Actual evapotranspiration calculation method based on multi-source remote sensing data

Abstract

The invention discloses an actual evapotranspiration calculation method based on multi-source remote sensing data. Comprising the following steps: S1, acquiring multi-source remote sensing data; S2, based on an energy balance principle; S3, adopting a dynamic data fusion algorithm based on machine learning; S4, constructing an actual evapotranspiration calculation model adapted by regions; the method can accurately estimate the actual evapotranspiration of the heterogeneous region through multi-source remote sensing data fusion and a dynamic calculation model, is suitable for scenes such as agricultural areas, wetlands, urban green lands and forests, keeps high calculation precision and adaptability under various weather conditions, and has a wide application prospect. The problems of low resolution, data missing, lack of regional adaptability, low calculation efficiency and the like in the prior art are effectively solved, the precision, adaptability and efficiency of evapotranspiration calculation are improved, and more efficient technical support is provided for large-scale water resource management, ecological assessment and agricultural production.

Description

Technical Field

[0001] The invention relates to the field of remote sensing and hydrological technology, and in particular to a method for calculating actual evapotranspiration based on multi-source remote sensing data. Background Art

[0002] Actual evapotranspiration (AET) is an important part of the surface hydrological cycle, which directly affects regional water resources management, agricultural production and ecosystem function assessment. Traditional AET calculation methods usually rely on meteorological station data (such as the Penman-Monteith formula) or single remote sensing data (such as MODIS inversion data).

[0003] However, the above methods have the following problems:

[0004] 1. Most remote sensing data have low resolution (such as MODIS 1km), which makes it difficult to capture the evapotranspiration characteristics of small-scale areas with strong heterogeneity (such as agricultural irrigation areas or urban green spaces);

[0005] 2. Remote sensing data are often missing due to cloud cover, sensor failure or other factors, especially in real-time estimation, which affects the continuity of time and space, resulting in reduced reliability of evapotranspiration estimation results;

[0006] 3. Existing computational models usually rely on unified parameters and cannot be adapted to different surface types (such as agricultural areas, wetlands, urban green spaces, etc.), which limits their application in complex surface environments. In addition, current methods have slow calculation speeds when processing large-scale areas and large-scale remote sensing data, making it difficult to meet the needs of real-time or quasi-real-time estimation.

[0007] Therefore, a new method is urgently needed to effectively improve the accuracy, adaptability and efficiency of actual evapotranspiration estimation, especially in large-scale areas and dynamic climate conditions. To solve the above problems, we propose a method for calculating actual evapotranspiration based on multi-source remote sensing data. Summary of the invention

[0008] In view of the deficiencies in the prior art, the present invention provides a method for calculating actual evapotranspiration based on multi-source remote sensing data to solve the problems raised in the background technology.

[0009] The above technical objectives of the present invention are achieved through the following technical solutions:

[0010] A method for calculating actual evapotranspiration based on multi-source remote sensing data comprises the following steps:

[0011] S1. Obtain multi-source remote sensing data, including surface temperature, vegetation index, soil moisture data and meteorological data;

[0012] S2. Based on the energy balance principle, calculate the surface net radiation flux, surface heat flux and latent heat flux;

[0013] S3. For the acquired multi-source remote sensing data, a dynamic data fusion algorithm based on machine learning is used to perform temporal and spatial interpolation processing to solve the problem of missing data;

[0014] S4. Construct a regionally adapted actual evapotranspiration calculation model, divide evapotranspiration into two parts: evaporation and plant transpiration, and perform dynamic calculations based on the dual-source energy balance model;

[0015] S5. Combine the fusion results of multi-source remote sensing data and the actual evapotranspiration calculation results to generate a high-resolution actual evapotranspiration spatial distribution map and time dynamic change curve.

[0016] Preferably, the multi-source remote sensing data acquired in S1 includes:

[0017] Optical remote sensing data for obtaining surface reflectance, vegetation index and surface temperature;

[0018] Microwave remote sensing data for obtaining surface soil moisture;

[0019] Meteorological data, including precipitation, temperature, wind speed, relative humidity, and shortwave radiation intensity;

[0020] High-resolution remote sensing images are used to improve the expression of spatial details after data fusion.

[0021] Preferably, the surface energy balance equation based on the energy balance principle in S2 is:

[0022]

[0023] in, is the net radiation, is the surface heat flux, is the sensible heat flux, is the latent heat flux; latent heat flux Used to characterize actual evapotranspiration.

