CN114912075A - Soil humidity data missing filling method fusing spatial-temporal three-dimensional information of site and satellite remote sensing observation - Google Patents

Soil humidity data missing filling method fusing spatial-temporal three-dimensional information of site and satellite remote sensing observation Download PDF

Info

Publication number
CN114912075A
CN114912075A CN202210498546.9A CN202210498546A CN114912075A CN 114912075 A CN114912075 A CN 114912075A CN 202210498546 A CN202210498546 A CN 202210498546A CN 114912075 A CN114912075 A CN 114912075A
Authority
CN
China
Prior art keywords
data
remote sensing
soil humidity
pls
site
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
Application number
CN202210498546.9A
Other languages
Chinese (zh)
Inventor
方秀琴
曹煜
郭晓萌
杨露露
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hohai University HHU
Original Assignee
Hohai University HHU
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Hohai University HHU filed Critical Hohai University HHU
Priority to CN202210498546.9A priority Critical patent/CN114912075A/en
Publication of CN114912075A publication Critical patent/CN114912075A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
    • G01N33/24—Earth materials
    • G01N33/246—Earth materials for water content

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Chemical & Material Sciences (AREA)
  • Computational Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Mathematical Physics (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Probability & Statistics with Applications (AREA)
  • Databases & Information Systems (AREA)
  • Software Systems (AREA)
  • Operations Research (AREA)
  • Evolutionary Biology (AREA)
  • Environmental & Geological Engineering (AREA)
  • General Life Sciences & Earth Sciences (AREA)
  • Geology (AREA)
  • Remote Sensing (AREA)
  • Algebra (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Food Science & Technology (AREA)
  • Medicinal Chemistry (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • General Health & Medical Sciences (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Radar Systems Or Details Thereof (AREA)
  • Testing Or Calibration Of Command Recording Devices (AREA)
  • Geophysics And Detection Of Objects (AREA)

Abstract

A soil humidity data missing filling method fusing space-time three-dimensional information of site and satellite remote sensing observation is characterized by comprising the following steps: 1. collecting ESA CCI remote sensing soil humidity SM _ S and site actual measurement soil humidity data SM _ O of 10cm of the surface layer; 2. matching the actually measured soil humidity SM _ O of the station to the soil humidity SM _ S of the satellite remote sensing by adopting an accumulative probability distribution function matching method; 3. based on a CDF matching method, establishing a piecewise linear equation for SM _ O and SM _ S in the non-loss period of the remote sensing data; 4. applying a piecewise linear equation to SM _ O in the remote sensing data loss period, and adding the SMObc obtained through calculation to SM _ S to obtain SM _ SO; 5. reconstructing data of SM _ SO based on DCT-PLS method; 6. and (3) quantitatively evaluating the filling effect by adopting the 3 evaluation indexes of RMSE, BIAS and CORR. The method can be suitable for areas and periods where satellite remote sensing SM data is widely lack of measurement, and accuracy of data missing filling is remarkably improved.

