Land surface evapotranspiration driving factor identification method, equipment and medium

Abstract

The invention relates to the field of carbon neutralization, and discloses a land surface evapotranspiration driving factor identification method and device and a medium, and the method comprises the steps: data collection; global land evapotranspiration space-time evolution; evaluating the model simulation performance of global land surface evapotranspiration; detecting and attributing global land surface evapotranspiration driving factors; global land surface evapotranspiration driving factor spatio-temporal characteristic quantification; the method has the advantages that evapotranspiration changes are explored, ecosystem model performance is evaluated, signals of different driving factors are identified, contribution of the signals is quantified, spatial-temporal characteristics of influences of the different driving factors on evapotranspiration are evaluated, influences of climate factors on evapotranspiration trends are separated, dominant contribution of the climate factors is quantified, and the ecological system model performance is evaluated. And scientific understanding of problems such as dynamic evolution of hydrological cycle and carbon sink management under climate change can be improved.

Description

Technical Field

[0001] The present invention relates to the field of carbon neutrality, and in particular to a method, device and medium for identifying driving factors of land surface evapotranspiration. Background Art

[0002] Land surface evapotranspiration (ET) is an important component of the global water and energy cycle. Changes in evapotranspiration affect runoff, temperature and other hydrometeorological variables through land-atmosphere interactions, thereby affecting water resource availability and even changing regional and global warming. Changes in evapotranspiration are mainly determined by climate change (CLI), land use change (LUC) and carbon dioxide fertilization effect (CO2). Climate change is the main factor controlling changes in the hydrological cycle, while precipitation and potential evapotranspiration are the main factors affecting climate change, affecting available water and available energy. Therefore, understanding the trend of global evapotranspiration changes and its potential driving mechanisms is crucial to more accurately understand and assess its impact on the Earth system and the hydrological cycle.

[0003] At present, there is no detailed study on the response of evapotranspiration changes to environmental changes at different temporal and spatial scales, and the impact of driving factors such as climate change, CO2 and land use change on evapotranspiration is still unclear. This method helps to better understand the hydrological cycle under climate change and provide a scientific basis for formulating policies to address climate change. Summary of the invention

[0004] The purpose of the present invention is to propose a method, device and medium for identifying driving factors of land surface evapotranspiration, so as to solve the technical problem that the impact of driving factors such as climate change, CO2 and land use change on evapotranspiration is still unclear.

[0005] The present invention provides a method for identifying driving factors of land surface evapotranspiration, comprising the following steps:

[0006] Step S1: Data collection; collect the latest satellite observation evapotranspiration data; collect the ecosystem model provided by the dynamic global vegetation model TRENDY-v7; collect the pre-industrial control run simulation experiment CTL of the sixth phase of the International Coupled Comparison Program CMIP6; collect precipitation and potential evapotranspiration data from the Climate Research Center CRU data set; collect forest, cultivated land and irrigated area ratio data from the land use coordination data LUH2;

[0007] Step S2: Spatiotemporal evolution of global land surface evapotranspiration: Combining the data obtained in step S1, using satellite observations and four experimental scenarios of the ecosystem model, the spatiotemporal trend changes of global land surface evapotranspiration on annual and seasonal scales are calculated;

[0008] Step S3: Evaluation of the model simulation performance of global land surface evapotranspiration; combining the satellite observations obtained in step S2 with the evapotranspiration trends simulated by the model, and using the Nash efficiency coefficient NSE to evaluate the simulation performance of the ecosystem model in simulating the annual average land surface evapotranspiration;

[0009] Step S4: Detection and attribution of driving factors of global land surface evapotranspiration; Combine the evapotranspiration trends obtained from satellite observations and model simulations in step S2 with the pre-industrial controlled run simulation experiment CTL in CMIP6 to detect and attribute signals of different driving factors;

[0010] Step S5: Quantify the spatiotemporal characteristics of driving factors of global land surface evapotranspiration; Combined with the changes in the spatiotemporal trend of simulated evapotranspiration obtained in step S2, calculate the spatiotemporal characteristics of the impact of different driving factors on the annual evapotranspiration trend, and quantify the spatial patterns of the dominant driving factors of global evapotranspiration in different regions; Based on the extended Budyko framework, further separate the contributions of precipitation and potential evapotranspiration under climate change in the driving factors, evaluate the effect of the Budyko framework in the simulation process, and determine the spatial distribution of the dominant factors of land surface evapotranspiration trend in climate change.

[0011] A storage medium stores instructions and data for implementing a method for identifying driving factors of land surface evapotranspiration.