[0024] Preferably, the dynamic data fusion algorithm based on machine learning is adopted in S3, including:

[0025] Temporal interpolation: Repairing missing remote sensing data due to cloud cover or sensor failure through a time series model based on a long short-term memory network (LSTM);

[0026] Spatial interpolation: Combine terrain features, vegetation coverage and meteorological data to perform regional feature fitting and repair of local missing points;

[0027] Fusion algorithm: Based on the complementary characteristics of optical remote sensing data and microwave remote sensing data, a multi-source data fusion set with higher consistency is generated.

[0028] Preferably, a dual-source energy balance model is used in S4 to decompose the actual evapotranspiration into:

[0029] Soil evaporation: calculated by combining surface temperature and soil moisture data with the evaporation rate model;

[0030] Transpiration: Dynamically calculate the contribution of vegetation transpiration through vegetation index, meteorological data and vegetation type distribution, and optimize regional model parameters.

[0031] Preferably, the actual evapotranspiration calculation model in S4 is combined with real-time meteorological driving data to improve the accuracy and adaptability of calculation under extreme climate conditions by updating dynamic data such as precipitation, temperature, wind speed, etc.

[0032] Preferably, the output result in S5 includes:

[0033] High-resolution spatial distribution maps of actual evapotranspiration, with spatial resolutions ranging from 30 to 10 meters;

[0034] The temporal dynamic change curve of evapotranspiration, with a time resolution of daily or hourly;

[0035] Total evapotranspiration statistics for different regions, summarized by vegetation type and surface characteristics.

[0036] Preferably, a parallel computing framework is used to process multi-source remote sensing data, and a large-scale data optimization algorithm is used to improve computing efficiency to meet the needs of real-time actual evapotranspiration estimation in a large area.

[0037] Preferably, the method is suitable for highly heterogeneous surface types such as agricultural areas, wetlands, urban green spaces and forests, and can enhance its adaptability to surface characteristics of different regions by optimizing parameter settings.

[0038] In summary, the present invention mainly has the following beneficial effects:

[0039] 1. By fusing multi-source remote sensing data (such as optical and microwave remote sensing data) and meteorological data, this method can achieve high computational accuracy in complex surface environments, and adopt a dynamic data fusion algorithm based on machine learning to effectively repair data missing problems caused by cloud coverage and sensor failures, ensuring data integrity and consistency.

[0040] 2. This method adopts a regional adaptive calculation model and optimizes parameter settings according to different surface types to improve its applicability under diverse surface types. It is especially suitable for highly heterogeneous surface types such as agricultural areas, wetlands and urban green spaces. Through a parallel computing framework and large-scale data optimization algorithms, this method significantly improves the calculation efficiency and meets the needs of real-time actual evapotranspiration estimation in a large area.

[0041] 3. Combined with real-time meteorological data updates, this method can flexibly adjust the calculation model under extreme climate conditions, generate high-resolution actual evapotranspiration distribution maps and time dynamic change curves, and provide a reliable basis for water resources management and ecological assessment. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution of the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the described embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work belong to the scope of protection of the present invention.

[0043] The following examples are used to illustrate the present invention, but they cannot be used to limit the scope of protection of the present invention. The conditions in the examples can be further adjusted according to specific conditions. Simple improvements to the method of the present invention under the premise of the concept of the present invention belong to the scope of protection claimed in the present invention.

[0044] Example 1: Method for estimating actual evapotranspiration in agricultural areas based on multi-source remote sensing data

[0045] Step S1: Acquiring multi-source remote sensing data

[0046] An agricultural area was selected as the research area, and the following multi-source remote sensing data were collected:

[0047] Optical remote sensing data: Using Sentinel-2 images, we can obtain surface reflectance, vegetation index (NDVI) and surface temperature with a spatial resolution of 10 meters;

[0048] Microwave remote sensing data: Soil moisture is obtained using the SMAP satellite, with an original spatial resolution of 36 kilometers. It is downscaled to 10 meters resolution by combining optical remote sensing data and terrain features;

[0049] Meteorological data: hourly precipitation, temperature, wind speed, relative humidity, and shortwave radiation intensity are obtained from weather stations and reanalysis datasets;

[0050] High-resolution remote sensing images: 1-meter resolution images are obtained through drone multispectral images to enhance the expression of spatial details.

