Optimization method for calculating carbon storage of terrestrial vegetation on a monthly scale based on terrestrial water storage

Abstract

The present invention discloses an optimization method for calculating the monthly scale terrestrial vegetation carbon storage based on terrestrial water storage. First, historical measured meteorological data, atmospheric carbon dioxide concentration data and satellite remote sensing product data in a typical study area are collected; a time series of soil moisture in the study area is constructed; atmospheric CO2 concentration data is collected and interpolated; and a quantitative estimation model of vegetation carbon storage based on hydrological and meteorological data and terrestrial moisture is constructed. The present invention proposes a method for estimating vegetation carbon storage, which takes into account the impact of terrestrial available water on plant carbon fixation capacity and carbon storage, and provides a scientific reference basis for resource planning and management departments to more accurately estimate regional carbon storage capacity, thereby proposing a more reasonable carbon neutrality management planning scheme.

Description

Technical Field

[0001] The present invention relates to the technical field of hydrology and water resources application, and in particular to a method for optimizing the calculation of vegetation carbon reserves at a monthly scale based on terrestrial water reserves. Background Art

[0002] Since the Industrial Revolution, the atmospheric CO2 concentration has increased from about 280ppm (parts per million) to about 415ppm, causing the global average surface temperature to continue to rise, leading to frequent extreme weather and climate events. In order to maintain the stability of the global carbon cycle, the ability of marine and terrestrial ecosystems to absorb CO2 in the air is also increasing, which has slowed down the rising trend of greenhouse gases to a certain extent. As one of the most critical variables in the terrestrial carbon cycle, the change process of vegetation carbon storage is closely related to climate change. Therefore, in order to mitigate the greenhouse effect and reduce the negative impact of extreme meteorological events on human social economy, it is very important to study the vegetation carbon storage capacity of terrestrial ecosystems. In addition, the analysis of changes in vegetation carbon storage is not only conducive to ensuring the ecological and environmental security of our country, but also helps resource planning and management departments to accurately estimate the regional vegetation carbon absorption and formulate reasonable and effective regional carbon neutrality planning schemes.

[0003] Due to the problems of unclear physical mechanism and overestimation of vegetation carbon storage in current carbon storage calculation methods, the inversion and estimation of regional vegetation carbon storage is still under research and exploration. At present, most vegetation carbon storage calculation methods mainly use historical climate and surface vegetation cover type and other factors combined with dynamic global vegetation models to estimate the time series of changes in surface vegetation carbon storage (Ecological Modelling, 2007, 200 (3-4): 279-307). However, previous calculation methods and models have not fully considered the impact of surface available water on vegetation carbon storage changes, and there are problems such as overestimation of surface vegetation carbon storage changes (Nature, 2018, 560 (7720): 628; Nature, 2021, 592 (7852): 65). In addition, the existing technology has failed to make full use of satellite remote sensing observation information. Based on satellite data and water balance methods, the time series of surface available water changes is established, and there is no literature report on the analysis of its impact on the change of terrestrial vegetation carbon storage. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a quantitative estimation method for vegetation carbon storage taking into account the available surface water, taking into account the impact of available terrestrial water on plant carbon fixation capacity and carbon storage, and providing a scientific reference basis for resource planning and management departments to more accurately estimate regional carbon storage capacity, thereby proposing a more reasonable carbon neutrality management planning scheme.

[0005] The present invention proposes a monthly-scale terrestrial vegetation carbon storage calculation optimization method based on terrestrial water storage, comprising the following steps:

[0006] Step 1: Select a typical research area, collect historical meteorological data and atmospheric carbon dioxide concentration data in the area, and obtain satellite remote sensing product data;

[0007] Among them, the typical study area is an area where the underlying surface coverage type is mainly forest grassland;

[0008] Step 2, reconstructing the time series of terrestrial water storage in the typical area and calculating the groundwater change;

[0009] Step 3, establishing a time series of atmospheric CO2 concentration in the typical area;

[0010] Step 4: Based on the measured meteorological data and available surface water in the study area, a response module of vegetation carbon storage to available water changes is constructed to quantify the impact of water changes on vegetation carbon storage.

