A method for developing a global MODIS land surface temperature product with consistent angles

Abstract

The application provides a kind of angle consistent global MODIS ground surface temperature product development method, comprising: according to the initial global scale MODIS remote sensing image data and initial reanalysis data, the database of MODIS remote sensing data and ERA5-Land reanalysis data is constructed;The MODIS zenith observation direction remote sensing data and reanalysis data in the database of MODIS remote sensing data and ERA5-Land reanalysis data are extracted using MODIS observation zenith angle information, and the training data set is constructed according to the MODIS zenith observation direction remote sensing data and reanalysis data;Based on the training data set, a global scale vertical ground surface temperature estimation model is established using a random forest machine learning algorithm;Based on the global scale vertical ground surface temperature estimation model, the corresponding MODIS vertical ground surface temperature data is estimated using MODIS non-vertical observation time remote sensing data and ERA5-Land reanalysis data.The application helps to improve the application value of ground surface temperature product in climate change, resource environment monitoring, ground surface energy balance and the like.

Description

Technical Field

[0001] This invention relates to the field of remote sensing surface temperature estimation technology, and in particular to a method for developing a globally consistent MODIS surface temperature product. Background Technology

[0002] Land surface temperature (LST) is an important physical quantity for the energy balance of the Earth's surface and a good indicator of the exchange of heat and water between the Earth's surface and the atmosphere at regional and global scales. It is widely used in many research fields such as climate change monitoring, surface evapotranspiration estimation, and urban thermal environment research.

[0003] Due to the prevalence of heterogeneous and non-isothermal mixed pixels in satellite observations, and the influence of the three-dimensional structure of vegetation canopy on the surface temperature of the illumination / shadow components under different observation geometries, surface temperatures acquired by wide-swath satellites typically exhibit strong angular effects. Therefore, there is currently no globally angularly consistent surface temperature remote sensing product. In practical applications, the thermal radiation directionality model is the most commonly used method for surface temperature angular normalization; however, this model suffers from low computational efficiency, numerous and difficult-to-obtain input parameters, and limited applicability, making it unsuitable for acquiring globally angularly consistent surface temperature products.

[0004] Therefore, developing a global-scale, angle-consistent surface temperature algorithm model to obtain global-scale vertical surface temperature remote sensing products is of great significance for improving the practical applications of surface temperature products in climate change, resource and environmental monitoring, and surface energy balance. Summary of the Invention

[0005] To overcome the shortcomings of the existing technology, the purpose of this invention is to provide a method for developing a globally consistent MODIS surface temperature product.

[0006] To achieve the above objectives, the present invention provides the following solution:

[0007] A method for developing a globally consistent MODIS surface temperature product, comprising:

[0008] A database of MODIS remote sensing data and ERA5-Land reanalysis data was constructed based on the acquired initial global-scale MODIS remote sensing image data and initial reanalysis data.

[0009] Using MODIS observation zenith angle information, remote sensing data and reanalysis data of the MODIS zenith observation direction are extracted from the database of MODIS remote sensing data and ERA5-Land reanalysis data, and a training dataset is constructed based on the remote sensing data and reanalysis data of the MODIS zenith observation direction.

[0010] Based on the training dataset, a global-scale vertical surface temperature estimation model was established using the random forest machine learning algorithm.

[0011] Based on the global-scale vertical surface temperature estimation model, the corresponding MODIS vertical surface temperature data are estimated using remote sensing data from MODIS non-vertical observation times and ERA5-Land reanalysis data.

[0012] Preferably, the step of constructing a database of MODIS remote sensing data and ERA5-Land reanalysis data based on the acquired initial global-scale MODIS remote sensing data and initial reanalysis data includes:

[0013] Acquire the initial global-scale MODIS remote sensing data and the initial ERA5-Land reanalysis data;

[0014] Spatiotemporal matching and preprocessing are performed on the initial global-scale remote sensing data and the initial reanalysis data to obtain preprocessed data;

[0015] A database of MODIS remote sensing data and ERA5-Land reanalysis data is constructed based on the preprocessed data.

