CN116738161A - Frozen soil active layer thickness inversion method, device, equipment and storage medium - Google Patents
Frozen soil active layer thickness inversion method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN116738161A CN116738161A CN202310499168.0A CN202310499168A CN116738161A CN 116738161 A CN116738161 A CN 116738161A CN 202310499168 A CN202310499168 A CN 202310499168A CN 116738161 A CN116738161 A CN 116738161A
- Authority
- CN
- China
- Prior art keywords
- soil
- active layer
- data
- thickness
- index
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 239000002689 soil Substances 0.000 title claims abstract description 272
- 238000000034 method Methods 0.000 title claims abstract description 62
- 238000003860 storage Methods 0.000 title claims abstract description 11
- 230000000694 effects Effects 0.000 claims abstract description 45
- 230000008014 freezing Effects 0.000 claims abstract description 31
- 238000007710 freezing Methods 0.000 claims abstract description 31
- 238000010801 machine learning Methods 0.000 claims abstract description 29
- 238000010257 thawing Methods 0.000 claims abstract description 23
- 230000008859 change Effects 0.000 claims abstract description 22
- 230000007613 environmental effect Effects 0.000 claims abstract description 17
- 238000012549 training Methods 0.000 claims description 42
- 238000012360 testing method Methods 0.000 claims description 24
- 238000007637 random forest analysis Methods 0.000 claims description 20
- 238000004458 analytical method Methods 0.000 claims description 13
- 230000008018 melting Effects 0.000 claims description 11
- 238000002844 melting Methods 0.000 claims description 11
- 238000007781 pre-processing Methods 0.000 claims description 11
- 238000012952 Resampling Methods 0.000 claims description 10
- 238000012847 principal component analysis method Methods 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 8
- 238000011160 research Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 4
- 230000002194 synthesizing effect Effects 0.000 claims description 4
- 239000000155 melt Substances 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 238000011426 transformation method Methods 0.000 claims description 3
- 239000002131 composite material Substances 0.000 claims description 2
- 238000004088 simulation Methods 0.000 abstract description 5
- 230000006866 deterioration Effects 0.000 abstract description 4
- 239000000126 substance Substances 0.000 abstract description 4
- 230000009466 transformation Effects 0.000 abstract description 4
- 238000005553 drilling Methods 0.000 description 10
- 238000004422 calculation algorithm Methods 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 238000012706 support-vector machine Methods 0.000 description 5
- 238000010792 warming Methods 0.000 description 5
- 238000000513 principal component analysis Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 230000015556 catabolic process Effects 0.000 description 2
- 238000006731 degradation reaction Methods 0.000 description 2
- 238000001556 precipitation Methods 0.000 description 2
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 1
- 241000282326 Felis catus Species 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 229910052799 carbon Inorganic materials 0.000 description 1
- 239000004927 clay Substances 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 230000010485 coping Effects 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 230000003090 exacerbative effect Effects 0.000 description 1
- 239000005431 greenhouse gas Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000011065 in-situ storage Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000005416 organic matter Substances 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 239000011435 rock Substances 0.000 description 1
- 239000004576 sand Substances 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000012876 topography Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/08—Thermal analysis or thermal optimisation
Abstract
The invention provides a frozen soil active layer thickness inversion method, a device, equipment and a storage medium. Considering the influence of factors such as continuous deterioration of climate conditions, aggravation of human activities, increasingly complex physical and chemical properties of soil on the freeze thawing change of a frozen soil area, utilizing noctilucent remote sensing data to represent human activity intensity, utilizing main component transformation to compress a soil data set so as to contain richer information in the same dimension, and fitting the relation between an environment prediction factor and the thickness of the frozen soil active layer based on the thickness point data of the soil active layer and multisource remote sensing data; 8 machine learning methods are selected to construct a best fit model between the environmental predictors and the frozen soil active layer thickness. And calculating the time sequence change condition of the layer thickness of the frozen soil activity layer from the beginning of the century to 2100 years based on the optimal result obtained by prediction by using the CMIP6 temperature data. Compared with the traditional inversion method, the method can obtain the thickness simulation result of the permafrost active layer with higher precision, thereby analyzing the permafrost freezing and thawing change rule more comprehensively and effectively.
Description
Technical Field
The invention relates to the field of analysis of frozen soil freezing and thawing change rules, in particular to a frozen soil activity layer thickness inversion method, device and equipment and a storage medium.
Background
The Qinghai-Tibet plateau is known as the world ridge and the world third pole, and is the highest and largest plateau in the world, and the altitude is from 87 meters to 8354 meters. The average air temperature in QTP is-6-20 ℃, the annual precipitation is 50-2000 mm, and the vegetation community transits from the subtropical rain forest in the southeast part to the alpine desert in the northwest part. The high altitude and severe weather form the highest and largest permanent frozen soil in low and medium latitude areas of the world, with permafrost and seasonal frozen soil areas accounting for 40.2% and 56.0%, respectively. Compared with other permafrost regions with the same latitude as the northern hemisphere, the permafrost in the Qinghai-Tibet plateau region has the characteristics of high temperature, thin thickness, poor thermal stability, limited distribution, non-uniformity and the like, so that the permafrost is more sensitive to climate change, and is one of the most sensitive regions for coping with global climate change. The frozen earth active layer refers to the rock and soil layers above the permanent frozen earth within a certain depth of the earth's surface, and the thickness of this layer is called the active layer thickness. The frozen earth moving layer is the most important place for heat and mass exchange between the permanent frozen earth and the atmosphere, and is also the medium for climate change and interaction of the permanent frozen earth. The characteristics of the active layer are primarily dependent on climate, geography and many other factors such as temperature, precipitation, water content, soil material composition, surface vegetation cover conditions and topography. Under the influence of global warming, the warming trend of the Qinghai-Tibet plateau is obvious according to meteorological observation analysis data, the warming speed in the past 40 years is about twice that of the synchronous global warming speed, and the near-surface heat condition of a permanent frozen soil area can be changed remarkably due to artificial climate warming. Research shows that the Qinghai-Tibet plateau frozen soil undergoes significant degradation, resulting in increased thickness of the active layer, which promotes decomposition of organic matter in the soil and increased emission of greenhouse gases, affecting local and even global climate change, thereby exacerbating the temperature rise and degradation of the permanent frozen soil in QTP. Further thawing of the subsurface ice in permanently frozen soil will also increase the risk of collapse and damage of the Tibet plateau infrastructure, road and railway infrastructure.
