspective of urban spatial structure and urban morphology is of great significance for optimizing urban planning and improving urban climate adaptability.

[0003] To systematically describe urban morphological characteristics, Stewart and Oke proposed the LCZ classification system, providing a unified framework for urban morphology and thermal environment research. By combining LCZ classification with LST inversion, researchers can finely analyze the thermal response characteristics of different urban morphologies, generally finding that compact, high-density built-up areas correspond to high temperatures, while vegetated and water-based areas have a cooling effect. In recent years, scholars have further introduced landscape pattern analysis into heat island research, revealing the formation mechanism of the thermal environment from a spatial organization perspective, and promoting research to deepen from "temperature differences" to "spatial structure." Simultaneously, with the deepening of research, some studies have begun to explore the differences in the relationship between LCZ and LST from a cross-city perspective, finding that even under the same LCZ type, different cities still exhibit significant differences in their thermal environment response due to differences in development stage, construction intensity, and spatial expansion patterns. These studies emphasize the important role of the overall urban development pattern in modulating the thermal environment effect of LCZ, expanding the application of LCZ theory in multi-city scale and regional comparative studies. In addition, in the study of long-term urban thermal environment evolution, scenario simulation methods based on models such as Markov Chain, cellular automata (CA) and Patch-Generating Land Use Simulation (PLUS) are widely used to characterize the evolution trend of urban spatial patterns and to assess the potential response of the thermal environment under different development scenarios. Summary of the Invention

[0004] To address the aforementioned problems in the existing technology, the purpose of this invention is to provide a method and system for urban thermal environment regulation that couples LCZ with spatial structure.

[0005] This invention provides the following technical solution: a method for urban thermal environment regulation coupled with LCZ and spatial structure, comprising the following steps: S1. Acquire remote sensing images, socio-economic data, and physical geographic data of the study area, and perform radiometric calibration, atmospheric correction, and resampling preprocessing on the acquired data; S2. Based on the preprocessed remote sensing images, multidimensional features are extracted, and machine learning classification methods are used to generate local climate zone LCZ classification maps, and the accuracy of the classification results is verified. S3. Based on the thermal infrared band of remote sensing images, land surface temperature (LST) is retrieved, temperature levels are divided and high-temperature zones are defined, the area proportion of each LCZ type in the high-temperature zone is calculated, and the high-temperature dominant LCZ type is identified. S4. Select multiple landscape pattern indices, calculate the comprehensive weight after dimensionless processing, construct the comprehensive thermal pattern index (HPI), and classify the thermal pattern levels. S5. Based on multi-period historical LCZ classification data and LST inversion data, establish a quantitative response function between LCZ type and LST for urban thermal environment evolution prediction. S6. Construct a transition matrix based on historical LCZ changes, integrate multi-source driving factors, and use expert system and neighborhood competition mechanism to simulate the future LCZ spatial pattern. S7. Based on Markov chains, predict the area requirement of future thermal environment levels, and use the future LCZ pattern predicted by S6 as a spatial constraint. Through the random patch seed cellular automaton module, predict the spatial distribution of future thermal environment under different scenarios. S8. Based on the high-temperature dominant LCZ type, HPI level, and the future spatial distribution of the thermal environment predicted by S7, identify the thermal environment control zones under various urban development scenarios, determine the spatial optimization priority of different LCZ types, and output a spatial control map to guide urban renewal, ventilation corridor layout, and blue-green infrastructure optimization.

[0006] Furthermore, in S1: Remote sensing images include multispectral and thermal infrared bands; Socioeconomic data includes nighttime light data, GDP, and Points of Interest (POIs). Natural geographic data includes road networks, digital elevation models (DEM), slope, distance from water bodies, vegetation index (NDVI), building index (NDBI), and water index (NDWI). Among them, the radiometric calibration method uses the radiative transfer equation to convert the DN value into the radiance of the top layer of the atmosphere, and the atmospheric correction uses the MODTRAN model and resamples to a uniform spatial resolution.

[0007] Furthermore, S2 specifically includes the following steps: S2.1. A stratified sampling strategy is used to collect training samples to ensure the spatial representativeness of various LCZ types; S2.2. Integrate spectral features, texture features, and land cover index to construct a multi-dimensional feature space; S2.3. The LCZ classification map is generated using the random forest algorithm, and the overall accuracy and Kappa coefficient are calculated using the confusion matrix to verify the accuracy.