[0016] Preferably, the initial global-scale remote sensing data comes from the MODIS / Terra+Aqua satellite, and includes MOD11A1 / MYD11A1 surface temperature data, MCD15A3H leaf area index data, MCD43A4 reflectance data, and MOD44B vegetation cover data; the initial reanalysis data comes from the ERA5-Land hourly dataset, and includes air temperature, radiation, and soil moisture data.

[0017] Preferably, the spatiotemporal matching step includes:

[0018] At the pixel scale, reanalysis data corresponding to the MODIS observation time is extracted from the initial reanalysis data based on the MODIS observation time;

[0019] The extracted reanalysis data were resampled from the original 9km resolution to a 1km resolution, the same as the MODIS surface temperature.

[0020] Preferably, the preprocessing includes quality control of MODIS surface temperature data and quality control between initial global-scale remote sensing image data and initial reanalysis data.

[0021] Preferably, the training dataset includes vertical reflectance data, leaf area index data, vegetation cover data, radiation / air temperature / soil moisture reanalysis data, and vertical surface temperature data.

[0022] Preferably, the method for constructing the training dataset includes:

[0023] Vertical reflectance data, leaf area index data, vegetation cover data, radiation / air temperature / soil moisture reanalysis data, and vertical surface temperature data were split into training and validation sets using random sampling; the ratio of the amount of data in the training set to the amount of data in the validation set was 8:2.

[0024] Preferably, the global-scale vertical surface temperature estimation model is a functional relationship f between remotely sensed surface temperature data from the satellite zenith observation direction and vegetation and reanalysis data:

[0025] T Nadir =f(TA) Nadir SM Nadir LAI Nadir NDVI Nadir ,SSRD Nadir ,....)

[0026] In this context, the subscript Nadir represents the MODIS vertical observation time, T represents the surface temperature of the remote sensing image at the MODIS vertical observation time, TA represents the air temperature, SM represents the soil moisture, LAI represents the leaf area index, NDVI represents the normalized difference vegetation index, and SSRD represents the downward solar shortwave radiation.

[0027] Preferably, the formula for estimating the corresponding MODIS vertical surface temperature data using MODIS non-vertical observation time remote sensing data and ERA5-Land reanalysis data based on the global-scale vertical surface temperature estimation model is as follows:

[0028] To ff Nadir =f(TAo ff SMo ff ,LAIo ff NDVIo ff Nadir ,SSRDo ff ,....);

[0029] Wherein, the subscript off represents the MODIS non-vertical observation time, the superscript Nadir represents the vertical remote sensing data, T represents the estimated vertical land surface temperature at the MODIS non-vertical observation time, TA is the air temperature, SM is the soil moisture, LAI is the leaf area index, NDVI is the normalized difference vegetation index, SSRD is the solar downwave radiation, and f is the global-scale vertical land surface temperature estimation model.

[0030] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0031] This invention provides a method for developing a globally consistent MODIS land surface temperature product, comprising: constructing a database of MODIS remote sensing data and ERA5-Land reanalysis data based on acquired initial global-scale MODIS remote sensing data and initial reanalysis data; extracting remote sensing data and reanalysis data in the MODIS zenith observation direction from the database of the MODIS remote sensing product and ERA5-Land reanalysis data using MODIS observation zenith angle information, and constructing a training dataset based on the remote sensing data and reanalysis data in the MODIS zenith observation direction; establishing a global-scale vertical land surface temperature estimation model based on the training dataset using a random forest machine learning algorithm; and estimating the corresponding MODIS vertical land surface temperature data based on the global-scale vertical land surface temperature estimation model using remote sensing data from MODIS non-vertical observation times and ERA5-Land reanalysis data. This invention addresses the problem that current satellite-acquired surface temperature products exhibit significant angular effects and lack globally scale vertical surface temperature products with consistent angles. The globally MODIS vertical surface temperature products with consistent angles will help enhance the application value of surface temperature products in climate change, resource and environmental monitoring, and surface energy balance. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 A flowchart of the method provided in an embodiment of the present invention;

[0034] Figure 2 This is a schematic diagram illustrating the implementation steps of an embodiment of the present invention. Detailed Implementation

[0035] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0036] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0037] The terms "first," "second," "third," and "fourth," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, including a series of steps, processes, methods, etc., is not limited to the steps listed, but may optionally include steps not listed, or may optionally include other steps inherent to these processes, methods, products, or devices.