In the research of the thickness of the existing simulated frozen soil active layer, the spatial interpolation method and the semi-empirical and semi-theoretical model are widely applied. These parameterized or spatially-modeled models only take into account the effect of a fraction of the factors on the thickness of the active layer of frozen earth, and have limitations for increasingly complex frozen earth conditions.
Disclosure of Invention
In order to solve the technical problems that the traditional method is insufficient in consideration of influence factors and poor in flexibility in parameter nonlinear fitting, and provide a high-precision high-resolution long-time sequence frozen soil active layer thickness product, the invention provides a frozen soil active layer thickness inversion method, device, equipment and storage medium.
According to a first aspect of the invention, there is provided a frozen earth active layer thickness inversion method comprising the steps of:
acquiring thickness point position data (namely field drilling point position data) of a soil active layer in a research area during a baseline period;
acquiring a feature set within a study area during a baseline, the feature set comprising: melting index, freezing index, annual average rainfall, annual average snowfall, normalized leaf area index, digital elevation model, noctilucent remote sensing data and soil data set, and preprocessing the feature set;
compressing the preprocessed soil data set by using a principal component analysis method, adding the compressed soil data set into a feature set, and selecting a plurality of variables with weight ratio larger than a preset value in the feature set by using a random forest weight analysis method as environment prediction factors;
calculating by using a Stefan formula in an inverse way to obtain a soil factor, and constructing a sample set in a base line period according to the thickness point location data of the soil active layer;
dividing a sample set in a baseline period into a training set and a testing set, training by using a plurality of machine learning models, selecting a model with the best training effect, and obtaining soil factor space distribution based on the best model;
and calculating the thickness of the frozen soil active layer until a preset year by using the predicted soil factors and the temperature data.
Further, the acquiring a feature set within the study area during the baseline, the feature set comprising: the method comprises the steps of melting index, freezing index, annual average rainfall, annual average snowfall, normalized leaf area index, digital elevation model, noctilucent remote sensing data and soil data set, and preprocessing the feature set, and comprises the following steps:
setting 2006-2020 as a baseline period, downloading an annual average rainfall, an annual average snowfall, a normalized leaf area index, a digital elevation model, noctilucent remote sensing data and a soil data set in a feature set in a study area during the baseline period by using a Google Earth Engine cloud platform, and projecting the annual average rainfall, the annual average snowfall, the normalized leaf area index, the digital elevation model, the noctilucent remote sensing data and the soil data set into an EPSG:4326″ in the coordinate system, resampling to 1km;
the melt index and freeze index in the feature set were calculated using the MODIS month composite product during baseline.
Further, the step of compressing the preprocessed soil data set by using a principal component analysis method, adding the compressed soil data set into a feature set, and then selecting a plurality of variables with weight ratio larger than a preset value in the feature set as environment prediction factors by using a random forest weight analysis method, wherein the method comprises the following steps:
compressing a soil data set containing 8 soil parameters by using a principal component transformation method, and selecting three principal components with the largest proportion as new soil data so as to contain richer information in the same dimension;
and according to the obtained feature set, combining the compressed three main component information, adopting a random forest weight method to perform feature screening, selecting features with the influence ratio on the thickness of the frozen soil active layer larger than a preset value as environment prediction factors, and participating in the prediction of the thickness of the frozen soil active layer.
Further, the step of calculating the soil factor by using the Stefan formula in an inverse way and constructing a sample set during a baseline according to the thickness point location data of the soil active layer comprises the following steps:
constructing a baseline LEP time sequence image set corresponding to the environmental prediction factor by using the selected environmental prediction factor and time sequence data of different factors in the combined baseline period characteristic set;
resampling on a baseline LEP time sequence image set of a corresponding time period according to the sampling time of the collected soil active layer thickness point positions (namely the field drilling point positions) to obtain sample data corresponding to the soil active layer thickness point positions; finally, synthesizing sample data corresponding to the thickness point positions of the soil active layer in all time periods, so as to obtain a sample set in a baseline period;
according to the obtained sample set in the baseline period, the soil factors of each point are calculated by utilizing the point location data corresponding to the frozen soil active layer thickness and the thawing index and utilizing the inverse operation of the Stefan formula, and are added into the sample set in the baseline period.
Further, the soil factor of each point location is calculated by using the inverse operation of the Stefan formula, and the specific formula is as follows:
where E represents the soil factor, ALT represents the soil active layer thickness, and TI represents the thawing index.
Further, the step of dividing the sample set during the baseline period into a training set and a testing set, training by using a plurality of machine learning models, selecting a model with the best training effect, and obtaining the spatial distribution of the soil factors based on the best model comprises the following steps:
dividing a sample set in a baseline period into a training set and a test set according to a ratio of 7:3, wherein the training set is used for fitting a model, and the test set is used for verifying the correctness of the fitting model;
using soil factors as labels, and fitting nonlinear relations between environment prediction factors and the soil factors by using 8 machine learning models;
checking the fitted model by using a test set, comparing the differences among different models by using root mean square error, average absolute error and Nash efficiency coefficient, and selecting the model with the best fitting effect;
and calculating to obtain a spatial distribution result of soil factors of the Qinghai-Tibet plateau frozen soil area during the baseline period by using the optimal model.
Further, the step of calculating the thickness of the frozen soil active layer up to a preset year by using the predicted soil factor and the temperature data comprises the following steps:
selecting CMIP6 temperature data, and calculating a freezing index from a baseline period to 2100 years in the future by using a formula for calculating the freezing index;
and calculating the time sequence change of the frozen soil activity layer thickness from the baseline period to 2100 years by using a Stefan formula by using the soil factor obtained by prediction and the freezing index obtained by calculation, wherein the concrete formula is as follows:
where E represents the soil factor, ALT represents the soil active layer thickness, and TI represents the thawing index.
