Surface urban heat island intensity calculation method for mountainous city complex terrain

Abstract

The invention discloses an earth surface city heat island intensity calculation method for complex terrains of mountainous cities, which comprises the following steps of: 1, acquiring data and preprocessing the data; step 2, acquiring urban area pixels; 3, taking the urban pixel as a center, and extracting to obtain a basic suburban background; 4, screening the pixels from the basic suburb background, wherein the distance difference between the pixels and the urban pixel elevation average value is within a threshold range, and taking the pixels as alternative suburb backgrounds; 5, screening the pixels with the night city light data pixels within the threshold range from the alternative suburban background as alternative suburban pixels; step 6, counting standard deviations of all pixels in the normalized vegetation index data set according to the surface temperature data, and taking the pixels of which the standard deviations are within a threshold range as suburb backgrounds; step 7, according to the suburb background, calculating by adopting an urban-rural bisection method to obtain the surface urban heat island intensity; the method provided by the invention has higher accuracy and good universality.

Description

Technical Field

[0001] The present invention relates to the field of environmental technology, and in particular to a method for calculating the surface urban heat island intensity for complex terrain in mountainous cities. Background Art

[0002] In recent years, urban climate has been recognized as one of the key factors that influence the urban ecological environment, and the most obvious characteristic of urban climate is the surface urban heat island effect. SUHI refers to urban overheating caused by factors such as large amounts of artificial heating, high heat storage bodies such as buildings and roads, and reduced green space. This phenomenon occurs when the surface temperature in the central urban area is higher than that in the surrounding suburbs. This exacerbates urban energy consumption and the frequency of high-temperature disasters, adversely affecting the health of urban residents and the sustainable development of the city. Therefore, timely and accurate understanding of the spatiotemporal distribution characteristics of the urban heat island effect is of great significance for alleviating the urban heat island effect and addressing related issues such as environmental pollution, energy consumption, and public health.

[0003] Remote sensing technology, with its macroscopic, rapid, and economical characteristics, can compensate for the limited spatial distribution of meteorological stations and is the primary method for regional or global-scale SUHI monitoring. However, the primary challenge in estimating SUHI using remote sensing data is how to define the suburban background. Currently, two common methods for delineating the suburban background are the buffer zone method and the scale method. The buffer zone method uses a buffer zone of 1 to 50 km outside the urban built-up area as the suburban background. The scale method, after determining the urban built-up area, uses an area 1.5 or 2 times the built-up area outside the urban area as the suburban background. However, Li et al. (2019) argue that simply using a fixed buffer zone or fixed area outside the city as the suburban background can lead to errors in monitoring urban heat island intensity.

[0004] Previous studies of surface urban heat islands (SUHI) have focused on tropical coastal cities, desert cities, and plain cities, with a lack of research on SUHI in mountainous cities. With the rapid development of urbanization, mountainous cities have also been impacted, with damage to their natural and human living environments. Therefore, monitoring SUHI in mountainous cities is of great significance. However, unlike plain cities, the complex topography of mountainous cities results in significant elevation differences between urban and suburban areas. The temperature difference between urban and suburban areas increases with increasing elevation differences. Simply using a fixed area outside the city as a suburban background will increase the urban-suburban temperature difference, resulting in inaccurate assessments of the mountainous urban heat island effect. Therefore, simply using buffer or scale methods to estimate the mountainous urban heat island effect is unreasonable. Furthermore, mountainous urban land cover types are primarily composed of mosaics of vegetation and crops and woodlands. Crop planting areas are scattered, with no large, contiguous planting areas. Furthermore, woodlands are often concentrated in areas with significant elevation differences from urban areas. Therefore, at medium-resolution scales (spatial resolution 1000 m), using only farmland or woodland as a suburban background to estimate the surface urban heat island in mountainous areas is inaccurate. Summary of the Invention

[0005] In view of the problems existing in the prior art, the present invention provides a method for calculating the surface urban heat island intensity in mountainous cities with complex terrain.

[0006] The technical solution adopted by the present invention is: a method for calculating the surface urban heat island intensity for complex terrain in mountainous cities, comprising the following steps:

[0007] Step 1: Obtain data and preprocess it;

[0008] Step 2: Extract the pixels marked as cities in the land use / cover change image to obtain urban pixels;

[0009] Step 3: Taking the urban pixel obtained in step 2 as the center, the buffer method is used to extract the basic suburban background;

[0010] Step 4: Calculate the average elevation of urban pixels, and select pixels whose elevation difference with the average elevation of urban pixels is within the threshold range from the basic suburban background obtained in step 3 as candidate suburban backgrounds;

[0011] Step 5: Filter the pixels of the nighttime city light data within the threshold range from the candidate suburban background obtained in step 4 as candidate suburban pixels;

[0012] Step 6: Calculate the standard deviation of all pixels in the normalized vegetation index dataset based on the surface temperature data, and obtain the pixels whose standard deviation is within the threshold range as the suburban background;

[0013] Step 7: Based on the suburban background obtained in step 6, the urban-rural dichotomy method is used to calculate the surface urban heat island intensity.