[0012] A device for identifying driving factors of land surface evapotranspiration comprises: a processor and the storage medium; the processor loads and executes instructions and data in the storage medium to implement a method for identifying driving factors of land surface evapotranspiration.

[0013] The beneficial effects provided by the present invention are:

[0014] (1) The present invention quantifies the temporal and spatial variation trends of global land surface evapotranspiration at different time scales and regional scales, explores the variation trends of evapotranspiration driven by different factors, quantitatively detects the driving factors of global evapotranspiration, and quantifies the contribution of different driving factors to the change of global evapotranspiration trend, providing a scientific basis for a deep understanding of the evolution of global land surface evapotranspiration.

[0015] (2) This paper analyzes the driving factors of global land surface evapotranspiration, clarifies the contribution of different driving factors to global land surface evapotranspiration, quantifies the dominant driving factors of global evapotranspiration, and further decomposes the driving factors to clarify their main physical mechanisms, providing theoretical support for the attribution of global land surface evapotranspiration. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic flow chart of the method of the present invention;

[0017] Figure 2It is the temporal and spatial variation of annual evapotranspiration over global land surface based on satellite observations and ecosystem simulations;

[0018] Figure 3 is a scatter plot of the correlation between model simulation and satellite observation of global evapotranspiration;

[0019] Figure 4 It is the time series of evapotranspiration under different simulation test scenarios in different seasons;

[0020] Figure 5 is the detection and attribution of drivers of global evapotranspiration trends at regional scales;

[0021] Figure 6 It is the detection and attribution of drivers of global evapotranspiration trends on a seasonal scale;

[0022] Figure 7 is the spatial distribution of the contribution of different driving factors to the global evapotranspiration trend;

[0023] Figure 8 It is the performance evaluation of the extended Budyko equation in evapotranspiration trend simulation and attribution;

[0024] Fig. 9 is the spatial distribution of the dominant driver of global evapotranspiration trends;

[0025] Fig.10 It is the spatial distribution of the dominant climate factor of global evapotranspiration trends;

[0026] Fig.11 It is a working schematic diagram of the hardware device of the present invention. DETAILED DESCRIPTION

[0027] In order to make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0028] Before formally describing the present invention, the scheme of the present invention is first generally described for easy understanding.

[0029] Please refer to Figure 1 , Figure 1 It is a schematic diagram of the process of the present invention.

[0030] The present invention provides a method for identifying driving factors of land surface evapotranspiration, comprising the following steps:

[0031] Step S1: Data collection; collect the latest satellite observation evapotranspiration data; collect the ecosystem model provided by the dynamic global vegetation model TRENDY-v7; collect the pre-industrial control run simulation experiment CTL of the sixth phase of the International Coupled Comparison Program CMIP6; collect precipitation and potential evapotranspiration data from the Climate Research Center CRU data set; collect forest, cultivated land and irrigated area ratio data from the land use coordination data LUH2;

[0032] It should be noted that the present invention collects satellite observed evapotranspiration data of the latest Amsterdam model global land evaporation data (GLEAMv3.3a); collects the ecosystem model provided by the dynamic global vegetation model TRENDY-v7; collects the pre-industrial controlled run simulation experiment (CTL) of the sixth phase of the International Coupled Comparison Program (CMIP6); collects precipitation and potential evapotranspiration data from the CRU data set of the Climate Research Center of the University of East Anglia; and collects forest, cultivated land and irrigated area ratio data in the coordinated land use data (LUH2).

[0033] Step S2: Spatiotemporal evolution of global land surface evapotranspiration: Combining the data obtained in step S1, using satellite observations and four experimental scenarios of the ecosystem model, the spatiotemporal trend changes of global land surface evapotranspiration on annual and seasonal scales are calculated;

[0034] It should be noted that in step S2, the four experimental scenarios in the ecosystem model include: S0 (control without forced change); S1 (change of CO2 only); S2 (change of CO2 and climate change (CLI)); S3 (change of CO2, CLI and land use (LULC)); based on the multi-model ensemble mean, the differences between the S1-S0, S2-S1 and S3-S2 scenarios are calculated to evaluate the impact of CO2, CLI and LUC on evapotranspiration trends; the common period of model simulation and satellite observation (GLEAM) is selected for analysis.