According to a second aspect of the invention, a frozen earth active layer thickness inversion device for implementing the method comprises the following modules:
the point data acquisition module is used for acquiring thickness point data of the soil active layer in the research area during the baseline period;
a feature set acquisition module for acquiring a feature set within a study area during a baseline, the feature set comprising: melting index, freezing index, annual average rainfall, annual average snowfall, normalized leaf area index, digital elevation model, noctilucent remote sensing data and soil data set, and preprocessing the feature set;
the environment prediction factor selection module is used for compressing the preprocessed soil data set by using a principal component analysis method, adding the compressed soil data set into the feature set, and then selecting a plurality of variables with the weight ratio of the feature set being larger than a preset value by using a random forest weight analysis method as environment prediction factors;
the baseline period sample set construction module is used for calculating soil factors by utilizing a Stefan formula in an inverse way and constructing a baseline period sample set according to the thickness point location data of the soil active layer;
the soil factor spatial distribution acquisition module is used for dividing a sample set in a baseline period into a training set and a testing set, training by using a plurality of machine learning models, and selecting a model with the best training effect to obtain soil factor spatial distribution;
the frozen soil active layer thickness prediction module is used for calculating the frozen soil active layer thickness until a preset year by using the predicted soil factors and temperature data.
According to a third aspect of the present invention there is also provided an electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the steps of the frozen soil activity layer thickness inversion method when executing the program.
According to a fourth aspect of the present invention there is also provided a storage medium having stored thereon a computer program which when executed by a processor performs the steps of the frozen earth moving layer thickness inversion method.
The technical scheme provided by the invention has the following beneficial effects:
considering the influence of factors such as continuous deterioration of climate conditions, aggravation of human activities, increasingly complex physical and chemical properties of soil on the freeze thawing change of a frozen soil area of a research area, the luminous remote sensing data are utilized to represent the human activity intensity, the main component is utilized to transform and compress a soil data set so as to contain richer information in the same dimension, and the relation between an environment prediction factor and the thickness of a frozen soil active layer is fitted based on drilling point position data and multi-source remote sensing data. Considering the flexibility and robustness of the machine learning method in multi-parameter nonlinear fitting compared with the parameterized formula method and the spatial interpolation method, 8 machine learning methods are selected to construct the relation between the environmental predictor and the layer thickness of the frozen soil activity layer, and the best fitting model is obtained. And calculating to obtain the time sequence change condition of the frozen soil active layer thickness from the beginning of the century to the preset year based on the optimal model obtained by prediction by utilizing the CMIP6 temperature data. And finally, compared with the traditional inversion method, the method can acquire a higher-precision frozen soil active layer thickness simulation result, so that the frozen soil freezing and thawing change rule can be more comprehensively and effectively analyzed.
Drawings
The invention will be further described with reference to the accompanying drawings and examples, in which:
FIG. 1 is a flow chart of a frozen soil active layer thickness inversion method provided by an embodiment of the invention;
FIG. 2 is a plot of DEM data and soil activity layer thickness profile for a study area provided by an embodiment of the present invention;
FIG. 3 shows the accuracy display result provided by the embodiment of the present invention, wherein GroupA is a result obtained by calculating by applying the method provided by the present invention, and Group B, group C and Group D are comparison groups for verifying superiority of the method provided by the present invention;
FIG. 4 shows the result of inversion of the thickness of the frozen soil active layer of the multi-source remote sensing data provided by the embodiment of the invention, wherein (A, B) is the spatial distribution of the frozen soil ALT in the Qinghai-Tibet plateau areas 2060s and 2100s under the SSP1-1.9 scene, (C, D) is the spatial distribution of the frozen soil ALT in the Qinghai-Tibet plateau areas 2060s and 2100s under the SSP2-4.5 scene, and (E, F) is the spatial distribution of the frozen soil ALT in the Qinghai-Tibet plateau areas 2060s and 2100s under the SSP5-8.5 scene;
FIG. 5 is a schematic structural diagram of a frozen soil active layer thickness inversion device according to an embodiment of the present invention;
fig. 6 is a schematic structural diagram of an electronic device according to the present invention.
Detailed Description
For a clearer understanding of technical features, objects and effects of the present invention, a detailed description of embodiments of the present invention will be made with reference to the accompanying drawings.
Referring to fig. 1, an embodiment of the present invention provides a method for inverting the thickness of a frozen soil active layer based on multi-source data of machine learning, which uses Google Earth Engine cloud platform to complete the proposed multi-source data inversion algorithm by using 8 machine learning algorithms such as RF, GDBT, SVM and the like by using Landsat8 data, administrative division vector data, soil active layer thickness point location data, modis Liu Dewen degree products, DMSP-OLS noctilucent remote sensing data, global soil data set and the like covering Qinghai-Tibet plateau areas, and a research area refers to fig. 2.
The method comprises the following steps:
s1: acquiring thickness point location data of a soil active layer in a study area during a baseline period;
in this embodiment, the thickness point location data of the soil active layer in the Qinghai-Tibet plateau region (2006-2020) during the baseline period, that is, the in-situ drilling point location data, is obtained, and the specific operations are as follows: and obtaining permanent frozen soil point location data of the Qinghai-Tibet plateau area provided in the literature published in the past.