[0011] Step 5: Based on the measured meteorological data of the study area and the vegetation carbon storage model of the land available water response module, a vegetation carbon storage quantitative estimation model is constructed, and the CO2 concentration is input into the vegetation carbon storage model as driving data.

[0012] Furthermore, the step 1 also includes preprocessing of the acquired data, specifically including:

[0013] Step 1.1: The meteorological data is a monthly scale dataset. Local observation data is used for downscaling. The meteorological data is converted from the monthly scale to the daily scale by linear interpolation.

[0014] Step 1.2: The satellite remote sensing product data include terrestrial water storage product data inverted by GRACE and GRACE-FO satellites. These data are preprocessed to unify the time scale and spatial resolution.

[0015] Furthermore, in step 2.1, the missing values ​​of the satellite inversion data are interpolated and filled using the long short-term memory method to reconstruct the time series of terrestrial water storage;

[0016] Step 2.2, unify the time scale and spatial resolution of the terrestrial water storage time series data, and then calculate the terrestrial water storage change based on the water balance principle.

[0017] Furthermore, in step 3, the CO2 data of the study area are interpolated and filled using extension interpolation.

[0018] Furthermore, satellite remote sensing data on land use types in the study area were collected to calculate the underlying surface vegetation cover types and coverage percentages;

[0019] The SEIB-DGVM model was optimized using soil moisture data, and a land available moisture response module was added to improve the calculation accuracy of vegetation carbon storage. The multivariate linear regression method was used to recalibrate the parameters, analyze the sensitivity of land carbon storage to changes in various components of surface moisture, and determine the response relationship of carbon storage to changes in surface moisture, as follows:

[0020]

[0021] in, is the vegetation carbon storage variable simulated by the SEIB-DGVM model, is the sensitivity coefficient of soil moisture content change, is the sensitivity coefficient of the change in snow water equivalent, is the sensitivity coefficient of groundwater change, is the sensitivity coefficient of the change in canopy interception water content, is the data residual, is the change in soil moisture content, is the change in snow water content equivalent, is the change in canopy interception water content, is the groundwater change.

[0022] Preferably, the specific process of step 5 is as follows:

[0023] The observed soil available moisture, meteorological data, CO2 data, vegetation data, and the moisture response module with calibrated parameters are input or coupled into the SEIB-DGVM model to construct a corrected vegetation carbon storage estimation model, as follows:

[0024]

[0025] in, The vegetation carbon storage calculated for the SEIB-DGVM model is to account for changes in available water.

[0026] The advantages of the present invention are:

[0027] Existing models mainly use atmospheric carbon dioxide concentration and meteorological data as driving data to simulate the changing trend of terrestrial vegetation carbon storage. Compared with the existing technology, the improvement of the present invention is to consider the impact of surface moisture limitation in the study area on the changing trend of terrestrial vegetation carbon storage. Combining remote sensing observation of terrestrial water storage data, atmospheric CO2 concentration data and meteorological data can simulate the changing trend of vegetation carbon storage under the influence of available land moisture. This method clarifies the mechanism of the influence of available soil moisture on vegetation carbon storage, and provides a scientific reference for resource planning and management departments to more accurately estimate regional carbon storage capacity, thereby proposing a more reasonable carbon neutrality management planning scheme. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a specific flow chart of the method of the present invention;

[0029] Figure 2 This is the GRACE gravity satellite interpolation processing map based on the long short-term memory method;

[0030] Figure 3 This is a graph showing the carbon storage results of vegetation in a typical study area simulated by the vegetation model constructed in the present invention taking into account the available terrestrial water. DETAILED DESCRIPTION

[0031] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and specific embodiments.

[0032] <Example>

[0033] like Figure 1 As shown, the present invention provides a monthly scale terrestrial vegetation carbon storage calculation optimization method based on terrestrial water storage, comprising the following steps:

[0034] Step 1: Select areas with forest or grassland as the main coverage type as typical study areas, collect historical meteorological data and atmospheric carbon dioxide (CO2) concentration data in the area, download a variety of satellite remote sensing products, mainly including terrestrial water storage products inverted by GRACE and GRACE-FO satellites and MCD12C1 underlying surface land use type products, evaporation, runoff, rainfall, soil moisture, canopy moisture, snow equivalent in high-cold areas from the Global Land Data Assimilation System GLDAS, and FLUXNET ground measurement data, pre-process these data, unify the time scale to the monthly scale and the spatial resolution to 0.5°.