[0035] Based on satellite observations and model simulation data, the temporal and spatial characteristics of global evapotranspiration trends and the probability density functions on annual and seasonal scales were calculated using Sen's slope; the time series of the contributions of three driving factors, CO2 (S1-S0), CLI (S2-S1) and LULC (S3-S2), to evapotranspiration trends were calculated;

[0036] Step S3: Evaluation of the model simulation performance of global land surface evapotranspiration; combining the satellite observations obtained in step S2 with the evapotranspiration trends simulated by the model, and using the Nash efficiency coefficient NSE to evaluate the simulation performance of the ecosystem model in simulating the annual average land surface evapotranspiration;

[0037] It should be noted that in step S3, the Nash efficiency (NSE) coefficient is used to compare the annual average evapotranspiration observed by satellite and simulated by the ecosystem model to evaluate the fitting performance of the ecosystem model; in order to reduce uncertainty, models with relatively low Nash efficiency coefficients are excluded, and only models with Nash efficiency coefficients greater than 0.80 are used for analysis;

[0038] Step S4: Detection and attribution of driving factors of global land surface evapotranspiration; Combine the evapotranspiration trends obtained from satellite observations and model simulations in step S2 with the pre-industrial controlled run simulation experiment CTL in CMIP6 to detect and attribute signals of different driving factors;

[0039] It should be noted that in step S4, the signal detection and attribution of different driving factors of global evapotranspiration based on the optimal fingerprint method are specifically calculated as follows:

[0040]

[0041] Where Y is the global evapotranspiration trend sequence based on satellite observations, g is the evapotranspiration change sequence driven by different factors based on model simulation, including CO2, climate change, land use and trend change sequence under the three experimental scenarios S1-S3, n is the number of models used, β is the scale factor, and ε is the internal climate variability estimated by CTL;

[0042] To construct the noise covariance of internal variability, multiple non-overlapping time blocks in the CTL of CMIP6 with the same length as the study period (38 years) are used to calculate the area-weighted mean annual evapotranspiration anomaly time series for each time block, which is used to determine the noise covariance and detect the consistency of residuals to detect the consistency of internal variability of evapotranspiration based on satellite observations and model simulations;

[0043] In order to reduce the time dimension, the obtained evapotranspiration time series is weighted averaged every three years; in addition, the global evapotranspiration in spring (March-May in the Northern Hemisphere, September-November in the Southern Hemisphere), summer (June-August in the Northern Hemisphere, December-February in the Southern Hemisphere), autumn (September-November in the Northern Hemisphere, March-May in the Southern Hemisphere), and winter (December-February in the Northern Hemisphere, June-August in the Southern Hemisphere) is detected and attributed; four latitudinal zones are also divided into the southern hemisphere temperate zone (90-25°S), tropical zone (25°S-25°N), northern hemisphere temperate zone (25-50°N), and northern hemisphere high latitude (50-90°N) to explore and quantify the changes in global evapotranspiration;

[0044] Step S5: Quantify the spatiotemporal characteristics of driving factors of global land surface evapotranspiration; Combined with the changes in the spatiotemporal trend of simulated evapotranspiration obtained in step S2, calculate the spatiotemporal characteristics of the impact of different driving factors on the annual evapotranspiration trend, and quantify the spatial patterns of the dominant driving factors of global evapotranspiration in different regions; Based on the extended Budyko framework, further separate the contributions of precipitation and potential evapotranspiration under climate change in the driving factors, evaluate the effect of the Budyko framework in the simulation process, and determine the spatial distribution of the dominant factors of land surface evapotranspiration trend in climate change.

[0045] It should be noted that in step S5, the spatial distribution of global evapotranspiration trends driven by various factors (CO2, climate change and land use) on annual and seasonal scales is calculated to quantify the spatial patterns of the dominant driving factors of global evapotranspiration in different regions;

[0046] Based on the extended Budyko framework, the response of evapotranspiration trends to recent climate change is quantified, and the individual contributions of precipitation (P) and potential evapotranspiration (PET) in the climate change (CLI) factor are further separated. The Budyko framework describes the long-term relationship between evapotranspiration and available energy (potential evapotranspiration) and available water (precipitation). Under this framework, the Budyko water-energy coupled balance equation is proposed, as follows:

[0047]

[0048] Where ET is evapotranspiration, P is precipitation, PET is potential evapotranspiration, and n is the underlying surface parameter, which is used to explain all factors affecting evapotranspiration except precipitation and potential evapotranspiration, including CO2-vegetation feedback (CO2 fertilization and changes in water use efficiency) and land use; the value of n can be calibrated by minimizing the least squares error, and the change in evapotranspiration can be expressed by total differential decomposition, as follows:

[0049]

[0050] Where d represents the changes in evapotranspiration, precipitation, potential evapotranspiration and underlying surface parameters. represents the partial derivative of a variable.

