Evapotranspiration satellite remote sensing inversion method directly considering soil humidity stress

Abstract

The invention discloses an evapotranspiration satellite remote sensing inversion method directly considering soil humidity stress, and the method comprises the steps: firstly constructing an improved canopy water vapor conductivity model, and then enabling the conductivity model to be coupled into a canopy transpiration module in a P-LSH algorithm, and obtaining an improved canopy transpiration algorithm; then analyzing the sensitivity and uncertainty of framework evaluation parameters, constructing a most effective humidity stress model suitable for bare soil evaporation in a bare soil area, and coupling the humidity stress model into a soil evaporation mode in a P-LSH algorithm to form an improved bare soil evaporation algorithm; and finally, coupling the improved canopy transpiration algorithm and the improved bare soil evaporation algorithm in combination with a water evaporation module in the P-LSH algorithm to obtain a global evapotranspiration remote sensing inversion algorithm directly considering soil humidity stress. According to the method, the stress effect of soil humidity on evapotranspiration is considered, a global evapotranspiration remote sensing inversion algorithm is developed, and a global evapotranspiration product with high precision and high temporal-spatial resolution is generated.

Description

Technical Field

[0001] The invention belongs to the field of remote sensing hydrology, and in particular relates to an evapotranspiration satellite remote sensing inversion method that directly considers soil moisture stress. Background Art

[0002] Evapotranspiration is an important link between the water cycle and the energy cycle. Accurately quantifying the impact of soil water supply on evapotranspiration can promote the development and improvement of evapotranspiration inversion models. However, current evapotranspiration algorithms based on satellite remote sensing inversion rarely directly consider soil moisture stress, which leads to great uncertainty in the inversion results in water-restricted areas. Considering the strong heterogeneity of global climate and land cover, how to develop a globally applicable soil moisture stress scheme has become a key difficulty in current research.

[0003] From the physical mechanism of transpiration, due to the water potential gradient between soil, leaves and air, water is absorbed by the soil into the plant roots, then transferred to the leaves, and finally dissipated into the atmosphere through the leaf stomata. Therefore, soil moisture is the most direct source of water supply for vegetation, which controls the amount of water that can be extracted from the soil by the plant roots. Some current studies believe that the saturated water vapor pressure difference and soil moisture are closely related through land-air interaction, so the current existing remote sensing evapotranspiration algorithms usually use the saturated water vapor pressure difference to reflect the humidity stress of canopy conductance. However, there is a time propagation and lag effect between atmospheric water vapor humidity and soil moisture, and in the context of global warming and climate change, there may be a partial decoupling between water vapor humidity and soil moisture. Therefore, it is necessary to adopt a more direct soil moisture stress structure in the estimation of daily evapotranspiration.

[0004] Although soil moisture stress has gradually gained attention, incorporating soil moisture stress into remote sensing evapotranspiration algorithms remains a great challenge. The difficulty lies in how to develop a robust and globally applicable soil moisture stress scheme to reasonably quantify water surplus or deficit. Due to the limitations of moisture stress schemes and soil moisture data availability, it is difficult for current studies to generate long-term evapotranspiration datasets. In addition, existing soil moisture stress schemes mostly use soil properties as stress thresholds, but the large deviations in global soil attribute data have a great impact on evapotranspiration estimates. Summary of the invention