[0014] Furthermore, the data in step 1 includes multi-source remote sensing data, administrative boundary vector data and socioeconomic factor data.

[0015] Furthermore, the multi-source remote sensing data includes: land use / cover change image data, surface temperature data, night city light data, vegetation cover data, and elevation data.

[0016] Furthermore, the socioeconomic factor data include population size, GDP, energy consumption, built-up area and number of civilian vehicles.

[0017] Furthermore, the vegetation cover index is a normalized index, and the calculation process is as follows:

[0018] NDVI=(ρ NIR -ρ R ) / (ρ NIR +ρ R )

[0019] Where: NDVI is the normalized vegetation index, ρ NIR is the reflectivity in the near-infrared band, ρ R is the reflectivity in the red light band.

[0020] Furthermore, the preprocessing in step 1 includes: data reading, mosaicking, resampling, cropping, and unified storage format.

[0021] Furthermore, the calculation method of the surface urban heat island intensity in step 7 is as follows:

[0022]

[0023] Where: SUHI i is the surface urban heat island intensity corresponding to the i-th pixel on the image, T i is the surface temperature value, n is the number of suburban background clear sky pixels, T crop is the average background surface temperature in the suburbs.

[0024] An evaluation method for calculating the surface urban heat island intensity in mountainous cities with complex terrain is proposed. The applicability of the calculation method to the urban heat island effect in mountainous areas is evaluated using the size of the correlation coefficient.

[0025] Furthermore, the correlation coefficient R is calculated as follows:

[0026]

[0027] Where: X i is the surface urban heat island area, n is the number of pixels, i is the pixel number, is the average surface urban heat island area, Y i is the socioeconomic factor value, is the average of all socioeconomic factors.

[0028] The beneficial effects of the present invention are:

[0029] (1) The present invention solves the problem of difficult suburban background division in current surface urban remote sensing monitoring and is applicable to the calculation of thermal conductivity intensity in mountainous cities with complex terrain;

[0030] (2) The present invention uses correlation coefficient to evaluate the applicability of the calculation method;

[0031] (3) The method of the present invention can more accurately extract the suburban background, so the calculation of thermal conductivity intensity has higher accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 Schematic diagram of the process of the present invention.

[0033] Figure 2 This is a city background result map extracted by the method of the present invention.

[0034] Figure 3 A schematic diagram of the spatial distribution of urban heat island intensity on the surface of a city is obtained by using the method of the present invention and the traditional buffer zone algorithm.

[0035] Figure 4 This is a scatter diagram of the relationship between the urban heat island area estimated from the mountainous urban suburbs background extracted by the method of the present invention and socioeconomic factors.

[0036] Figure 5 This is a scatter diagram of the relationship between the urban heat island area estimated from the plain city suburbs background extracted by the method of the present invention and socioeconomic factors. DETAILED DESCRIPTION

[0037] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0038] like Figure 1 As shown in Figure 1, a method for calculating the surface urban heat island intensity for complex terrain in mountainous cities is Figure 1 As shown, the following steps are included:

[0039] Step 1: Obtain data and preprocess it;

[0040] The data includes multi-source remote sensing data, administrative boundary vector data and socioeconomic factor data.

[0041] Multi-source remote sensing data include: land use / cover change image data LUCC, surface temperature data LST, nighttime urban light data NTL, vegetation cover data (Landsat NDVI, an NDVI product inverted based on Landsat L1 surface reflectance data under clear sky conditions), and elevation data DEM.

[0042] The NDVI calculation method is as follows:

[0043] NDVI=(ρ NIR -ρ R ) / (ρ NIR +ρ R )

[0044] Where: NDVI is the normalized vegetation index, ρ NIR is the reflectivity in the near-infrared band, ρ R is the reflectivity in the red light band.

[0045] Socioeconomic factor data include population size, GDP, energy consumption, built-up area, and number of civilian vehicles. The data are sourced from the statistical yearbooks published by the statistical bureaus of the study areas.

[0046] Preprocessing includes: data reading, mosaicking, resampling, and cropping, and the unified storage format is GeoTIFF.