[0051] To meet the requirements of the steady-state assumption, a 15-year sliding anomaly analysis is used to quantify the impact of precipitation and potential evapotranspiration anomalies on evapotranspiration anomalies every 15 years, as follows:

[0052]

[0053] in, and are the effects of precipitation and potential evapotranspiration on evapotranspiration trends, respectively;

[0054] The dominant contributions of precipitation and potential evapotranspiration to evapotranspiration trends in climate change were determined through the following three steps: 1) using a multi-model ensemble mean of ecosystem models to identify regions where evapotranspiration trends are dominated by climate change; 2) using an extended Budyko framework to quantify the impact of precipitation and potential evapotranspiration on global evapotranspiration changes; and 3) comparing the individual contributions of precipitation and evapotranspiration in regions dominated by climate change and partitioning the climate change dominated regions into specific climate factors (i.e., precipitation and potential evapotranspiration).

[0055] As an embodiment, the present invention further describes the present invention by taking the global period from 1980 to 2017 as an example. The implementation case is used to illustrate the present invention, but this case is not used to limit the scope of application of the present invention. It is also applicable to different regions and other time periods.

[0056] The implementation flow chart of the method of the present invention for identifying the driving factors of land surface evapotranspiration is as follows: Figure 1 As shown, the specific steps are as follows:

[0057] (1) Collection of test data;

[0058] In this implementation case, the satellite-observed monthly evapotranspiration data with a spatial resolution of 0.25°×0.25° from 1980 to 2018 of the latest Amsterdam model global land evaporation data (GLEAM v3.3a) were collected;

[0059] We collected 13 ecosystem models provided by the dynamic global vegetation model TRENDY-v7, with a time series of 1980-2017. We interpolated all model data and satellite observation data to a uniform 0.5°×0.5° using bilinear interpolation. The specific information is shown in Table 1 below:

[0060] Table 1 Basic information of the selected TRENDY-v7 model output

[0061]

[0062] The evapotranspiration data of the pre-industrial controlled run simulation experiment (CTL) of the sixth phase of the International Coupled Comparison Program (CMIP6) were collected, and the specific model used is shown in Table 2 below;

[0063] Table 2 Models used

[0064]

[0065]

[0066] Collect monthly precipitation and potential evapotranspiration data with a spatial resolution of 0.5°×0.5° from 1980 to 2017 from the Climate Research Unit (CRU) dataset of the University of East Anglia; collect forest, cultivated land and irrigated area proportion data with a spatial resolution of 0.25°×0.25° from 1980 to 2017 from the Land Use Coordination Data (LUH2);

[0067] (2) Temporal and spatial evolution of global land surface evapotranspiration;

[0068] This implementation case is based on the global scope, and based on the evapotranspiration data of satellite observations and ecosystem models under four experimental scenarios from 1980 to 2017, calculates the temporal and spatial trend changes of global land surface evapotranspiration on annual and seasonal scales;

[0069] The four experimental scenarios in the ecosystem model include: S0 (control without forced change); S1 (change of CO2 only); S2 (change of CO2 and climate change (CLI)); S3 (simultaneous change of CO2, CLI and land use (LULC)); Based on the multi-model ensemble mean, the differences of S1-S0, S2-S1 and S3-S2 scenarios are calculated to evaluate the impact of CO2, CLI and LUC on evapotranspiration trends, and the time series of the contribution of the three driving factors of CO2 (S1-S0), CLI (S2-S1) and LULC (S3-S2) to evapotranspiration trends are calculated;

[0070] In this implementation case, Figure 2 As shown in the data, in the 38 years from 1980 to 2017, global satellite-based evapotranspiration showed an increasing trend, with an area-weighted average increase of 0.82 mm per year. The probability density functions of annual and seasonal evapotranspiration trends are asymmetric, with positive anomalies covering a wider range than negative anomalies, indicating that global evapotranspiration shows an increasing trend over a wider range. Overall, 41.6% of the world's land area shows a significant increase in annual evapotranspiration.

[0071] For seasonal evapotranspiration changes ( Figure 2 c), 32.2% of the global land area shows a positive trend in summer evapotranspiration, followed by spring (31.2%), autumn (31.0%) and winter (22.5%). In contrast, the annual evapotranspiration of regions with less than 5% of the global land area shows a significant downward trend, and except for winter (less than 10%), all other seasons are less than 5%. The spatial distribution of evapotranspiration trends is similar on the annual and seasonal scales, with evapotranspiration in Europe, southern Africa, southern China and India showing a significant increase, while evapotranspiration in western America and southern Brazil shows a significant decrease.

[0072] (3) Evaluation of model simulation performance of global land surface evapotranspiration;

[0073] This implementation case uses the Nash efficiency coefficient (NSE) to evaluate the simulation performance of the ecosystem model in simulating the annual average land surface evapotranspiration. The Nash efficiency coefficient is a widely used and reliable statistical data for evaluating model performance. It is used to evaluate the fitting effect of the annual average evapotranspiration simulated by the ecosystem model.