S2: acquiring a feature set within a study area during a baseline, the feature set comprising: melting index, freezing index, annual average rainfall, annual average snowfall, normalized leaf area index, digital elevation model, noctilucent remote sensing data and soil data set, and preprocessing the feature set;
in this embodiment, a feature set of a Qinghai-Tibet plateau region during a baseline period (2006-2020) is obtained, and includes preprocessing operations such as melt index (TI), freezing Index (FI), annual average rainfall (Prain), annual average rainfall (Psnow), normalized leaf area index (NDVI), digital Elevation Model (DEM), noctilucent remote sensing data and soil data set, and performing cutting, projection, resampling and the like, and the specific operation steps are as follows:
s2.1: using Google Earth Engine cloud platform, using administrative division vector data, download and obtain the annual average rainfall (Prain), annual average snowfall (Psnow), digital Elevation Model (DEM), DMSP-OSL noctilucent remote sensing data and soil dataset (clay content, sand content, soil group, soil moisture content, soil bulk density, soil organic carbon content, soil texture classification and soil PH) during baseline period (2006-2020) of the tibetan plateau region, crop, project to "EPSG:4326″ under a coordinate system, resampling to 1000m resolution;
s2.2: using the Tibet plateau MOD11-A1.061 Liu Dewen degree product data, resampled to 1000m resolution, the melt index (TI), freeze Index (FI) was calculated, which refers to the cumulative sum of monthly ground or air temperatures above (below) 0deg.C multiplied by the number of days of the corresponding month. The specific calculation formula is as follows:
wherein TI is a thawing index (. Degree. C. Day), FI is a freezing index (. Degree. C. Day), T i Is the average surface temperature (DEG C) of the month, D i Day of the month (. Degree. C.);
s2.3: utilizing Google Earth Engine cloud platform, using administrative division vector data, cutting and obtaining Landsat8 image data of Qinghai-Tibet plateau region, calculating normalized leaf area index (NDVI) during the period of the Qinghai-Tibet plateau baseline (2006-2020), and calculating the following formula:
NDVI＝(nir-red)/(nir+red)
where nir denotes the near infrared band and red denotes the red band.
S3: compressing the preprocessed soil data set by using a principal component analysis method, adding the compressed soil data set into a feature set, and selecting a plurality of variables with weight ratio larger than a preset value in the feature set by using a random forest weight analysis method as environment prediction factors;
in this embodiment, the specific operation of S3 is as follows:
s3.1: in order to obtain richer data information in lower dimensionality, after obtaining 8 pieces of soil parameter data preprocessed in S2.1, performing dimensionality reduction compression on the data by using a Principal Component Analysis (PCA), and selecting the first three principal components (more than 95%) with the largest proportion as new soil data, namely compressing the information of the original 8 dimensionalities into 3 dimensionalities;
s3.2: according to the feature set selected in the step S2, combining the three main components obtained in the step S3.1 after the soil data set is compressed, and calculating and determining the weight of each feature by adopting a Random Forest (RF) weight method;
s3.3: according to the weights of different parameters in the calculation of S3.2, the first ten features with the largest influence ratio on the thickness of the frozen soil active layer are selected as environment prediction factors (LEPs) to participate in the prediction of the thickness (ALT) of the frozen soil active layer, and a model is constructed.
S4: calculating by using a Stefan formula in an inverse way to obtain soil factors, and constructing a sample set in a baseline period according to spot location data of the field drilling points;
in this embodiment, S4 specifically includes:
s4.1: constructing a baseline LEP time sequence image set corresponding to the environmental predictors (LEPs) by using the environmental predictors (LEPs) obtained in the S3.3 and combining time sequence data of different influencing factors in the baseline period obtained in the S2;
s4.2: resampling on a baseline LEP time sequence image set of a corresponding time period according to the sampling time of the thickness point position (namely the field drilling point position) of the soil active layer collected in the S1 to obtain sample data corresponding to the thickness point position of the soil active layer;
s4.3: synthesizing sample data corresponding to the thickness point positions of the soil active layer in all time periods, so as to obtain a sample set in a baseline period;
s4.4: and (3) calculating the soil factor value of each drilling point by using the inverse operation of a Stefan formula according to the sample set during the baseline period obtained in the step (4.3), the point data measured value corresponding to the collected frozen soil active layer thickness and the thawing index in the corresponding grid unit, and adding the soil factor value into the sample set. The soil factor E is calculated as follows:
where E represents the soil factor, ALT represents the soil active layer thickness, and TI represents the thawing index.
S5: dividing a sample set in a baseline period into a training set and a testing set, training by using a plurality of machine learning models, selecting a model with the best training effect, and obtaining soil factor space distribution based on the best model;
in this embodiment, the specific operation of step S5 is as follows:
s5.1: the sample set during the baseline period obtained in the step S4.4 is divided into a training set (70%) and a test set (30%) in proportion, wherein the training set is used for fitting a model, and the test set is used for verifying the correctness of the fitting model;
s5.2: 8 machine learning models (algorithms) of Random Forest (RF), gradient lifting tree (GDBT), support Vector Machine (SVM), minimum Distance (MD), classification regression tree (Cart), elastic, LASSO regression and Naive Bayes (NBM) are adopted, soil factor E is used as a label, the model is trained, and a nonlinear relation between LEP and the soil factor E is fitted;
s5.3: using the remaining 30% of the test set obtained in the step S5.1 to test the model obtained by fitting 8 machine learning algorithms in the step S5.2;
s5.4: and respectively calculating the Root Mean Square Error (RMSE), the Mean Absolute Error (MAE) and the Nash efficiency coefficient (NSE) of the model to compare the accuracy of different algorithms so as to obtain the model with the best training effect. The calculation formula is as follows:
y i is the value of ALT at the drilling point,is the average of all ALT measurements, +.>Is a mapping value;
s5.5: the spatial distribution of E values of soil factors during the baseline period of the Qinghai-Tibet plateau permafrost region (2006-2020) is gridded by using the best model obtained in S5.4 and the local environment predictors of the corresponding grid cells.
S6: and calculating the thickness of the frozen soil active layer until a preset year by using the predicted soil factors and the temperature data.