[0047] Step 2: Extract the pixels marked as cities in the land use / cover change image LUCC to obtain urban pixels;

[0048] Step 3: Taking the urban pixel obtained in step 2 as the center, the buffer method is used to extract the basic suburban background;

[0049] The buffer zone method is a traditional buffer zone method, i.e., a suburban background division method, which refers to using a buffer zone range of 1 to 50 km outside the urban built-up area as the suburban background required for surface urban heat island monitoring. In the present invention, a buffer zone within a range of 25 km is extracted as the basic suburban background.

[0050] Step 4: Calculate the average elevation of urban pixels, and select pixels whose path difference with the average elevation of urban pixels is within a threshold range from the basic suburban background obtained in step 3 as alternative suburban backgrounds; in the present invention, pixels with a path difference of ≤50 meters are selected as alternative suburban backgrounds.

[0051] Step 5: Filter pixels whose NTL pixels of night city light data are within the threshold range from the candidate suburban background obtained in step 4 as candidate suburban pixels; in the present invention, pixels whose NTL pixels of night city light data are ≤ 15 are selected as candidate suburban pixels.

[0052] Step 6: Calculate the standard deviation of all pixels in the normalized vegetation index dataset based on the land surface temperature data LST, and obtain pixels with a standard deviation within the threshold range as the suburban background; in the present invention, calculate the standard deviation of all Landsat NDVI pixels within the LST pixel range, and select pixels with an NDVI standard deviation ≤ 0.05 as the final suburban background.

[0053]

[0054] Where: σ is the standard deviation, x i is the i-th data point, μ is the mean of the data, and N is the total number of data points.

[0055] Step 7: Based on the suburban background obtained in step 6, the urban-rural dichotomy method is used to calculate the surface urban heat island intensity.

[0056]

[0057] Where: SUHI i is the surface urban heat island intensity corresponding to the i-th pixel on the image, T i is the surface temperature value, n is the number of suburban background clear sky pixels, T crop is the average background surface temperature in the suburbs.

[0058] The size of the correlation coefficient is used to evaluate the applicability of the calculation method in the urban heat island effect in mountainous areas.

[0059] The calculation method of correlation coefficient R is as follows:

[0060]

[0061] Where: X i is the surface urban heat island area, n is the number of pixels, i is the pixel number, is the average surface urban heat island area, Y i is the socioeconomic factor value, is the average of all socioeconomic factors.

[0062] A correlation coefficient will be calculated for each different socioeconomic factor, and the applicability will be judged based on the size of the correlation coefficient. The larger the value, the better the applicability.

[0063] To illustrate the effect of the present invention, a typical mountain city and a plain city were selected using the traditional buffer zone method and the method of the present invention. The SUHI of the two methods was calculated using the method of the present invention, and their correlation with socioeconomic factors was evaluated to assess the applicability of the two methods in urban heat island monitoring of mountain city landmarks.

[0064] The process of the present invention is as follows:

[0065] Step 1: Data collection. Remote sensing data includes: DEM, LUCC, NTL, Landsat L1 data, and LST remote sensing products for the main urban areas of mountainous cities and plain cities from 2005 to 2022. Ancillary data includes administrative division vector data of the two cities and statistical yearbook data from 2005 to 2022. The DEM is SRTM1 30-meter resolution data; the LUCC is the 1000-meter MODIS Land Use / Cover Change annual product (MCD12Q1); the NTL is the 2005-2022 annual nighttime light dataset integrating DMSP OLS and NPP VIIRS; the Landsat L1 data is the clear-sky Landsat 5-9 dataset from 2005 to 2022; and the LST is the 1000-meter MODIS Land Surface Temperature 8-day composite product (MYD11A2). The statistical yearbook data includes: population size, total GDP, energy consumption, built-up area and number of civilian vehicles in the main urban areas of mountainous cities and plain cities from 2005 to 2022.

[0066] Step 2: Extract the pixels marked as cities in the MCD12Q1 image to obtain the urban pixels of the two cities.

[0067] Step 3: Using the urban pixel obtained in step 2 as the center, a buffer zone method is used to extract the basic suburban background. Specifically, a 5-25 km buffer zone outside the urban area is defined as the basic suburban background, and the basic suburban background of the two cities is obtained.

[0068] Step 4: Calculate the average elevation DEM of urban pixels city , filter the pixel DEM in the basic suburban background rural Value and urban pixel DEM city Pixels with an absolute value of the difference |ΔDEM| ≤ 50 meters are selected as alternative suburban background pixels to reduce the estimation error of the surface urban heat island intensity caused by topographic factors. The calculation formula of ΔDEM is as follows:

[0069] |ΔDEM|=|DEM rural -DEM city |;

[0070] Where: |ΔDEM| is the absolute value of the difference between the DEM of the suburban pixel and the urban pixel; DEM rural is the DEM value of the suburban pixel; DEM city is the DEM value of urban pixels.

