A method and device for dynamic inversion of global snow water equivalent impact of forests

Abstract

The present application discloses a method for dynamic inversion of forest impact on global snow water equivalent, comprising the following steps: calculating the microwave forest transmittance of the pixel based on the microwave forest transmittance of the coniferous forest, the microwave forest transmittance of the broad-leaved forest, and the proportion of the coniferous forest and the broad-leaved forest; inverting the exponential correlation length of the snow layer at the ground meteorological station, and then establishing a fitting model based on the exponential correlation length of the snow at the ground meteorological station, the snow depth at the ground station, and the brightness temperature gradient data of the satellite radiometer; performing difference on the exponential correlation length of the station and extending it to any position in the global space; making each snow pixel have an exponential correlation length value; determining the ground snow depth of the snow pixel based on the fitting model, using the exponential correlation length value of the snow pixel, and the brightness temperature gradient of the satellite radiometer, and then obtaining the snow water equivalent. The present application also includes a device for implementing the method. The present application solves the problem of low accuracy in the inversion of snow water equivalent in the prior art.

Description

Technical Field

[0001] The present application relates to the field of satellite remote sensing technology, and in particular to a global snow water equivalent dynamic inversion method and device taking into account the impact of forests. Background Art

[0002] The cryosphere plays an important role in the global climate system and is an indicator of global climate change. Snow is one of the main elements of the cryosphere and is widely distributed. At least 80% of the Eurasian and North American continents are covered by snow in winter, covering an area of ​​about 46.5 million square kilometers, and 3.8 million square kilometers in August (Frei and Robinson 1999). Snow has a high albedo, with the albedo of new snow reaching more than 0.9 and the albedo of old snow also above 0.4. Therefore, snow will reflect a large amount of solar short-wave incident radiation reaching the surface, profoundly affecting the energy balance of the earth's surface. Snow is a poor conductor of heat, which significantly reduces the heat exchange between the surface and the atmosphere, protects the growth of plants under the snow cover, and affects the development of frozen soil and other natural processes on the surface. Snow also plays a prominent role in the hydrological cycle and is very important for human production and life. In many mountainous areas, snowmelt runoff is a reliable and stable water source used for irrigation and life. At the same time, snow disasters and unexpected avalanches in pastoral areas can also pose a major threat to the safety of life and property of local people. Therefore, it is very important to obtain high-precision snow distribution. Passive microwave remote sensing has high temporal resolution and can work all day and all weather, making it a commonly used remote sensing data for large-scale snow monitoring. The key factors restricting the accuracy of passive microwave remote sensing snow depth inversion are the shielding attenuation effect of forests and the temporal and spatial changes in snow characteristics.

[0003] The existing snow water equivalent inversion algorithm seriously underestimates the inversion results in forest areas. The main reasons are the attenuation of microwave signals from snow under the forest by the forest and the radiation effect of the forest itself. The complexity of the interaction between the vegetation canopy and electromagnetic waves comes from the coexistence of attenuation and emission. The canopy attenuates the microwave radiation from the snow under the forest, while also emitting some energy itself. Especially in densely forested areas, high-frequency microwaves cannot penetrate the canopy. At this time, the radiation characteristics are completely dominated by the forest, and the snow information cannot be reflected. There are three main solutions to the impact of forests: the first research direction is to directly establish the relationship between the brightness temperature difference at 18.7 GHz and 36.5 GHz and the snow water equivalent under pure forest pixels (Goita et al. 2003). This method cannot solve the problem of non-pure forest pixels, nor does it distinguish between broad-leaved forests and coniferous forests. Different types of forest radiation have large differences; the second research direction is to introduce forest cover factors through coefficient correction. The US Snow and Ice Data Center (NSIDC) product and China Western Data Center (WESTDC) product both use this method. This method also does not distinguish between broad-leaved forests and coniferous forests. At the same time, because the scaling factor of the coefficient is difficult to control, the impact of forests is overestimated, resulting in a serious overestimation of snow depth; the third research direction is semi-empirical models, microwave transmittance zero-order semi-empirical τ-ω model and empirical function based on timber volume (European Space Agency GlobSnow product uses this method). Semi-empirical models require the actual forest biomass in the study area to calibrate parameters, otherwise the empirical parameters will have large errors on a large regional scale. This method is not suitable for application on a global scale because it requires a large number of field forest biomass parameters.