Description

Soil humidity data missing filling method fusing spatial-temporal three-dimensional information of site and satellite remote sensing observation
Technical Field
The invention relates to a remote sensing soil humidity data missing filling method, and belongs to the technical field of remote sensing data product fusion processing by using geographical space-time information.
Background
Soil water is the core of Moisture conversion and circulation to connect the atmosphere, the earth surface, the Soil and the underground, and the world meteorological organization writes Soil humidity (SM) as a 'basic climate variable' into a global climate condition statement for the first time in 2012, which marks the importance of Soil humidity on a global scale. On a regional scale, the spatial and temporal dynamics of soil humidity have important application values for agricultural management such as irrigation and crop estimation, flood and drought prediction, water quality management, natural protection and the like.
Due to differences in soil characteristics, land use, vegetation, terrain, and climatic conditions, the spatial and temporal variation in soil humidity is very large. In the traditional site observation, large-area continuous monitoring cannot be obtained due to the fact that the number of sites is small, satellite remote sensing becomes a main way for obtaining large-scale soil humidity information, and space-time continuous soil humidity observation can be provided. Among them, microwaves have the characteristics of strong penetrating power to cloud layers, all-weather working, close correlation with soil humidity and dielectric constant, and the like, and have been used for monitoring the surface soil humidity under all climatic conditions. At present, a plurality of domestic and foreign satellite research and application organizations release own global soil moisture remote sensing products, however, the satellite orbit change, the radio frequency interference and the physical limitation of a satellite sensor generally cause the problem of data loss of all satellite remote sensing soil moisture products. The absence of data limits the practical application of remote sensing soil moisture products to a number of fields requiring spatially and temporally complete SM data sets, such as climate simulation, drought monitoring, water resource management.
Several methods have emerged to fill the gap in satellite telemetry SM data sets. From the aspect of dimension, the reconstruction methods can be divided into a remote sensing image reconstruction method based on space and a time series reconstruction method based on time. The remote sensing image reconstruction method based on space usually assumes that the data non-missing region and the missing region have the same or similar spatial variability, and mainly includes an inverse distance weight algorithm (IDW), a Kriging (Kriging) model, a BSHADE (binary sensitive area-based area estimation) and a point estimation model (pbshield), and the methods consider spatial autocorrelation and spatial heterogeneity of data distribution to some extent. The time-based reconstruction method is to interpolate missing data by using a time series prediction method, such as an exponential smoothing model (SES), an autoregressive integrated moving average (ARIMA), and the like. However, when the time series length is short or the spatial correlation of the target variable is stronger than the temporal correlation, it is difficult for the time series reconstruction method to obtain a good effect.
Besides, reconstructing SM remote sensing data by using a machine learning algorithm (RF, SVM, ANN, etc.) is one of the hot spots studied in recent years, the machine learning algorithm can process a system with many influencing factors and complex relationships, but a model established based on the machine learning algorithm depends on the selection of input variables and the relationships between the input variables and target variables to a great extent, and the types of variables and the structure of the model are highly correlated with the target variables, the study region, the study period, the data types, and the like, so that the model has almost no migratability.
For geophysical datasets, it is important to use spatio-temporal variation information to predict missing values. A penalty least square regression method (DCT-PLS) based on discrete cosine transform can fill in the data blank of a geophysical data set by using full three-dimensional information of time and space, and the method can achieve good effect under the condition that missing data are uniformly distributed in an original data set.
However, for many telemetric SM data sets, the data loss does not feature a uniform distribution. Taking the ESA CCI SM in the Liaohe area in northeast china as an example, remote sensing SM is continuously missing in vegetation non-growing seasons, and the missing data covers the whole Liaohe area, which results in the loss of SM full three-dimensional information on a time and space scale. When the continuous missing range of the remote sensing SM exceeds the spatial-temporal autocorrelation range of the SM variable, the result of directly reconstructing remote sensing SM data inevitably has great deviation from an actual value, especially in the non-growing season of continuous missing.
Disclosure of Invention
The invention aims to solve the problems, the invention provides a method for fusing space-time full three-dimensional information obtained by utilizing the time integrity and space representativeness of site observation data with the existing satellite remote sensing SM product, and improves the application of the existing DCT-PLS method.
The object of the invention is achieved by a solution comprising the following steps.
A soil humidity data missing filling method fusing space-time three-dimensional information of site and satellite remote sensing observation is characterized by comprising the following steps:
step 1, collecting ESA CCI remote sensing soil humidity and site actual measurement soil humidity data of 10cm of a surface layer, and recording the data as SM _ S and SM _ O respectively;
step 2, matching the site actually-measured soil humidity SM _ O to the satellite remote sensing soil humidity SM _ S by adopting a cumulative probability distribution function (CDF) matching method, so that the matched SM _ O _ bc and SM _ S have the same value range and cumulative probability distribution condition, and the integral deviation between the actually-measured soil humidity data and the satellite remote sensing soil humidity data is reduced;
step 3, based on a CDF matching method, establishing a piecewise linear equation for SM _ O and SM _ S in the non-loss period of the remote sensing data;
step 4, applying the piecewise linear equation to SM _ O in the remote sensing data loss period, and adding the SM _ O _ bc obtained through calculation into SM _ S to obtain SM _ SO;
step 5, reconstructing data of SM _ SO based on DCT-PLS method;
and 6, adopting 3 evaluation indexes of RMSE, BIAS and CORR to quantitatively evaluate the error between the actual measurement data and the remote sensing data before and after the deviation correction and the deviation between the remote sensing data and the reference data before and after the reconstruction.
The CDF curves of the measured soil humidity data and the ESA CCI soil humidity data of the site are divided into 10 sections by taking 0, 10, 20, 30, 40, 50, 60, 70, 80, 90 and 100 as dividing nodes.
The method comprises the steps of obtaining ESA CCI remote sensing soil humidity data, wherein the time resolution is day, the space resolution is 0.25 degrees and the unit is m 3 /m 3 。
Respectively taking SM _ O and SM _ S as verification data, and quantitatively evaluating errors between the measured data and the remote sensing data before and after deviation correction and deviations between the remote sensing data and the reference data before and after reconstruction by adopting 3 evaluation indexes of RMSE, BIAS and CORR; the calculation formula of the 3 evaluation indexes is as follows:
Figure BDA0003634350780000031
Figure BDA0003634350780000032
Figure BDA0003634350780000033
wherein n represents the number of samples, S i Representative of the data being evaluated, and,
Figure BDA0003634350780000034
represents the mean of the evaluation data, R i Which represents the reference data, is,
Figure BDA0003634350780000035
is the reference data mean; when the deviation correction result of the measured data is evaluated, S represents SM _ O or SM _ O _ bc, R represents SM _ S, if compared with SM _ O, the RMSE between SM _ O _ bc and SM _ S is smaller, the BIAS absolute value is smaller, and CORR is larger, the SM _ O _ bc is more matched with SM _ S, and the deviation correction method can effectively match the measured data to the remote sensing data; when evaluating the ESACCI SM reconstruction result, S represents SM _ DCT-PLS or SM _ ODCT-PLS, and SM _ R represents SM _ O or SM _ S, and if RMSE between SM _ ODCT-PLS and SM _ O and SM _ S is smaller, BIAS absolute value is smaller, and CORR is larger than SM _ DCT-PLS, the superiority of ODCT-PLS relative to DCT-PLS can be explained.
The invention has the beneficial effects that:
the invention adopts a DCT-PLS (odd computed tomography-PLS) method fusing site measured data and satellite remote sensing data, and the main idea is to combine three-dimensional information provided by the site measured data to reconstruct data with higher precision and more reliability on the basis of the DCT-PLS. The improved method can be suitable for areas and periods where satellite remote sensing SM data is widely absent, and accuracy of data missing filling is remarkably improved.
Drawings
Fig. 1 is a flow chart of a method provided by an embodiment of the invention.
Fig. 2 is a 2019 station ESACCI SM and station measured SM time sequence diagram in the embodiment of the present invention.
Fig. 3 is a time sequence diagram of the 2019 annual site SM _ O, SM _ O _ bc and SM _ S in the embodiment of the invention.
FIG. 4 is a time series diagram of SM _ DCT-PLS, SM _ ODCT-PLS and SM _ O in 2019 in the embodiment of the present invention.
Detailed Description
The invention is further described below with reference to the drawings and examples.
The invention utilizes the time integrity and space representativeness of the site observation data to obtain space-time full three-dimensional information to fill the data missing of the widely-missing remote sensing humidity product.
As shown in fig. 1.
A soil humidity data missing filling method fusing space-time three-dimensional information of site and satellite remote sensing observation comprises the following steps:
1. recording collected ESACCI remote sensing soil humidity and site actual measurement soil humidity data of 10cm of the surface layer as SM _ S and SM _ O respectively;
2. matching the site actual measurement soil humidity SM _ O to the satellite remote sensing soil humidity SM _ S by adopting a cumulative probability distribution function (CDF) matching method, so that the matched SM _ O _ bc and SM _ S have the same value range and cumulative probability distribution condition, and the integral deviation between the actual measurement soil humidity data and the satellite remote sensing soil humidity data is reduced;
3. and respectively drawing CDF curves of the site measured soil humidity data and the ESA CCI soil humidity data, and dividing the curves into 10 sections by taking 0, 10, 20, 30, 40, 50, 60, 70, 80, 90 and 100 as dividing nodes.
4. And performing piecewise linear regression on each piecewise node to obtain a piecewise linear regression curve and a corresponding piecewise linear regression equation.
5. And (3) utilizing a piecewise linear regression equation to perform CDF adjustment on the actually-measured soil humidity data SM _ O of the station to be adjusted point by point to obtain the adjusted SM _ O _ bc.