[0071] Step 5: Filter pixels with NTL ≤ 15 among the alternative suburban background pixels to determine areas not affected by human activities, that is, rural areas that are generally not affected by the urban heat island effect, reduce the surface urban heat island intensity error caused by the scattered satellite towns of mountainous cities, and further obtain the alternative suburban backgrounds of the two cities.

[0072] Step 6: Based on the extracted candidate suburban background pixels, NDVI products were inverted using Landsat L1 data. The standard deviation of all Landsat NDVI pixels within the LST pixel range was calculated, and pixels with an NDVI standard deviation ≤ 0.05 were selected to minimize the impact of pixel scale and mixed pixels on the intensity of the mountainous urban heat island effect. Based on these criteria, suburban backgrounds for the two cities were extracted with a spatial resolution of 1 km and a temporal resolution of 1 year.

[0073] Figure 2 The spatial distribution maps of suburban backgrounds extracted by the traditional buffer zone algorithm (TBM) and the multi-factor decision-making algorithm (MFD) are shown in Tables 1 and 2. a is the suburban background of plain cities extracted based on the TBM algorithm at a buffer scale of 5 to 25 km, b is the plain city background extracted based on the MFD algorithm at a buffer scale of 5 to 25 km, c is the suburban background of mountainous cities based on the TBM algorithm at a buffer scale of 5 to 25 km, and d is the suburban background of mountainous cities based on the MFD algorithm at a buffer scale of 5 to 25 km. Tables 1 and 2 are the statistical results of suburban background information of plain cities and mountainous cities, respectively. Figure 2 As shown in Tables 1 and 2, the LST average value of suburban background pixels extracted by the multi-factor decision-making algorithm varies less with the buffer zone than the traditional buffer algorithm, indicating that the multi-factor decision-making algorithm is more stable in SUHI estimation at different buffer scales than the traditional buffer algorithm. Furthermore, as the buffer scale changes, the LST average value of suburban background pixels extracted by the traditional buffer algorithm in mountainous urban areas (1.9K) varies more than that of Chengdu (1.6K), while the LST average value of suburban background pixels extracted by the multi-factor decision-making algorithm in mountainous urban areas (0.4K) varies less than that of Chengdu (0.5K), indicating that the suburban background extracted by the multi-factor decision-making algorithm is more suitable for SUHI monitoring in mountainous cities.

[0074] Table 1. Statistical results of the number of background pixels, urban-suburban altitude difference, and average surface temperature in plain urban suburbs

[0075]

[0076] Table 2. Statistical results of the number of background pixels, urban-suburban altitude difference, and average surface temperature in the suburbs of the main urban area of ​​mountainous cities

[0077]

[0078] Step 7: Based on the suburban background obtained in step 6, the urban-rural dichotomy method is used to calculate the surface urban heat island intensity.

[0079] Figure 3 The spatial distribution of SUHI estimated using the traditional buffer algorithm and the suburban background extracted using the multi-factor decision-making algorithm is shown. The SUHI estimated using the traditional buffer algorithm contains some "false heat island" pixels, and the range and intensity of the heat island increase significantly with increasing buffer size. However, the SUHI estimated using the multi-factor decision-making algorithm does not change significantly, indicating that the suburban background extracted using the multi-factor decision-making algorithm is more suitable for SUHI monitoring in mountainous cities.

[0080] The evaluation method of the present invention is used to compare and analyze the correlation between the surface urban heat island area estimated based on the suburban background of the traditional buffer zone algorithm and the multi-factor decision-making algorithm and the socioeconomic factors, so as to evaluate the applicability of the multi-factor decision-making algorithm in the monitoring of the surface urban heat island effect in mountainous cities. The correlation verification results are as follows: Figure 4 and Figure 5As shown, the correlations between SUHI estimated using the suburban context extracted using the multi-factor decision-making algorithm and socioeconomic factors are greater than those using the traditional buffer zone algorithm, indicating that the suburban context extracted using the multi-factor decision-making algorithm is more suitable for SUHI monitoring in mountainous cities. Furthermore, the R values ​​between the SUHI area estimated using the multi-factor decision-making algorithm and various economic factors in plain cities are also greater than those using the traditional buffer zone algorithm. This demonstrates that the multi-factor decision-making algorithm (the method for extracting suburban context in this paper) is not only applicable to mountainous cities but can also improve SUHI monitoring accuracy in plain cities, demonstrating its good universality. Figure 4 and Figure 5 In the figure, a, b, c, d, and e are the correlation coefficients calculated from different socioeconomic factor data.

