Global sea and land temperature data production method and system based on MODIS data

Abstract

The present invention relates to a method and system for generating global land surface temperature data based on MODIS data. The method comprises: performing mean processing on global land surface temperature LST data and global sea surface temperature SST data; resampling SST data according to the resolution of LST data; splicing the resampled SST data with LST data; interpolating the spliced ​​data, including interpolation of mid-latitude missing value areas and interpolation of Arctic missing value areas; wherein for missing values ​​in the Arctic region, the temperature value of the region is reassigned; and generating global land surface temperature GST data according to the interpolation result. The method and system for generating global land surface temperature data based on MODIS data provided by the present invention can not only take into account the information of clear sky images, but also avoid the problem of low data accuracy caused by the warming / cooling effect of cloud cover on the surface temperature.

Description

Technical Field

[0001] The present invention relates to the field of remote sensing technology, and in particular to a method and system for producing global land and sea temperature data based on MODIS data. Background Art

[0002] Global land surface temperature (LST) data is controlled by land-atmosphere interaction and energy flux, and is an important parameter of surface energy balance and water cycle physical processes at regional and global scales. It plays an important role in agricultural meteorological research, regional climate change, surface energy interaction and global circulation patterns. Therefore, regional and global scale surface temperature research is of great significance for further improving and perfecting natural disaster monitoring, urban climate and environmental research, and hydroclimate prediction models. How to accurately obtain the important geophysical parameter of surface temperature has always been a hot topic for many scholars.

[0003] The purpose of obtaining global sea surface temperature (Sea Surface Temperature, abbreviated as SST) data is to understand regional and global climate changes and enable people to understand large-scale ocean current changes. The upper ocean has the ability to regulate the global climate system, and the temporal distribution of atmospheric water vapor and heat flux is a function of the sea surface temperature. The gradient distribution of sea surface temperature is related to the flow system, eddy, air flow and upwelling area. With the improvement of people's understanding and the development of satellite detection technology, satellite sea surface temperature inversion methods are also constantly developing. Sea surface temperature inversion methods are mainly divided into two categories: one is to use the regression of buoys synchronized with the satellite to obtain the sea surface temperature inversion coefficient, and the other is to use the atmospheric radiation transfer model to simulate and calculate the radiation transfer from the sea surface to the satellite altitude to obtain the sea surface temperature inversion coefficient.

[0004] The frequent and random cloud cover makes it very difficult to reconstruct the surface temperature under cloud cover. At any time, about 65% of the global surface is covered by clouds, which directly leads to large-scale unevenly distributed missing values ​​in thermal infrared remote sensing images. At the same time, thin clouds are difficult to detect through cloud mask products or cloud detection methods. However, the presence of thin clouds will cause the partial loss of surface information obtained by the sensor, resulting in the acquisition of the brightness temperature of the cloud top or the brightness temperature value with the cloud effect added, resulting in many abnormal values ​​in the inverted remote sensing images. Since the 1980s, many researchers at home and abroad have begun to study the reconstruction of the surface temperature of cloud-covered pixels from thermal infrared data. These methods mainly include data reconstruction methods based on quality files; reconstruction methods based on spatiotemporal information, and regression model reconstruction methods combined with multi-source auxiliary information.

[0005] In order to generate global surface temperature GST (Globe Surface Temperature) data, there are the following technical solutions in the prior art:

[0006] The first is a data reconstruction method based on quality files. To produce global MODIS temperature data, ocean and land temperature data are required. The two types of data can be spliced ​​and the missing values ​​can be filled to obtain global surface temperature GST data. This article (Ye Yuhan, Zhang Wenjuan, Guan Yunlan, et al. Global temperature data generation method based on MODIS products [J]. Jiangxi Science, 2018, 036(002): 209-216.) first used product quality control (Quality Control, abbreviated as QC) data for positioning and repair, repaired invalid pixels in the original data, and made the data to be synthesized have better quality. Then interpolation was used to fill in the missing values, and a unified attribute standard was established for land surface temperature data and sea surface temperature data, and on this basis, global large-scale data synthesis was achieved.

[0007] The second is a method based on reconstruction of spatiotemporal information: including methods that directly reconstruct cloud-covered pixels using neighboring information that is highly similar in time and space. The missing values ​​are filled in by borrowing from the time series harmonic analysis interpolation algorithm. A filter based on a moving window weighted average is used in the reconstruction process to obtain MODIS surface temperature data with high consistency. It is believed that pixels with closer distances have higher similarity. Similar pixels are defined based on spatial distance, and closer pixels contribute more to the reconstruction process. A similarity function is established to identify similar pixels, and a transfer function between similar pixels and missing pixels is established to reconstruct missing pixels.