6. Applying a piecewise linear equation to the SM _ O in the remote sensing data loss period, and adding the SM _ O _ bc obtained through calculation into the SM _ S to obtain SM _ SO;
7. carrying out data reconstruction on the SM _ SO based on a DCT-PLS method to obtain SM _ ODCT-PLS;
8. and (3) quantitatively evaluating errors between the measured data and the remote sensing data before and after deviation correction and deviations between the remote sensing data and the reference data before and after reconstruction by adopting the 3 evaluation indexes of RMSE, BIAS and CORR.
The invention is further illustrated below with reference to a specific example:
1. acquiring ESA CCI remote sensing soil humidity data with time resolution of day, space resolution of 0.25 degrees and unit of m 3 /m 3 . According to the European space agency CCI SM product, a complete global SM data set is provided by a weather change initiative project of the European space agency, and a western Liaoriver basin area is cut out as required and is recorded as SM _ O. Collecting data of actually measured soil humidity of a site of the national weather service, wherein the actually measured soil humidity is measured as volume water content (m) 3 /m 3 ) In units, SM was measured at 6 depths (10, 20, 30, 40, 50 and 60cm) with a time resolution of 1 day. The number of the measured stations in the western Liaohe area is 22, and SM _ S is recorded as the measured soil humidity of the stations.
2. Taking 2019 as an example, the soil humidity (SM _ O) and the ESA CCI SM (SM _ S) are measured at 3 sites located in the western Liaoh river basin, the time sequence of the sites is checked, and missing data of the SM _ S are mainly concentrated in 1-4 months and 11-12 months in one year (as shown in FIG. 2). Respectively drawing CDF curves of actually measured soil humidity data and ESA CCI soil humidity data of the site by adopting a cumulative probability distribution function (CDF) matching method, dividing the curves into 10 sections by taking 0, 10, 20, 30, 40, 50, 60, 70, 80, 90 and 100 as dividing nodes, and performing piecewise linear regression on each piecewise node to obtain a piecewise linear regression curve and a corresponding piecewise linear regression process.
3. And applying the piecewise linear equation to the SM _ O in the remote sensing data loss period, and matching the SM _ O _ bc obtained by calculation with the SM _ O _ bc to the SM _ S of the satellite remote sensing soil humidity to obtain SM _ SO. The matched SM _ O _ bc and SM _ S have the same value range and the same accumulative probability distribution condition so as to reduce the integral deviation between the actually measured soil humidity data and the satellite remote sensing soil humidity data, and the result is shown in figure 3.
4. And applying the piecewise linear equation to the SM _ O in the remote sensing data loss period, and matching the SM _ O _ bc obtained by calculation with the SM _ O _ bc to the SM _ S of the satellite remote sensing soil humidity to obtain SM _ SO. And enabling the matched SM _ O _ bc and SM _ S to have the same value range and cumulative probability distribution condition so as to eliminate the system deviation between the actually measured soil humidity data and the satellite remote sensing soil humidity data.
5. And reconstructing data of the SM _ SO based on a DCT-PLS method to obtain the SM _ ODCT-PLS, wherein the result is shown in FIG. 4. The time sequence result shows that the time sequence information of the actually measured data is fully applied, and the problem that the DCT-PLS method is poor in performance in a continuous time missing mode is solved.
6. And respectively taking SM _ O and SM _ S as verification data, and quantitatively evaluating errors between the measured data and the remote sensing data before and after deviation correction and deviations between the remote sensing data and the reference data before and after reconstruction by adopting 3 evaluation indexes of RMSE, BIAS and CORR. The calculation formula of the 3 evaluation indexes is as follows:
Figure BDA0003634350780000051
Figure BDA0003634350780000052
Figure BDA0003634350780000061
wherein n represents the number of samples, S i Representative of the data being evaluated, and,
Figure BDA0003634350780000062
represents the mean of the evaluation data, R i Which represents the reference data, is,
Figure BDA0003634350780000063
is the reference data mean. When the deviation correction result of the measured data is evaluated, S represents SM _ O or SM _ O _ bc, R represents SM _ S, and if compared with SM _ O, the RMSE between SM _ O _ bc and SM _ S is smaller, the BIAS absolute value is smaller, and CORR is larger, the result shows that SM _ O _ bc is more matched with SM _ S, and the deviation correction method can effectively match the measured data to the remote sensing data. When evaluating the ESACCI SM reconstruction result, S represents SM _ DCT-PLS or SM _ ODCT-PLS, and SM _ R represents SM _ O or SM _ S, and if RMSE between SM _ ODCT-PLS and SM _ O and SM _ S is smaller, BIAS absolute value is smaller, and CORR is larger than SM _ DCT-PLS, the superiority of ODCT-PLS relative to DCT-PLS can be explained.
From the results of RMSE, CORR and BIAS of SM _ DCT-PLS, SM _ ODCT-PLS and SM _ O of different sites, the CORR (p <0.01) of SM _ ODCTPLS and SM _ O is larger than that of SM _ DCTPLS, and the numerical values are all more than 0.2, which shows that ODCT-PLS can remarkably improve the correlation between remote sensing data and measured data after deletion filling reconstruction. The RMSE of SM _ ODCTPLS and SM _ O is smaller for most sites, indicating that the average error between SM _ ODCTPLS and SM _ O is smaller. The absolute value of BIAS for most sites is less than 0.01.
Results of 3 quantitative evaluation indexes show that the remote sensing soil humidity missing data filling method provided by the invention shows better filling performance in a continuous time missing mode.
The present invention is not concerned with parts which are the same as or can be implemented using prior art techniques.

Claims (4)

1. A soil humidity data missing filling method fusing space-time three-dimensional information of site and satellite remote sensing observation is characterized by comprising the following steps:
step 1, collecting ESA CCI remote sensing soil humidity and site actual measurement soil humidity data of 10cm of a surface layer, and recording the data as SM _ S and SM _ O respectively;
step 2, matching the site actually-measured soil humidity SM _ O to the satellite remote sensing soil humidity SM _ S by adopting a cumulative probability distribution function (CDF) matching method, so that the matched SM _ O _ bc and SM _ S have the same value range and cumulative probability distribution condition, and the integral deviation between the actually-measured soil humidity data and the satellite remote sensing soil humidity data is reduced;
step 3, based on a CDF matching method, establishing a piecewise linear equation for SM _ O and SM _ S in the non-loss period of the remote sensing data;
step 4, applying the piecewise linear equation to SM _ O in the remote sensing data loss period, and adding the SM _ O _ bc obtained through calculation into SM _ S to obtain SM _ SO;
step 5, reconstructing data of SM _ SO based on DCT-PLS method;
and 6, adopting the 3 evaluation indexes of RMSE, BIAS and CORR to quantitatively evaluate the error between the measured data and the remote sensing data before and after deviation correction and the deviation between the remote sensing data and the reference data before and after reconstruction.
2. The method as claimed in claim 1, wherein the CDF curve of the measured soil moisture data and ESA CCI soil moisture data at the site is divided into 10 segments with 0, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100 as dividing nodes.
3. The method as claimed in claim 1, wherein said obtaining ESA CCI remote sensing soil moisture data has a time resolution of day, a spatial resolution of 0.25 ° and a unit of m 3 /m 3 。
4. The method as claimed in claim 1, wherein the error between the measured data and the remote sensing data before and after the deviation correction and the deviation between the remote sensing data and the reference data before and after the reconstruction are quantitatively evaluated using 3 evaluation indexes of RMSE, BIAS and CORR, respectively, SM _ O and SM _ S as the verification data; the calculation formula of the 3 evaluation indexes is as follows:
Figure FDA0003634350770000011
Figure FDA0003634350770000012
Figure FDA0003634350770000013
wherein n represents the number of samples, S i Representative of the data being evaluated, and,
Figure FDA0003634350770000014
represents the mean of the evaluation data, R i Which represents the reference data, is,
Figure FDA0003634350770000015
is the reference data mean; when the deviation correction result of the measured data is evaluated, S represents SM _ O or SM _ O _ bc, R represents SM _ S, if compared with SM _ O, the RMSE between SM _ O _ bc and SM _ S is smaller, the BIAS absolute value is smaller, and CORR is larger, the SM _ O _ bc is more matched with SM _ S, and the deviation correction method can effectively match the measured data to the remote sensing data; when evaluating the ESACCI SM reconstruction result, S represents SM _ DCT-PLS or SM _ ODCT-PLS, and SM _ R represents SM _ O or SM _ S, and if RMSE between SM _ ODCT-PLS and SM _ O and SM _ S is smaller, BIAS absolute value is smaller, and CORR is larger than SM _ DCT-PLS, the superiority of ODCT-PLS relative to DCT-PLS can be explained.
CN202210498546.9A 2022-05-09 2022-05-09 Soil humidity data missing filling method fusing spatial-temporal three-dimensional information of site and satellite remote sensing observation Pending CN114912075A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210498546.9A CN114912075A (en) 2022-05-09 2022-05-09 Soil humidity data missing filling method fusing spatial-temporal three-dimensional information of site and satellite remote sensing observation

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210498546.9A CN114912075A (en) 2022-05-09 2022-05-09 Soil humidity data missing filling method fusing spatial-temporal three-dimensional information of site and satellite remote sensing observation

Publications (1)

Publication Number Publication Date
CN114912075A true CN114912075A (en) 2022-08-16

Family

ID=82767608

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210498546.9A Pending CN114912075A (en) 2022-05-09 2022-05-09 Soil humidity data missing filling method fusing spatial-temporal three-dimensional information of site and satellite remote sensing observation

Country Status (1)

Country Link
CN (1) CN114912075A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116028480A (en) * 2023-03-30 2023-04-28 中国科学院空天信息创新研究院 Filling method for improving space-time coverage of remote sensing soil moisture product

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116028480A (en) * 2023-03-30 2023-04-28 中国科学院空天信息创新研究院 Filling method for improving space-time coverage of remote sensing soil moisture product

Similar Documents

Publication Publication Date Title
Yuan et al. Estimating surface soil moisture from satellite observations using a generalized regression neural network trained on sparse ground-based measurements in the continental US
Beck et al. Evaluation of 18 satellite-and model-based soil moisture products using in situ measurements from 826 sensors
Wu et al. Evaluation of soil moisture climatology and anomaly components derived from ERA5-land and GLDAS-2.1 in China
Mishra et al. Reconstruction of droughts in India using multiple land-surface models (1951–2015)
Wang et al. Validation and trend analysis of ECV soil moisture data on cropland in North China Plain during 1981–2010
Albergel et al. Soil moisture analyses at ECMWF: Evaluation using global ground-based in situ observations
Wang et al. Assessment of water budget for sixteen large drainage basins in Canada
Tait et al. An assessment of the accuracy of interpolated daily rainfall for New Zealand
Aryee et al. Development of high spatial resolution rainfall data for Ghana
CN111666656A (en) Rainfall estimation method and rainfall monitoring system based on microwave rainfall attenuation
CN111401602B (en) Assimilation method for satellite and ground rainfall measurement values based on neural network
Cammalleri et al. Comparing soil moisture anomalies from multiple independent sources over different regions across the globe
Ko et al. Strategies to improve and evaluate physics‐based hyperresolution hydrologic simulations at regional basin scales
Wang et al. Evaluation of remotely sensed precipitation and its performance for streamflow simulations in basins of the southeast Tibetan Plateau
Baik et al. Agricultural drought assessment based on multiple soil moisture products
Sridhar et al. Evaluating bias‐corrected AMSR‐E soil moisture using in situ observations and model estimates
Tong et al. Spatial gap-filling of SMAP soil moisture pixels over Tibetan plateau via machine learning versus geostatistics
Chen et al. A deep learning hybrid predictive modeling (HPM) approach for estimating evapotranspiration and ecosystem respiration
CN114912075A (en) Soil humidity data missing filling method fusing spatial-temporal three-dimensional information of site and satellite remote sensing observation
CN116029162B (en) Flood disaster inundation range monitoring method and system by using satellite-borne GNSS-R data
Tran et al. Analysing Drought Intensity in the Mekong River Delta using Time Series Analysis and Google Earth Engine.
Xu Mapping soil moisture from remotely sensed and in-situ data with statistical methods
Ma et al. Long-term changes in surface soil moisture based on CCI SM in Yunnan Province, Southwestern China
Su et al. Preliminary assessment of regional moderate-resolution atmospheric reanalysis for Australia
Wang et al. Determining relative errors of satellite precipitation data over the Netherlands

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