[0038] The purpose of this invention is to provide a method for developing globally consistent MODIS land surface temperature products, which solves the problem that current satellite-acquired land surface temperature products have obvious angular effects and lack globally consistent vertical land surface temperature products. Globally consistent MODIS vertical land surface temperature products can help enhance the application value of land surface temperature products in climate change, resource and environmental monitoring, and land surface energy balance.

[0039] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0040] Figure 1 The method flowchart provided in the embodiments of the present invention is as follows: Figure 1 As shown, this invention provides a method for developing a globally consistent MODIS surface temperature product, comprising:

[0041] Step 100: Construct a database of MODIS remote sensing data and ERA5-Land reanalysis data based on the acquired initial global-scale MODIS remote sensing data and initial reanalysis data;

[0042] Step 200: Using MODIS observation zenith angle information, extract the remote sensing data and reanalysis data of the MODIS zenith observation direction from the database of MODIS remote sensing data and ERA5-Land reanalysis data, and construct a training dataset based on the remote sensing data and reanalysis data of the MODIS zenith observation direction.

[0043] Step 300: Based on the training dataset, establish a global-scale vertical surface temperature estimation model using the random forest machine learning algorithm;

[0044] Step 400: Based on the global-scale vertical surface temperature estimation model, estimate the corresponding MODIS vertical surface temperature data using remote sensing data from MODIS non-vertical observation times and ERA5-Land reanalysis data.

[0045] Preferably, step 100 specifically includes:

[0046] Acquire the initial global-scale MODIS remote sensing data and the initial ERA5-Land reanalysis data;

[0047] Spatiotemporal matching and preprocessing are performed on the initial global-scale remote sensing data and the initial reanalysis data to obtain preprocessed data;

[0048] A database of MODIS remote sensing data and ERA5-Land reanalysis data is constructed based on the preprocessed data.

[0049] like Figure 2 As shown, in this embodiment, step S1 first acquires global-scale MODIS remote sensing data and ERA5-Land hourly reanalysis data; then performs spatiotemporal matching and preprocessing of the MODIS remote sensing data and ERA5-Land reanalysis data; and constructs a database of MODIS remote sensing data and ERA5-Land reanalysis data.

[0050] Preferably, the initial global-scale remote sensing data comes from the MODIS / Terra+Aqua satellite, and includes MOD11A1 / MYD11A1 surface temperature data, MCD15A3H leaf area index data, MCD43A4 reflectance data, and MOD44B vegetation cover data; the initial reanalysis data comes from the ERA5-Land hourly dataset, and includes air temperature, radiation, and soil moisture data.

[0051] Preferably, the spatiotemporal matching step includes:

[0052] At the pixel scale, reanalysis data corresponding to the MODIS observation time is extracted from the initial reanalysis data based on the MODIS observation time;

[0053] The extracted reanalysis data were resampled from the original 9km resolution to a 1km resolution, the same as the MODIS surface temperature.

[0054] Preferably, the preprocessing includes quality control of MODIS surface temperature data and quality control between initial global-scale remote sensing image data and initial reanalysis data.

[0055] Specifically, in this embodiment, the initial global-scale remote sensing data comes from the MODIS / Terra+Aqua satellite, including daily 1000-meter spatial resolution MOD11A1 / MYD11A1 surface temperature data, every 4 days 500-meter spatial resolution MCD15A3H leaf area index data, daily 500-meter resolution MCD43A4 reflectance data corrected by the bidirectional reflectance distribution function, and annual 250-meter resolution MOD44B vegetation cover data. The initial remote sensing image data preprocessing process includes vector clipping and resampling to construct the MODIS remote sensing product database. The initial reanalysis data comes from the ERA5-Land hourly dataset, including hourly 9-kilometer resolution air temperature, radiation, and soil moisture data. Data preprocessing includes vector clipping and resampling to construct the ERA5-Land reanalysis data database. By constructing a database of MODIS remote sensing products and ERA5-Land reanalysis data, spatiotemporal matching and preprocessing were performed on the data. Temporal matching specifically involved extracting meteorological and radiation reanalysis data corresponding to the observation time from the hourly ERA5-Land data at the pixel scale, based on the MODIS observation time. Spatial matching involved resampling the ERA5-Land reanalysis data from its original 9km resolution to a 1km resolution, the same as the MODIS land surface temperature data. Data preprocessing included quality control of the MODIS land surface temperature data and quality control between the MODIS and ERA5-Land data.