[0035] The study first uses the changes in spherical harmonic coefficients and mass concentrations provided by gravity satellites, and performs correction processing through C20 correction, first-order correction, spatial smoothing and other methods to calculate the change in terrestrial water storage. The monthly data of spherical harmonic coefficients and mass concentrations mainly come from RL06 data. This embodiment selects data from April 2000 to April 2021. The RL06 version uses the latest time-varying gravity field model and standards, which is consistent with the model and processing standards used by the subsequent GRACEFollow-On gravity satellite, which greatly reduces the uncertainty brought by the method. The change in terrestrial water storage is expressed by equivalent water height (in cm), and the calculation formula is as follows:

[0036] (1)

[0038] (2)

[0040] Where: and Respectively remove the background field Spherical harmonic coefficients of the time-varying gravity field model of order m and month; is the average radius of the Earth, which is 6371 km; the average density of the solid Earth ≈5.517×10 3 kg / m 3 ;∆ is the change in bulk density of the substance; ()for Normalized Legendre function of order m; and Represent longitude and latitude respectively; is the distance between the particle and the differential unit; is the equivalent water height of land water storage change; and represent longitude and latitude respectively is the order of spherical harmonic expansion; m is the degree of spherical harmonic expansion; is the average radius of the earth; water density ≈10 3 kg / m 3 ; and are respectively the background field removed ( Love coefficients of earth model order loading obtained for linear interpolation; is the fan-shaped filter weight function, and the calculation formula is as follows:

[0041] (3)

[0042] It is obtained by the recursive formula, where r is the radius of the Gaussian smoothing function:

[0043] (4)

[0044] (5)

[0045] (6)

[0046] Similarly, Replace it with m, and we can get recursively This formula reflects that within the 70th order, as the smoothing radius increases, the curve converges faster and the weight of the low-order terms increases. In this embodiment, the smoothing radius is 1000 km.

[0047] According to the monthly average values ​​of precipitation, evaporation and runoff at 0.5°, the average values ​​of the 1°×1° grid containing four 0.5°×0.5° grids are calculated to obtain the precipitation, evaporation and runoff values ​​with a resolution of 1°. The calculation formula based on the water balance equation is shown in formula (7): The change in terrestrial water storage per unit time is equal to precipitation minus evaporation minus runoff. The change values ​​of terrestrial water storage at 0.5° and 1° (S 0.5° and S 1° ), establish S 1° And the empirical regression function relationship of GRACE:

[0048] (7)

[0049] (8)

[0050] Where: , , They are respectively the precipitation, evaporation and runoff of the i-th month of the assimilated data GLDAS on a 1° grid, in mm / month. The 1° grid terrestrial water storage value is obtained by GRACE satellite inversion, in mm / month. Compared with the water balance calculation There is a time lag effect of time t. In this example, t is 2, that is, the terrestrial water storage calculated based on the water balance method is shifted back by two months, and then the regression function equation is established.

[0051] The regression function with the highest significance test (P < 0.05) and the highest coefficient of determination was selected for downscaling. The simulated surface water storage value calculated based on the regression equation was Surface water storage observations obtained from satellite inversion Compare and calculate the residual value ( ), the calculation formula is as follows:

[0052] (9)

[0053] Based on formulas (8) and (9), S 0.5° and the empirical regression function of GRACE, and the residual value is used as a correction term for the S 0.5° Calculated GRACE 0.5° The results are corrected to obtain the final downscaling function relationship.