[0008] The third is a method that combines multi-source auxiliary information reconstruction. This method (Hengl T, Heuvelink G, Tadi MP, et al. Spatio-temporal prediction of daily temperatures using time-series of MODIS LST images [J]. Theoretical & Applied Climatology, 2012, 107 (1-2): 265-277.) solves the problem of missing data by establishing a correlation regression model between cloud-contaminated pixels and corresponding auxiliary data pixels. An 8-day composite surface temperature product was estimated by establishing a regression model containing multiple auxiliary prediction factors (latitude, longitude, altitude and normalized difference vegetation index (NDVI)). More prediction factors were added, and the MODIS surface temperature time series was reconstructed using solar radiation, time, altitude, ocean distance and other factors as input parameters of the regression model, which also achieved good accuracy. Surface temperature data were reconstructed in flat areas using multiple auxiliary images, including land cover, NDVI and the 7th band of MODIS. Other similar algorithms also obtain surface temperature data with continuous spatiotemporal information by using many factors that affect the surface temperature for the reconstruction of the surface temperature. Considering the complexity of the terrain and the uneven spatial distribution of large-scale surface temperature, the model reconstructed by the regression model combined with multi-source auxiliary information can provide rich reference information for missing pixels, thereby improving the accuracy of the interpolation results.

[0009] The first technical solution requires other data for reference correction. The missing values ​​near the poles are filled with temperature data from nearby areas. The operation process is cumbersome, the calculation amount is large, and the computer processing speed is slow. It cannot meet the requirements of global surface temperature data production.

[0010] The limited information of the second technical solution cannot meet the information requirements of large areas of missing values, resulting in different degrees of smoothness in the reconstruction results, making it difficult to capture extreme temperatures and sudden and drastic changes. Moreover, these methods only consider the information of clear sky pixels and ignore the warming / cooling effect of cloud cover on surface temperature under different conditions.

[0011] The third technical solution has the problem of being unable to obtain the surface temperature value under the cloud layer when the Terra / Aqua-MODIS satellite passes over, and cannot meet the requirements for global surface temperature data production. Summary of the invention

[0012] The technical problem to be solved by the present invention is to provide a method and system for producing global land and sea temperature data based on MODIS data, which can not only take into account the information of clear sky images, but also avoid the problem of low data accuracy caused by the warming / cooling effect of cloud cover on the surface temperature.

[0013] To solve the above technical problems, the present invention provides a method for generating global surface temperature data based on MODIS data, the method comprising: averaging global land surface temperature LST data and global sea surface temperature SST data; resampling SST data according to the resolution of LST data; splicing the resampled SST data with LST data; interpolating the mid-latitude missing value area and the Arctic missing value area for the spliced ​​data respectively; for missing values ​​in the Arctic region, reassigning the temperature value of the region; and generating day and night global surface temperature GST data according to the interpolation results.

[0014] In some embodiments, generating day and night global surface temperature GST data according to the interpolation results includes: generating 8 days of day and night global surface temperature GST data according to the interpolation results.

[0015] In some embodiments, generating diurnal global land surface temperature GST data according to the interpolation result further includes: generating monthly diurnal global land surface temperature GST data according to 8 days of diurnal global land surface temperature GST data.

[0016] In some embodiments, generating monthly daytime global surface temperature GST data based on 8 days of daytime global surface temperature GST data includes: generating monthly daytime global surface temperature GST data according to the following formula:

[0017]

[0018] in, is the i-th 8-day day and night global surface temperature GST data, n is the data belonging to the data processing of this month, i is the number of data in the month, It is the monthly day and night global surface temperature GST data of the kth month, and the maximum value of k is 12.

[0019] In some embodiments, the global land surface temperature LST data and the global sea surface temperature SST data are averaged, including: averaged according to the following formula:

[0020]

[0021] Among them, S x，y The value of the (x,y) position in the i-th year, n is the number of years used for averaging, and n is the number of valid values ​​of (x,y) in 5 years. The output is the mean of the point (x, y). The output result is the 5-year mean data of LST and SST, where x is latitude and y is longitude.

[0022] In some embodiments, the spliced ​​data are interpolated for the missing value areas in the mid-latitudes and the missing value areas in the Arctic, respectively, including: for the missing values ​​in the mid-latitude areas, a linear interpolation function in the fillmissing interpolation function is used to interpolate the missing values ​​in the digital image; for the interpolation of the missing value areas in the Arctic, the transplanted data uses the Antarctic temperature data that is 6 months different from that month, and the interpolation results of the missing value areas near the Antarctic are used. By analyzing the trend of temperature changes in the Antarctic, the temperature change trend is transplanted to the interpolation of the Arctic region.