[0056] Furthermore, step S2 in this embodiment includes: using the spatiotemporal matching and preprocessed MODIS remote sensing data and reanalysis data, and utilizing MODIS observation zenith angle information, extracting data such as zenith observation (vertical observation) surface temperature from the database.

[0057] Preferably, the training dataset includes vertical reflectance data, leaf area index data, vegetation cover data, radiation / air temperature / soil moisture reanalysis data, and vertical surface temperature data.

[0058] Preferably, the method for constructing the training dataset includes:

[0059] Vertical reflectance data, leaf area index data, vegetation cover data, radiation / air temperature / soil moisture reanalysis data, and vertical surface temperature data were split into training and validation sets using random sampling; the ratio of the amount of data in the training set to the amount of data in the validation set was 8:2.

[0060] Specifically, step S2 in this embodiment further includes: constructing a training dataset for a global-scale vertical surface temperature random forest model. The dataset includes five types of data: vertical reflectance data, leaf area index data, vegetation cover data, reanalysis data of radiation / air temperature / soil moisture, and vertical surface temperature data. The training dataset is constructed by randomly sampling the five types of datasets into a training dataset and a validation dataset, with a training data:validation data ratio of 8:2. The training data is used to train the random forest model and test the optimal parameters of the model. The validation data is used to verify the accuracy of the trained random forest model. When the model accuracy remains stable and meets the requirements, the model training is complete.

[0061] As an optional implementation, step S3 in this embodiment includes: establishing a global-scale vertical land surface temperature estimation model based on a training dataset using a random forest machine learning algorithm. This model represents the functional relationship f between satellite zenith observation direction (vertical observation) remote sensing land surface temperature data and vegetation (leaf area index, vegetation cover) and reanalysis data (air temperature, solar radiation, soil moisture, etc.).

[0062] T Nadir =f(TA) Nadir SM Nadir LAI Nadir NDVI Nadir ,SSRD Nadir (2)

[0063] In the above formula, the subscript Nadir represents the MODIS vertical observation time, T represents the surface temperature of the remote sensing image at the MODIS vertical observation time, TA is the air temperature, SM is the soil moisture, LAI is the leaf area index, NDVI is the normalized vegetation index, and SSRD is the solar downwave radiation.

[0064] Preferably, the MODIS vertical surface temperature estimation model based on the global scale is used to estimate the corresponding MODIS vertical surface temperature data using remote sensing data from MODIS non-vertical observation times and ERA5-Land reanalysis data, i.e.:

[0065] Based on the global-scale vertical surface temperature estimation model, the MODIS vertical surface temperature data corresponding to the MODIS non-vertical observation time is estimated using remote sensing data and reanalysis data from MODIS non-vertical observation times.

[0066] Specifically, in this embodiment, step S4 includes: estimating the corresponding MODIS vertical surface temperature data based on the global-scale vertical surface temperature estimation model f, using remote sensing data and reanalysis data from MODIS non-vertical observation times, as shown in the following formula:

[0067] To ff Nadir =f(TAo ff SMo ff ,LAIo ff NDVIo ff Nadir ,SSRDo ff ,....)(3)

[0068] In the above formula, the subscript off represents the MODIS non-vertical observation time, the superscript Nadir represents the vertical remote sensing data, T represents the estimated vertical surface temperature at the MODIS non-vertical observation time, TA is the air temperature, SM is the soil moisture, LAI is the leaf area index, NDVI is the normalized vegetation index, SSRD is the solar downwave radiation, and f is the global-scale vertical surface temperature estimation model.

[0069] Based on the above steps, this embodiment obtains global-scale MODIS vertical observation land surface temperature products, that is, obtains global MODIS land surface temperature products with consistent angles.

[0070] The beneficial effects of this invention are as follows:

[0071] (1) This invention fills the gap in the current lack of global-scale vertical surface temperature products;

[0072] (2) This invention solves the scientific problem of the angle effect in current satellite observations of Earth's surface temperature;

[0073] (3) This invention enhances the application value of surface temperature products in climate change, resource and environmental monitoring, and surface energy balance.