[0004] The spatiotemporal heterogeneity of snow microstructure parameters is one of the main factors affecting the inversion of snow water equivalent. Due to changes in meteorological factors (temperature, wind speed, water vapor, etc.), snow parameters will evolve over time, and the radiation characteristics of snow will change significantly. In view of the impact of changes in snow characteristic parameters, there are certain considerations in empirical algorithms and algorithms based on physical models. In the inversion algorithm, the prior information of snow structure parameters is crucial for snow depth inversion, and the snow microstructure parameters will gradually change with the evolution of time. The various regional linear statistical empirical methods currently used set various parameters to fixed values ​​and cannot solve the impact of parameter evolution. At present, there are some problems in the assumptions of the research on eliminating the influence of snow structure parameters. In the AMSR2 empirical algorithm, the polarization difference index is introduced to establish a time empirical function to characterize the evolution of snow parameters. This method has been verified to have poor global applicability and the algorithm has serious overestimation. Dai Liyun (2012) and Che Tao (2016) fixed the prior information of field observations by month in time, but there will be a snowfall offset in the time of each snow season, so there will be obvious errors between years. At the same time, the multi-layer prior information is limited to a small area and cannot be applied on a global scale. Pulliainen (2006) (GlobSnow product) calculated the particle size by assimilating the snow depth of the station in the Arctic region. This type of assimilation algorithm has high requirements on the accuracy of the brightness temperature observation operator. Otherwise, the optimized effective snow particle size contains the errors of the input data and the model, distorting the evolution trend of the particle size. At the same time, the assimilation algorithm has high computational cost.

[0005] Existing technologies do not consider the impact of forest types (coniferous forests, broad-leaved forests). The shielding effect of forests on snow water equivalent will lead to a significant underestimation of the inversion results. Passive microwave snow depth and snow water equivalent inversion uses the brightness temperature difference between 18.7 GHz and 36.5 GHz, but the attenuation effect of forests on the two frequencies is different, and the radiation of the forest itself also has a large difference at these two frequencies.

[0006] The existing technology does not consider the relationship between the exponential correlation length and snow depth. The evolution of the snow microstructure parameter (exponential correlation length) will lead to the inversion results of snow water equivalent being overestimated in the early snow season and underestimated in the late snow season. Summary of the invention

[0007] The present application proposes a method and device for dynamically inverting the global snow water equivalent of forest impact, which solves the technical problem of low accuracy in snow water equivalent inversion in the prior art.

[0008] The present application embodiment provides a method for dynamically inverting the impact of forests on global snow water equivalent, comprising the following steps:

[0009] The microwave forest transmittance of the pixel is calculated according to the microwave forest transmittance of the coniferous forest, the microwave forest transmittance of the broad-leaved forest, and the proportion of the coniferous forest and the broad-leaved forest; preferably, for the forest area other than the coniferous forest and the broad-leaved forest, the microwave forest transmittance uses the average value of the microwave forest transmittance of the coniferous forest and the microwave forest transmittance of the broad-leaved forest;

[0010] Invert the snow cover index correlation length of the ground meteorological station, and then establish a fitting model between the snow cover index correlation length, the snow depth of the ground station, and the satellite radiometer brightness temperature gradient data according to the snow cover index correlation length of the ground meteorological station, the snow depth of the ground station, and the satellite radiometer brightness temperature gradient data;

[0011] The exponential correlation length of the station is interpolated and extended to any position in the global space; each snow pixel has an exponential correlation length value;

[0012] Determine the ground site snow depth of the snow pixel according to the fitting model, using the exponential correlation length value of the snow pixel and the satellite radiometer brightness temperature gradient of the snow pixel;

[0013] The snow water equivalent is calculated based on the snow density and snow depth.