S6.1: selecting a soil factor E in a baseline period (2006-2020) obtained by projection of the optimal model in the step S5, and calculating an average value of each grid unit as the soil factor E;
s6.2: selecting CMIP6 temperature data, and calculating freezing index TI from the baseline period to the future 2100 years by using a formula for calculating freezing index in S2.2;
s6.3: combining the soil factor E obtained in the step S6.1 with the freezing index TI obtained in the step S6.2, and calculating to obtain the time sequence change of the frozen soil activity layer thickness from the baseline period (2006-2020) to 2100 by using a simplified Stefan model, wherein the calculation formula is as follows:
the invention provides a frozen soil activity layer thickness inversion method of multisource remote sensing data based on machine learning. Considering the influence of factors such as continuous deterioration of climate conditions, aggravation of human activities, increasingly complex physical and chemical properties of soil on the freeze thawing change of a Qinghai-Tibet plateau frozen soil region, utilizing noctilucent remote sensing data to represent human activity intensity, utilizing principal component transformation (PCA) to compress a soil data set so as to contain richer information in smaller dimension, and fitting the relation between an environment prediction factor and the thickness of a frozen soil active layer based on drilling point position data and multisource remote sensing data. Considering the flexibility and robustness of the machine learning method in multi-parameter nonlinear fitting compared with the parameterized formula method and the spatial interpolation method, 8 machine learning methods are selected to construct the relation between the environmental predictor and the layer thickness of the frozen soil activity layer, and the best fitting model is obtained. And calculating the time sequence change condition of the layer thickness of the frozen soil activity layer from the beginning of the century to 2100 years based on the optimal result obtained by prediction by utilizing the CMIP6 temperature data. And finally, compared with the traditional inversion method, the method can acquire a higher-precision frozen soil active layer thickness simulation result, so that the frozen soil freezing and thawing change rule can be more comprehensively and effectively analyzed.
Referring to fig. 3, fig. 3 shows the result of the present invention, where Group pa is a result obtained by calculating by applying the method of the present invention, group B, group C, and Group D are comparison groups, and the performance of the soil dataset after transformation using the light index and the main component is compared respectively, so as to verify the superiority of the method of the present invention. Wherein, groupA has the best result, RMSE is 19.37cm, R 2 0.97; 26.11cm and 47.83cm far above the RMSE in Group B and Group C; group D without any improvement had the worst effect and the RMSE was 57.59cm.
Referring to fig. 4, fig. 4 shows the result of the inversion of the thickness of the frozen soil active layer of the multi-source remote sensing data according to the present invention, wherein (A, B) is the spatial distribution of the frozen soil ALT in the Qinghai-Tibet plateau region 2060s and 2100s under the SSP1-1.9 scenario, (C, D) is the spatial distribution of the frozen soil ALT in the Qinghai-Tibet plateau region 2060s and 2100s under the SSP2-4.5 scenario, and (E, F) is the spatial distribution of the frozen soil ALT in the Qinghai-Tibet plateau region 2060s and 2100s under the SSP5-8.5 scenario. According to the invention, the best fitting model of the relation between the environmental prediction factor and the frozen soil active layer thickness is obtained through the combination of the multi-source remote sensing data, and the simulation result of the frozen soil active layer thickness with higher precision is obtained, so that the comprehensive analysis of the frozen soil freezing and thawing change rule in the following years is facilitated.
The frozen soil active layer thickness inversion device provided by the invention is described below, and the frozen soil active layer thickness inversion device described below and the frozen soil active layer thickness inversion method described above can be correspondingly referred to each other.
Referring to fig. 5, a frozen soil active layer thickness inversion device includes the following modules:
the point data acquisition module 001 is used for acquiring thickness point data of the soil active layer in the research area during the baseline period;
a feature set acquisition module 002 for acquiring a feature set within a study area during a baseline, the feature set comprising: melting index, freezing index, annual average rainfall, annual average snowfall, normalized leaf area index, digital elevation model, noctilucent remote sensing data and soil data set, and preprocessing the feature set;
the environmental predictor selecting module 003 is configured to compress the preprocessed soil dataset by using a principal component analysis method, add the compressed soil dataset into the feature set, and select, by using a random forest weight analysis method, a plurality of variables with weights in the feature set being larger than a preset value as environmental predictors;
the baseline period sample set construction module 004 is used for obtaining soil factors by means of inverse operation calculation of a Stefan formula and constructing a baseline period sample set according to the thickness point location data of the soil active layer;
the soil factor spatial distribution acquisition module 005 is used for dividing a sample set during a baseline period into a training set and a testing set, training by using a plurality of machine learning models, and selecting a model with the best training effect to obtain soil factor spatial distribution;
the frozen soil active layer thickness prediction module 006 is configured to calculate the frozen soil active layer thickness up to a preset year by using the predicted soil factor and the temperature data.
Based on but not limited to the above apparatus, the feature set obtaining module 002 is specifically configured to:
setting 2006-2020 as a baseline period, downloading an annual average rainfall, an annual average snowfall, a normalized leaf area index, a digital elevation model, noctilucent remote sensing data and a soil data set in a feature set in a study area during the baseline period by using a Google Earth Engine cloud platform, and projecting the annual average rainfall, the annual average snowfall, the normalized leaf area index, the digital elevation model, the noctilucent remote sensing data and the soil data set into an EPSG:4326″ under a coordinate system, resampling to 1000m resolution;
resampling the data of MOD11-A1.061 Liu Dewen degrees in Qinghai-Tibet plateau to 1000m resolution, and calculating a melting index (TI) and a Freezing Index (FI);
and utilizing a Google Earth Engine cloud platform, cutting and acquiring Landsat8 image data of the Qinghai-Tibet plateau region by using administrative division vector data, and calculating normalized leaf area index (NDVI) during a baseline period of the Qinghai-Tibet plateau (2006-2020).
Based on but not limited to the above apparatus, the environmental predictor selection module 003 is specifically configured to:
compressing a soil data set containing 8 soil parameters by using a principal component transformation method, and selecting three principal components with the largest proportion as new soil data so as to contain richer information in the same dimension;
and according to the obtained feature set, combining the compressed three main component information, adopting a random forest weight method to perform feature screening, selecting features with the influence ratio on the thickness of the frozen soil active layer larger than a preset value as environment prediction factors, and participating in the prediction of the thickness of the frozen soil active layer.