[0054] (10)

[0055] Where: is the surface water storage at 0.5° resolution obtained by fitting, mm / month; is the surface water storage at a resolution of 0.5°, based on the GLADS assimilated data, mm / month;

[0056] Step 2: Use the long short-term memory method to interpolate and fill in the missing values ​​of satellite inversion data to reconstruct the time series of terrestrial water storage. Due to technical problems, GRACE / GRACE-FO products lost data in the data gap of 11 months (July 2017 to May 2018). First, use the long short-term memory method (LSTM) to train the model, and input the temperature data and precipitation data obtained from ground observations and the surface water storage data calculated based on GLDAS as training samples into the model to reconstruct the time series of surface water storage changes and interpolate to fill in the missing values ​​of 11 months. Figure 2 As shown, a schematic diagram of a memory unit of the long short-term memory (LSTM) model constructed in this embodiment is given.

[0057] The spatial resolution of the interpolated and filled GRACE gravity satellite inverted terrestrial water storage sequence data was first adjusted to be consistent with the spatial resolution of the GLDAS assimilated data, which is 0.5°. In order to obtain the trend of the components of available surface water, especially the change in groundwater. The downscaled GRACE data and the resampled GLDAS data were combined to calculate soil moisture according to the water balance equation. The soil moisture change data was spatially filtered to eliminate outliers, and a high-precision soil available moisture data set was generated. The equation for calculating groundwater change is as follows:

[0058] (11)

[0059] in, is the change in terrestrial water storage, is the change in soil moisture content, is the change in snow water content equivalent, is the change in canopy interception water content, is the groundwater change.

[0060] Step 3: Establish the time series data of atmospheric CO2 concentration based on literature records and actual observation data, use extension interpolation to interpolate the CO2 data of the study area, and analyze the temporal variation trend of CO2;

[0061] Step 4: Collect satellite remote sensing data on land use types in the study area and calculate the underlying vegetation coverage type and coverage percentage. This example uses the MCD12C1 product of the MODIS satellite with a resolution of 1km to extract vegetation information, vegetation types, growth conditions, etc. Combined with the field survey, determine the vegetation function type, area, and growth degree in the study area; prepare a vegetation information map based on the underlying vegetation coverage.

[0062] Based on the measured meteorological data of the study area and the available land moisture rate, the parameters of the vegetation carbon storage quantitative estimation model were determined. The multivariate linear regression method was used to recalibrate the parameters, analyze the sensitivity of land carbon storage to changes in various components of surface moisture, and determine the response relationship of carbon storage to changes in surface moisture:

[0063] (12)

[0064] in, is the vegetation carbon storage variable simulated by the SEIB-DGVM model, is the sensitivity coefficient of soil moisture content change, is the sensitivity coefficient of the change in snow water equivalent, is the sensitivity coefficient of groundwater change, is the sensitivity coefficient of the change in canopy interception water content, is the data residual. The sensitivity coefficients in this step can be calculated according to the linear regression method. In this embodiment, the parameters are k1 is 0.3042, k2 is 0.0509, k3 is 0.022, and k4 is 0.0033.

[0065] Step 5: The observed soil available moisture, meteorological data, CO2 data, vegetation data, and the calibrated moisture response module are input or coupled into the SEIB-DGVM model to construct a corrected vegetation carbon storage estimation model:

[0066] (13)

[0067] in, The vegetation carbon storage calculated for the SEIB-DGVM model is to account for changes in available water.

[0068] Figure 3This is the temporal variation trend of the vegetation carbon storage results of the typical study area simulated by the vegetation model constructed by the present invention taking into account the available terrestrial water. The left ordinate in the figure is the CO2 concentration, and the right ordinate is the terrestrial vegetation carbon storage. The annual fluctuation pattern of vegetation carbon storage is basically the same, but the interannual variation trend is obvious, and vegetation carbon storage has a slow upward trend.

[0069] The inventive idea of ​​this embodiment is: first, collect historical meteorological data of the study area, including daily-scale temperature, precipitation, radiation, relative humidity and horizontal wind speed, collect atmospheric carbon dioxide concentration data, extract the terrestrial water storage series inverted by GRACE and GRACE-FO satellites covering the study area, and collect the global land surface data assimilation system GLDAS, including evaporation, runoff, rainfall, soil moisture, canopy moisture, snow equivalent in alpine areas, and unify the time scale and spatial resolution; use the long short-term memory method to interpolate and fill the GRACE satellite data to obtain a complete terrestrial water storage sequence; based on multi-source remote sensing meteorological data and atmospheric CO2 concentration data, use the SEIB-DGVM model to calculate and realize preliminary terrestrial vegetation carbon storage simulation; combine the data of GRACE satellite and GLDAS to obtain the changing trend of each surface water equivalent, and construct a response module of vegetation carbon storage to water equivalent changes; finally, combine the SEIB-DGVM model with the water change response module to simulate the surface vegetation carbon storage for a long time series.