[0008] Furthermore, S3 specifically includes the following steps: S3.1. Based on the thermal infrared band, radiometric calibration, atmospheric correction and surface emissivity estimation are performed sequentially to obtain the LST. S3.2. The mean-standard deviation method is used to divide LST into multiple levels, and the two highest levels are combined to define the high temperature zone. S3.3 Calculate the area percentage (PLAND) of each LCZ type in the high-temperature zone, sort them in descending order of percentage, and output a sorted list of LCZ types dominated by high temperature.

[0009] Furthermore, S4 specifically includes the following steps: S4.1 Select multiple landscape pattern indices, use range standardization to perform dimensionless processing on each index, calculate the weight of each index based on the entropy weight method, and obtain the comprehensive thermal pattern index HPI by weighted summation; S4.2. Using the natural breakpoint method, HPI is divided into several thermal pattern levels, and the spatial distribution map of urban thermal pattern intensity is output.

[0010] Furthermore, the specific process of S5 is as follows: S5.1 Based on multi-period LCZ classification data and LST inversion data from historical time phases, calculate the average LST and its standard deviation for each LCZ type; S5.2. Use cubic polynomial fitting to establish the LCZ–LST response function, and output the quantitative response curve from LCZ type to LST and the fitting parameters.

[0011] Furthermore, S6 specifically includes the following steps: S6.1. Based on the at least two periods of LCZ classification data obtained in the historical time phase of the study area, perform overlay analysis to construct the LCZ type transition matrix and calculate the transition probability; S6.2 Integrate multiple sources of natural geographical and socio-economic driving factors to construct a set of driving factors for LCZ evolution; S6.3. Establish an expert rule base based on the driving factor set, and simulate the spatial expansion and transformation trend of LCZ type by combining the neighborhood competition mechanism; S6.4. The Monte Carlo simulation method is used for multiple iterations, and the mode of the multiple simulation results is used as the prediction result of the future LCZ spatial pattern.

[0012] Furthermore, S7 specifically includes the following steps: S7.1 Construct a Markov transition matrix based on historical thermal environment level transition data to predict the area demand for each thermal environment level in the future; S7.2 Set up multiple urban development scenarios and adjust the transition probabilities between different thermal environment levels to obtain the thermal environment evolution trend under each scenario; S7.3. Using the land expansion analysis strategy module in the PLUS model, based on the expansion patches of historical thermal environment level and various driving factors, the random forest algorithm is used to generate the development probability map of each thermal environment level. S7.4. Using a cellular automaton module based on random patch seeds, and taking the future LCZ spatial pattern predicted in step S6 as a spatial constraint, the area requirement is allocated to spatial units, and the predicted distribution map of the future thermal environment under different scenarios is output.

[0013] Furthermore, in S8, the specific rules for identifying the thermal environment control zone and determining the spatial optimization priority are as follows: 1) Spatial units that simultaneously meet the following criteria are classified as first-level control zones: belonging to the high-temperature-dominated LCZ type, having a high HPI level, and being predicted to be in the high-temperature region. Corresponding output recommendations: Reduce building density in the first-level control zone, add vertical greening, or convert to permeable paving; 2) Spatial units that meet the following conditions are classified as secondary control zones: belonging to the high-temperature dominant LCZ type or having a high HPI level, but not simultaneously meeting the three conditions of high-temperature dominant LCZ type, high HPI level, and predicted as a high-temperature zone. Corresponding output recommendation: Within the secondary control zone, reserve ventilation corridors or control the height of new buildings; 3) Areas with existing water bodies and high vegetation cover of the LCZ type that are predicted to remain at low temperatures are designated as cold source protection zones; Corresponding output recommendation: Maintain the existing ecological functions of the cold source protection area and prohibit development activities.

[0014] The urban thermal environment control system coupled with the spatial structure of the LCZ includes: The data acquisition module is used to acquire remote sensing images, socio-economic data and natural geographic data of the study area, and to perform radiometric calibration, atmospheric correction and resampling preprocessing. The LCZ classification module is used to generate LCZ classification maps based on preprocessed remote sensing images; The LST inversion module is used to invert surface temperature LST and identify high-temperature dominant LCZ types. The HPI construction module is used to construct the comprehensive thermal pattern index (HPI) and classify its levels. The LCZ prediction module is used to simulate the future LCZ spatial pattern based on historical LCZ changes and driving factors. The thermal environment prediction module is used to predict the future spatial distribution of the thermal environment based on Markov chains and PLUS models. The regulation map generation module is used to identify thermal environment regulation zones and output spatial regulation maps based on the high-temperature dominant LCZ type, HPI level, and future thermal environment prediction results.