[0014] Preferably, the microwave forest transmittance of coniferous forests or the microwave forest transmittance of broad-leaved forests is inverted using a snow radiation transmission model, which specifically includes the following steps: using snow profile information provided by field observation data of snow, snow surface observation brightness temperature, satellite-observed atmospheric precipitable water data, and surface cover data, at a set frequency, the microwave forest transmittance changes from 0 to 1, and traverses according to a set step size, and forward simulates the simulated satellite brightness temperature of the snow profile point reaching the sensor; extracting the satellite observation brightness temperature corresponding to the field observation data from the satellite sensor, and then obtaining the microwave forest transmittance when the error between the simulated satellite brightness temperature and the satellite observation brightness temperature is the smallest.

[0015] Preferably, the snow radiation transfer model is used to invert the snow layer index correlation length of the ground meteorological station, which specifically includes the following steps: using the snow depth data and surface cover data of the ground meteorological station, at a set frequency, the snow layer index correlation length is traversed within a set range and according to a set step size, and the simulated satellite brightness temperature of the ground meteorological station reaching the sensor is forward simulated; the satellite observed brightness temperature corresponding to the location of the meteorological station is extracted from the satellite sensor, and then the snow layer index correlation length is obtained when the error between the simulated satellite brightness temperature and the satellite observed brightness temperature is minimized.

[0016] Preferably, the 18.7 GHz and 36.5 GHz microwave forest transmittances corresponding to the snow profile information provided by the snow field observation data are obtained by inverting the snow radiation transfer model.

[0017] Preferably, the 18.7 GHz and 36.5 GHz snow layer index correlation lengths corresponding to the snow depth at the ground meteorological station are obtained by inverting the snow radiation transfer model.

[0018] Preferably, the optimal microwave forest transmittances at vertical polarizations of 18.7 GHz and 36.5 GHz are 0.723 and 0.601, respectively; and the results for broadleaf forest at vertical polarizations of 18.7 GHz and 36.5 GHz are 0.821 and 0.615, respectively.

[0019] To implement the method of the present application, before each of the steps, the present application further includes a data acquisition step, including: satellite radiometer brightness temperature data, snow accumulation field observation data, snow surface observation brightness temperature; atmospheric precipitable water data, ground meteorological station snow depth, longitude and latitude data, surface cover data, and at least one of the following data.

[0020] The present application also proposes a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the method described in any embodiment of the present application is implemented.

[0021] Furthermore, the present application also proposes an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method described in any embodiment of the present application is implemented.

[0022] At least one of the above technical solutions adopted in the embodiments of the present application can achieve the following beneficial effects:

[0023] The present invention considers the influence of forest types, combines ground observation data to establish the microwave radiation transmission process of forest snow, and optimizes and iterates the microwave forest transmittance at 18.7GHz and 36.5GHz frequencies. The solution of the present invention reduces the influence of forests on the radiation brightness temperature of snow under the forest, and improves the accuracy of snow depth inversion in forest areas.

[0024] The present invention brings the optimized microwave forest transmittance into the snow radiation transmission model containing the forest, inverts the snow exponential correlation length, and thus can dynamically calculate the impact of the evolution of snow characteristic parameters. By dynamically inverting the exponential correlation length, the degree of snow particle size evolution can be dynamically determined, and the inversion model of the corresponding stage can be found. The algorithm can be implemented in a large area, thereby improving the accuracy of global snow water equivalent inversion. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0026] Figure 1 is a flow chart of an embodiment of the method of the present invention;

[0027] Figure 2 Calculation process for transmittance of coniferous forests;

[0028] Figure 3 The vertical polarization results of coniferous forest transmittance at 18.7 GHz and 36.5 GHz;

[0029] Figure 4 This is the flow chart for the inversion of global snow water equivalent;

[0030] Figure 5 This is the accuracy verification result from 2019 to 2020;

[0031] Figure 6 is the root mean square error distribution diagram. DETAILED DESCRIPTION

[0032] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0033] The present invention proposes a global snow water equivalent dynamic inversion algorithm taking forest influence into consideration, thereby improving the accuracy of snow water equivalent inversion.