[0051] Step S2: Surface flux calculation based on energy balance principle

[0052] The net radiation flux on the surface is calculated according to the following formula , surface heat flux , sensible heat flux and latent heat flux :

[0053] Surface net radiation calculation:

[0054]

[0055] in is the surface reflectivity, is the shortwave radiation intensity, is the long-wave radiation flux;

[0056] Sensible heat flux calculation:

[0057]

[0058] in is the difference between the surface temperature and the air temperature, is the vertical temperature gradient, , is a constant;

[0059] Latent heat flux calculation:

[0060]

[0061] Step S3: Dynamic data fusion based on machine learning

[0062] Temporal interpolation: Use the long short-term memory network (LSTM) model to complete the time series of optical remote sensing data and repair image missing caused by cloud cover;

[0063] Spatial interpolation: Generate higher-precision spatial data by combining vegetation cover, soil moisture and meteorological characteristics through support vector machine regression model;

[0064] Data fusion: Based on the complementary characteristics of optical and microwave data, a Bayesian optimization algorithm is used to generate a more consistent multi-source data set.

[0065] Step S4: Regional evapotranspiration calculation model

[0066] Actual evapotranspiration is decomposed into soil evaporation and vegetation transpiration :

[0067] Soil evaporation: calculated from surface temperature and soil moisture data:

[0068]

[0069] Vegetation transpiration: , meteorological data and vegetation type distribution to calculate dynamic transpiration:

[0070]

[0071] Dynamically update model parameters by combining real-time meteorological driving data (precipitation, temperature, wind speed).

[0072] Step S5: Result output

[0073] Generate actual evapotranspiration distribution map with 10-meter spatial resolution;

[0074] Output hourly dynamic change curves to analyze the time trend of evapotranspiration;

[0075] Statistics on the total amount of evapotranspiration in different regions and vegetation types provide data support for the optimization of agricultural water use.

[0076] Example 2: Estimation of actual evapotranspiration from wetlands

[0077] According to the unique water reflection characteristics of wetlands, the dual-source energy balance model is optimized:

[0078] Step S2: Introduce the wetland water reflectivity correction factor and adjust the net radiation calculation formula:

[0079]

[0080] Step S4: adding a stomatal conductance calculation module to dynamically simulate the transpiration characteristics of wetland vegetation;

[0081] The actual evapotranspiration results of wetlands show the evapotranspiration distribution of water bodies and vegetation at a higher resolution, providing a reference for wetland protection and management.

[0082] Example 3: Estimation of evapotranspiration from urban green spaces

[0083] For urban areas:

[0084] Step S3: Optimize the data interpolation algorithm by using high-resolution remote sensing data (1 meter resolution) combined with urban thermal effect parameters;

[0085] Step S4: Construct a dual-source energy balance model that includes the thermal effect of urban buildings and dynamically calculate the evapotranspiration of urban green spaces and bare surfaces.

[0086] Example 4: Estimation of evapotranspiration from forest ecosystems

[0087] Forests have a complex hierarchical structure and high evapotranspiration potential:

[0088] Step S1: Combining high-resolution optical images and lidar data to obtain forest canopy height and structure information;

[0089] Step S4: Decompose vegetation transpiration into contributions from different layers (canopy layer, surface layer), and dynamically optimize the parameters of the forest ecosystem model;

[0090] The results show the distribution characteristics of evapotranspiration at different levels of the forest, providing support for forest ecological management.

[0091] Example 5: Adaptability analysis based on extreme climate conditions

[0092] In extreme weather conditions (such as drought or heavy rainfall):

[0093] Step S4: Modify the model parameters by introducing real-time meteorological driving data (precipitation, wind speed, temperature) to enhance the adaptability of the model;

[0094] The outputs show the changes in evapotranspiration under extreme conditions on an hourly and daily basis, assisting disaster management and water resources planning.

[0095] Working principle: The present invention can accurately estimate the actual evapotranspiration in heterogeneous areas through multi-source remote sensing data fusion and dynamic calculation model. It is suitable for scenes such as agricultural areas, wetlands, urban green spaces and forests, and maintains high calculation accuracy and adaptability under various climatic conditions. The present invention effectively solves the problems of low resolution, missing data, lack of regional adaptability, low calculation efficiency and so on in the prior art, improves the accuracy, adaptability and efficiency of evapotranspiration calculation, and provides more efficient technical support for large-scale water resource management, ecological assessment and agricultural production.