[0023] In some embodiments, for interpolation of missing value areas in the Arctic, the transplanted data uses Antarctic temperature data that is 6 months different from the current month, and uses the interpolation results of the missing value areas near the Antarctic. By analyzing the trend of temperature changes in the Antarctic, the temperature change trend is transplanted into the interpolation of the Arctic region, including: searching pixel by pixel for the point where a valid value first appears in the Arctic region according to the latitude change, recording its longitude and longitude, and then synchronously searching for the temperature values ​​of the points in the Antarctic region and the latitude symmetric to the equator with respect to the latitude of the point, and when the temperature difference is less than 0.5K, it can be considered that the longitude temperature change is similar to the temperature change trend of the missing value area to be filled; transplanting the relevant data to the missing value area in the Arctic, and adding the temperature difference.

[0024] In some embodiments, adding the temperature difference includes: calculating according to the following formula:

[0025] N(x a ,y b )=S(x i ,y u )+△t

[0026] Among them, N(x a ,y b ) is the area in the Arctic that needs to be filled in, S(x i ,y u ) is based on the screening conditions, and S(x i ,y u ) time and N(x a ,y b ) data differ by 6 months, which is consistent with the Antarctic region value for transplantation data. △t is S(x i ,y u ) and N(x a ,y b ), and the absolute value of △t is less than 0.5K.

[0027] In some embodiments, for missing values ​​in the Arctic region, reassigning the temperature value of the region includes: reassigning the value according to the following formula:

[0028]

[0029] in, is the longitude x a The mean of the i elements at different latitudes, is the corrected temperature data value.

[0030] In addition, the present invention also provides a global surface temperature data generation system based on MODIS data, the system comprising: one or more processors; a storage device for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the global surface temperature data generation method based on MODIS data as described above.

[0031] After adopting such a design, the present invention has at least the following advantages:

[0032] Other data are needed for reference correction, the operation process is cumbersome, the amount of calculation is large, and the computer processing speed is slow; only the information of clear sky pixels is considered while ignoring the warming / cooling effect of cloud cover on the surface temperature, resulting in low data accuracy; the surface temperature value under the cloud layer at the time of satellite transit cannot be obtained, and the requirements for global surface temperature data production cannot be met. The present invention can quickly and conveniently produce global-scale earth surface temperature products while avoiding the above shortcomings. After splicing the sea and land temperature products, the missing value areas can be further quickly repaired to obtain complete global surface temperature data, improve data availability, greatly improve production efficiency, and save manpower and material resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The above is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0034] Figure 1 This is a schematic diagram of the global surface temperature (GST) data production process;

[0035] Figure 2 This is a graph of the 5-year average of the 8-day LST data;

[0036] Figure 3 It is a mosaic of 8-day LST data and SST;

[0037] Figure 4 It is a differentiated quick repair missing value map. DETAILED DESCRIPTION

[0038] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0039] Based on MODIS data, global surface temperature data is generated. First, the average of multi-year LST and SST data is used as the initial value to repair missing data, and then differentiated rapid repair of missing values ​​is performed. The steps are to interpolate and repair data for mid- and low-latitudes and the Arctic region respectively, so that the global daytime and nighttime sea and land temperature products can fully reflect the temporal and spatial distribution of temperature on the earth's surface. Then, 8-day and monthly earth surface temperature data can be produced, providing a complete reference for global surface temperature changes.

[0040] The global surface temperature GST data flow consists of the splicing of sea surface temperature SST data and global land surface temperature LST data. The processing flow is as follows: Figure 1 shown.

[0041] (1) Mean processing of LST and SST data

[0042] First, the LST data and SST data are averaged, and the formula is as follows:

[0043]

[0044] Among them, S x，y The value of the (x, y) position in the i-th year, n is the number of years used for mean processing, and this paper selects n = 5 years of data, n a is the number of valid values ​​of (x,y) within 5 years, is the mean of the point (x,y), and the output result is the 5-year mean data of LST and SST data.