[0034] Snow water equivalent is defined as the height of liquid water after the snow melts, and is one of the most critical parameters for describing seasonal snow reserves. The inversion of snow water equivalent has been developed for more than 30 years, but there are still many challenges in extracting the distribution of snow water equivalent with high precision on a large scale around the world. Among them, the spatiotemporal changes in forest canopy attenuation and snow microstructure parameters (exponential correlation length) are the two most important issues affecting the accurate inversion of snow water equivalent. Therefore, this patent mainly focuses on the two major factors that affect the accuracy of snow water equivalent inversion, namely forest and particle size, and conducts research to weaken the impact of these two factors and improve the accuracy of global snow water equivalent inversion.

[0035] The technical solutions provided by various embodiments of the present application are described in detail below in conjunction with the accompanying drawings.

[0036] Figure 1The flowchart of the embodiment of the method of the present invention is as follows. The present invention first combines field observation data to simulate brightness temperature through a model to match satellite observation brightness temperature, optimizes iterative coniferous forest microwave forest transmittance, and then calculates microwave forest transmittance of each microwave pixel through a linear mixed pixel decomposition method according to forest percentage data; then, the microwave forest transmittance of each microwave pixel is substituted into the snow forest radiation transmission model, the exponential correlation length is dynamically inverted, and then combined with the passive microwave radiometer brightness temperature gradient, a new dynamic inversion algorithm for global forest and non-forest areas is established.

[0037] The embodiment of the present application specifically includes the following steps 10 to 70:

[0038] Step 10: Data acquisition, including: satellite radiometer brightness temperature data, snow accumulation field observation data, snow surface observation brightness temperature; atmospheric precipitable water data, ground meteorological station snow depth, longitude and latitude data, and surface cover data.

[0039] Here we explain the content of each data and its use in the subsequent steps of this method.

[0040] Satellite radiometer brightness temperature data: Preferably, the Advanced Microwave Scanning Radiometer 2 (AMSR2) is used. AMSR2 is a passive microwave radiometer carried on the Global Change Observation Mission–Water 1 (GCOM-W1). In step 30, preferably, AMSR2 radiometer brightness temperature data is used as the brightness temperature observed by the satellite, with frequencies of 18.7 GHz and 36.5 GHz as comparison data in the objective function.

[0041] Snow field observation data: used to provide snow profile information for step 30, including: snow stratification, snow layer depth, snow density, snow particle size, snow temperature, etc. For example, in an embodiment of the present application, in order to obtain the microwave forest transmittance of coniferous forests, the experimental area for snow field observation data is located in Colorado, USA. The field observation experiment is named SnowEx, and the field observation time is from 20170206 to 20170225. There are a total of 255 snow profiles (forest area + forest gaps). The profiles with wet snow inside the snow layer and incomplete data records are eliminated, and 85 dry snow profile data are screened out. In step 30, the data is brought into the MEMLS snow emission model to calculate the radiated brightness temperature of the snow surface, which is used as the simulated understory brightness temperature for input into the snow radiation transmission model.

[0042] Snow surface brightness temperature observation: The snow brightness temperature observation at the snow profile point comes from the ground-based microwave radiometer (SBR). In step 30, it is used as the observed understory brightness temperature to input the snow radiation transfer model.

[0043] Atmospheric precipitable water data: preferably, ECWMF TCWV data is used. In the radiation transfer process, in order to calculate the influence of the atmosphere, atmospheric precipitable water data is needed to calculate parameters such as atmospheric optical thickness and atmospheric transmittance. In step 30, it is used to input the snow radiation transfer model.

[0044] Snow depth at ground meteorological stations: Global meteorological station snow depth dataset (GHCN) and snow cover station data in China provided by the China Meteorological Administration (quality control of the station is required: the water area in the pixel is less than 30% and the temperature is less than zero). In step 40, the snow depth at the ground meteorological station is used to invert the snow layer index correlation length.

[0045] Surface cover data, preferably MODIS surface cover data MOD12Q1, uses the IGBP surface type classification data in the data to calculate the coverage of underlying surfaces such as forests, farmlands, bare soil, grasslands, and water bodies within the microwave brightness temperature grid in order to develop algorithms (eliminate the impact of large areas of water and reduce the impact of forests). In step 30, forest coverage is used to input the snow radiation transfer model.