[0074] like Figure 3 As shown in the figure, the evapotranspiration simulated by the ecosystem model under the S3 scenario matches well with the evapotranspiration based on satellite observations, and the Nash efficiency coefficients of most models are greater than 0.80 (10 out of 13 models). To reduce uncertainty, models with lower Nash efficiency coefficients were excluded, and only models with Nash efficiency coefficients greater than 0.80 were used for analysis. The ensemble average Nash efficiency coefficient and determination coefficient of the 10 selected ecosystem models were both 0.92 ( Figure 3 The spatial distribution of the mean annual evapotranspiration is similar to that of satellite observations. Although the trend of satellite-observed evapotranspiration is greater than that of the simulated evapotranspiration driven by the multi-model average S3 scenario, its variability is generally within the range of simulated evapotranspiration variability and has similar interannual fluctuations on the annual and seasonal scales ( Figure 4 ).

[0075] (4) Detection and attribution of drivers of global land surface evapotranspiration;

[0076] This implementation case is based on the optimal fingerprint method for global evapotranspiration signal detection and attribution of different driving factors. The specific calculation is as follows:

[0077]

[0078] Where T is the global evapotranspiration trend sequence based on satellite observations, g is the evapotranspiration change sequence driven by different factors based on model simulation, including CO2, climate change, land use change and trend change sequence under three experimental scenarios S1 to S3, n is the number of models used, β is the scale factor, and ε is the internal climate variability estimated by the pre-industrial controlled run simulation experiment (CTL);

[0079] To construct the noise covariance of internal variability, the area-weighted mean annual evapotranspiration anomaly time series of 21 non-overlapping time blocks of the same length as the study period (38 years) in the CTL simulation experiments in CMIP6 were used to calculate each time block, with a total of 328 38-year blocks. These time series were divided into two parts, and two independent (time × number) matrices were obtained to construct the noise covariance, one matrix was used to determine the noise covariance, and the other was used for residual consistency detection to detect the consistency of internal variability of evapotranspiration based on satellite observations and model simulations;

[0080] In order to reduce the time dimension, the obtained evapotranspiration time series was weighted averaged every three years; the driving factors of global evapotranspiration trends were detected and attributed in spring (March-May in the Northern Hemisphere, September-November in the Southern Hemisphere), summer (June-August in the Northern Hemisphere, December-February in the Southern Hemisphere), autumn (September-November in the Northern Hemisphere, March-May in the Southern Hemisphere), and winter (December-February in the Northern Hemisphere, June-August in the Southern Hemisphere); four latitudinal zones were divided into the Southern Hemisphere temperate zone (90-25°S), tropical zone (25°S-25°N), Northern Hemisphere temperate zone (25-50°N), and Northern Hemisphere high latitude (50-90°N) to explore and quantify the changes in global evapotranspiration;

[0081] like Figure 5 As shown, globally, the scaling factors and their confidence intervals for the S2 (CO2 + climate change) and S3 (CO2 + climate change + land use change) scenarios are greater than zero, indicating that the impact of these two composite signals is detectable on a global scale. For individual driving factors, only the signal of climate change is detected, which is consistent with the significant increase in annual evapotranspiration based on satellite observations and climate change (CLI) simulations ( Figure 2 b). The global signal of CO2 and land use change cannot be detected. The linear trend of each driving factor multiplied by its scale factor is further used to attribute the trend of evapotranspiration ( Figure 5 f), the global evapotranspiration change caused by climate change is 0.55 mm / year, which is similar and comparable to the satellite-based evapotranspiration trend (0.82 mm / year). Overall, climate change is the main driver of the increase in global evapotranspiration.

[0082] At the regional scale, climate change in the tropics and high latitudes of the Northern Hemisphere is the main driver of global evapotranspiration changes. Climate change signals can be detected in these regions, indicating that their contribution to evapotranspiration is reliable ( Figure 5 In the temperate zones of the southern hemisphere and the northern hemisphere, the dominant factor becomes land use change ( Figure 5 gi in ), but the driving factor signal was not significantly detected. Figure 6 ), climate change signals in evapotranspiration changes can be detected in different seasons. Climate change is the dominant factor controlling evapotranspiration trends except in autumn, where the contribution of climate change is smaller than that of land use change. CO2 reduces evapotranspiration in all seasons, but its signal cannot be detected.