Based on, but not limited to, the above-described apparatus, the baseline period sample set construction module 004 is specifically configured to:
constructing a baseline LEP time sequence image set corresponding to the environmental prediction factor by using the selected environmental prediction factor and time sequence data of different factors in the combined baseline period characteristic set;
resampling on a baseline LEP time sequence image set of a corresponding time period according to the collected sampling time of the thickness point position of the soil active layer to obtain sample data corresponding to the thickness point position of the soil active layer; finally, synthesizing sample data corresponding to the thickness point positions of the soil active layer in all time periods, so as to obtain a sample set in a baseline period;
according to the obtained sample set in the baseline period, the soil factors of each point are calculated by utilizing the point location data corresponding to the frozen soil active layer thickness and the thawing index and utilizing the inverse operation of the Stefan formula, and are added into the sample set in the baseline period, wherein the concrete formula is as follows:
where E represents the soil factor, ALT represents the soil active layer thickness, and TI represents the thawing index.
Based on but not limited to the above device, the soil factor spatial distribution obtaining module 005 is specifically configured to:
dividing a sample set during a baseline into a training set and a testing set, training by using a plurality of machine learning models, selecting a model with the best training effect, and obtaining the spatial distribution of soil factors based on the best model, wherein the method comprises the following steps:
dividing a sample set in a baseline period into a training set and a test set according to a ratio of 7:3, wherein the training set is used for fitting a model, and the test set is used for verifying the correctness of the fitting model;
using soil factors as labels, and fitting nonlinear relations between environment prediction factors and the soil factors by using 8 machine learning models;
checking the fitted model by using a test set, comparing the differences among different models by using root mean square error, average absolute error and Nash efficiency coefficient, and selecting the model with the best fitting effect;
and calculating to obtain a spatial distribution result of soil factors of the Qinghai-Tibet plateau frozen soil area during the baseline period by using the optimal model.
It should be noted that the machine learning model is based on, but not limited to, 8 machine learning models (algorithms) using Random Forest (RF), gradient-lifted tree (GDBT), support Vector Machine (SVM), minimum Distance (MD), classification regression tree (cat), elastic, LASSO regression, naive Bayes (NBM).
Based on but not limited to the above-mentioned apparatus, the frozen soil active layer thickness prediction module 006 is specifically configured to:
selecting CMIP6 temperature data, and calculating a freezing index from a baseline period to 2100 years in the future by using a formula for calculating the freezing index;
and calculating the time sequence change of the frozen soil activity layer thickness from the baseline period to 2100 years by using a Stefan formula by using the soil factor obtained by prediction and the freezing index obtained by calculation, wherein the concrete formula is as follows:
where E represents the soil factor, ALT represents the soil active layer thickness, and TI represents the thawing index.
The invention provides a frozen soil activity layer thickness inversion device of multisource remote sensing data based on machine learning. Considering the influence of factors such as continuous deterioration of climate conditions, aggravation of human activities, increasingly complex physical and chemical properties of soil on the freeze thawing change of a Qinghai-Tibet plateau frozen soil region, utilizing noctilucent remote sensing data to represent human activity intensity, utilizing principal component transformation (PCA) to compress a soil data set so as to contain richer information in smaller dimension, and fitting the relation between an environment prediction factor and the thickness of a frozen soil active layer based on drilling point position data and multisource remote sensing data. Considering the flexibility and robustness of the machine learning method in multi-parameter nonlinear fitting compared with the parameterized formula method and the spatial interpolation method, 8 machine learning methods are selected to construct the relation between the environmental predictor and the layer thickness of the frozen soil activity layer, and the best fitting model is obtained. And calculating the time sequence change condition of the layer thickness of the frozen soil activity layer from the beginning of the century to 2100 years based on the optimal result obtained by prediction by utilizing the CMIP6 temperature data. And finally, compared with the traditional inversion method, the method can acquire a higher-precision frozen soil active layer thickness simulation result, so that the frozen soil freezing and thawing change rule can be more comprehensively and effectively analyzed.
Referring to fig. 6, a schematic diagram of an entity structure of an electronic device is illustrated, where the electronic device may include: processor (processor) 610, communication interface (Communications Interface) 620, memory (memory) 630, and communication bus 640, wherein processor 610, communication interface 620, memory 630 communicate with each other via communication bus 640. Processor 610 may invoke logic instructions in memory 630 to perform the steps of the frozen earth active layer thickness inversion method described above, including: acquiring thickness point location data of a soil active layer in a study area during a baseline period; acquiring a feature set within a study area during a baseline, the feature set comprising: melting index, freezing index, annual average rainfall, annual average snowfall, normalized leaf area index, digital elevation model, noctilucent remote sensing data and soil data set, and preprocessing the feature set; compressing the preprocessed soil data set by using a principal component analysis method, adding the compressed soil data set into a feature set, and selecting a plurality of variables with weight ratio larger than a preset value in the feature set by using a random forest weight analysis method as environment prediction factors; calculating by using a Stefan formula in an inverse way to obtain a soil factor, and constructing a sample set in a base line period according to the thickness point location data of the soil active layer; dividing a sample set in a baseline period into a training set and a testing set, training by using a plurality of machine learning models, selecting a model with the best training effect, and obtaining soil factor space distribution based on the best model; and calculating the thickness of the frozen soil active layer until a preset year by using the predicted soil factors and the temperature data.
Further, the logic instructions in the memory 630 may be implemented in the form of software functional units and stored in a computer-readable storage medium when sold or used as a stand-alone product. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk, etc., which can store program codes.
In still another aspect, an embodiment of the present invention further provides a storage medium having stored thereon a computer program, which when executed by a processor, implements the steps of the above method for inverting a layer thickness of a frozen soil active layer, including: acquiring thickness point location data of a soil active layer in a study area during a baseline period; acquiring a feature set within a study area during a baseline, the feature set comprising: melting index, freezing index, annual average rainfall, annual average snowfall, normalized leaf area index, digital elevation model, noctilucent remote sensing data and soil data set, and preprocessing the feature set; compressing the preprocessed soil data set by using a principal component analysis method, adding the compressed soil data set into a feature set, and selecting a plurality of variables with weight ratio larger than a preset value in the feature set by using a random forest weight analysis method as environment prediction factors; calculating by using a Stefan formula in an inverse way to obtain a soil factor, and constructing a sample set in a base line period according to the thickness point location data of the soil active layer; dividing a sample set in a baseline period into a training set and a testing set, training by using a plurality of machine learning models, selecting a model with the best training effect, and obtaining soil factor space distribution based on the best model; and calculating the thickness of the frozen soil active layer until a preset year by using the predicted soil factors and the temperature data.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or system. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or system that comprises the element.