[0015] By employing the above-described technology, the beneficial effects of the present invention compared to the prior art are as follows: 1) By statistically analyzing the area proportion of each LCZ type under different city levels and its corresponding LST distribution, this invention can output the predicted value of the differential response of the same LCZ type in different cities from the perspective of multi-city level comparison. The prediction results can be used to identify urban thermal environment sensitive patterns and provide quantitative basis for cross-city climate adaptation planning.

[0016] 2) This invention differs from the traditional approach that only focuses on average temperature differences. It divides LST levels and defines high-temperature zones using the natural breakpoint method, calculates the area proportion (PLAND) of each LCZ type in the high-temperature zone, and can output the dominant LCZ type ranking in the high-temperature zone. The prediction results can be used to locate urban morphological areas that need priority for heat mitigation.

[0017] 3) This invention selects multiple landscape pattern indices, and after dimensionless processing, uses the entropy weight method to assign weights to construct a comprehensive thermal pattern index (HPI), which solves the problem of insufficient representation power of a single index. This index can output a comprehensive score of the scale, fragmentation and spatial aggregation of high-temperature patches. The prediction results can be used to assess the potential of urban spatial pattern to regulate the thermal environment and guide the optimization of ventilation corridors and green space layout.

[0018] 4) This invention embeds LCZ as a key spatial variable into the thermal environment evolution prediction process. It combines the LCZ spatial pattern prediction model with the thermal environment simulation model based on Markov chains and cellular automata to predict the future thermal environment distribution under different urban development scenarios. This can help compare urban planning schemes and assess the thermal environment risks under different development paths, providing decision support for climate-adaptive urban design.

[0019] 5) This application generates a spatial control map that can be directly applied to urban planning and management by constructing a multi-source spatial dataset that includes high-temperature dominant LCZ types, thermal pattern comprehensive index (HPI), and future thermal environment prediction. Based on the set rules, the map associates the thermal environment assessment results with specific spatial units, outputs differentiated engineering transformation instructions for different plots, and transforms the thermal environment analysis results into a spatial control basis that can directly guide planning and management.

[0020] 6) This application improves the system performance and accuracy of thermal environment prediction by embedding the LCZ type as a key driving variable into the coupled simulation framework. At the same time, based on the first-level control zone identified by the spatial control map, the local surface temperature can be reduced by implementing engineering measures such as reducing building density, adding vertical greening and permeable paving, and utilizing physical mechanisms such as increasing surface albedo, enhancing evapotranspiration and reducing heat storage. Attached Figure Description

[0021] Figure 1 This is a flowchart of the method in this invention; Figure 2 This is a schematic diagram of a typical example of LCZ classification in this invention; Figure 3 This is a comparison chart of the proportions of LCZ classification results for cities A, B, C, and D in 2025 in this invention. Figure 4 This is a heatmap of the LCZ and LST levels in 2025 in this invention; taking city A as an example, it shows the LST percentage of each LCZ type in 2025. Figure 5 This is a schematic diagram showing the ranking of the contribution of the LCZ high-temperature zone in the study area in 2025 in this invention; taking city A as an example, it shows the ranking of the area proportion of each LCZ type in the high-temperature zone in 2025. Figure 6 This is a schematic diagram showing the distribution ratio of each LCZ type in 2025 in this invention, representing the area proportion of each LCZ in the four cities in 2025; Figure 7 This is a schematic diagram showing the ranking of high-temperature-dominated LCZ contributions in 2030 according to the present invention; taking city A as an example, it represents the high-temperature contribution results of each LCZ type in 2030. Figure 8 This is a box plot showing the statistical distribution of LST-LCZ in 2025 in this invention; taking city A as an example, it shows the box plot of land surface temperature (LST) distribution for each LCZ type; Figure 9 The LCZ-LST response curve for 2025 in this invention; Figure 10 This is a statistical chart showing the HPI distribution of four cities in 2025, as described in this invention. Figure 10(a) in the figure is the HPI box plot for city A. Figure 10 (b) in the figure is the HPI box plot for city B. Figure 10 (c) in the figure represents the HPI box plot for city C. Figure 10 (d) in the figure represents the HPI box plot for city D; Figure 11 This is a statistical chart showing the predicted LCZ area for the four cities in 2030, representing the area proportion of each LCZ in 2030. Figure 12 This is a statistical chart of the LST prediction results under multiple scenarios in the future 2030 of this invention. Taking city A as an example, it shows the LST classification percentage under the three future scenarios in 2030. Detailed Implementation

[0022] The present invention will be further described in detail below with reference to specific embodiments. The specific parameters, data sources, and years in this embodiment are merely preferred embodiments and are used to explain the present invention, but are not intended to limit the scope of protection of the present invention.