[0070] The embodiments are only for illustrating the technical idea of ​​the present invention and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the present invention.

Claims

1. A monthly scale terrestrial vegetation carbon storage calculation optimization method based on terrestrial water storage, characterized in that: The following steps are involved: Step 1: Select a typical study area, collect historical meteorological data and atmospheric carbon dioxide concentration data in the area, obtain satellite remote sensing product data, and find out the land water storage data obtained by GRACE satellite inversion and the soil moisture content calculated based on water balance. There is a time lag effect of time t, and a regression function equation is established to obtain the surface water storage; Among them, the typical study area is the area where the underlying surface coverage type is mainly forest grassland; Step 2, reconstructing the time series of terrestrial water storage in the typical area and calculating the groundwater change; Step 3, establishing a time series of atmospheric CO2 concentration in the typical area; Step 4: Based on the measured meteorological data and available surface water in the study area, a response module of vegetation carbon storage to available water changes is constructed to quantify the impact of water changes on vegetation carbon storage. Step 5: Based on the measured meteorological data of the study area and the vegetation carbon storage model of the land available water response module, the observed soil available water, meteorological data, CO2 data and vegetation data, as well as the calibrated water change response module are input or coupled into the SEIB-DGVM model to construct a corrected vegetation carbon storage estimation model, as follows: in, The vegetation carbon storage calculated for the SEIB-DGVM model is To account for changes in available water, actual vegetation carbon storage is the sensitivity coefficient of soil moisture content change, is the sensitivity coefficient of the change in snow water equivalent, is the sensitivity coefficient of groundwater change, is the sensitivity coefficient of the change in canopy interception water content, is the change in soil moisture content, is the change in snow water content equivalent, is the change in canopy interception water content, is the groundwater change.

2. The monthly scale terrestrial vegetation carbon storage calculation optimization method based on terrestrial water storage according to claim 1 is characterized by: The step 1 also includes preprocessing of the acquired data, specifically including: Step 1.1: The meteorological data is a monthly scale dataset. Local observation data is used for downscaling. The meteorological data is converted from the monthly scale to the daily scale by linear interpolation. Step 1.2: The satellite remote sensing product data include terrestrial water storage product data inverted by GRACE and GRACE-FO satellites. These data are preprocessed to unify the time scale and spatial resolution.

3. The method for optimizing the calculation of terrestrial vegetation carbon storage on a monthly scale based on terrestrial water storage according to claim 2, characterized in that: Step 2.1, use the long short-term memory method to interpolate and fill in the missing values ​​of the satellite inversion data to reconstruct the time series of terrestrial water storage; Step 2.2, unify the time scale and spatial resolution of the terrestrial water storage time series data, and then calculate the terrestrial water storage change based on the water balance principle.

4. The method for optimizing the calculation of terrestrial vegetation carbon storage at a monthly scale based on terrestrial water storage according to claim 3, characterized in that: In step 3, the CO2 data of the study area are interpolated and filled using extension interpolation.

5. The method for optimizing the calculation of terrestrial vegetation carbon storage at a monthly scale based on terrestrial water storage according to claim 4, characterized in that: Collect satellite remote sensing data on land use types in the study area and calculate the underlying vegetation cover types and coverage percentages; The SEIB-DGVM model was optimized using soil moisture data, and a land available moisture response module was added to improve the calculation accuracy of vegetation carbon storage. The multivariate linear regression method was used to recalibrate the parameters, analyze the sensitivity of land carbon storage to changes in various components of surface moisture, and determine the response relationship of carbon storage to changes in surface moisture, as follows: in, is the vegetation carbon storage variable simulated by the SEIB-DGVM model, is the data residual.

Images