The foregoing embodiment numbers of the present invention are merely for the purpose of description, and do not represent the advantages or disadvantages of the embodiments. In the unit claims enumerating several means, several of these means may be embodied by one and the same item of hardware. The use of the terms first, second, third, etc. do not denote any order, but rather the terms first, second, third, etc. are used to interpret the terms as labels.
The foregoing description is only of the preferred embodiments of the present invention, and is not intended to limit the scope of the invention, but rather is intended to cover any equivalents of the structures or equivalent processes disclosed herein or in the alternative, which may be employed directly or indirectly in other related arts.
Claims (10)
1. The frozen soil active layer thickness inversion method is characterized by comprising the following steps of:
acquiring thickness point location data of a soil active layer in a study area during a baseline period;
acquiring a feature set within a study area during a baseline, the feature set comprising: melting index, freezing index, annual average rainfall, annual average snowfall, normalized leaf area index, digital elevation model, noctilucent remote sensing data and soil data set, and preprocessing the feature set;
compressing the preprocessed soil data set by using a principal component analysis method, adding the compressed soil data set into a feature set, and selecting a plurality of variables with weight ratio larger than a preset value in the feature set by using a random forest weight analysis method as environment prediction factors;
calculating by using a Stefan formula in an inverse way to obtain a soil factor, and constructing a sample set in a base line period according to the thickness point location data of the soil active layer;
dividing a sample set in a baseline period into a training set and a testing set, training by using a plurality of machine learning models, selecting a model with the best training effect, and obtaining soil factor space distribution based on the best model;
and calculating the thickness of the frozen soil active layer until a preset year by using the predicted soil factors and the temperature data.
2. The method of claim 1, wherein the acquiring a feature set within a study area during baseline, the feature set comprising: the method comprises the steps of melting index, freezing index, annual average rainfall, annual average snowfall, normalized leaf area index, digital elevation model, noctilucent remote sensing data and soil data set, and preprocessing the feature set, and comprises the following steps:
setting 2006-2020 as a baseline period, downloading an annual average rainfall, an annual average snowfall, a normalized leaf area index, a digital elevation model, noctilucent remote sensing data and a soil data set in a feature set in a study area during the baseline period by using a Google Earth Engine cloud platform, and projecting the annual average rainfall, the annual average snowfall, the normalized leaf area index, the digital elevation model, the noctilucent remote sensing data and the soil data set into an EPSG:4326″ in the coordinate system, resampling to 1km;
the melt index and freeze index in the feature set were calculated using the MODIS month composite product during baseline.
3. The method for inverting the thickness of a frozen soil activity layer according to claim 1, wherein the steps of compressing the preprocessed soil data set by a principal component analysis method, adding the compressed soil data set to the feature set, and selecting several variables with the weight ratio of the feature set larger than a preset value as environmental predictors by a random forest weight analysis method comprise:
compressing a soil data set containing 8 soil parameters by using a principal component transformation method, and selecting three principal components with the largest proportion as new soil data so as to contain richer information in the same dimension;
and according to the obtained feature set, combining the compressed three main component information, adopting a random forest weight method to perform feature screening, selecting features with the influence ratio on the thickness of the frozen soil active layer larger than a preset value as environment prediction factors, and participating in the prediction of the thickness of the frozen soil active layer.
4. The method for inverting the thickness of a frozen earth moving bed according to claim 1, wherein the step of calculating the soil factor by using a Stefan formula for inverse operation and constructing a sample set during a base line according to the thickness point location data of the moving bed comprises the steps of:
constructing a baseline LEP time sequence image set corresponding to the environmental prediction factor by using the selected environmental prediction factor and time sequence data of different factors in the combined baseline period characteristic set;
resampling on a baseline LEP time sequence image set of a corresponding time period according to the collected sampling time of the thickness point position of the soil active layer to obtain sample data corresponding to the thickness point position of the soil active layer; finally, synthesizing sample data corresponding to the thickness point positions of the soil active layer in all time periods, so as to obtain a sample set in a baseline period;
according to the obtained sample set in the baseline period, the soil factors of each point are calculated by utilizing the point location data corresponding to the frozen soil active layer thickness and the thawing index and utilizing the inverse operation of the Stefan formula, and are added into the sample set in the baseline period.
5. The method for inverting the thickness of a frozen soil activity layer according to claim 4, wherein the soil factor of each point location is calculated by using the inverse operation of the Stefan formula, and the concrete formula is as follows:
where E represents the soil factor, ALT represents the soil active layer thickness, and TI represents the thawing index.
6. The method for inverting the thickness of a frozen soil activity layer according to claim 1, wherein the step of dividing the sample set during the baseline period into a training set and a test set, training by using a plurality of machine learning models, selecting a model with the best training effect, and obtaining the spatial distribution of soil factors based on the best model comprises the following steps:
dividing a sample set in a baseline period into a training set and a test set according to a ratio of 7:3, wherein the training set is used for fitting a model, and the test set is used for verifying the correctness of the fitting model;
using soil factors as labels, and fitting nonlinear relations between environment prediction factors and the soil factors by using 8 machine learning models;
checking the fitted model by using a test set, comparing the differences among different models by using root mean square error, average absolute error and Nash efficiency coefficient, and selecting the model with the best fitting effect;
and calculating to obtain a spatial distribution result of soil factors of the Qinghai-Tibet plateau frozen soil area during the baseline period by using the optimal model.