[0074] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0075] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for developing a globally consistent MODIS surface temperature product, characterized in that, include: A database of MODIS remote sensing data and ERA5-Land reanalysis data was constructed based on the acquired initial global-scale MODIS remote sensing data and initial reanalysis data. Using MODIS observation zenith angle information, remote sensing data and reanalysis data of the MODIS zenith observation direction are extracted from the database of MODIS remote sensing data and ERA5-Land reanalysis data, and a training dataset is constructed based on the remote sensing data and reanalysis data of the MODIS zenith observation direction. Based on the training dataset, a global-scale vertical surface temperature estimation model was established using the random forest machine learning algorithm. Based on the global-scale vertical surface temperature estimation model, the corresponding MODIS vertical surface temperature data are estimated using remote sensing data from MODIS non-vertical observation times and ERA5-Land reanalysis data. The global-scale vertical surface temperature estimation model is a functional relationship f between satellite zenith observation direction remotely sensed surface temperature data and vegetation and reanalysis data: ； Wherein, the subscript Nadir represents the MODIS vertical observation time, T represents the surface temperature of the remote sensing image at the MODIS vertical observation time, TA is the air temperature, SM is the soil moisture, LAI is the leaf area index, NDVI is the normalized vegetation index, and SSRD is the solar downwave radiation. The global-scale vertical surface temperature estimation model uses remote sensing data from MODIS non-vertical observation times and ERA5-Land reanalysis data to estimate the corresponding MODIS vertical surface temperature data, expressed by the following formula: ； Wherein, the subscript off represents the MODIS non-vertical observation time, the superscript Nadir represents the vertical remote sensing data, T represents the estimated vertical land surface temperature at the MODIS non-vertical observation time, TA is the air temperature, SM is the soil moisture, LAI is the leaf area index, NDVI is the normalized difference vegetation index, SSRD is the solar downwave radiation, and f is the global-scale vertical land surface temperature estimation model.

2. The method for developing a globally consistent MODIS surface temperature product according to claim 1, characterized in that, The database of MODIS remote sensing data and ERA5-Land reanalysis data, constructed based on the acquired initial global-scale MODIS remote sensing image data and initial reanalysis data, includes: Acquire the initial global-scale remote sensing image data and the initial reanalysis data; Spatiotemporal matching and preprocessing are performed on the initial global-scale remote sensing image data and the initial reanalysis data to obtain preprocessed data; A database of MODIS remote sensing data and ERA5-Land reanalysis data is constructed based on the preprocessed data.

3. The method for developing a globally consistent MODIS surface temperature product according to any one of claims 2, characterized in that, The initial global-scale remote sensing image data comes from the MODIS / Terra+Aqua satellite and includes MOD11A1 / MYD11A1 surface temperature data, MCD15A3H leaf area index data, MCD43A4 reflectance data, and MOD44B vegetation cover data. The initial reanalysis data comes from the hourly dataset of ERA5-Land and includes air temperature, radiation, and soil moisture data.

4. The method for developing a globally consistent MODIS surface temperature product according to claim 2, characterized in that, The spatiotemporal matching steps include: At the pixel scale, reanalysis data corresponding to the MODIS observation time is extracted from the initial reanalysis data based on the MODIS observation time; The extracted reanalysis data were resampled from the original 9km resolution to a 1km resolution, the same as the MODIS surface temperature.

5. The method for developing a globally consistent MODIS surface temperature product according to claim 2, characterized in that, The preprocessing includes quality control of MODIS surface temperature data and quality control between initial global-scale remote sensing image data and initial reanalysis data.

6. The method for developing a globally consistent MODIS surface temperature product according to claim 1, characterized in that, The training dataset includes vertical reflectance data, leaf area index data, vegetation cover data, radiation / air temperature / soil moisture reanalysis data, and vertical surface temperature data.

7. The method for developing a globally consistent MODIS surface temperature product according to claim 1, characterized in that, The method for constructing the training dataset includes: Vertical reflectance data, leaf area index data, vegetation cover data, radiation / air temperature / soil moisture reanalysis data, and vertical surface temperature data were split into training and validation sets using random sampling; the ratio of the amount of data in the training set to the amount of data in the validation set was 8:2.

Images