[0005] In view of the problems existing in the prior art, the present invention provides a satellite remote sensing inversion method for evapotranspiration that directly considers soil moisture stress. Starting from the vegetation canopy conductance algorithm and the bare soil evaporation algorithm, a soil moisture stress scheme suitable for various ecosystems and climate zones around the world is developed, and the optimal mathematical expression of soil evaporation water supply constraints is improved, thereby forming a robust satellite remote sensing inversion method for evapotranspiration that directly considers soil moisture stress. Based on the algorithm, a global 1 / 12° resolution daily-scale evapotranspiration dataset from 1982 to 2022 is provided.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a satellite remote sensing inversion method for evapotranspiration that directly considers soil moisture stress, comprising the following steps: S1. An improved canopy water vapor conductivity model is constructed based on the quantile canopy conductivity soil moisture stress factor, and then the canopy water vapor conductivity model is coupled into the canopy transpiration module in the P-LSH algorithm to obtain an improved canopy transpiration algorithm; S2. A Bayesian-based uncertainty analysis framework is used to quantify parameter uncertainty in the improved canopy transpiration algorithm. S3, constructing the most effective humidity stress model for bare soil evaporation in bare soil areas, coupling the humidity stress model into the soil evaporation mode in the P-LSH algorithm, and forming an improved bare soil evaporation algorithm; S4. Combine the water evaporation module in the P-LSH algorithm, couple the improved canopy evaporation algorithm in step S1 and the improved bare soil evaporation algorithm in step S3 to obtain a global evapotranspiration remote sensing inversion algorithm that directly considers soil moisture stress.

[0007] Furthermore, in the aforementioned step S1, based on the actual soil moisture content, a quantile-based canopy conductance soil moisture stress factor is constructed as follows: (1) (2) in: m ( SM ) is the soil moisture stress factor, SM is the actual surface soil volume water content; SM min is the minimum soil moisture content, SM c is the critical soil moisture content; n is the quantile reflecting the critical soil moisture content, which is obtained by calibration of the observed value; f is the quantile mapping function.

[0008] Furthermore, in the aforementioned step S1, the constructed soil moisture stress factor is incorporated into the Jarvis-Stewart equation to construct an improved canopy water vapor conductivity model, as shown in the following formula: (3) (4) in: is the canopy water vapor conductivity; Determined by the ecosystem type The maximum value of is defined as the sigmoid function of NDVI; b1, b2, and b3 are parameters affected by ecosystem type; m(i) is the stress factor of the i-th climate variable; i is the Climate variables that produce stressful effects.

[0009] Furthermore, in the aforementioned step S1, the improved canopy transpiration algorithm is as follows: (5) in: is the canopy latent heat flux, is the latent heat of evaporation; A is the slope of the air saturated water vapor pressure and temperature curve; c is the effective energy component allocated to the canopy; VPD is the saturated vapor pressure difference; is the air density; C p is the specific heat capacity of air; is the psychrometric constant; is the canopy aerodynamic conductivity; is the canopy water vapor conductivity.

[0010] Furthermore, the aforementioned step S2 includes the following sub-steps: S2.1. The Sobol' parameter sensitivity analysis method based on the variance decomposition principle was used to quantify the sensitivity of each parameter. The total effect index greater than 10% was used as an indicator to identify the sensitive parameters in the improved algorithm, as follows: (6) (7) Where: V i is the parameter X i The first-order sensitivity of the partial variance; V ij is the partial variance of the second-order sensitivity of the interaction between the i-th and j-th parameters; is the d-order sensitivity partial variance of all parameter interactions; is a parameter set consisting of d parameters, and Y is the selected univariate model output; S2.2. Use the differential evolution algorithm-Markov chain model to perform uncertainty analysis on the sensitive parameters in the improved algorithm and the non-sensitive parameters whose prior values ​​cannot be determined, determine the posterior distribution of each parameter, and take the median of the posterior distribution as the optimal parameter value, as follows: (8) (9) Where: x is the sample; X is the parameter set; is the prior distribution, which is set to uniform distribution; is the likelihood function; T is the total number of observed samples; O is the observed value; S is the simulated value; σ is the standard deviation of the model error.

[0011] Furthermore, the aforementioned step S3 includes the following sub-steps: S3.1. Based on the actual soil moisture content, the most effective humidity stress model for bare soil evaporation in bare soil areas is constructed as follows: (10) in: f is the humidity stress on soil evaporation; SM max (m 3 m -3 ) and SM min (m 3 m -3 ) are the maximum and minimum values ​​of the SM sequence during the study period, respectively; S3.2. The constructed bare soil moisture stress scheme is coupled into the soil evaporation model in the P-LSH algorithm to form an improved bare soil evaporation algorithm as follows: (11) Among them: A s is the effective energy distributed to the soil, is the aerodynamic conductivity of the soil surface; is the total aerodynamic conductivity corrected for actual air temperature and pressure, is the psychrometric constant, is the air density.