[0045] (2) SST data resampling

[0046] The spatial resolution of LST data is different from that of SST data. The resolution of LST data is 1km, and that of SST data is 4km. SST data should be resampled to make the spatial resolution of SST data the same as that of LST data. In order to complete accurate splicing, the nearest neighbor interpolation method is used to obtain the global sea surface temperature data SST with a resolution of 1km. The formula is as follows:

[0047] X=DX*(SW / DW)

[0048] Y=DY*(SH / DH)

[0049] In the above formula, DX and DY are the horizontal and vertical coordinates of a pixel in the target image, DW and DH are the length and width of the target image; SW and SH are the width and height of the original image. X, Y are the coordinates of the original image corresponding to the target image at that point (DX, DY). If it is an enlarged image, the (SW / DW) value is less than 1, and DX is also mapped to X of the original image in the same proportion. If SW / DW=1, it is equivalent to copying the pixel.

[0050] (3) Global surface temperature (GST) data splicing

[0051] The LST data and the resampled SST data have the same spatial resolution of 1 km, and can be spliced ​​together to obtain a spliced ​​map of global surface temperature data including ocean and land. The spliced ​​data still have missing values, which need to be further processed.

[0052] (4) Interpolation of missing values ​​in mid-latitude areas

[0053] The steps are as follows: The missing values ​​in the 60°Nˉ60°S interval are generally the junction of the ocean and the land, and the missing values ​​affected by the cloud factors in the middle and low latitudes. The missing values ​​in the above areas can be filled by the nearest neighbor interpolation method, so that the spatial resolution of the two data is consistent and can be subsequently spliced. First, for the missing values ​​in the mid-latitude area, the linear interpolation function in the fillmissing interpolation function is used for processing.

[0054] (5) Interpolation of missing values ​​in the Arctic region:

[0055] Since the boundary constraints in the area near the North Pole are basically missing, the interpolation results are poor, and areas with large numerical deviations such as the North Pole need to be replaced. The specific operation process is as follows:

[0056] For interpolation of missing value areas in the Arctic, the transplanted data uses Antarctic temperature data that is 6 months different from that month. Using the interpolation results of the missing value areas near the South Pole, by analyzing the trend of temperature change in the South Pole, the temperature change trend is transplanted to the interpolation of the Arctic region. The specific method is: search for the point where the first valid value appears in the Arctic region pixel by pixel according to the latitude change, record its longitude and latitude, and then synchronously search for the temperature values ​​of the points in the Antarctic region and the latitude symmetric to the equator. When the temperature difference (△t) is less than 0.5K, it can be considered that the temperature change trend of the longitude is similar to that of the missing value area to be filled. Transplant the relevant data to the missing value area in the Arctic, and add the temperature difference (△t). The formula is as follows:

[0057] N(x a ,y b )=S(x i ,y u )+△t

[0058] N(xa ,y b ) is the area in the Arctic that needs to be filled in, S(x i ,y u ) is based on the screening conditions, and S(x i ,y u ) time and N(x a ,y b ) data differ by 6 months, which is consistent with the Antarctic region value for transplantation data. △t is S(x i ,y u ) and N(x a ,y b ) and the absolute value of △t is less than 0.5K.

[0059] (6) Interpolation area correction:

[0060] After completing the preliminary interpolation of the Arctic region, some areas may have missing values, and the temperature difference at the same latitude may be large. Usually, the Arctic region is covered with sea ice in winter, and the temperature difference at the same latitude is not large, so the temperature data for the missing values ​​is taken as the average value of the same latitude. The temperature value of the Arctic region is reassigned, and the formula is as follows:

[0061]

[0062] is the longitude at x a， The mean of the i elements under latitude, is the corrected temperature data value. The data correction can be completed according to the above formula.

[0063] (7) Monthly day and night global surface temperature (GST) data:

[0064] The 8-day day and night global surface temperature GST data is used to produce monthly average data. The formula is as follows:

[0065]

[0066] in, is the i-th 8-day day and night global surface temperature GST data, n is the data belonging to the data processing of the month, this paper selects n as the maximum of 4, i is the number of data in the month, It is the monthly day and night global surface temperature GST data of the kth month, and the maximum value of k is 12.

[0067] The present invention also provides a global surface temperature data generation system based on MODIS data. For example, the global surface temperature data generation system based on MODIS data can be used to serve as a data product generation host in a remote sensing satellite system. As described herein, the global surface temperature data generation system based on MODIS data can be used to implement the generation function of data products in a remote sensing satellite imaging system. The global surface temperature data generation system based on MODIS data can be implemented in a single node, or the function of the global surface temperature data generation system based on MODIS data can be implemented in multiple nodes in a network. Those skilled in the art should be aware that the term global surface temperature data generation system based on MODIS data includes equipment in a broad sense, and the global surface temperature data generation system based on MODIS data mentioned here is only one example. Including a global surface temperature data generation system based on MODIS data is for clarity of expression, and is not intended to limit the application of the present invention to a specific global surface temperature data generation system based on MODIS data or a certain type of global surface temperature data generation system based on MODIS data. At least some of the features / methods described in the present invention can be implemented in a network device or component, for example, in a global surface temperature data generation system based on MODIS data. For example, the features / methods of the present invention can be implemented in hardware, firmware and / or software installed and run on hardware. The global surface temperature data generation system based on MODIS data can be any device that processes, stores and / or forwards data frames through a network, such as a server, a client, a data source, etc. The processor can include one or more multi-core processors and / or memory devices, which can be used as data storage, buffers, etc. The processor can be implemented as a general-purpose processor, or can be part of one or more application specific integrated circuits (ASICs) and / or digital signal processors (DSPs).

[0068] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Those skilled in the art may make some simple modifications, equivalent changes or modifications using the technical contents disclosed above, which all fall within the protection scope of the present invention.

Claims

1. A method for generating global surface temperature data based on MODIS data, characterized in that: include: The global land surface temperature LST data and the global sea surface temperature SST data are averaged; Resample the SST data according to the resolution of the LST data; The resampled SST data are spliced ​​with the LST data; The spliced ​​data are interpolated in the mid-latitude missing area and the Arctic missing area, including: For missing values ​​in mid-latitude areas, the linear interpolation function in the fillmissing interpolation function is used to interpolate the missing values ​​in the digital image; For interpolation of the missing value area in the Arctic, the transplanted data uses the Antarctic temperature data that is 6 months different from the corresponding month of the missing value area in the Arctic. Using the interpolation results of the missing value area near the Antarctic, by analyzing the trend of Antarctic temperature change, the temperature change trend is transplanted to the interpolation of the Arctic region, including: According to the change in latitude, search pixel by pixel for the point where the valid value first appears in the Arctic region, record its longitude and latitude, and then synchronously search for the temperature value of the point in the Antarctic region that is symmetrical to the latitude of the point about the equator. When the temperature difference is less than 0.5K, it can be considered that the temperature change trend of the latitude is similar to that of the area to be filled in. Transplant the relevant data to the missing areas in the Arctic and add the temperature difference; Plus the temperature difference, including: Calculate according to the following formula: N(x a ,and b )=S(x i ,and u )+△t Among them, N(x a ,y b ) is the area in the Arctic that needs to be filled in, S(x i ,y u ) is based on the screening conditions, and S(x i ,y u ) time and N(x a ,y b ) data differ by 6 months, which is consistent with the Antarctic region value for transplantation data. △t is S(x i ,y u ) and N(x a ,y b ), and the absolute value of △t is less than 0.5K; For missing values ​​in the Arctic region, the temperature values ​​in this region are reassigned, including: Re-assign values ​​according to the following formula: in, is the longitude x a The mean of the i elements at different latitudes, is the corrected temperature data value; Generate day and night global surface temperature GST data based on the interpolation results.

2. The method for generating global surface temperature data based on MODIS data according to claim 1, characterized in that: Generate day and night global surface temperature GST data based on the interpolation results, including: Based on the interpolation results, 8-day day and night global surface temperature GST data are generated.

3. The method for generating global surface temperature data based on MODIS data according to claim 2, characterized in that: Generate day and night global surface temperature GST data based on the interpolation results, including: Based on 8 days of daytime global surface temperature GST data, monthly daytime global surface temperature GST data are generated.

4. The method for generating global surface temperature data based on MODIS data according to claim 3, characterized in that: Based on 8 days of daytime global surface temperature GST data, monthly daytime global surface temperature GST data are generated, including: The monthly day and night global surface temperature GST data is generated according to the following formula: in, is the i-th 8-day day and night global surface temperature GST data, n is the data belonging to the data processing of this month, i is the number of data in the month, It is the monthly day and night global surface temperature GST data of the kth month, and the maximum value of k is 12.

5. The method for generating global surface temperature data based on MODIS data according to claim 1, characterized in that: The global land surface temperature LST data and the global sea surface temperature SST data are averaged, including: The mean value is processed according to the following formula: Among them, s x，y The value of the position (x, y) in the i-th year, n is the number of years used for averaging, and n is the number of valid values ​​of (x, y) in 5 years. The output is the mean of the point (x, y). The output result is the 5-year mean data of LST and SST, where x is latitude and y is longitude.

6. A global surface temperature data generation system based on MODIS data, characterized in that: include: one or more processors; A storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method for generating global surface temperature data based on MODIS data according to any one of claims 1 to 5.

Images