[0023] Reference Figure 1 The flowchart of the urban thermal environment control method coupled with LCZ and spatial structure is shown. In this embodiment, the above steps S1 to S7 are executed sequentially, and the method is illustrated by taking cities A, B, C and D as the study areas.

[0024] S1. Data Collection and Preprocessing: The data mainly includes three categories: remote sensing imagery, socioeconomic data, and physical geographic data. The remote sensing imagery used was Landsat 8 / 9 data from April 2015, 2020, and 2025, taken on clear days with cloud cover <10%. After radiometric calibration and atmospheric correction, the data was uniformly resampled to a 30m resolution. The 2015 and 2020 data were used for model training, while the 2025 data represents the current stage.

[0025] As an optional implementation, the 2018 LCZ baseline data was derived from the Global Local Climate Zone Classification (LCZ) product published by Demuzere et al. (e.g., Figure 2 The diagram shown illustrates typical examples of LCZ classification in this embodiment, illustrating typical land cover and morphological characteristics of different LCZ types. This dataset is based on Stewart and Oke's (2012) LCZ classification framework and was generated using the WUDAPT method. Its spatial resolution is 100m, but it can be resampled to 30m during implementation to maintain consistency with other datasets. Other data information is shown in Table 1. Table 1 contains detailed information on the data used.

[0026] S2, LCZ Classification: In step S2, based on the 2025 LCZ classification results, the area proportion of different LCZ types in the four cities is statistically analyzed (e.g., Figure 3 The figure shown is a comparison chart of the proportion of LCZ classification results for cities A, B, C, and D in 2025 in this embodiment, demonstrating the spatial differences in the composition of LCZ types among different cities.

[0027] The LCZ area proportions output by this method show that: City C is mainly a mixture of natural underlying surface and industrial land; City A exhibits a high-intensity urban structure, with a high proportion of low-lying vegetation, water bodies, and industrial land, while the proportion of high-coverage vegetation is relatively low; City D is dominated by natural landscapes, with the highest proportion of dense trees; and City B has a combination of a high proportion of vegetation and a certain scale of medium-density built-up area. The LCZ spatial composition gradient output in this step can be used as a morphological input parameter for subsequent thermal environment prediction.

[0028] S3. Based on the thermal infrared band of remote sensing imagery, the Land Surface Temperature (LST) is retrieved using the atmospheric radiative transfer equation method: First, radiometric calibration is performed to convert the DN value into the radiance of the upper atmosphere. Then, atmospheric correction is performed using the MODTRAN model, and the surface emissivity is estimated using the NDVI threshold method. Subsequently, the LST is divided into five levels using the mean-standard deviation method, as shown in Table 2. Table 2 LST Classification Criteria

[0029] in, μ = average value , σ = Standard deviation 。

[0030] Ultimately, the two highest levels were merged to define a high-temperature zone (levels 4 and 5 were merged to define a high-temperature zone). Further analysis was conducted to determine the area proportion of each LCZ type within the high-temperature zone, in order to identify the dominant morphological type of high-temperature zone formation in different cities. The calculation formula is as follows: ; in, This represents the area of ​​type LCZ i that falls within the high-temperature region. This represents the total area of ​​the urban high-temperature zone. This indicator is used to quantify the direct contribution of different LCZ types to the spatial structure of high temperatures.

[0031] Applying step S3 of this method, the area proportion (PLAND) of each LCZ type within the high-temperature zone is calculated, yielding a ranking of high-temperature zone contributions. Built-up LCZs have a higher high-temperature zone contribution than natural LCZs. The dominant high-temperature LCZ types differ across cities; for example, the top four PLAND rankings for city A are LCZ8, LCZ2, LCZ10, and LCZ3; for city B, they are LCZ2, LCZ8, LCZ1, and LCZ3; for city C, they are LCZ3, LCZ2, LCZ7, and LCZ5; and for city D, they are LCZ3, LCZ2, LCZ8, and LCZ10. This method can quantitatively identify the key thermal morphologies in different cities based on their high-temperature zone contributions.

[0032] like Figure 4 As shown, taking city A as an example, this represents the LST percentage results for each LCZ type in 2025; like Figure 5 As shown, the PLAND ranking of city A is used as an example for illustration.