[0083] (5) Quantification of the spatiotemporal characteristics of global land surface evapotranspiration driving factors;

[0084] In this implementation case, the spatial distribution of global evapotranspiration trends driven by various factors (CO2, climate change and land use) on annual and seasonal scales is calculated, and the spatial patterns of the dominant driving factors of global evapotranspiration in different regions are quantified;

[0085] like Figure 7 As shown in Figure 2, based on the ensemble mean of 10 selected ecosystem models, elevated CO2 concentrations reduce annual evapotranspiration over most of the world, mainly in forested areas ( Figure 7 In a few regions (accounting for 2.5% of the global land area), such as the Amazon Basin, rising CO2 concentrations will also lead to significant increases in evapotranspiration. In summer and autumn, the areas where CO2-driven evapotranspiration changes show a significant increase trend are larger than those in spring and winter.

[0086] Climate change leads to increased evapotranspiration in most parts of the world, which is due to the increase of potential evapotranspiration and precipitation in most parts of the world. This implementation case is based on the extended Budyko framework to quantify the response of evapotranspiration trends to recent climate change, and further separate the individual contributions of precipitation (P) and potential evapotranspiration (PET) in the climate change (CLI) factor. The Budyko framework describes the long-term relationship between evapotranspiration and available energy (potential evapotranspiration) and available water (precipitation). Under this framework, the Budyko water-energy coupled balance equation is proposed, as follows:

[0087]

[0088] Where ET is evapotranspiration, P is precipitation, PET is potential evapotranspiration, and n is the underlying surface parameter, which is used to explain all factors affecting evapotranspiration except precipitation and potential evapotranspiration, including CO2-vegetation feedback (CO2 fertilization and changes in water use efficiency) and land use; the value of n can be calibrated by minimizing the least squares error, and the change in evapotranspiration can be expressed by total differential decomposition, as follows:

[0089]

[0090] Where d represents the changes in evapotranspiration, precipitation, potential evapotranspiration and underlying surface parameters. represents the partial derivative of a variable.

[0091] To meet the requirements of the steady-state assumption, a 15-year sliding anomaly analysis is used to quantify the impact of precipitation and potential evapotranspiration anomalies on evapotranspiration anomalies every 15 years, as follows:

[0092]

[0093] in, and are the effects of precipitation and potential evapotranspiration on evapotranspiration trends, respectively;

[0094] In this implementation case, the simulated evapotranspiration trend based on the extended Budyko framework can generally capture the global evapotranspiration trend, and its Nash efficiency coefficient is 0.48 ( Figure 8 ). In addition, the impact of climate change on evapotranspiration calculated using the extended Budyko framework is comparable to the results calculated using the ecosystem model ( Figure 8 (b) However, the Budyko framework assesses a wider range of climate change impacts than the ecosystem models.

[0095] The dominant contributions of precipitation and potential evapotranspiration to evapotranspiration trends in climate change were determined through the following three steps: 1) using a multi-model ensemble mean of ecosystem models to identify regions where evapotranspiration trends are dominated by climate change; 2) using an extended Budyko framework to quantify the impact of precipitation and potential evapotranspiration on global evapotranspiration changes; and 3) comparing the individual contributions of precipitation and evapotranspiration in regions dominated by climate change and partitioning the climate change dominated regions into specific climate factors (i.e., precipitation and potential evapotranspiration).

[0096] In this implementation case, climate change dominates the contribution to annual evapotranspiration change over 65.5% of the global land area ( Fig. 9 ), with precipitation being the dominant factor driving evapotranspiration trends over 48.4% of global land area, while the remaining 17.1% is dominated by potential evapotranspiration ( Fig.10 Significant increases in evapotranspiration due to climate change are attributable to increases in precipitation over northeastern North America, southern North Africa, southern Africa, northeastern Russia, and northern Australia ( Fig.10 ), while in Brazil, Europe, southern China and Japan, the significant increase in evapotranspiration caused by climate change is caused by the increase in potential evapotranspiration. In addition, the evapotranspiration caused by climate change in 16.1% of the global land area has significantly decreased, mainly in western America, southern Brazil, northern Middle East and eastern Australia, mainly due to the decrease in precipitation in these regions. Usually, precipitation changes in arid areas dominate the change in evapotranspiration trend, while potential evapotranspiration in humid areas controls the change in evapotranspiration trend.

[0097] Land use change has strong local effects on evapotranspiration changes, although these effects are relatively small at the global scale. Overall, land use change significantly reduced (increased) evapotranspiration ( Figure 7c). The significant negative contribution of land use change to evapotranspiration is mainly concentrated in areas of tropical rainforest deforestation, such as southern Brazil and Southeast Asia, where rainforest has been replaced by farmland. The positive contribution of land use change to evapotranspiration is mainly concentrated in areas where farmland has been replaced by forest (such as the eastern United States, southern Europe, and New Zealand). Other areas that are positively affected by land use change, such as the Middle East, northern India, and North China, may be attributed to the significant increase in farmland irrigation.