[0096] Although the embodiments of the present invention have been shown and described, it is understood by those skilled in the art that, unless otherwise defined, the technical terms or scientific terms used in the present invention should have the usual meanings understood by persons with ordinary skills in the field to which the present invention belongs, and the words "include" or "comprise" and the like used in the present invention mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents.

[0097] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for calculating actual evapotranspiration based on multi-source remote sensing data, characterized in that: The following steps are involved: S1. Obtain multi-source remote sensing data, including surface temperature, vegetation index, soil moisture data and meteorological data; S2. Based on the energy balance principle, calculate the surface net radiation flux, surface heat flux and latent heat flux; S3. For the acquired multi-source remote sensing data, a dynamic data fusion algorithm based on machine learning is used to perform temporal and spatial interpolation processing to solve the problem of missing data; S4. Construct a regionally adapted actual evapotranspiration calculation model, divide evapotranspiration into two parts: evaporation and plant transpiration, and perform dynamic calculations based on the dual-source energy balance model; S5. Combine the fusion results of multi-source remote sensing data and the actual evapotranspiration calculation results to generate a high-resolution actual evapotranspiration spatial distribution map and time dynamic change curve.

2. The actual evapotranspiration calculation method based on multi-source remote sensing data according to claim 1, characterized in that: The multi-source remote sensing data acquired in S1 includes: Optical remote sensing data for obtaining surface reflectance, vegetation index and surface temperature; Microwave remote sensing data for obtaining surface soil moisture; Meteorological data, including precipitation, temperature, wind speed, relative humidity, and shortwave radiation intensity; High-resolution remote sensing images are used to improve the expression of spatial details after data fusion.

3. The method for calculating actual evapotranspiration based on multi-source remote sensing data according to claim 1, characterized in that: The surface energy balance equation based on the energy balance principle in S2 is: ； in, is the net radiation, is the surface heat flux, is the sensible heat flux, is the latent heat flux; Latent heat flux Used to characterize actual evapotranspiration.

4. The method for calculating actual evapotranspiration based on multi-source remote sensing data according to claim 1, characterized in that: The S3 uses a dynamic data fusion algorithm based on machine learning, including: Temporal interpolation: Repairing missing remote sensing data due to cloud cover or sensor failure through a time series model based on a long short-term memory network (LSTM); Spatial interpolation: Combine terrain features, vegetation coverage and meteorological data to perform regional feature fitting and repair of local missing points; Fusion algorithm: Based on the complementary characteristics of optical remote sensing data and microwave remote sensing data, a multi-source data fusion set with higher consistency is generated.

5. The method for calculating actual evapotranspiration based on multi-source remote sensing data according to claim 1, characterized in that: The dual-source energy balance model is used in S4 to decompose the actual evapotranspiration into: Soil evaporation: calculated using surface temperature and soil moisture data combined with an evaporation rate model; Plant transpiration: Dynamically calculate the contribution of vegetation transpiration through vegetation index, meteorological data and vegetation type distribution, and optimize regional model parameters.

6. The method for calculating actual evapotranspiration based on multi-source remote sensing data according to claim 1, characterized in that: The actual evapotranspiration calculation model in S4 is combined with real-time meteorological driving data to improve the accuracy and adaptability of calculations under extreme climate conditions by updating dynamic data such as precipitation, temperature, and wind speed.

7. The method for calculating actual evapotranspiration based on multi-source remote sensing data according to claim 1, characterized in that: The output results in S5 include: High-resolution spatial distribution maps of actual evapotranspiration, with spatial resolutions ranging from 30 to 10 meters; The temporal dynamic change curve of evapotranspiration, with a time resolution of daily or hourly; Total evapotranspiration statistics for different regions, summarized by vegetation type and surface characteristics.

8. The method for calculating actual evapotranspiration based on multi-source remote sensing data according to claim 1, characterized in that: A parallel computing framework is used to process multi-source remote sensing data, and the computing efficiency is improved based on large-scale data optimization algorithms to meet the needs of real-time actual evapotranspiration estimation in large areas.

9. The method for calculating actual evapotranspiration based on multi-source remote sensing data according to claim 1, characterized in that: This method is suitable for highly heterogeneous surface types such as agricultural areas, wetlands, urban green spaces and forests, and can enhance its adaptability to surface characteristics of different regions by optimizing parameter settings.

Images