[0033] S4. Spatial Distribution of HPI: Five landscape pattern indices—Patch Density (PD), Aggregation Index (AI), Maximum Patch Index (LPI), Spread Index (CONTAG), and Shannon Diversity Index (SHDI)—were selected to construct the HPI. The natural breakpoint method was used to classify HPI values ​​into high, medium, and low levels, and spatial representation was performed on a grid scale (100m × 100m) to characterize the spatial differentiation features of thermal pattern intensity (e.g., ...). Figure 10 As shown in the table). The natural breakpoint method, based on the statistical distribution of HPI values, automatically calculates two optimal breakpoint values ​​(denoted as the first breakpoint value and the second breakpoint value, where the first breakpoint value is less than the second breakpoint value) by maximizing the between-group variance and minimizing the within-group variance. This divides the HPI into three levels: high (greater than the second breakpoint value), medium (between the first and second breakpoint values), and low (less than the first breakpoint value). The breakpoint values ​​are determined by the input data and do not require a preset fixed threshold. The weight calculation results are shown in Table 3. Patch density (PD) has the highest weight (0.3540). In the HPI constructed by this method, the degree of fragmentation of high-temperature patches contributes the most to the thermal pattern. Table 3 Weights of Each Index

[0034] The spatial distribution of HPI (grid scale 100m×100m) output by this method reveals different thermal pattern structures in different cities: High-value HPI areas in city A exhibit a multi-center distribution, are numerous, and have a high degree of spatial segmentation; high-value HPI areas in city B show a continuous, plate-like distribution along the main urban area and development axis; high-value HPI areas in cities C and D show a discrete, patchy distribution, with city C exhibiting an embedded distribution due to topographic constraints. The HPI constructed by this method can characterize the spatial differentiation features of thermal pattern intensity.

[0035] like Figure 6 The diagram shown is a schematic representation of the distribution ratio of each LCZ type in this embodiment in 2025, representing the area ratio of each LCZ in the four cities in 2025.

[0036] S5. Using LCZ as the key driving variable, the average LST of each LCZ type from 2015 to 2025 was statistically analyzed, and the LCZ-LST response relationship was established by fitting a cubic polynomial.

[0037] like Figure 7 The diagram shown is a ranking of the contribution of high temperature-dominant LCZ in 2030 in this embodiment; taking city A as an example, it shows the high temperature contribution results of each LCZ type in 2030.

[0038] Figure 9 The LCZ-LST response curves output by this method are shown, which can be used to predict the thermal response range of different LCZ types in a given urban context. The urban morphology-thermal environment coupled simulation framework constructed by this method realizes the collaborative prediction of LCZ and LST in a unified spatiotemporal dimension.

[0039] S6. Predict the spatial pattern of the LCZ in 2030 using the ECOBRIDGE model. First, based on the overlay analysis of LCZ data from 2018 and 2025, a transition matrix is ​​constructed, and the transition probabilities for each type are calculated. Then, 12 driving factors (including NDVI, NDBI, NDWI, distance to water body, distance to government, distance to tourist attractions, DEM, slope, GDP, nighttime light, distance to highway, and distance to railway) are integrated and uniformly resampled to a 100m resolution. An expert rule base is established, for example: "IF (distance to road ≤ 500m AND nighttime light value > 30 AND slope ≤ 5°) THEN (LCZ A - LCZ 6) WITH certainty = 0.72". The Monte Carlo method is used for 100 iterations to output the 2030 LCZ prediction results (e.g., ...). Figure 11 The figure shown is a statistical chart of the LCZ prediction results for 2030 in this embodiment, representing the area proportion of each LCZ in the four cities in 2030. Backtesting of the LCZ in 2025 using this method shows an overall accuracy of 81.2% and a Kappa coefficient of 0.78, meeting the prediction requirements. This prediction result can be used as a spatial constraint input for subsequent thermal environment simulations.

[0040] S7. Predicting LST distribution under multiple scenarios in 2030 using the Markov-PLUS model. First, based on thermal environment level transition data from 2020 to 2025, a Markov transition matrix is ​​constructed to predict the area demand for each level in 2030. Then, three scenarios are set: Natural development (ND) scenario: Continuing the probability of historical transition; High-intensity development (HD) scenario: The probability of a shift to high-temperature regions increases by 20%; Low-intensity development (LD) scenario: Increases the probability of transition to the low-temperature region by 20%, while decreasing the probability of transition to the high-temperature region by 20%.

[0041] The LEAS module of the PLUS model extracts expanding patches for each thermal environment level from 2015 to 2025. Combining 12 driving factors, a random forest algorithm (in this embodiment, 500 decision trees, feature subset size is the square root of the total number of features, and minimum leaf node sample size is 5) is used to generate a development probability map. Then, through the CARS module, using the 2030 LCZ spatial pattern predicted by S6 as a spatial constraint, area requirements are allocated to spatial units, outputting 2030 LST prediction maps under different scenarios (e.g., ...). Figure 12 The figure shown is a statistical chart of the LST prediction results under multiple scenarios in this embodiment for 2030. Taking city A as an example, it represents the percentage of LST levels under the three future scenarios in 2030. Model validation uses the historical backtesting method: 2015 data is used to simulate the 2020 pattern. The Kappa coefficient is 0.82 and the FOM is 0.14, which is within the acceptable range of similar studies. This method can output the predicted results of thermal environment evolution under different urban development scenarios.

[0042] like Figure 8 The figure shows the LST-LCZ box plot for a city in 2025; taking city A as an example, it shows the LST distribution of each LCZ type.

[0043] like Figure 10 The figure shown is a statistical chart of the HPI distribution in four cities in 2025 in this embodiment; Figure 10 (a) in the figure is the HPI box plot for city A. Figure 10 (b) in the figure is the HPI box plot for city B. Figure 10 (c) in the figure represents the HPI box plot for city C. Figure 10 (d) in the figure represents the HPI box plot for city D.

[0044] This method is achieved through... Figure 8 LST-LCZ box plot (taking the LST distribution box plot of land surface temperature for each LCZ type in city A as an example) and Figure 9 The LCZ-LST response curves quantify the temperature response differences of the same LCZ type in different cities.

[0045] S8, Output Space Control Graph Based on the high-temperature dominant LCZ types identified in step S3 (taking city A as an example, the high-temperature dominant LCZ types are ranked as follows: LCZ8, LCZ2, LCZ10, LCZ3), the HPI level distribution map constructed in S4 (the high HPI value area of ​​city A is multi-center distributed), and the multi-scenario LST distribution map predicted in step S7 for 2030 (the high-temperature area accounts for 30.01% under the natural development scenario), this embodiment further identifies the thermal environment regulation zones of city A under different development scenarios.

[0046] The specific identification rules are as follows: Level 1 Control Zone: Spatial units that simultaneously meet the following criteria are classified as Level 1 Control Zones: belonging to the top three high-temperature dominant LCZ types, having a high HPI level, and being predicted to be a high-temperature zone in 2030 (LST level 4 or 5). Corresponding engineering measures: It is recommended to prioritize the implementation of building density reduction (e.g., reducing the existing building density by more than 20%), adding vertical greening or roof greening (green coverage rate ≥ 30%), and converting to permeable paving (permeability rate ≥ 70%).

[0047] Level 2 control zone: Spatial units that meet the following conditions are classified as Level 2 control zones: those that are dominated by high temperature LCZ type or have a high HPI level, but do not meet all three conditions at the same time. Corresponding engineering measures: It is recommended to reserve ventilation corridors (width ≥ 50m, direction parallel to the prevailing summer wind direction), control the height of new buildings to not exceed 24 meters (equivalent to 8 floors or less), and increase multi-layered greening of trees, shrubs and grasses.

[0048] Cold source protection area: The existing water body (LCZ G) and the high vegetation cover LCZ type (LCZ A, LCZ B) that are predicted to remain at low temperature (LST level 1 or 2, i.e. lower than the regional average LST minus 0.5 times the standard deviation) are identified as cold source protection areas. Corresponding engineering measures: strict protection, prohibition of any development and construction activities, and maintenance of existing ecological functions.

[0049] Based on the above rules, spatial regulation maps for each city under natural development scenarios are generated. These maps can be directly used to guide the prioritization of land redevelopment in urban renewal, the selection of linear spaces for ventilation corridors, and the optimization of blue-green infrastructure layout.

[0050] This embodiment also provides an urban thermal environment control system coupled with LCZ and spatial structure. This system is deployed in a computer device or cloud server to execute the steps described in the above method embodiments. The system mainly includes the following modules: The data acquisition module is used to acquire remote sensing images, socio-economic data, and natural geographic data of the study area, and to perform unified preprocessing on the above data to provide a standardized data foundation for subsequent analysis.

[0051] The LCZ classification module is used to generate local climate zone LCZ classification maps based on preprocessed remote sensing data, verify the accuracy of the classification results, and output the spatial distribution of different LCZ types in the study area.

[0052] The LST inversion module is used to invert land surface temperature based on remote sensing images, identify high-temperature zones, statistically analyze the contribution of each LCZ type to the high-temperature zones, and output the ranking results of the high-temperature dominant LCZ types.

[0053] The HPI construction module is used to construct a comprehensive thermal pattern index (HPI) based on the landscape pattern index, classify thermal pattern levels, and output the spatial distribution characteristics of urban thermal environment intensity.

[0054] The LCZ prediction module is used to simulate and output the spatial pattern of LCZ at different future stages based on historical LCZ change patterns and multi-source driving factors, combined with expert rules and neighborhood competition mechanisms.

[0055] The thermal environment prediction module is used to predict the spatial distribution of the future thermal environment under different urban development scenarios, constrained by the future LCZ spatial pattern and combining Markov chain and cellular automata models.

[0056] The regulation map generation module is used to integrate the high-temperature dominant LCZ type, HPI level and future thermal environment prediction results to identify key areas and optimization priorities for thermal environment regulation, and output a spatial regulation map that can directly guide urban renewal, ventilation corridor layout and blue-green infrastructure construction.

[0057] Each module transmits data through a standardized spatial data interface. The system outputs spatial control maps, converting thermal environment prediction and assessment results into instructions for the transformation of urban physical space, which are used to guide urban planning and management.

[0058] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for urban thermal environment control coupled with LCZ and spatial structure, characterized in that, Includes the following steps: S1. Acquire remote sensing images, socio-economic data, and physical geographic data of the study area, and perform radiometric calibration, atmospheric correction, and resampling preprocessing on the acquired data; S2. Based on the preprocessed remote sensing images, multidimensional features are extracted, and machine learning classification methods are used to generate local climate zone LCZ classification maps, and the accuracy of the classification results is verified. S3. Based on the thermal infrared band of remote sensing images, land surface temperature (LST) is retrieved, temperature levels are divided and high-temperature zones are defined, the area proportion of each LCZ type in the high-temperature zone is calculated, and the high-temperature dominant LCZ type is identified. S4. Select multiple landscape pattern indices, calculate the comprehensive weight after dimensionless processing, construct the comprehensive thermal pattern index (HPI), and classify the thermal pattern levels. S5. Based on multi-period historical LCZ classification data and LST inversion data, establish a quantitative response function between LCZ type and LST for urban thermal environment evolution prediction. S6. Construct a transition matrix based on historical LCZ changes, integrate multi-source driving factors, and use expert system and neighborhood competition mechanism to simulate the future LCZ spatial pattern. S7. Based on Markov chains, predict the area requirement of future thermal environment levels, and use the future LCZ pattern predicted by S6 as a spatial constraint. Through the random patch seed cellular automaton module, predict the spatial distribution of future thermal environment under different scenarios. S8. Based on the high-temperature dominant LCZ type, HPI level, and the future spatial distribution of the thermal environment predicted by S7, identify the thermal environment control zones under various urban development scenarios, determine the spatial optimization priority of different LCZ types, and output a spatial control map to guide urban renewal, ventilation corridor layout, and blue-green infrastructure optimization.

2. The urban thermal environment control method coupled with LCZ and spatial structure according to claim 1, characterized in that, In S1: Remote sensing images include multispectral and thermal infrared bands; Socioeconomic data includes nighttime light data, GDP, and Points of Interest (POIs). Natural geographic data includes road networks, digital elevation models (DEM), slope, distance from water bodies, vegetation index (NDVI), building index (NDBI), and water index (NDWI). Among them, the radiometric calibration method uses the radiative transfer equation to convert the DN value into the radiance of the top layer of the atmosphere, and the atmospheric correction uses the MODTRAN model and resamples to a uniform spatial resolution.

3. The urban thermal environment control method coupled with LCZ and spatial structure according to claim 1, characterized in that, S2 specifically includes the following steps: S2.

1. A stratified sampling strategy is used to collect training samples to ensure the spatial representativeness of various LCZ types; S2.

2. Integrate spectral features, texture features, and land cover index to construct a multi-dimensional feature space; S2.

3. The LCZ classification map is generated using the random forest algorithm, and the overall accuracy and Kappa coefficient are calculated using the confusion matrix to verify the accuracy.

4. The urban thermal environment control method coupled with LCZ and spatial structure according to claim 1, characterized in that, S3 specifically includes the following steps: S3.

1. Based on the thermal infrared band, radiometric calibration, atmospheric correction and surface emissivity estimation are performed sequentially to obtain the LST. S3.

2. The mean-standard deviation method is used to divide LST into multiple levels, and the two highest levels are combined to define the high temperature zone. S3.3 Calculate the area percentage (PLAND) of each LCZ type in the high-temperature zone, sort them in descending order of percentage, and output a sorted list of LCZ types dominated by high temperature.