[0012] Furthermore, the aforementioned evapotranspiration satellite remote sensing inversion method that directly considers soil moisture stress includes: taking the evapotranspiration data of flux tower observations and water balance reconstruction as a benchmark, and evaluating the evapotranspiration accuracy of the global evapotranspiration remote sensing inversion algorithm that directly considers soil moisture stress at the station and basin scales on a daily, monthly, annual and average scale of several years.

[0013] Furthermore, the aforementioned satellite remote sensing inversion method for evapotranspiration that directly considers soil moisture stress includes inverting a global 1 / 12° resolution daily-scale evapotranspiration dataset from 1982 to 2022 using a satellite remote sensing inversion method for evapotranspiration that directly considers soil moisture stress, specifically: Step A: Collect and preprocess various global meteorological driving data, including temperature, specific humidity, radiation, soil moisture, and NDVI data, and resample them to a 1 / 12° resolution; Step B: A global evapotranspiration remote sensing inversion algorithm that directly considers soil moisture stress is driven by various meteorological data to generate a global evapotranspiration dataset with a daily scale and a resolution of 1 / 12° from 1982 to 2022.

[0014] Compared with the prior art, the beneficial technical effects of the present invention using the above technical solution are as follows: The present invention develops a satellite remote sensing inversion method for evapotranspiration that directly considers soil moisture stress, which specifically includes the construction of soil moisture stress schemes for canopy conductance under various ecosystems, the improvement of the most effective moisture stress scheme in bare soil evaporation, and the realization of remote sensing inversion of global evapotranspiration. Based on the flux observation data of various ecosystems around the world, a soil moisture stress scheme for canopy conductance is constructed, and evapotranspiration is estimated in combination with the P-LSH algorithm; the sensitivity and uncertainty of parameters are evaluated using an uncertainty analysis framework based on Bayesian theory, and the sensitive parameters of the improved algorithm are identified and optimized; the optimal mathematical expression of soil evaporation water supply constraints in the bare soil evaporation algorithm is improved, and more direct soil moisture is used instead of atmospheric humidity to quantify the stress effect of humidity on bare soil evaporation; the degree of soil moisture stress in each partition is quantified, and the improved canopy evapotranspiration, improved bare soil evaporation, and water evaporation based on wind speed and improved Penman equation are coupled to form the P-LSHv2 algorithm; the remote sensing inversion of global evapotranspiration is realized based on the P-LSHv2 algorithm, and a global 1 / 12° resolution daily scale evapotranspiration data set from 1982 to 2022 is generated. This method takes into account the stress of soil moisture on evapotranspiration, and uses remote sensing and reanalysis data as basic data, with a stable and reliable data source. A satellite remote sensing inversion method for evapotranspiration that directly considers soil moisture stress has been developed. Based on the classic remote sensing model, the objectivity and rationality of the results are guaranteed, and the remote sensing inversion of global evapotranspiration is realized to obtain evapotranspiration products with high precision and high temporal and spatial resolution. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a flow chart of a satellite remote sensing inversion method (P-LSHv2) of evapotranspiration that directly considers soil moisture stress in the present invention.

[0016] Figure 2 It is a schematic diagram for calculating the soil moisture stress factor in the present invention.

[0017] Figure 3 It is a box plot of the posterior distribution of parameters of the method of the present invention.

[0018] Figure 4 It is a diagram of the evaporation result estimated by the flux tower in the present invention. DETAILED DESCRIPTION

[0019] In order to better understand the technical content of the present invention, specific embodiments are given and described as follows in conjunction with the accompanying drawings.

[0020] Various aspects of the invention are described herein with reference to the accompanying drawings, in which many illustrative embodiments are shown. The embodiments of the invention are not limited to those described in the accompanying drawings. It should be understood that the invention is implemented by any of the various concepts and embodiments described above, as well as the concepts and embodiments described in detail below, because the concepts and embodiments disclosed in the invention are not limited to any implementation. In addition, some aspects disclosed in the invention may be used alone or in any appropriate combination with other aspects disclosed in the invention.