[0081] The present invention first collects and preprocesses multi-source remote sensing data to calculate the Landsat NDVI product. Then, combining DEM, LUCC, NTL, and Landsat NDVI datasets, a multi-factor decision-making algorithm for suburban background delineation in mountainous cities, based on the traditional buffer algorithm, is proposed. Suburban background data is extracted using both the traditional buffer algorithm and the multi-factor decision-making algorithm. Combining suburban background information with the Landsat Stress Test (LST) product, the SUHI (Suspension Heating Index) of mountainous cities is estimated using a rural-urban dichotomy. Finally, the applicability of both algorithms for surface urban heat island monitoring in mountainous cities is evaluated by comparing the correlations between SUHI estimated using the traditional buffer method and the multi-factor decision-making algorithm, extracted using socioeconomic factors. The suburban background data derived from the proposed multi-factor decision-making algorithm for mountainous cities demonstrates higher accuracy than that derived from the traditional buffer method in SUHI monitoring in mountainous cities. Furthermore, the multi-factor decision-making algorithm exhibits good universality and is applicable not only to mountainous cities but also to improving SUHI monitoring accuracy in plain cities.

Claims

1. A method for calculating the surface urban heat island intensity for complex terrain in mountainous cities, characterized by: The following steps are involved: Step 1: Obtain data and preprocess it; Step 2: Extract the pixels marked as cities in the land use / cover change image to obtain urban pixels; Step 3: Taking the urban pixel obtained in step 2 as the center, the buffer method is used to extract the basic suburban background; Step 4: Calculate the average elevation of urban pixels, and select pixels whose elevation difference with the average elevation of urban pixels is within the threshold range from the basic suburban background obtained in step 3 as candidate suburban backgrounds; Step 5: Filter the pixels of the nighttime city light data within the threshold range from the candidate suburban background obtained in step 4 as candidate suburban pixels; Step 6: Calculate the standard deviation of all pixels in the normalized vegetation index dataset based on the surface temperature data, and obtain the pixels whose standard deviation is within the threshold range as the suburban background; Step 7: Based on the suburban background obtained in step 6, the urban-rural dichotomy method is used to calculate the surface urban heat island intensity.

2. The method for calculating the surface urban heat island intensity for complex terrain in mountainous cities according to claim 1 is characterized in that: The data in step 1 include multi-source remote sensing data, administrative boundary vector data and socioeconomic factor data.

3. The method for calculating the surface urban heat island intensity for complex terrain in mountainous cities according to claim 2 is characterized in that: The multi-source remote sensing data includes: land use / cover change image data, surface temperature data, nighttime city light data, vegetation cover data, and elevation data.

4. The method for calculating the surface urban heat island intensity for complex terrain in mountainous cities according to claim 2 is characterized in that: The socioeconomic factor data include population size, GDP, energy consumption, built-up area and number of civilian vehicles.

5. The method for calculating the surface urban heat island intensity for complex terrain in mountainous cities according to claim 3 is characterized in that: The vegetation cover index is a normalized index, and the calculation process is as follows: NDVI=(r NIR -r R ) / (ρ NIR +r R ) Where: NDVI is the normalized vegetation index, ρ NIR is the reflectivity in the near-infrared band, ρ R is the reflectivity in the red light band.

6. The method for calculating the surface urban heat island intensity for complex terrain in mountainous cities according to claim 1 is characterized in that: The preprocessing in step 1 includes: data reading, mosaicking, resampling, cropping, and unified storage format.

7. The method for calculating the surface urban heat island intensity for complex terrain in mountainous cities according to claim 1 is characterized in that: The calculation method of the surface urban heat island intensity in step 7 is as follows: Where: SUHI i is the surface urban heat island intensity corresponding to the i-th pixel on the image, T i is the surface temperature value, n is the number of suburban background clear sky pixels, T crop is the average background surface temperature in the suburbs.

8. The evaluation method for calculating the surface urban heat island intensity of a mountainous city with complex terrain according to any one of claims 1 to 7, characterized in that: The size of the correlation coefficient is used to evaluate the applicability of the calculation method in the urban heat island effect in mountainous areas.

9. The evaluation method for calculating the surface urban heat island intensity of a mountainous city with complex terrain according to claim 8 is characterized in that: The calculation method of the correlation coefficient R is as follows: Where: X i is the surface urban heat island area, n is the number of pixels, i is the pixel number, is the average surface urban heat island area, Y i is the socioeconomic factor value, is the average of all socioeconomic factors.

Images