[0098] See also Fig.11 , Fig.11 4 is a schematic diagram of the working of the hardware device of an embodiment of the present invention, wherein the hardware device specifically comprises: a land surface evapotranspiration driving factor identification device 401, a processor 402 and a storage medium 403.

[0099] A land surface evapotranspiration driving factor identification device 401: the land surface evapotranspiration driving factor identification device 401 implements the land surface evapotranspiration driving factor identification method.

[0100] Processor 402: The processor 402 loads and executes the instructions and data in the storage medium 403 to implement the method for identifying driving factors of land surface evapotranspiration.

[0101] Storage medium 403: the storage medium 403 stores instructions and data; the storage medium 403 is used to implement the method for identifying driving factors of land surface evapotranspiration.

[0102] The beneficial effects of the present invention are:

[0103] (1) The present invention quantifies the temporal and spatial variation trends of global land surface evapotranspiration at different time scales and regional scales, explores the variation trends of evapotranspiration driven by different factors, quantitatively detects the driving factors of global evapotranspiration, and quantifies the contribution of different driving factors to the change of global evapotranspiration trend, providing a scientific basis for a deep understanding of the evolution of global land surface evapotranspiration.

[0104] (2) This paper analyzes the driving factors of global land surface evapotranspiration, clarifies the contribution of different driving factors to global land surface evapotranspiration, quantifies the dominant driving factors of global evapotranspiration, and further decomposes the driving factors to clarify the main physical mechanisms of the driving factors, providing theoretical support for the attribution of global land surface evapotranspiration.

[0105] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for identifying driving factors of land surface evapotranspiration, characterized by: The method comprises the following steps: Step S1: Data collection; collect the latest satellite observation evapotranspiration data; collect the ecosystem model provided by the dynamic global vegetation model TRENDY-v7; collect the pre-industrial control run simulation experiment CTL of the sixth phase of the International Coupled Comparison Program CMIP6; collect precipitation and potential evapotranspiration data from the Climate Research Center CRU data set; collect forest, cultivated land and irrigated area ratio data from the land use coordination data LUH2; Step S2: Spatiotemporal evolution of global land surface evapotranspiration: Combining the data obtained in step S1, using satellite observations and four experimental scenarios of the ecosystem model, the spatiotemporal trend changes of global land surface evapotranspiration on annual and seasonal scales are calculated; Step S3: Evaluation of the model simulation performance of global land surface evapotranspiration; combining the satellite observations obtained in step S2 with the evapotranspiration trends simulated by the model, and using the Nash efficiency coefficient NSE to evaluate the simulation performance of the ecosystem model in simulating the annual average land surface evapotranspiration; Step S4: Detection and attribution of driving factors of global land surface evapotranspiration; Combine the evapotranspiration trends obtained from satellite observations and model simulations in step S2 with the pre-industrial controlled run simulation experiment CTL in CMIP6 to detect and attribute signals of different driving factors; Step S5: Quantification of spatiotemporal characteristics of global land surface evapotranspiration driving factors; Combined with the changes in the spatiotemporal trend of simulated evapotranspiration obtained in step S2, the spatiotemporal characteristics of the impact of different driving factors on the annual evapotranspiration trend are calculated, and the spatial patterns of the dominant driving factors of global evapotranspiration in different regions are quantified; based on the extended Budyko framework, the contributions of precipitation and potential evapotranspiration under climate change in the driving factors are further separated, the effect of the Budyko framework in the simulation process is evaluated, and the spatial distribution of the dominant factors of land surface evapotranspiration trend in climate change is determined.

2. A method for identifying driving factors of land surface evapotranspiration according to claim 1, characterized in that: The different driving factors in step S5 include: climate change CLI, carbon dioxide fertilization effect CO2 and land use change LUC; the dominant factors in step S5 include: precipitation and potential evapotranspiration.

3. The method for identifying driving factors of land surface evapotranspiration according to claim 1, characterized in that: Step S2 is specifically as follows: S21: The four experimental scenarios in the ecosystem model include: Scenario S0: Control without forced change; Scenario S1: CO2 only changes; Scenario S2: Changing CO2 and climate change CLI; Scenario S3: Changing CO2, CLI and land use change LULC; Based on the multi-model ensemble mean, the differences among scenarios S1-S0, S2-S1 and S3-S2 were calculated to assess the impacts of CO2, climate change and land use change on evapotranspiration trends; S22: Based on satellite observations and model simulation data, the temporal and spatial characteristics of the global evapotranspiration trend and the probability density function on the annual and seasonal scales are calculated using Sen's slope; the time series of the contribution of the three driving factors of CO2 in scenarios S1-S0, climate change in scenarios S2-S1 and land use change in scenarios S3-S2 to the evapotranspiration trend are calculated.