[0021] like Figure 1 As shown, the present invention provides a satellite remote sensing inversion method for evapotranspiration that directly considers soil moisture stress, comprising the following steps: S1. Based on the classic Jarvis-Stewart equation, an improved canopy water vapor conductivity model is constructed based on the quantile canopy conductivity soil moisture stress factor, and then the canopy water vapor conductivity model is coupled into the canopy transpiration module in the P-LSH algorithm to obtain an improved canopy transpiration algorithm; Based on the actual soil moisture content, the quantile-based canopy conductance soil moisture stress factor was constructed as follows: (1) (2) in: m ( SM ) is the soil moisture stress factor, SM is the actual surface soil volume water content; SM min is the minimum soil moisture content, SM c is the critical soil moisture content; n is the quantile reflecting the critical soil moisture content, which is obtained by calibration of the observed value; f is the quantile mapping function.

[0022] The constructed soil moisture stress factor is incorporated into the Jarvis-Stewart equation to construct an improved canopy water vapor conductivity model as follows: (3) (4) in: is the canopy water vapor conductivity; Determined by the ecosystem type The maximum value of is defined as the sigmoid function of NDVI; b1, b2, and b3 are parameters affected by ecosystem type; m(i) is the stress factor of the i-th climate variable; i is the The present invention adds the constructed soil moisture stress factor to the climate variable that produces stress, so as to characterize the stress process of soil moisture on canopy conductance.

[0023] Then the canopy water vapor conductivity model is coupled into the canopy transpiration module in the P-LSH algorithm to obtain the improved canopy transpiration algorithm: (5) in: is the canopy latent heat flux, is the latent heat of evaporation; A is the slope of the air saturated water vapor pressure and temperature curve; c is the effective energy component allocated to the canopy; VPD is the saturated vapor pressure difference; is the air density; C p is the specific heat capacity of air; is the psychrometric constant; is the canopy aerodynamic conductivity; is the canopy water vapor conductivity.

[0024] S2. Quantify the parameter uncertainty in the improved canopy transpiration algorithm using a Bayesian uncertainty analysis framework, including the following sub-steps: S2.1. The Sobol' parameter sensitivity analysis method based on the variance decomposition principle was used to quantify the sensitivity of each parameter. The total effect index greater than 10% was used as an indicator to identify the sensitive parameters in the improved algorithm, as follows: (6) (7) Where: V i is the parameter X i The first-order sensitivity of the partial variance; V ij is the partial variance of the second-order sensitivity of the interaction between the i-th and j-th parameters; is the d-order sensitivity partial variance of all parameter interactions; is a parameter set consisting of d parameters, and Y is the selected univariate model output.

[0025] S2.2. The differential evolution algorithm-Markov chain model is used to perform uncertainty analysis on the sensitive parameters in the improved algorithm and the non-sensitive parameters whose prior values ​​cannot be determined. Ten Markov chains are set to run in parallel, and each chain is set to have 30,000 iterations. The proportional reduction coefficient of multiple chains is less than 1.2 as the stable convergence indicator. Finally, the posterior distribution of each parameter is determined, and the median of the posterior distribution is taken as the optimal parameter value, such as Figure 3 shown.

[0026] The details are as follows: (8) (9) Where: x is the sample; X is the parameter set; is the prior distribution, which is set to uniform distribution; is the likelihood function; T is the total number of observed samples; O is the observed value; S is the simulated value; σ is the standard deviation of the model error.

[0027] S3, constructing the most effective humidity stress model for bare soil evaporation in bare soil areas, coupling the humidity stress model into the soil evaporation mode in the P-LSH algorithm, and forming an improved bare soil evaporation algorithm, including the following sub-steps: S3.1. Based on the actual soil moisture content, the most effective humidity stress model for bare soil evaporation in bare soil areas is constructed as follows: (10) in: f is the humidity stress on soil evaporation; SM max (m 3 m -3 ) and SM min (m 3 m -3 ) are the maximum and minimum values ​​of the SM sequence during the study period; Step S3.1 is the formula (1) at the quantile n The special case of reaching the maximum value, when SM c SM max .