[0046] Step 20: pre-process the acquired data;

[0047] The acquired satellite radiometer brightness temperature data is preprocessed. For example, the acquired AMSR2 data is preprocessed and divided into radiation calibration. The radiation calibration includes: using the radiation calibration formula Tb = Gain × DN + Intercpt, where Tb is the brightness temperature value after radiation calibration, Gain and Intercpt are radiation calibration parameters, DN is the pixel value in the data provided by AMSR2, and according to the scientific data set provided by AMSR2, Gain = 0.01 and Intercpt = 0.

[0048] Preprocess the acquired snow characteristics data, because the passive microwave radiometer can only invert the snow water equivalent under dry snow conditions, so it is necessary to extract the snow profile information and snow surface brightness temperature with snow humidity greater than zero from the snow field observation data;

[0049] Pre-process the snow depth data of ground meteorological stations, and remove the data of meteorological stations where the water area of ​​the microwave pixel exceeds 30%;

[0050] The acquired surface cover data were preprocessed, and the land classification data were converted into grid data with a resolution of 0.25° to match the microwave pixels according to the conversion method of geographic coordinates. MOD12Q1 provided five forest cover types, including evergreen coniferous forest, evergreen broad-leaved forest, deciduous coniferous forest, deciduous broad-leaved forest and mixed forest, and the forest coverage rate data with a resolution of 0.25° was calculated.

[0051] Step 30: Use the snow radiation transmission model to invert the microwave forest transmittance of the coniferous forest and / or the microwave forest transmittance of the broad-leaved forest, and then calculate the microwave forest transmittance of the pixel based on the proportion of the coniferous forest and the broad-leaved forest.

[0052] Figure 2 This is the calculation process of coniferous forest transmittance, which is used to illustrate how to use the snow radiation model to invert the coniferous forest microwave transmittance in step 30.

[0053] In order to invert the microwave forest transmittance through the snow radiation transfer model, the data acquisition step includes satellite radiometer brightness temperature data, snow field observation data, snow surface observation brightness temperature; atmospheric precipitable water data, longitude and latitude data, and surface cover data.

[0054] Specifically, on the one hand, the satellite-observed brightness temperatures at 18.7 GHz and 36.5 GHz of the corresponding field-observed snow profiles are extracted from the passive microwave brightness temperature data on satellite sensors. On the other hand, the simulated satellite brightness temperature of the snow profile point reaching the sensor is forward simulated. During the simulation process, other parameters are determined (the snow profile information, snow surface observation brightness temperature, atmospheric precipitable water data, and surface cover data provided by the snow field observation data in steps 10 to 20 are used. Other related parameters can be set to typical values, such as the forest temperature can be set to 258K, the forest single scattering albedo can be set to 0.064, the snow density is set to 240kg / m3, the reflectivity of the 18.7GHz snow-soil interface in vertical polarization and horizontal polarization are 0.05 and 0.06 respectively, the reflectivity of the 36.5GHz snow-soil interface in vertical polarization and horizontal polarization are 0.03 and 0.05 respectively, the 18.7GHz sky background radiation is 15K, and the 36.5GHz sky background radiation is 25K), and the forest transmittances at 18.7GHz and 36.5GHz vary from 0 to 1, respectively, and are traversed according to the set step size (for example, 0.001). Finally, an objective function is established between the brightness temperature observed by the satellite and the simulated brightness temperature. The forest transmittance corresponding to the minimum objective function value is the optimal transmittance. In other words, the transmittance result is when the error between the brightness temperature simulated by the model and the brightness temperature observed by the satellite is the smallest.

[0055] The following is an example of 18.7GHz and 36.5GHz microwave forest transmittance, the objective function is:

[0056]

[0057] The above formula is the objective function, which includes the error between the model-simulated brightness temperature and the satellite-observed brightness temperature. Where N is the number of snow profiles, and The root mean square error (RMSE) of the brightness temperature at 18.7 GHz and 36.5 GHz, is the RMS error of the brightness temperature difference between 18.7 GHz and 36 GHz.