4. The method for identifying driving factors of land surface evapotranspiration according to claim 1, characterized in that: Step S3 is as follows: The Nash efficiency (NSE) coefficient was used to compare the annual average evapotranspiration from satellite observations and that simulated by the ecosystem model to evaluate the fitting performance of the ecosystem model. Among them, models with relatively low Nash efficiency coefficients were excluded, and only models with Nash efficiency coefficients greater than the preset value were used for analysis.

5. The method for identifying driving factors of land surface evapotranspiration according to claim 1, characterized in that: Step S4 is specifically as follows: S41: Detection and attribution of different driving factor signals of global evapotranspiration based on the optimal fingerprint method. The specific calculation is as follows: Where Y is the global evapotranspiration trend sequence based on satellite observations, g is the evapotranspiration trend sequence driven by different factors based on model simulation, including the trend change sequence under three experimental scenarios of CO2, climate change, land use and scenarios S1-S3, n is the number of models used, β is the scale factor, and ε is the internal climate variability estimated by the CTL simulation experiment; S42: To construct the noise covariance of internal variability, use multiple non-overlapping time blocks of the same length as the study period in the CTL simulation experiment of the pre-industrial control run simulation experiment in CMIP6, calculate the area-weighted mean annual evapotranspiration anomaly time series of each time block, divide these time series into two, and obtain two independent time × number matrices to construct the noise covariance, one matrix is ​​used to determine the noise covariance, and the other is used for residual consistency detection to detect the consistency of the internal variability of evapotranspiration based on satellite observations and model simulations; S43: To reduce the temporal dimension, the obtained evapotranspiration time series was subjected to a three-year area-weighted average. In addition, the global evapotranspiration was detected and attributed in spring, summer, autumn and winter respectively. Four latitudinal zones were divided into the southern hemisphere temperate zone, tropical zone, northern hemisphere temperate zone and northern hemisphere high latitudes to explore and quantify the changes in global evapotranspiration.

6. A method for identifying driving factors of land surface evapotranspiration according to claim 2, characterized in that: Step S5 is specifically as follows: S51: Calculate the spatial distribution of global evapotranspiration trends on annual and seasonal scales driven by different driving factors, and quantify the spatial patterns of the dominant driving factors of global evapotranspiration in different regions; S52: Based on the extended Budyko framework, quantify the response of evapotranspiration trends to recent climate change and further separate the individual contributions of precipitation P and potential evapotranspiration PET in the climate change CLI factor; S53: Determine the dominant contributions of precipitation and potential evapotranspiration to evapotranspiration trends in a changing climate through the following three steps: 1) Use a multi-model ensemble mean of ecosystem models to identify regions where evapotranspiration trends are dominated by climate change; 2) Use an extended Budyko framework to quantify the impact of precipitation and potential evapotranspiration on global evapotranspiration changes; 3) Compare the individual contributions of precipitation and evapotranspiration in regions dominated by climate change and classify regions dominated by climate change into specific climate factors.

7. A method for identifying driving factors of land surface evapotranspiration according to claim 6, characterized in that: Step S52 is specifically: The Budyko framework describes the long-term relationship between evapotranspiration and available energy potential evapotranspiration and precipitation. Under this framework, the Budyko water-energy coupled balance equation is proposed as follows: Where ET is evapotranspiration, P is precipitation, PET is potential evapotranspiration, and n is the underlying surface parameter, which is used to explain all factors affecting evapotranspiration except precipitation and potential evapotranspiration, including CO2-vegetation feedback and land use; the value of n is calibrated by minimizing the least squares error, and the change in evapotranspiration is expressed by total differential decomposition, as follows: Where d represents the changes in evapotranspiration, precipitation, potential evapotranspiration and underlying surface parameters. represents the partial derivative of a variable; To meet the requirements of the steady-state assumption, a 15-year sliding anomaly analysis is used to quantify the impact of precipitation and potential evapotranspiration anomalies on evapotranspiration anomalies every 15 years, as follows: in, and are the effects of precipitation and potential evapotranspiration on evapotranspiration trends, respectively.

8. A storage medium, characterized in that: The storage medium stores instructions and data for implementing a method for identifying driving factors of land surface evapotranspiration as described in any one of claims 1 to 7.

9. A device for identifying driving factors of land surface evapotranspiration, characterized in that: include: A processor and a storage medium; the processor loads and executes instructions and data in the storage medium to implement a method for identifying driving factors of land surface evapotranspiration as described in any one of claims 1 to 7.

Images