[0028] S3.2. The constructed bare soil moisture stress scheme is coupled into the soil evaporation model in the P-LSH algorithm to form an improved bare soil evaporation algorithm as follows: (11) Among them: A s is the effective energy distributed to the soil, is the aerodynamic conductivity of the soil surface; is the total aerodynamic conductivity corrected for actual air temperature and pressure, is the psychrometric constant, is the air density.

[0029] S4. Combine the water evaporation module in the P-LSH algorithm, couple the improved canopy evaporation algorithm in step S1 and the improved bare soil evaporation algorithm in step S3 to obtain a global evapotranspiration remote sensing inversion algorithm that directly considers soil moisture stress.

[0030] The global evapotranspiration remote sensing inversion algorithm that directly considers soil moisture stress is named P-LSHv2 algorithm. The evapotranspiration data of flux tower observation and water balance reconstruction are used as the benchmark to evaluate the evapotranspiration accuracy of P-LSHv2 algorithm at the site and basin at the daily, monthly, annual and multi-year average scales. Figure 4 shown.

[0031] In addition, a satellite remote sensing inversion method for evapotranspiration that directly considers soil moisture stress is used to invert the global 1 / 12° resolution daily-scale evapotranspiration dataset from 1982 to 2022. The specific steps are as follows: Step A: Collect and preprocess various global meteorological driving data, including temperature, specific humidity, radiation, soil moisture, and NDVI data, and resample them to a 1 / 12° resolution; Step B: A global evapotranspiration remote sensing inversion algorithm that directly considers soil moisture stress is driven by various meteorological data to generate a global evapotranspiration dataset with a daily scale and a resolution of 1 / 12° from 1982 to 2022.

[0032] Although the present invention has been described above with preferred embodiments, it is not intended to limit the present invention. A person skilled in the art of the present invention may make various modifications and improvements without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention shall be determined by the definition of the claims.

Claims

1. A satellite remote sensing inversion method for evapotranspiration that directly considers soil moisture stress, characterized in that: The following steps are involved: S1. An improved canopy water vapor conductivity model is constructed based on the quantile canopy conductivity soil moisture stress factor, and then the canopy water vapor conductivity model is coupled into the canopy transpiration module in the P-LSH algorithm to obtain an improved canopy transpiration algorithm; S2. A Bayesian-based uncertainty analysis framework is used to quantify parameter uncertainty in the improved canopy transpiration algorithm. S3, constructing the most effective humidity stress model for bare soil evaporation in bare soil areas, coupling the humidity stress model into the soil evaporation mode in the P-LSH algorithm, and forming an improved bare soil evaporation algorithm; S4. Combine the water evaporation module in the P-LSH algorithm, couple the improved canopy evaporation algorithm in step S1 and the improved bare soil evaporation algorithm in step S3 to obtain a global evapotranspiration remote sensing inversion algorithm that directly considers soil moisture stress.

2. The satellite remote sensing inversion method for evapotranspiration directly considering soil moisture stress according to claim 1, characterized in that: In step S1, based on the actual soil moisture content, a quantile-based canopy conductance soil moisture stress factor is constructed as follows: （1） （2） in: m ( SM ) is the soil moisture stress factor, SM is the actual surface soil volume water content; SM min is the minimum soil moisture content, SM c is the critical soil moisture content; n is the quantile reflecting the critical soil moisture content, which is obtained by calibration of the observed value; f is the quantile mapping function.

3. The satellite remote sensing inversion method for evapotranspiration directly considering soil moisture stress according to claim 1, characterized in that: In step S1, the constructed soil moisture stress factor is incorporated into the Jarvis-Stewart equation to construct an improved canopy water vapor conductivity model, as shown in the following formula: （3） （4） in: is the canopy water vapor conductivity; Determined by the ecosystem type The maximum value of is defined as the sigmoid function of NDVI; b1, b2, and b3 are parameters affected by ecosystem type; m(i) is the stress factor of the i-th climate variable; i is the Climate variables that produce stressful effects.