[0058] Figure 3 The vertical polarization results of the coniferous forest transmittance at 18.7 GHz and 36.5 GHz. Studies have shown that the microwave forest transmittance of coniferous forests is sensitive to vertical polarization, so the microwave forest transmittance in the vertical polarization direction can be used. The gray scale in the figure, the smaller the value, the smaller the objective function value.

[0059] The microwave forest transmittance of the pixel is calculated based on the microwave forest transmittance of coniferous forests, the microwave forest transmittance of broad-leaved forests, and the proportion of coniferous forests and broad-leaved forests; preferably, for forest areas other than coniferous forests and broad-leaved forests, the microwave forest transmittance uses the average value of the microwave forest transmittance of coniferous forests and the microwave forest transmittance of broad-leaved forests.

[0060] T=f needle ×t needle +f broad ×t broad +(ff needle -f broad )×(t needle +t broad ) / 2 (2) Where T is the microwave pixel forest transmittance, f represents the pixel forest percentage, and f needle represents the percentage of coniferous forest in the microwave pixel, f broad Indicates the percentage of broad-leaved forest in the microwave pixel; t needle represents the microwave forest transmittance of coniferous forest, t broad Represents the microwave forest transmittance of broad-leaved forest.

[0061] In the embodiments of the present application, for example, using the field observation data of SnowEx, according to Figure 2 For the steps of the illustrated embodiment, the optimum transmittances for vertical polarization at 18.7 GHz and 36.5 GHz are 0.723 and 0.601, respectively.

[0062] Therefore, the microwave forest transmittance of vertical polarization of a microwave pixel at a set frequency is:

[0063] T f_18.7V =f needle×0.723+f broad ×0.821+(ff needle -f broad )×(0,723+0.821) / 2 (3.1)

[0064] T f_36.5V =f needle ×0.601+f broad ×0.615+(ff needle -f broad )×(0,601+0.615) / 2 (3.2)

[0065] Among them, T f_18.7V and T f_36.5V The microwave pixel forest transmittance at 18.7 GHz and 36.5 GHz respectively. The microwave forest transmittance of the broad-leaved forest with vertical polarization at 18.7 GHz and 36.5 GHz is 0.821 and 0.615 respectively, and the results are cited from the prior art (Che et al. 2016).

[0066] Step 40, invert the snow layer index correlation length of the ground meteorological station, and then establish a fitting model between the snow index correlation length, the snow depth of the ground station, and the satellite radiometer brightness temperature gradient data based on the snow index correlation length of the ground meteorological station, the snow depth of the ground station, and the satellite radiometer brightness temperature gradient.

[0067] Figure 4 The invention is a flowchart of global snow water equivalent inversion, and the optimal iterative method inversion of the snow layer index correlation length of the meteorological station includes: using the snow depth data and surface cover data of the ground meteorological station, at a set frequency, the snow layer index correlation length is traversed within a set range and according to a set step size, and the simulated satellite brightness temperature of the ground meteorological station reaching the sensor is forward simulated; the satellite observed brightness temperature corresponding to the position of the meteorological station is extracted from the satellite sensor, and then the snow layer index correlation length is obtained when the error between the simulated satellite brightness temperature and the satellite observed brightness temperature is minimized.

[0068] In order to invert the snow layer index correlation length of the meteorological station through the snow radiation transfer model, the data acquisition step includes at least one of the following data: satellite radiometer brightness temperature data, ground meteorological station snow depth, longitude and latitude data, and surface cover data. Other data can be set to typical values, such as snow density of 240kg / m 3 , the snow temperature is 258K, and the snow-soil interface temperature is 258K.

[0069] In the embodiment of the present application, the vertical polarization difference (18.7-36.5) GHz is used as the inversion operator. When the error between the model simulation and the satellite data is the smallest, the result of the correlation length at this time is output as the inversion value of the site. The objective function is:

[0070]

[0071] Among them, P ex is the exponential correlation length, ρ is the snow density, SD is the snow depth, Tb model is the brightness temperature simulated by the model, Tb AMSR2 is the brightness temperature observed by satellite.

[0072] According to previous studies, the impact of forests is mainly affected by the forest percentage f, microwave forest transmittance Tf_18.7V and Tf_36.5V, and forest temperature. By bringing in the microwave forest transmittance of each pixel calculated in step 3, the radiation process of the forest can be accurately described, that is, the impact of the forest is effectively reduced. Through optimal iteration, the inversion result of the exponential correlation length of the site is obtained.

[0073] The exponential correlation length inversion algorithm is used to calculate the exponential correlation length results of meteorological stations over the years, which are then segmented and screened. The snow depth at the station is regressed and fitted with the corresponding satellite brightness temperature gradient to obtain a new inversion model. The modeling data are the snow stations from 2012 to 2018. The dynamic inversion model is shown in Table 1.

[0074] Table 1 Dynamic snow depth inversion model

[0075]

[0076] Tb18V represents the brightness temperature of the 18.7 GHz V polarization channel; Tb37V represents the brightness temperature of the 36.5 GHz V polarization channel.

[0077] As shown in Table 1, we established snow depth inversion models for forested and non-forested areas under different exponential correlation lengths. In the process of snow depth inversion, the corresponding snow depth inversion algorithm was found based on the exponential correlation length results of each pixel in space.

[0078] Step 50, interpolate the exponential correlation lengths of the sites and extend them to any position in the global space;

[0079] Spatial exponential correlation length inversion: Perform Kriging interpolation on the exponential correlation length of the station, expand the result to space, and obtain the effective correlation length distribution;

[0080] Step 60, by identifying snow accumulation in global pixels and combining effective correlation length distribution, each snow accumulation pixel has an exponential correlation length value;

[0081] Kelly snow recognition algorithm is used to identify dry snow pixels, and then the interpolation result is masked to obtain the dry snow area. At the same time, each snow pixel has an exponential correlation length value.

[0082] Step 70: Determine the ground snow depth of the snow pixel according to the fitting model, using the exponential correlation length value of the snow pixel and the satellite radiometer brightness temperature gradient of the snow pixel, and further calculate the snow water equivalent;

[0083] In this step, the fitting model is used as a semi-empirical dynamic physical model, as shown in Table 1, where threshold division is performed according to the results of the exponential correlation length of the meteorological station. Therefore, combined with the brightness temperature gradient of the passive microwave radiometer, the ground snow depth of the snow pixel can be determined, thereby realizing the dynamic inversion of the global snow water equivalent.

[0084] The snow water equivalent results are calculated based on the snow density and snow depth (SD, i.e. y in Table 1):

[0085] Snow pressure (g / cm 2 ) = snow density (g / cm 3 )×snow depth (cm) (5.1)

[0086] Snow water equivalent (mm) = snow pressure (g / cm 2 )×10 / (water density (g / cm 3 )) (5.2)

[0087] The global snow water equivalent total technical route is as follows Figure 4 shown.

[0088] Figure 5 This is the accuracy verification result from 2019 to 2020. Regardless of whether it is a forest area or a non-forest area, the total accuracy of the snow water equivalent is within 40 mm, which can meet the requirements of the climate hydrological model. At the same time, for the case of a wider spatial distribution of snow water equivalent less than 150 mm, the accuracy is higher and the error is less than 20 mm. The specific accuracy is as follows: the accuracy of the snow water equivalent less than 150 mm in non-forest areas is 13.1 mm, the accuracy of the snow water equivalent less than 250 mm is 19.5 mm, and the total accuracy is 26.5 mm; the accuracy of the snow water equivalent less than 150 mm in forest areas is 18.8 mm, the accuracy of the snow water equivalent less than 250 mm is 26 mm, and the total accuracy is 35.1 mm.

[0089] Figure 6 This is the root mean square error distribution diagram. It can be seen from the figure that in low and medium latitudes, whether in forest areas or non-forest areas, the accuracy of snow water equivalent is basically below 30 mm. Places with errors above 30 mm are mainly distributed in high-altitude forest areas.

[0090] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0091] Therefore, the present application also proposes a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the method described in any embodiment of the present application is implemented.

[0092] Furthermore, the present application also proposes an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in any embodiment of the present application when executing the computer program.

[0093] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0094] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0095] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1The steps for the functions specified in one or more boxes.

[0096] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0097] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0098] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0099] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0100] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A dynamic inversion method for forest impact on global snow water equivalent, characterized in that: The following steps are involved: The microwave forest transmittance of coniferous forest or broad-leaved forest is inverted by using the snow radiation transmission model, specifically including: using the snow profile information provided by the snow field observation data, the snow surface observation brightness temperature, the atmospheric precipitable water data observed by satellite, and the surface cover data; at the set frequency, the microwave forest transmittance changes from 0 to 1, and traverses according to the set step size, and forward simulates the simulated satellite brightness temperature of the snow profile point reaching the sensor; extracts the satellite observation brightness temperature corresponding to the field observation data from the satellite sensor, and then obtains the microwave forest transmittance when the error between the simulated satellite brightness temperature and the satellite observation brightness temperature is the smallest; The microwave forest transmittance of the pixel is calculated based on the microwave forest transmittance of the coniferous forest, the microwave forest transmittance of the broad-leaved forest, and the proportion of the coniferous forest and the broad-leaved forest. For forest areas other than the coniferous forest and the broad-leaved forest, the microwave forest transmittance uses the average value of the microwave forest transmittance of the coniferous forest and the microwave forest transmittance of the broad-leaved forest. The snow radiation transfer model is used to invert the snow layer index correlation length of the ground meteorological station, specifically including: using the snow depth data and surface cover data of the ground meteorological station, at a set frequency, the snow layer index correlation length is within a set range, and traverses according to a set step size, and forward simulates the simulated satellite brightness temperature of the ground meteorological station reaching the sensor; extracting the satellite observed brightness temperature corresponding to the position of the meteorological station from the satellite sensor, and then obtaining the snow layer index correlation length when the error between the simulated satellite brightness temperature and the satellite observed brightness temperature is the smallest; Then, according to the snow layer index correlation length of the ground meteorological station, the snow depth of the ground station, and the satellite radiometer brightness temperature gradient data, a fitting model between the snow layer index correlation length, the snow depth of the ground station, and the satellite radiometer brightness temperature gradient is established; The exponential correlation length of the station is interpolated and extended to any position in the global space; each snow pixel has an exponential correlation length value; Determine the ground snow depth of the snow pixel according to the fitting model, using the exponential correlation length value of the snow pixel and the satellite radiometer brightness temperature gradient of the snow pixel; The snow water equivalent is calculated based on the snow density and snow depth.

2. The method for dynamic inversion of forest impact on global snow water equivalent as claimed in claim 1, characterized in that: The 18.7 GHz and 36.5 GHz microwave forest transmittances corresponding to the snow profile information provided by the snow field observation data were obtained through the inversion of the snow radiation transfer model.

3. The method for dynamic inversion of forest impact on global snow water equivalent according to claim 1, characterized in that: The 18.7 GHz and 36.5 GHz snow layer index correlation lengths corresponding to the snow depth at the ground meteorological station were obtained by inverting the snow radiation transfer model.

4. The method for dynamic inversion of global snow water equivalent of forest impact as claimed in claim 1, characterized in that: The optimal microwave forest transmittances at vertical polarizations of 18.7 GHz and 36.5 GHz were 0.723 and 0.601, respectively; the results for broad-leaved forests at vertical polarizations of 18.7 GHz and 36.5 GHz were 0.821 and 0.615, respectively.

5. The method for dynamic inversion of forest impact on global snow water equivalent according to claim 1, characterized in that: In order to invert the microwave forest transmittance through the snow radiation transmission model, the method further includes a data acquisition step, including: at least one of satellite radiometer brightness temperature data, snow field observation data, snow surface observation brightness temperature, atmospheric precipitable water data, longitude and latitude data, and surface cover data.

6. The method for dynamic inversion of forest impact on global snow water equivalent according to claim 1, characterized in that: In order to invert the snow layer exponential correlation length of the meteorological station through the snow radiation transfer model, the method further includes a data acquisition step, including: satellite radiometer brightness temperature data, ground meteorological station snow depth, longitude and latitude data, and at least one of surface cover data.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

8. An electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 6 is implemented.

Images