5. The urban thermal environment control method coupled with LCZ and spatial structure according to claim 1, characterized in that, S4 specifically includes the following steps: S4.1 Select multiple landscape pattern indices, use range standardization to perform dimensionless processing on each index, calculate the weight of each index based on the entropy weight method, and obtain the comprehensive thermal pattern index HPI by weighted summation; S4.

2. Using the natural breakpoint method, HPI is divided into several thermal pattern levels, and the spatial distribution map of urban thermal pattern intensity is output.

6. The urban thermal environment control method coupled with LCZ and spatial structure according to claim 1, characterized in that, The specific process of S5 is as follows: S5.1 Based on multi-period LCZ classification data and LST inversion data from historical time phases, calculate the average LST and its standard deviation for each LCZ type; S5.

2. Use cubic polynomial fitting to establish the LCZ–LST response function, and output the quantitative response curve from LCZ type to LST and the fitting parameters.

7. The urban thermal environment control method coupled with LCZ and spatial structure according to claim 1, characterized in that, S6 specifically includes the following steps: S6.

1. Based on the at least two periods of LCZ classification data obtained in the historical time phase of the study area, perform overlay analysis to construct the LCZ type transition matrix and calculate the transition probability; S6.2 Integrate multiple sources of natural geographical and socio-economic driving factors to construct a set of driving factors for LCZ evolution; S6.

3. Establish an expert rule base based on the driving factor set, and simulate the spatial expansion and transformation trend of LCZ type by combining the neighborhood competition mechanism; S6.

4. The Monte Carlo simulation method is used for multiple iterations, and the mode of the multiple simulation results is used as the prediction result of the future LCZ spatial pattern.

8. The urban thermal environment control method coupled with LCZ and spatial structure according to claim 1, characterized in that, S7 specifically includes the following steps: S7.1 Construct a Markov transition matrix based on historical thermal environment level transition data to predict the area demand for each thermal environment level in the future; S7.2 Set up multiple urban development scenarios and adjust the transition probabilities between different thermal environment levels to obtain the thermal environment evolution trend under each scenario; S7.

3. Using the land expansion analysis strategy module in the PLUS model, based on the expansion patches of historical thermal environment level and various driving factors, the random forest algorithm is used to generate the development probability map of each thermal environment level. S7.

4. Using a cellular automaton module based on random patch seeds, and taking the future LCZ spatial pattern predicted in step S6 as a spatial constraint, the area requirement is allocated to spatial units, and the predicted distribution map of the future thermal environment under different scenarios is output.

9. The urban thermal environment control method coupled with LCZ and spatial structure according to claim 1, characterized in that, In step S8, the specific rules for identifying the thermal environment control zone and determining the spatial optimization priority are as follows: 1) Spatial units that simultaneously meet the following criteria are classified as first-level control zones: belonging to the high-temperature-dominated LCZ type, having a high HPI level, and being predicted to be in the high-temperature region. Corresponding output recommendations: Reduce building density in the first-level control zone, add vertical greening, or convert to permeable paving; 2) Spatial units that meet the following conditions are classified as secondary control zones: belonging to the high-temperature dominant LCZ type or having a high HPI level, but not simultaneously meeting the three conditions of high-temperature dominant LCZ type, high HPI level, and predicted as a high-temperature zone. Corresponding output recommendation: Within the secondary control zone, reserve ventilation corridors or control the height of new buildings; 3) Areas with existing water bodies and high vegetation cover of the LCZ type that are predicted to remain at low temperatures are designated as cold source protection zones; Corresponding output recommendation: Maintain the existing ecological functions of the cold source protection area and prohibit development activities.

10. An urban thermal environment control system coupled with spatial structure by LCZ, used to implement the method of any one of claims 1 to 9; characterized in that, include: The data acquisition module is used to acquire remote sensing images, socio-economic data and natural geographic data of the study area, and to perform radiometric calibration, atmospheric correction and resampling preprocessing. The LCZ classification module is used to generate LCZ classification maps based on preprocessed remote sensing images; The LST inversion module is used to invert surface temperature LST and identify high-temperature dominant LCZ types. The HPI construction module is used to construct the comprehensive thermal pattern index (HPI) and classify its levels. The LCZ prediction module is used to simulate the future LCZ spatial pattern based on historical LCZ changes and driving factors. The thermal environment prediction module is used to predict the future spatial distribution of the thermal environment based on Markov chains and PLUS models. The regulation map generation module is used to identify thermal environment regulation zones and output spatial regulation maps based on the high-temperature dominant LCZ type, HPI level, and future thermal environment prediction results.

Images