4. The satellite remote sensing inversion method for evapotranspiration directly considering soil moisture stress according to claim 1, characterized in that: In step S1, the improved canopy transpiration algorithm is as follows: （5） in: is the canopy latent heat flux, is the latent heat of evaporation; A is the slope of the air saturated water vapor pressure and temperature curve; c is the effective energy component allocated to the canopy; VPD is the saturated vapor pressure difference; is the air density; C p is the specific heat capacity of air; is the psychrometric constant; is the canopy aerodynamic conductivity; is the canopy water vapor conductivity.

5. The satellite remote sensing inversion method for evapotranspiration directly considering soil moisture stress according to claim 1, characterized in that: Step S2 includes the following sub-steps: S2.

1. The Sobol' parameter sensitivity analysis method based on the variance decomposition principle was used to quantify the sensitivity of each parameter. The total effect index greater than 10% was used as an indicator to identify the sensitive parameters in the improved algorithm, as follows: （6） （7） Where: V i is the parameter X i The first-order sensitivity of the partial variance; V ij is the partial variance of the second-order sensitivity of the interaction between the i-th and j-th parameters; is the d-order sensitivity partial variance of all parameter interactions; is a parameter set consisting of d parameters, and Y is the selected univariate model output; S2.

2. Use the differential evolution algorithm-Markov chain model to perform uncertainty analysis on the sensitive parameters in the improved algorithm and the non-sensitive parameters whose prior values ​​cannot be determined, determine the posterior distribution of each parameter, and take the median of the posterior distribution as the optimal parameter value, as follows: （8） （9） Where: x is the sample; X is the parameter set; is the prior distribution, which is set to uniform distribution; is the likelihood function; T is the total number of observed samples; O is the observed value; S is the simulated value; σ is the standard deviation of the model error.

6. The satellite remote sensing inversion method for evapotranspiration directly considering soil moisture stress according to claim 1, characterized in that: Step S3 includes the following sub-steps: S3.

1. Based on the actual soil moisture content, the most effective humidity stress model for bare soil evaporation in bare soil areas is constructed as follows: （10） in: f is the humidity stress on soil evaporation; SM max (m 3 m -3 ) and SM min (m 3 m -3 ) are the maximum and minimum values ​​of the SM sequence during the study period, respectively; S3.

2. The constructed bare soil moisture stress scheme is coupled into the soil evaporation model in the P-LSH algorithm to form an improved bare soil evaporation algorithm as follows: （11） Among them: A s is the effective energy distributed to the soil, is the aerodynamic conductivity of the soil surface; is the total aerodynamic conductivity corrected for actual air temperature and pressure, is the psychrometric constant, is the air density.

7. The satellite remote sensing inversion method for evapotranspiration directly considering soil moisture stress according to claim 1, characterized in that: include: Based on the evapotranspiration data from flux tower observations and water balance reconstruction, the evapotranspiration accuracy of the global evapotranspiration remote sensing inversion algorithm that directly considers soil moisture stress at the station and basin level is evaluated at daily, monthly, annual and average scales of several years.

8. The satellite remote sensing inversion method for evapotranspiration directly considering soil moisture stress according to claim 1, characterized in that: This includes using a satellite remote sensing evapotranspiration inversion method that directly considers soil moisture stress to invert a global 1 / 12° resolution daily-scale evapotranspiration dataset from 1982 to 2022. Specifically: Step A: Collect and preprocess various global meteorological driving data, including temperature, specific humidity, radiation, soil moisture, and NDVI data, and resample them to a 1 / 12° resolution; Step B: A global evapotranspiration remote sensing inversion algorithm that directly considers soil moisture stress is driven by various meteorological data to generate a global evapotranspiration dataset with a daily scale and a resolution of 1 / 12° from 1982 to 2022.

Images