7. The method of claim 1, wherein the step of calculating the frozen soil activity layer thickness up to a preset year using the predicted soil factor and the temperature data comprises:
selecting CMIP6 temperature data, and calculating a freezing index from a baseline period to 2100 years in the future by using a formula for calculating the freezing index;
and calculating the time sequence change of the frozen soil activity layer thickness from the baseline period to 2100 years by using a Stefan formula by using the soil factor obtained by prediction and the freezing index obtained by calculation, wherein the concrete formula is as follows:
where E represents the soil factor, ALT represents the soil active layer thickness, and TI represents the thawing index.
8. A frozen earth active layer thickness inversion device for implementing the method of any one of claims 1-7, comprising the following modules:
the point data acquisition module is used for acquiring thickness point data of the soil active layer in the research area during the baseline period;
a feature set acquisition module for acquiring a feature set within a study area during a baseline, the feature set comprising: melting index, freezing index, annual average rainfall, annual average snowfall, normalized leaf area index, digital elevation model, noctilucent remote sensing data and soil data set, and preprocessing the feature set;
the environment prediction factor selection module is used for compressing the preprocessed soil data set by using a principal component analysis method, adding the compressed soil data set into the feature set, and then selecting a plurality of variables with the weight ratio of the feature set being larger than a preset value by using a random forest weight analysis method as environment prediction factors;
the baseline period sample set construction module is used for calculating soil factors by utilizing a Stefan formula in an inverse way and constructing a baseline period sample set according to the thickness point location data of the soil active layer;
the soil factor spatial distribution acquisition module is used for dividing a sample set in a baseline period into a training set and a testing set, training by using a plurality of machine learning models, and selecting a model with the best training effect to obtain soil factor spatial distribution;
the frozen soil active layer thickness prediction module is used for calculating the frozen soil active layer thickness until a preset year by using the predicted soil factors and temperature data.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor performs the steps of the frozen earth active layer thickness inversion method according to any one of claims 1-7.
10. A storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the frozen earth moving layer thickness inversion method according to any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310499168.0A CN116738161A (en) | 2023-05-05 | 2023-05-05 | Frozen soil active layer thickness inversion method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310499168.0A CN116738161A (en) | 2023-05-05 | 2023-05-05 | Frozen soil active layer thickness inversion method, device, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116738161A true CN116738161A (en) | 2023-09-12 |
Family
ID=87917685
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310499168.0A Pending CN116738161A (en) | 2023-05-05 | 2023-05-05 | Frozen soil active layer thickness inversion method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116738161A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117540132A (en) * | 2024-01-09 | 2024-02-09 | 中国科学院精密测量科学与技术创新研究院 | Permafrost active layer thickness estimation method based on star-earth observation |
-
2023
- 2023-05-05 CN CN202310499168.0A patent/CN116738161A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117540132A (en) * | 2024-01-09 | 2024-02-09 | 中国科学院精密测量科学与技术创新研究院 | Permafrost active layer thickness estimation method based on star-earth observation |
| CN117540132B (en) * | 2024-01-09 | 2024-04-02 | 中国科学院精密测量科学与技术创新研究院 | Permafrost active layer thickness estimation method based on star-earth observation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ge et al. | Modeling alpine grassland cover based on MODIS data and support vector machine regression in the headwater region of the Huanghe River, China | |
| Sun et al. | Reconstructing daily clear-sky land surface temperature for cloudy regions from MODIS data | |
| CN110751094B (en) | Crop yield estimation method based on GEE comprehensive remote sensing image and deep learning method | |
| Duan et al. | Dynamic monitoring of aeolian desertification based on multiple indicators in Horqin Sandy Land, China | |
| Xie et al. | A comparison of two models with Landsat data for estimating above ground grassland biomass in Inner Mongolia, China | |
| WO2018145229A1 (en) | Accurate large-area inversion method for near-surface air temperature | |
| Du et al. | Impact of urban expansion on land surface temperature in Fuzhou, China using Landsat imagery | |
| Fu et al. | Estimating landscape net ecosystem exchange at high spatial–temporal resolution based on Landsat data, an improved upscaling model framework, and eddy covariance flux measurements | |
| Zhou et al. | Remote sensing inversion of grassland aboveground biomass based on high accuracy surface modeling | |
| Choudhary et al. | Agriculture phenology monitoring using NDVI time series based on remote sensing satellites: a case study of Guangdong, China | |
| Qu et al. | Inter-comparison of several soil moisture downscaling methods over the Qinghai-Tibet Plateau, China | |
| Oaida et al. | A high-resolution data assimilation framework for snow water equivalent estimation across the Western United States and validation with the airborne snow observatory | |
| Zhao et al. | Downscaling of soil moisture products using deep learning: Comparison and analysis on Tibetan Plateau | |
| Ramdani et al. | Spatio-temporal analysis of urban temperature in Bandung City, Indonesia | |
| Fu et al. | Assessment of approaches for monitoring forest structure dynamics using bi-temporal digital aerial photogrammetry point clouds | |
| Guo et al. | Forest cover change in China from 2000 to 2016 | |
| Oleson et al. | The effects of remotely sensed plant functional type and leaf area index on simulations of boreal forest surface fluxes by the NCAR land surface model | |
| Huang et al. | A novel fusion method for generating surface soil moisture data with high accuracy, high spatial resolution, and high spatio‐temporal Continuity | |
| CN114819737B (en) | Method, system and storage medium for estimating carbon reserves of highway road vegetation | |
| CN116738161A (en) | Frozen soil active layer thickness inversion method, device, equipment and storage medium | |
| CN114241331B (en) | Remote sensing modeling method for ground biomass of reed in wetland by taking UAV as ground and Septinel-2 medium | |
| Rampheri et al. | Use of remotely sensed data to estimate tree species diversity as an indicator of biodiversity in Blouberg Nature Reserve, South Africa | |
| Chen et al. | Assessment of urban growth in Guangzhou using multi-temporal, multi-sensor Landsat data to quantify and map impervious surfaces | |
| Lou et al. | An effective method for canopy chlorophyll content estimation of marsh vegetation based on multiscale remote sensing data | |
| Guo et al. | Vegetation cover changes and their relationship to climate variation in the source region of the Yellow River, China, 1990–2000 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |