Früher|Nächster

Methoden und Systeme zur Qualitätskontrolle von Bodendaten für die Validierung der Fernerkundungsgenauigkeit

  • role： First author第一作者
  • Affiliation：

    National Engineering Research Center for Satellite Remote Sensing Applications/State Key Laboratory of Remote Sensing and Digital Earth, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100101, China

    University of Chinese Academy of Sciences, Beijing 100049, China

  • Email：liququ21@mails.ucas.ac.cn
  • Introduction：李曲曲，研究方向为遥感产品真实性检验。E-mail： liququ21@mails.ucas.ac.cn
LI Ququ12,  
  • role： Corresponding author通信作者
  • Affiliation：

    National Engineering Research Center for Satellite Remote Sensing Applications/State Key Laboratory of Remote Sensing and Digital Earth, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100101, China

    University of Chinese Academy of Sciences, Beijing 100049, China

  • Email：wenjg@aircas.ac.cn
  • Introduction：闻建光，研究方向为定量遥感。E-mail： wenjg@aircas.ac.cn
WEN Jianguang12*,  
  • Affiliation：

    National Engineering Research Center for Satellite Remote Sensing Applications/State Key Laboratory of Remote Sensing and Digital Earth, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100101, China

    University of Chinese Academy of Sciences, Beijing 100049, China

XIAO Qing12,  
  • Affiliation：

    Faculty of Geosciences and Environmental Engineering, Southwest Jiaotong University, Chengdu 611756, China

WU Xiaodan3,  
  • Affiliation：

    National Engineering Research Center for Satellite Remote Sensing Applications/State Key Laboratory of Remote Sensing and Digital Earth, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100101, China

YOU Dongqin1,  
  • Affiliation：

    National Engineering Research Center for Satellite Remote Sensing Applications/State Key Laboratory of Remote Sensing and Digital Earth, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100101, China

TANG Yong1,  
  • Affiliation：

    National Engineering Research Center for Satellite Remote Sensing Applications/State Key Laboratory of Remote Sensing and Digital Earth, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100101, China

LIAN Ting1,  
  • Affiliation：

    National Engineering Research Center for Satellite Remote Sensing Applications/State Key Laboratory of Remote Sensing and Digital Earth, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100101, China

PIAO Sen1,  
  • Affiliation：

    National Engineering Research Center for Satellite Remote Sensing Applications/State Key Laboratory of Remote Sensing and Digital Earth, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100101, China

ZHAO Na1

Resümee

Der Kern der Validierung der Fernerkundungsgenauigkeit ist die Gewinnung von Wahrheitsdaten im Pixelmaßstab. Bodendaten sind die Hauptquelle für referenzierte Wahrheitsdaten im Pixelmaßstab bei der Validierung der Fernerkundungsgenauigkeit. Allerdings werden Bodendaten während der Erfassung, Übertragung und Verarbeitung von verschiedenen Faktoren beeinflusst, was zu bestimmten Fehlern führt, die die Genauigkeit der Pixelwahrheit und die Zuverlässigkeit der Fernerkundungsvalidierung beeinträchtigen. Daher systematisiert diese Studie, die sich an den Anforderungen zur Erfassung von Pixelwahrheitsdaten für die Validierung der Fernerkundungsgenauigkeit orientiert, die Hauptquellen von Fehlern und die bestehenden Hauptprobleme der Bodendaten. Aus den Perspektiven Normativität, Vollständigkeit, Aktualität, Genauigkeit, Konsistenz und raum-zeitlicher Repräsentativität fasst sie die aktuellen Methoden und Systeme zur Qualitätskontrolle von Bodendaten zusammen und untersucht am Beispiel von Bodendaten zur Gewinnung der Pixelwahrheit der Oberflächenalbedo die Anwendungseffekte von Qualitätskontroll- und Bewertungsmethoden. Diese Studie kann technische Unterstützung zur Förderung der standardisierten Verwaltung von Bodendaten für die Validierung der Fernerkundungsgenauigkeit und zur Verbesserung der Qualität der Pixelwahrheitsdaten der Fernerkundung bieten.

Schlüsselwort

Bodendaten;Fernerkundungsprodukte;Validierung der Genauigkeit;Fehler;Qualitätskontrolle;Oberflächenalbedo

1 Introduction

Remote sensing ground observation data, as the input of model algorithms and the main source of reference truth values in authenticity verification, is the foundation of scientific research and application（Doldirina，2015；Balsamo等，2018；Xu等，2024）. With the rapid development of remote sensing and ground observation technology, remote sensing products have gradually become important sources of information in fields such as resources, environment, ecology, hydrology, etc. The Chinese high-resolution Earth observation system has built multiple types of Earth observation capabilities, including visible light, infrared, and microwave, by carrying multi spectral and multi-mode observation payloads. It can continuously provide common remote sensing products and provide strong data support for applications in multiple fields（柳钦火 等，2023）. However, remote sensing data may introduce errors through radiometric calibration, geometric correction, algorithm inversion, and other processes, resulting in a certain degree of uncertainty in the remote sensing products themselves（闻建光 等，2023b）Authenticity testing is an important means of evaluating its quality（晋锐 等，2017）. The reference truth values used for verification mainly come from ground observation data（吴小丹 等，2019；张圆 等，2020）Accurate and reliable ground observation data is an important reference for verifying the authenticity of remote sensing products（张仁华 等，2010）. Due to errors in the observation instruments themselves, interference from external environmental factors, differences in the understanding of observation standards by station staff at various locations, and variations in observation methods, the quality of observation data varies greatly. Quality control and evaluation of ground observation data are of great significance for objectively evaluating the accuracy of quantitative remote sensing products（李新，2013；Chave等，2019）.
Many publicly released datasets around the world, such as the FLUXNET 2015 dataset provided by the Global Flux Observing Network（Pastorello等，2020）Global Historical Climate Dataset（Peterson和Vose，1997）CLPX (the Cold Land Processes Field Experiment)（Parsons等，2004）The International Soil Moisture Network (ISMN) dataset（Dorigo等，2021）We have implemented strict quality control measures.
China has established a large number of ground observation stations in the fields of meteorology, ecology, remote sensing, hydrology, etc（闻建光 等，2023a）The construction of station networks at different observation sites is the responsibility of different scientific research institutions and government departments, lacking unified quality control standards, resulting in a lack of consistent quality benchmarks between data and affecting the cross regional and cross platform joint use of data. Therefore, quality control of ground observation data has become an important issue in improving data quality, promoting scientific research, and data sharing.
Quality Control (QC) refers to the analysis and labeling of suspicious and unreasonable data based on the spatiotemporal variation characteristics of observation elements and the interrelationships between them（刘小宁和任芝花，2005）. Quality control, as a core component of the data governance process, aims to ensure the standardization, completeness, timeliness, accuracy, consistency, and representativeness of data（国家市场监督管理总局和中国国家标准化管理委员会，2018）. Traditional quality control methods typically include boundary range checks based on physical models, consistency checks, etc（Guo和Liu，2015；任芝花 等，2015；林兴稳，2018；Bian等，2019）. These methods can effectively identify outliers by rigorously verifying and screening the data. In recent years, with the rapid development of computer science, especially machine learning and artificial intelligence technology, many scholars have devoted themselves to combining automation technology with artificial intelligence methods, proposing many innovative quality control frameworks. Anomaly detection methods based on deep learning have been widely applied in the fields of agriculture, meteorology, and environmental monitoring（Dandrifosse等，2024）Especially in terms of quality control of time series data, such asBandaru等（2024）By developing a bidirectional long short-term memory (LSTM) model, it is possible to effectively identify and eliminate outliers in soil moisture time series data. Some studies rely on comparing measured values with meteorological models（Båserud等，2020）Or machine learning algorithms（Kim等，2020；Sha等，2021）Comparing predictions can effectively identify and correct anomalies in observed data.Toreti等（2022）We have developed a near real-time quality control system called QuackMe (Quality Checks of Meteorological Data) to obtain high-quality daily meteorological information for crop yield forecasting systems and European drought monitoring. These studies indicate that improvements in quality control methods can not only improve data processing efficiency, but also enhance data accuracy and efficient utilization, providing more reliable basic data support for subsequent applications such as meteorological forecasting and agricultural decision-making.
Although these quality control methods have been applied in different fields, most of them only focus on the integrity and consistency of the observed data itself. The existing quality control methods have not yet formed a unified system, and there are many methods in the industry but lack systematicity, which makes the comparability and universality of different research results low. Especially in remote sensing authenticity verification, quality control not only ensures the validity of data, but also pays attention to whether it is representative at the spatiotemporal scale. Previous studies have shown that there is a significant correlation between the error of authenticity verification results and the representative error of ground observations. Especially when the representative error is large, the verification error mainly comes from the insufficient representativeness of ground observations at the spatial scale（Wu等，2019）. Therefore, how to summarize and standardize existing quality control methods, increase constraints on the spatiotemporal representativeness of data at the pixel scale, and propose a more systematic and standardized ground observation data quality control process for authenticity verification, in order to enhance the effectiveness of ground observation data for remote sensing authenticity verification, has become an urgent problem to be solved in practical applications.
Therefore, based on the Web of Science Core Collection and the China Science Citation Database, this study searched for 320 articles published between 2000 and 2025 using keywords such as ground observation data, quality control, and authenticity verification. Firstly, the main sources of errors in remote sensing ground observation data were identified, and the main problems currently faced were analyzed; Then, based on the characteristics of ground observation data, summarize various methods for quality control and evaluation; Finally, this study takes surface albedo as an example to explore the application effect and feasibility of quality control methods, in order to provide theoretical support and practical guidance for the quality control of ground observation data.

2. Main sources of error in ground observation data

The overall error sources of ground observation data for authenticity verification mainly include observation errors and representative errors（李新，2013）.
In the equation,Represents the overall error of ground observation data,Indicating observation error,Representing errors. Among them, observation errors include instrument measurement errors and measurement errors, which are inherent types of errors in the observation process, mainly derived from the limited accuracy of instruments and operational errors in the measurement process. Representative error is caused by the fact that the information of sampling points cannot effectively represent the information of surface elements at limited spatial resolution.

2.1 Instrument measurement error

Instrument measurement error refers to the error caused by the limited accuracy of the instrument during the measurement process, mainly affected by factors such as wear and aging of the observation instrument, inaccurate zero point, incorrect use of instrument calibration factors, accumulation of sensor dirt, problems with sensor physical positioning, equipment energy supply, damage to data transmission media, network latency, and changes in the external natural environment（刘丰和郭建文，2013；Pastorello等，2014）. Measurement errors often have obvious regularity and accumulation, which have a significant impact on the measurement results. Measures such as precise calibration, regular maintenance and calibration of the instrument before use, reasonable operation and inspection correction during use, and adding correction factors to the data for calibration after observation can be taken to reduce the accumulation of deviations caused by long-term operation. In addition, the authenticity verification field network can record and analyze the operating status of instruments and environmental metadata, making it easy to trace the source of abnormal information. Adopting redundant observations and multi-sensor cross validation to enhance the ability to identify outliers; Manual inspection and on-site recording also play an irreplaceable role in discovering equipment failures or human interference, and these measures can reduce the impact of measurement errors.

2.2 Measurement Error

Measurement error refers to the method error caused by incomplete measurement methods or inadequate theoretical basis, as well as the error caused by improper operation during the measurement process. It is not related to the instrument itself and can usually be divided into two categories: systematic error and random error. Systematic errors are caused by systematic factors such as measurement methods and theoretical models, and have regularity and repeatability. They can be controlled by optimizing measurement methods, improving theoretical models, or introducing statistical correction methods. Random errors are mainly caused by accidental factors during the measurement process, resulting in characteristics such as disorder, abnormality, and inaccuracy in the data. The size and sign of the errors often do not have a certain pattern, making it difficult to estimate directly. However, in a sufficient number of measurements, the overall distribution generally follows a Gaussian distribution, which has a certain degree of compensation（常国宾 等，2024）Therefore, the impact can be reduced by taking the average of multiple measurements. In addition, the accuracy and reliability of observation data can be improved by standardizing operating procedures and strengthening observer training.

2.3 Representative Error

Representative error mainly refers to the inference error caused by insufficient representativeness in the process of inferring the spatial population represented by sampled data. Representative error is closely related to spatial heterogeneity（李新，2013）. The main reasons for its occurrence include: insufficient spatial representativeness of sampling points, limited number of sampling points that cannot cover the overall spatial structure within the area, and so on. Representative errors cannot be completely eliminated, but can be effectively controlled by optimizing spatial sampling（Wen等，2022；Wu等，2021）. The core of optimizing sample layout lies in ensuring its optimal global representativeness. At present, the methods for optimizing sampling mainly include three categories: classical sampling methods (such as simple random sampling, stratified sampling, etc.)（Cochran，1977）Spatial sampling methods (such as spatial system sampling, spatial sandwich sampling, etc.)（Cressie，2015；Wang等，2012）And model-based sampling methods (such as geostatistical Kriging sampling, cumulative representativeness sampling, etc.)（Gruijter等，2006；Wu等，2016）. After determining the sampling points, repeat sampling at multiple points within the sample plot to obtain surface spatial heterogeneity while reducing the impact of accidental errors.
In authenticity verification, time representativeness error also needs to be considered. When there is a time difference between ground observation and remote sensing observation, whether the ground data can effectively reflect the surface state at the time of remote sensing transit is the focus of time representativeness evaluation. The key to time representativeness lies in the temporal stability of the terrain conditions: for example, the leaf area index changes slowly within a few days, and even if the observation time is different, its time representativeness is still good; If the surface temperature is significantly affected by radiation, strict time matching (such as controlling within half an hour) is required to ensure representativeness. Therefore, time representativeness should be evaluated in conjunction with the dynamic characteristics of land cover variables, and synchronous observation is an important means to improve time representativeness.

3 main problems faced

Due to the unpredictability of observation data quality issues, data quality control for authenticity verification still faces many challenges（Pastorello等，2017）This is mainly reflected in the following two aspects.

3.1 Multi source heterogeneous data, complex automated processing

Remote sensing ground observation data is multi-source heterogeneous and diverse in types, mainly consisting of semi-structured and unstructured data（王桥，2021）The original documents of manually observed data are mainly handwritten records and captured images, while the organized data is mainly in Word and Excel formats. The automatically observed data includes txt, xlsx, XML, dat files, images, and log documents. Different sources and formats of data often lack unified standards, resulting in differences in data structure, accuracy, timeliness, etc., further increasing the complexity of data integration and bringing difficulties to subsequent data organization and quality control.

3.2 Lack of quality control system for authenticity verification

The quality of remote sensing ground observation data, as the main source of reference truth values for authenticity verification, directly determines the reliability of the verification results. However, currently most quality control schemes only focus on the consistency and physical rationality within the data, lacking a complete quality control system for authenticity verification at the pixel scale. In authenticity verification, in addition to focusing on the quality of the data itself, attention is also paid to the spatiotemporal representativeness of ground data at the remote sensing pixel scale. Therefore, building a complete quality control process that covers data preprocessing, quality discrimination, and representative analysis has become an important direction for improving the reliability of remote sensing authenticity testing.

4 Quality Control Method System

Remote sensing ground observation data includes both continuous observations from wireless sensor networks, flux towers, and manually collected data from multiple points（李新 等，2023）. The quality of ground data can be described using quality elements, which refer to the basic characteristics of a product that meet user requirements and usage purposes（曾衍伟和龚健雅，2004）. According to the characteristics of ground observation data, primary quality elements can be classified into normativity, completeness, timeliness, accuracy, consistency, and spatiotemporal representativeness（Fig. 1）The second level quality element is a specific description of the first level quality element, that is, the second level quality element is the most basic carrier of data quality information（刘丰和郭建文，2013）.
figure

Fig. 1 Quality control framework for in situ observation data in remote sensing validation

4.1 Normative Inspection

Normative inspection mainly checks whether the format of the obtained observation data complies with the requirements of national or industry standards related to ground observation. It checks whether the metadata is standardized, whether the units of the same type of data are unified, whether the variable names are standardized, and whether the storage format of the same type of data complies with the specifications. Non standardized data is sorted and processed into standardized and formatted data for further application.

4.2 Integrity check

4.2.1 Data element integrity

The content of data element integrity check includes the integrity of file naming, observation variables, data observation periods, etc（马明国，2010）. The first step is to check the integrity of the file, checking whether the field information of the observation data is complete and whether the recorded observation data is missing. If there is a missing data, the corresponding observation personnel should be contacted to verify the reason for the missing data. If there is missing data during recording, it should be supplemented in a timely manner. If there is redundancy in the data, duplicate data should be removed. Secondly, check whether each variable has submitted data according to the actual observation period. For example, if the observation period for radiation data is 10 minutes, there should be new observation records every 10 minutes in the submitted data. If the frequency of data observation is not sufficient, check whether the instrument has malfunctioned. Finally, verify whether the provided data corresponds to the agreed upon variables.

4.2.2 Data Record Integrity

Conduct quality checks on the completeness and repeatability of data. In data processing, there are different types of null values, including transmission null values, calculation null values, and instrument failure null values (data records such as "N/A", "NAN", "‒ 999", or null). The transmission of null values is caused by data loss or format conversion failure during the data transmission process; Calculating null values refers to missing values in data calculated from other attribute data due to calculation failures; Instrument failure null values are caused by abnormal or interrupted observation equipment, resulting in the inability to collect and transmit data for a continuous period of time, leading to long-term data loss. These null values will cause systematic deviations in subsequent analysis and need to be identified and corrected in quality control. The automatically observed data can be filtered based on the quality identification or null attribute settings provided by the instrument. For null values in manually observed data, it is necessary to verify whether data was underreported during the organization of paper data.

4.3 Timeliness check

The timeliness check of data refers to the process of evaluating the frequency, novelty, and effectiveness of data updates. It is an important aspect of ensuring data quality and a prerequisite for timely and accurate data support for decision-making（李默涵和李建中，2016）. Data timeliness can be divided into timestamp based timeliness and rule-based timeliness（Gong等，2023）. Time stamp based timeliness refers to checking the timestamp of data to determine whether it has been updated within the expected time. For automatic observation of network transmission data timeliness,刘丰和郭建文（2013）pass式（2）Conduct timeliness assessment,
In the equation,Represents the time of data collection,Represents the time of data reception, andIt represents the frequency of data collection. whenWhen approaching 0, it indicates that the timeliness of the data is stronger; And whenApproaching 1 means that the timeliness of the data is poor.
Rule based timeliness refers to determining whether data is updated in a timely manner based on business needs or application scenarios. For manually observed ground data, timeliness evaluation can be conducted from the aspects of data aggregation frequency, timely updating of data storage, etc., such as monitoring the leaf area index of vegetation growth status. Manual observation needs to reflect the true state of the target time period, usually recording data on a weekly basis to ensure that the number or frequency distribution of records within the date range meets business needs（Gong等，2023）. It is possible to determine the data missing rate by specifying the observation periodMake a judgment:
In the equation,Indicating the planned amount of observation data within the specified date range,Indicates the actual amount of data collected.The closer it is to 0, the stronger the timeliness of the data; And whenWhen approaching 1, the data timeliness is poor.

4.4 Accuracy check

4.4.1 Physical Possible Value Check

Physical possibility value check is used to ensure that observed data is within a reasonable physical range, by setting a set of reasonable physical ranges or thresholds in advance to filter or mark abnormal data that exceeds the reasonable range. As for surface albedo data, its physical possible value range is limited to (0,1), and the drought index calculated based on vegetation index has a value range of (‒ 1,1). If the observed value is not within this range, the data is discarded or marked as abnormal data and cannot participate in the calculation of subsequent related parameters.Table 1Listed the physical possible ranges of several common remote sensing ground observation parameters.

Table 1 Physically possible value ranges for common remote sensing in situ observation parameters

参数物理可能值单位参数物理可能值单位
地表反射率（0， 1）植被指数（‒1， 1）
归一化离水辐射率（0， 1）叶面积指数（0， 10）m2/m2
地表发射率（0， 1）植被覆盖度（0， 1）
气溶胶光学厚度（0， 10）光合有效辐射吸收比例（0， 1）
地表反照率（0， 1）森林地上生物量（0， 1000）t/hm2
光合有效辐射（0， 1000）W/m2蒸散发（0， 10）mm/d
地表温度（150， 350）K土壤水分（0， 1）m3/m3

4.4.2 Extreme value check

Extreme value checking is mainly based on prior knowledge to establish the "soft boundary" of ground observation data range. If the observation data exceeds the boundary range established by prior knowledge, it can be considered as suspicious data. The key to extreme value inspection lies in setting the maximum and minimum extreme values reasonably（赵虹，2015）If the maximum extreme value is set too high or the minimum extreme value is set too low, it may not be possible to effectively detect and exclude abnormal data. On the contrary, if the maximum extreme value is set too low or the minimum extreme value is set too high, it may erroneously exclude normal ground observation data, making the quality control results unreasonable.
In traditional extreme value testing, the maximum and minimum extreme values are usually determined by adding or subtracting four times the standard deviation from the average value of the data（任芝花等，2006）. La ï da criterion (also known as 3)Guidelines)（Wang等，2018；Wang等，2021）Further narrowing down the possible range of extreme values, assuming there is a series of sample dataThe mean ofXisμ, and the standard deviation isIf the following conditions are met式（4）When the deviation of a certain data from the mean is greater than or equal to 3 times the standard deviation (3)）At that time, the data is considered suspicious.
The assumption of this method is that there is a sufficient amount of data, that isWhen the data is large enough and follows a normal distribution, normal data may be deleted when the amount of data is small, and this possibility decreases as the amount of data increases. In addition, when a certain outlier or some outliers are particularly large, it will result in an extreme value range of 3 for determining the outlierIt is also particularly large, at this time, only obviously abnormal data can be removed, and slightly abnormal data may fall within the range of extreme values, making it difficult to be effectively detected, thus making the efficiency of using the La ï da criterion method relatively low.
Guo和Liu（2015）Improved by 3The algorithm of the criterion extends the time series data and performs polynomial fitting and calculates the standard deviationCompare the absolute residual value between each data value and the fitting function to see if it is greater than or equal to 3To determine if the data is abnormal. Improved 3The use of fitting values of data sequences in calculations can accurately simulate the trend of data, and identifying abnormal data through residuals can improve the detection accuracy of suspicious data.

4.4.3 Environmental Change Inspection

Environmental change inspection is mainly aimed at identifying the impact of cloud changes on the observation results of all atmospheric related radiation parameters in ground data, effectively identifying cloud changes, providing signs of cloud presence, and correcting their effects.
The most direct method to determine whether there are clouds in the weather is visual observation. Experimenters can record the weather conditions in real time by directly observing the cloud cover in the sky. This method is feasible in a short period of time, but it is difficult to achieve for field stations that require long-term continuous observation（吴小丹 等，2019）. With the development of modern artificial intelligence technology and the use of deep learning methods, cloud images automatically captured by the all sky imaging system are classified into clear and cloudy categories, and the cloud cover is determined to correspond to whether the data for the day is available（Kazantzidis等，2012；Kalkan等，2022）.Li等（2011）By combining the advantages of fixed and adaptive thresholding methods through a hybrid thresholding algorithm, the color cloud map is normalized into a blue/red channel ratio image, which is processed by different thresholding algorithms based on its standard deviation, thereby improving the accuracy of cloud detection.
When cloud information is not available or unavailable, radiation data can be used for judgment. Due to the fact that in cloudy weather conditions, the downward shortwave radiation value is lower than that in sunny weather, this feature can be used to label or exclude data from cloudy weather（吴小丹 等，2019）； Theoretical clear sky downward radiation is calculated through radiative transfer simulation. When the actual incident solar radiation on the surface is greater than or equal to 0.8, it can be considered clear sky（Papachristopoulou等，2024）.

4.4.4 Geographical location change inspection

Due to urban construction and development, natural disasters, climate change, changes in land use types, and other human activities, the observation environment of the original site has significantly changed, which cannot meet the observation requirements and affects the accuracy of the data, making the data unrepresentative. This has led to the relocation of the site to a new location to ensure the accuracy of the observation data. The observation environment has undergone significant changes before and after the site migration, which has a significant impact on the uniformity of the data. The commonly used method for testing the non-uniformity of observation data caused by changes in geographical location is based on the maximum penaltyFtest (PMFT)（Wang，2008）Maximum Punishment Test (PMT)（Wang等，2007）Standard Normalization Test (SNHT) method（Pandžić等，2020；韩海涛 等，2021）.
The SNHT method can effectively identify breakpoints using reference sequences, but the confidence level during testing is easily affected by the breakpoint position. When the breakpoint is located at the beginning and end of the time series, the probability of false detection is higher, that is, the confidence level is lower（Toreti等，2011）. The PMFT and PMT methods detect change points by comparing the variance and mean before and after the change point. By introducing a penalty term, false positives are effectively reduced. However, in cases of high data noise, changes in variance or mean may mask the true change point. The PMT method requires a time series as a reference sequence, but it is inevitably affected by non-uniform elements in the reference sequence, which may result in false breakpoints. The PMFT method is suitable for uniformity testing without a reference station（Wan等，2017）Avoiding the influence of unreasonable and non-uniform factors in the reference sequence, but ignoring the trend characteristics of the station itself（肖晶晶 等，2018）.

4.5 Consistency check

4.5.1 Internal Consistency Check

Internal consistency check, also known as data comparison check, aims to identify whether multiple related variables observed at the same time or during the same time period follow established physical laws（Dandrifosse等，2024）. For example, the daily minimum air pressure should be lower than the daily average air pressure, and the daily average air pressure should be lower than the daily maximum air pressure; Similarly, the daily minimum ground temperature should be lower than the daily average ground temperature, while the daily average ground temperature should be lower than the daily maximum ground temperature. The above relationship reflects the physical consistency between variables. If the observed data does not meet these basic physical logical relationships, it may reflect the existence of anomalies or measurement errors in the observed data. Therefore, internal consistency checks are often used as a key link in the data quality control process to quickly screen and label suspicious data.

4.5.2 Time Consistency Check

Time consistency check refers to the observation data within a continuous time period that should maintain a stable and gradual change, rather than sudden changes, that is, the observed elements should be stable over time（赵虹，2015）. The purpose of time consistency check is to determine whether the changes of an observed element over time conform to a specific pattern within a certain time range. If the data exceeds the maximum allowed change threshold during the observation period, this study considers the data to be suspicious.
Using data from observation stations for time interpolation estimation, comparing estimated values with measured values for time-varying checks, effectively filtering out random noise in long-term time series（Steinacker，2023）Alternatively, a subsequence anomaly detection method can be used to select multiple fixed length subsequences from time series data through a sliding window. Statistical features such as mean, standard deviation, skewness, and kurtosis of the subsequences can be extracted separately, and dynamic thresholds can be set to determine whether the subsequences are abnormal. The advantage of this method is that it can preserve the original time series attributes within each relative sliding window（杨寒雨 等，2023）.

4.5.3 Spatial Consistency Check

The spatial consistency check mainly checks whether the changes in observed data within a certain spatial range conform to their spatial patterns. According to the first law of geography, many ground observation variables are correlated in geographic space. If the spatial distance between ground observation stations is close, their observation results are often similar（卢宾宾 等，2025）.
Parameters unrelated to the underlying surface (temperature, humidity, etc.) can be determined by calculating the difference between estimated and observed values. Common methods for calculating estimated values include inverse distance weighted interpolation（Lu和Wong，2008；赵虹，2015）Polynomial interpolation（Guidoum，2025）Optimal interpolation（Miatselskaya等，2022）Spatial regression（Cerlini等，2020）Wait. It is worth noting that in the selection of methods, the density of the observation network and the spatiotemporal variation characteristics of the surface need to be considered. For example, the inverse distance weighted interpolation method is computationally efficient and suitable for real-time business systems, while the spatial regression method has better interpretability by introducing geographic covariates such as elevation and slope direction.

4.6 Spatiotemporal Representative Evaluation

4.6.1 Time representativeness evaluation

Time representativeness is defined as a set of ground measurements within a given time and space domain, reflecting the actual situation at a specific application scale at what scale（Nappo等，1982）. Time representativeness mainly evaluates whether the measurement period of observation parameters is sufficient to capture natural changes in climate and biological conditions. However, the rationality of time representativeness cannot rely solely on the length of observation time. It needs to comprehensively consider the time series characteristics of specific parameters, the completeness and temporal correlation of data, and the overall time coverage（Chu等，2017；Baca-López等，2021）. For certain observation parameters, such as soil heat flux, evapotranspiration, etc., the observation data needs to cover at least one complete growing season to ensure the representativeness of the validation time.
In recent years, various time representative evaluation methods have been proposed, including trend analysis（Chang等，2021）Cycle analysis（Assfalg等，2009）Comparison of sample mean and standard deviation（Anttila等，2012）Monte Carlo simulation（Sharp等，2015）Sliding Window Analysis（杨寒雨 等，2023）And correlation analysis, etc（Chu等，2017）. These methods each have their own focus and can evaluate the temporal representativeness of data from different perspectives. For example, trend analysis determines the representativeness of a certain period by analyzing long-term trends, and determines whether the average status and interannual variability of driving variables can be fully captured during the observation period of the station（Chu等，2017；Xu等，2024）； Periodic analysis focuses on capturing periodic features in data; The comparison method of sample mean and standard deviation analyzes the representativeness of data from a statistical perspective（Anttila等，2012）； In addition, Monte Carlo simulations evaluate time representativeness through extensive random sampling calculations.

4.6.2 Evaluation of spatial representativeness

Spatial representativeness refers to the extent to which sample data measured in a specific spatiotemporal dimension can reflect the true situation within that specific range（徐保东 等，2015）. Due to the fact that the design of observation towers has never maximized spatial coverage, people usually consider extending from the observation location to a specific range, that is, spatial representativeness（Schimel等，2015）. When using ground observation data for modeling, inversion, verification, and other applications of remote sensing products, spatial representativeness evaluation should be conducted to ensure that the data has good spatial representativeness. The main factor affecting the spatial representativeness of observation data is the mismatch of point surface features caused by spatial heterogeneity.
The semi variance function model is a commonly used method for measuring surface spatial heterogeneity, which demonstrates the spatial representativeness of station observations at the pixel scale（Curran，1988；Román等，2009；Wang等，2014；Wu等，2022；Li等，2025；吴小丹 等，2025）As long as the spatial data satisfies the assumption of stationarity, the semi variance function can be used to describe and quantify its spatial variability and autocorrelation. When the spatial heterogeneity in the study area is small, it indicates that the distribution of land cover in the area is relatively uniform and the representativeness of the stations is good (Ma Jin, 2021).
In the equation,Representing a semi variance function,Indicate locationThe sample values at,Represents the spatial distance between any two samples,The distance is represented asThe number of point pairs in time.
In order to eliminate the influence of the numerical value of the target variable itself,Xu等（2016）Defined the benchmark coefficient to describe relative heterogeneity（）：
In the equation,The size representing the spatial resolution of pixels,Represents the average value of the target variable within the statistical region,andThey are block gold value and biased base station value, respectively.The smaller the size, the weaker the spatial heterogeneity of the surface, and the better the spatial representativeness of the site（徐保东，2018）.
The point surface feature comparison method compares the relationship between station observations and statistical indicators (such as mean) representing the data within the region. If it is less than a certain threshold, it is considered that station observations can represent the corresponding regional features（Wu等，2022）. Based on the characteristics of vegetation parameters,Xu等（2016）Designed an indicator that can quantitatively measure point surface consistency: the proportion of major vegetation types（）Relative absolute error（）, used to evaluate the spatial representativeness of ground observation LAI, the expressions for the above indicators are
In the equation,Representing pixel size,The area of vegetation types observed at the representative station within low resolution pixels,Representing theThe area of vegetation type,Representing the total number of vegetation types within a pixel,The pixel values corresponding to the vegetation parameter trend surface represent the location of the station,Represents the mean of vegetation parameters at the pixel scale. whenThe higherThe smaller the size, the better the consistency between points and surfaces, and the better the spatial representativeness.
When evaluating the representativeness of ground data, the dimensions of time and space are often inseparable, and spatiotemporal representativeness reflects the degree of representativeness of the entire research area and its temporal evolution process. Common methods include Euclidean distance（Su等，2022；He等，2015；Wu等，2021）Cosine similarity（Wu等，2016）Wait, it can quantitatively analyze the correlation and differences of data in spatial and temporal dimensions.

5 Application Examples

5.1 Research Area and Data

Surface albedo is one of the key parameters regulating surface radiation budget and ground atmosphere interaction（Dickinson，1995）It is widely used in various climate prediction models, surface energy balance, and global climate change research（Han等，2024）. This study takes surface albedo data as an example to introduce its quality control and evaluation.
The study area is the woodland observation area (40.3478 ° N, 115.7768 ° E) of Huailai Remote Sensing Comprehensive Experimental Station of the Chinese Academy of Sciences. The objects in the observation area are artificial forests with flat terrain, mainly including evergreen coniferous vegetation Chinese pine, broad-leaved deciduous vegetation Begonia and ornamental vegetation maple. Seven radiation data ground observation stations are arranged in the woodland, such asFig. 2As shown. The specific geographical location of the site is as followsTable 2As shown.
figure

Fig. 2 True-color composite image（bands 4-3-2）of Huailai Station from Landsat 8

Table 2 Geographical locations of the 7 in situ observation sites at Huailai station

站点名经度/（°E）纬度/（°N）
HH1115.790340.35093
HH2115.792240.34953
HH3115.788740.34929
HH4115.790240.34866
LH115.795640.35059
SS115.794740.35169
YS115.794840.34846
The observation time range of the ground data used is from May 15, 2020 to November 25, 2021. The data type is up and down shortwave radiation data, which is automatically measured by the CMP3 radiation meter with a time resolution of 10 minutes. The radiation meter has undergone strict calibration, and the uncertainty caused by the observation instrument itself can be ignored. The installation height of the radiation meter is 5 meters, and the diameter range of the observation footprint is 30 meters. The surface albedo can be obtained by comparing the measured upstream shortwave radiation with the downstream shortwave radiation. In this study, data from half an hour before and after noon, when the solar zenith angle is the smallest, were selected to represent the observed values of surface albedo for that day.
The quality control of ground observation data in this case is aimed at verifying the authenticity of ground albedo products at a scale of 500 meters at noon, such as MODIS albedo product MCD43A3（Schaaf等，2002）Therefore, in terms of time scale, the ground data corresponds to the observation data at noon. Due to the large difference in pixel scale between the ground station observation scale and the 500m resolution product, high-resolution surface albedo products are used（游冬琴 等，2023）As an intermediate reference, characterize the spatial heterogeneity within coarse resolution scale pixels.

5.2 Data Quality Control

5.2.1 Normative Inspection

Conduct standardization checks on the format, content, and variable names of surface albedo data, and ensure that the observed data meets the standards.

5.2.2 Integrity check

A quality review was conducted on the completeness, repeatability, observation period, and other aspects of the data. After inspection, it was found that there was a data interruption at the Red Haitang 3 (HH3) station, and the data was missing from July 3, 2020 to July 8, 2020. After verification with the data observation personnel, it was determined that there was a power supply problem at the HH3 station during this period, resulting in the missing observation data. The observation data from other sites meet the requirements for file integrity, cycle integrity, and variable integrity.

5.2.3 Timeliness check

Due to the fact that the data obtained is historical data provided by the site and not real-time observation data, the timeliness check is skipped.

5.2.4 Accuracy check

The physical range of possible values for surface albedo is limited to (0,1), and once it exceeds this range, it is considered abnormal data. In the extreme value check, based on the actual situation of the observed elements, the average value ± 2.5 times the standard deviation of the data is calculated as the maximum/minimum extreme value（Peterson和Vose，1997）.Fig. 3Displayed the extreme value check results of one of the sites HH1. After the above inspection, it was found that a total of 17 days of data from 7 sites did not pass the extreme value check.
figure

Fig. 3 Outlier inspection results of albedo at HongHaitang 1（HH1） site

5.2.5 Consistency check

In this study, the experimental data used were uplink shortwave radiation and downlink shortwave radiation data. According to the requirements of internal consistency check, the physical condition that the uplink shortwave radiation is less than the downlink shortwave radiation should be met. The ratio of uplink shortwave radiation to downlink shortwave radiation is the surface albedo. In the process of physical possibility value check, by defining the physical range of surface albedo as 0-1, data that meets the condition of uplink shortwave radiation ≤ downlink shortwave radiation have been screened out. Therefore, no repeated check will be conducted, and the data will pass the internal consistency check.
The discrimination method used for time consistency check is median average filtering, which takes the relative deviation of surface albedo within half an hour before and after noon on the same day, which does not exceed 5%, to determine whether the data is stable within half an hour before and after noon on the same day. Through inspection, it was found that from May 2020 to November 2020, except for the SS station, there were outliers in the data of the other six ground observation stations. A total of 29 time periods in this inspection did not pass the quality inspection, with a data anomaly rate of 0.715%. Based on the analysis of historical meteorological data, the reason for the occurrence of outliers is due to cloud cover at the observation time. The data marked with this anomaly time is averaged from the daily observation data that has passed the time consistency check as the ground measured value of the daily surface albedo.
Select observation data from two or more sites with the same or similar underlying surface as the site to be inspected, calculate their average value, and compare the average value with the observation data value of the site to be inspected. If the deviation exceeds the given threshold (set as 15% in this study), the data is considered suspicious. The data after consistency check is as followsFig. 4As shown.
figure

Fig. 4 Reference dataset of surface albedo at Huailai station

5.2.6 Representative evaluation

In order to evaluate whether ground station observations can represent high-resolution albedo with a resolution of 16 meters, this study used higher resolution data (high-resolution GF-2 multispectral data with a resolution of 4 meters) to examine the spatial distribution characteristics of DN values in the near-infrared band around the ground observation station to verify whether the station is uniformly located within the high-resolution 16 meter pixels. The main reason for using this band is to consider that the near-infrared band is less affected by the atmosphere. Except for March 8, 2021, which has a spatial resolution of 8 meters, all other times have a spatial resolution of 4 meters. On March 8, 2021, a high-resolution 16m pixel should consist of 2 × 2 pixels. Considering possible geometric deviations, 3 × 3 pixel values will be counted; The remaining time corresponds to 4 × 4 high-resolution pixels for 16 m pixels, taking into account geometric deviations, and counting 5 × 5 high-resolution pixels.
Fig. 5The DN value distribution of the high resolution 16m pixel at the Red Haitang 3 (HH3) site during four typical surface conditions was displayed. The selected dates cover June 2, 2020 (early stage of forest growth, flowering period of crabapple), August 10, 2020 (prosperous growth period, complete vegetation coverage), November 22, 2020 (withering period, mixed forest and bare soil), and March 8, 2021 (mainly bare soil at the end of winter). The results show that under different surface conditions, the DN value distribution within the high resolution 16 m pixel where the station is located is still relatively stable and has good spatial uniformity.
figure
figure
figure
figure

Fig. 5 Distribution of DN values within the 16 m pixels of Gaofen pixels containing the HH3 site during four representative periods

The spatial representativeness is calculated using high-resolution 16m albedo data. Three typical high-resolution No.1 albedo data images from different periods are selected as representatives, and 33 × 33 pixels are expanded outward from the high-resolution pixels where each station is located, corresponding to a spatial range of 500m × 500 m on the surface. The semi variance is calculated using a semi variogram model to evaluate the spatial representativeness of ground station observations.Fig. 6The results of the semi variogram simulation at the Red Haitang 3 (HH3) site were presented,Table 3Relevant quantitative indicators for spatial representativeness evaluation of 7 sites were provided. combineFig. 6andTable 3It can be seen that the spatial heterogeneity of different stations in the study area varies and changes over time. In the early stage of vegetation growth (June 4, 2020), the base station value and base station coefficient are relatively large, and the spatial representativeness is relatively weak. In the peak season of vegetation growth (July 15, 2020), the surface is completely covered by forest land, and the base station value and base station coefficient of each station are relatively small, indicating that the spatial heterogeneity of the stations is weak and the spatial representativeness is good; In the late stage of vegetation growth (2020-11-02), leaves wither, and the forest and bare soil mix, resulting in a decrease in spatial representativeness compared to the peak season of vegetation growth. In addition, among the 7 ground observation stations, HH1, HH3, LH, and SS clearly follow the above growth patterns; For HH2, HH4, and YS, they exhibit stronger spatial heterogeneity during the peak season of vegetation growth compared to other periods. Based on the actual situation in the study area, there are two main reasons: firstly, the flowering period of the two Haitang stations is relatively late, resulting in uneven growth during the 2020-07-15 period; Secondly, the planting area of Pinus tabulaeformis is relatively small, and spatial heterogeneity is more pronounced during the peak growth season. Overall, during the vegetation growth season, the base station values at each station are relatively small, the surface spatial heterogeneity is weak, the surface is relatively uniform, and the spatial representativeness is good.
figure
figure
figure

Fig. 6 Simulation results of semivariogram at HongHaitang 3（HH3） site

Table 3 Quantitative metrics of spatial representativeness in the study area during the experimental period

站点日期块金值偏基台值基台值变程平均值基台系数
HH12020-06-047.21E‒054.41E‒051.16E‒0449.5840.167796.4228
2020-07-152.17E‒057.50E‒059.67E‒0567.5270.159236.1754
2020-11-026.31E‒062.55E‒042.61E‒04119.2600.167109.6710
HH22020-06-047.79E‒058.96E‒051.68E‒0489.8350.163237.9296
2020-07-151.79E‒052.09E‒042.27E‒0493.7210.158949.4781
2020-11-027.11E‒069.73E‒051.04E‒0493.3930.159826.3949
HH32020-06-046.50E‒051.03E‒041.68E‒04116.2500.161808.0201
2020-07-152.57E‒059.34E‒051.19E‒0484.7710.159676.8342
2020-11-025.95E‒061.36E‒041.41E‒04111.0600.163827.2612
HH42020-06-045.24E‒051.20E‒041.73E‒0473.8510.161528.1331
2020-07-152.98E‒051.78E‒042.08E‒0497.0620.161288.9376
2020-11-024.84E‒069.02E‒059.51E‒05104.1500.160266.0841
LH2020-06-045.43E‒058.25E‒051.37E‒0436.3260.162967.1788
2020-07-151.04E‒059.50E‒051.05E‒0460.6990.153226.7001
2020-11-021.53E‒059.59E‒051.11E‒0473.4460.158006.6743
SS2020-06-043.78E‒059.46E‒051.32E‒0431.6200.166176.9235
2020-07-158.52E‒067.18E‒058.03E‒0561.1170.153265.8481
2020-11-021.39E‒051.59E‒041.73E‒0490.9890.157638.3530
YS2020-06-045.85E‒057.92E‒051.38E‒0450.5370.159517.3581
2020-07-151.11E‒052.10E‒042.21E‒0480.8440.156509.4953
2020-11-021.50E‒058.46E‒059.96E‒05108.5700.152066.5641
fromTable 3It can be seen that the results of the semi variogram simulation show that the base station value and base station coefficient are mostly consistent, manifested as the larger the base station value, the larger the base station coefficient. However, at the Green Haitang (LH) site, the above pattern is not satisfied. The base station value on July 15, 2020 is smaller than that of other times, while the corresponding base station coefficient during this period is greater than that of November 2, 2020. The reason for this situation is that in the process of fitting the semi variogram function, the calculation of the base station value will be affected by the surface albedo value itself. Calculating the base station coefficient eliminates the influence of the magnitude of the surface albedo value itself, thereby ensuring comparability of observations at different sites.

6 Conclusion

The quality control of ground observation data is a key link in ensuring data reliability and scientificity. With the development of observation technology, observation methods continue to enrich, and the amount of data increases significantly, which puts higher demands on quality control. This study summarizes the current research progress by exploring the methods, techniques, and application examples of ground observation data quality control. The core of quality control lies in ensuring the standardization, completeness, timeliness, accuracy, consistency, and representativeness of data. Traditional quality control methods mainly rely on manual inspection and basic statistical analysis. With the development of quality control technology, it is possible to quickly identify abnormal and inconsistent data points in massive amounts of data, thereby reducing the need for manual intervention.
(1) The quality control of ground observation data for authenticity verification is a deepening and expansion of traditional ground observation data quality control. In addition to routine checks for standardization and consistency, this is mainly reflected in the spatiotemporal representativeness of the data and the uncertainty introduced in the process of calculating from the site scale to the pixel scale. This process not only involves processing the observed data itself, but also comprehensively considers various factors such as the spatial distribution of the site, the distribution characteristics of variables in specific areas, and the time span of the data, to ensure the comprehensiveness and scientificity of data quality control, and to provide a reliable pixel scale reference truth value for authenticity testing.
(2) The standard specifications for quality control of ground observation data need to be further strengthened. Although a quality control technology system for ground observation data aimed at authenticity verification has been established, there are differences in the quality requirements for ground observation data among different industries. To define the quality of ground observation data that meets application needs, data quality standards and control methods need to be developed according to the requirements. At the same time, promote the adoption of unified data format specifications and metadata standards among various observation networks to enhance data interoperability and integrability.
(3) Promoting the integration and implementation of the quality control methodology system on the authenticity verification system platform is a key direction for achieving large-scale and efficient data management and application. On the basis of further improving the data quality control standard system, with the help of existing data management systems and high-performance computing frameworks, a quality control module with configurable and adaptive capabilities is developed according to different surface parameters and observation conditions to achieve automatic verification and analysis of ground observation data, solve the observation requirements of ground parameters from point to panel for authenticity verification, reduce manual intervention costs, and improve the timeliness and reliability of data processing.
(4) The quality control index system for authenticity verification constructed in this study focuses on identifying and eliminating observation data that does not meet the requirements of authenticity verification to ensure the reliability and representativeness of the validation data. It should be pointed out that this study focuses on quality control rather than quality evaluation, and has not yet involved the fusion and weighted scoring of various indicator results. Future work will further develop data quality evaluation methods based on this system, establish a unified weighted scoring mechanism, and achieve quantitative evaluation and grading of the overall quality of ground data for remote sensing authenticity testing.

References

  1. 1.
    Anttila S, Ketola M, Vakkilainen K and Kairesalo T. 2012. Assessing temporal representativeness of water quality monitoring data. Journal of Environmental Monitoring, 14(2): 589-595
  2. 2.
    Assfalg J, Bernecker T, Kriegel H P, Kröger P and Renz M. 2009. Periodic pattern analysis in time series databases//14th International Conference on Database Systems for Advanced Applications. Brisbane: Springer: 354-368
  3. 3.
    Baca-López K, Fresno C, Espinal-Enríquez J, Martínez-García M, Camacho-López M A, Flores-Merino M V and Hernández-Lemus E. 2021. Spatio-temporal representativeness of air quality monitoring stations in Mexico City: implications for public health. Frontiers in Public Health, 8: 536174
  4. 4.
    Balsamo G, Agusti-Panareda A, Albergel C, Arduini G, Beljaars A, Bidlot J, Blyth E, Bousserez N, Boussetta S, Brown A, Buizza R, Buontempo C, Chevallier F, Choulga M, Cloke H, Cronin M F, Dahoui M, De Rosnay P, Dirmeyer P A, Drusch M, Dutra E, Ek M B, Gentine P, Hewitt H, Keeley S P E, Kerr Y, Kumar S, Lupu C, Mahfouf J F, McNorton J, Mecklenburg S, Mogensen K, Muñoz-Sabater J, Orth R, Rabier F, Reichle R, Ruston B, Pappenberger F, Sandu I, Seneviratne S I, Tietsche S, Trigo I F, Uijlenhoet R, Wedi N, Woolway R I and Zeng X B. 2018. Satellite and in situ observations for advancing global earth surface modelling: a review. Remote Sensing, 10(12): 2038
  5. 5.
    Bandaru L, Irigireddy B C, Pvnr K and Davis B. 2024. DeepQC: a deep learning system for automatic quality control of in-situ soil moisture sensor time series data. Smart Agricultural Technology, 8: 100514
  6. 6.
    Båserud L, Lussana C, Nipen T N, Seierstad I A, Oram L and Aspelien T. 2020. TITAN automatic spatial quality control of meteorological in-situ observations. Advances in Science and Research, 17: 153-163
  7. 7.
    Bian Z Q, Chong W, Ding L and Yang W. 2019. Analysis and research on quality control method of global radiation observation data. The Journal of Engineering, 2019(23): 8975-8979
  8. 8.
    Cerlini P B, Silvestri L and Saraceni M. 2020. Quality control and gap-filling methods applied to hourly temperature observations over central Italy. Meteorological Applications, 27(3): e1913
  9. 9.
    Chang G B, Zhang S B and Liu Z P. 2024. Understanding adjustment from an information theoretic perspective. Geomatics and Information Science of Wuhan University, 49(2): 313-323
  10. 10.
    Chang K L, Schultz M G, Lan X, McClure-Begley A, Petropavlovskikh I, Xu X B and Ziemke J R. 2021. Trend detection of atmospheric time series: incorporating appropriate uncertainty estimates and handling extreme events. Elementa: Science of the Anthropocene, 9(1): 00035
  11. 11.
    Chave J, Davies S J, Phillips O L, Lewis S L, Sist P, Schepaschenko D, Armston J, Baker T R, Coomes D, Disney M, Duncanson L, Hérault B, Labrière N, Meyer V, Réjou-Méchain M, Scipal K and Saatchi S. 2019. Ground data are essential for biomass remote sensing missions. Surveys in Geophysics, 40(4): 863-880
  12. 12.
    Chu H S, Baldocchi D D, John R, Wolf S and Reichstein M. 2017. Fluxes all of the time? A primer on the temporal representativeness of FLUXNET. Journal of Geophysical Research: Biogeosciences, 122(2): 289-307
  13. 13.
    Cochran W G. 1977. Sampling Techniques. 3rd ed. New York: John Wiley and Sons
  14. 14.
    Cressie N A C. 2015. Statistics for Spatial Data. Hoboken: John Wiley and Sons
  15. 15.
    Curran P J. 1988. The semivariogram in remote sensing: an introduction. Remote Sensing of Environment, 24(3): 493-507
  16. 16.
    Dandrifosse S, Jago A, Huart J P, Michaud V, Planchon V and Rosillon D. 2024. Automatic quality control of weather data for timely decisions in agriculture. Smart Agricultural Technology, 8: 100445
  17. 17.
    Dickinson R E. 1995. Land processes in climate models. Remote Sensing of Environment, 51(1): 27-38
  18. 18.
    Doldirina C. 2015. Open data and earth observations: the case of opening up access to and use of earth observation data through the global earth observation system of systems. Journal of Intellectual Property, Information Technology and Electronic Commerce Law, 6(1): 73-85
  19. 19.
    Dorigo W, Himmelbauer I, Aberer D, Schremmer L, Petrakovic I, Zappa L, Preimesberger W, Xaver A, Annor F, Ardö J, Baldocchi D, Bitelli M, Blöschl G, Bogena H, Brocca L, Calvet J C, Camarero J J, Capello G, Choi M, Cosh M C, van de Giesen N, Hajdu I, Ikonen J, Jensen K H, Kanniah K D, de Kat I, Kirchengast G, Kumar Rai P, Kyrouac J, Larson K, Liu S X, Loew A, Moghaddam M, Martínez Fernández J, Mattar Bader C, Morbidelli R, Musial J P, Osenga E, Palecki M A, Pellarin T, Petropoulos G P, Pfeil I, Powers J, Robock A, Rüdiger C, Rummel U, Strobel M, Su Z B, Sullivan R, Tagesson T, Varlagin A, Vreugdenhil M, Walker J, Wen J, Wenger F, Wigneron J P, Woods M, Yang K, Zeng Y J, Zhang X, Zreda M, Dietrich S, Gruber A, van Oevelen P, Wagner W, Scipal K, Drusch M and Sabia R. 2021. The International Soil Moisture Network: serving Earth system science for over a decade. Hydrology and Earth System Sciences, 25(11): 5749-5804
  20. 20.
    Gong Y D, Liu G Z, Xue Y Z, Li R and Meng L Z. 2023. A survey on dataset quality in machine learning. Information and Software Technology, 162: 107268
  21. 21.
    Gruijter J J, Bierkens M F P, Brus D J and Knotters M. 2006. Sampling for Natural Resource Monitoring. Berlin, Heidelberg: Springer
  22. 22.
    Guidoum A. 2025. Statistical modeling and mapping of rainfall in the endorheic basins of Northern Algeria: a comparison of spatial interpolation methods. Acta Geophysica, 73(2): 1679-1699
  23. 23.
    Guo J W and Liu F. 2015. Automatic data quality control of observations in wireless sensor network. IEEE Geoscience and Remote Sensing Letters, 12(4): 716-720
  24. 24.
    Han H T, Wang Y P and Zhang J. 2021. Research progress on the effect of station relocation on climate data homogeneity in Northwest China. Plateau Meteorology, 40(2): 448-454
  25. 25.
    Han Y, Wen J G, Xiao Q, You D Q, Meng L, Wu S B, Hao D L, Tang Y, Chen X, Liu Q H and Zhao C C. 2024. Asymmetry in the diurnal variation of land surface albedo and its impacts on daily mean albedo estimation. Journal of Geophysical Research: Atmospheres, 129(14): e2023JD039728
  26. 26.
    He H L, Zhang L Y, Gao Y Z, Ren X L, Zhang L, Yu G R and Wang S Q. 2015. Regional representativeness assessment and improvement of eddy flux observations in China. Science of the Total Environment, 502: 688-698
  27. 27.
    Jin R, Li X, Ma M G, Ge Y, Liu S M, Xiao Q, Wen J G, Zhao K, Xin X P, Ran Y H, Liu Q H and Zhang R H. 2017. Key methods and experiment verification for the validation of quantitative remote sensing products. Advances in Earth Science, 32(6): 630-642
  28. 28.
    Kalkan M, Bostancı G E, Güzel M S, Kalkan B, Özsarı Ş, Soysal Ö and Köse G. 2022. Cloudy/clear weather classification using deep learning techniques with cloud images. Computers and Electrical Engineering, 102: 108271
  29. 29.
    Kazantzidis A, Tzoumanikas P, Bais A F, Fotopoulos S and Economou G. 2012. Cloud detection and classification with the use of whole-sky ground-based images. Atmospheric Research, 113: 80-88
  30. 30.
    Kim H J, Park S M, Choi B J, Moon S H and Kim Y H. 2020. Spatiotemporal approaches for quality control and error correction of atmospheric data through machine learning. Computational Intelligence and Neuroscience, 2020: 7980434
  31. 31.
    Li M H and Li J Z. 2016. Data currency determination: key theories and technologies. Intelligent Computer and Applications, 6(6): 72-75
  32. 32.
    Li Q Q, Wen J G, Wu X D, Xiao Q, You D Q and Tang R Q. 2025. The optimal deployment of ground samples: whether spatial heterogeneity is dominated by randomness or structure factors? IEEE Geoscience and Remote Sensing Letters, 22: 2500505
  33. 33.
    Li Q Y, Lu W T and Yang J. 2011. A hybrid thresholding algorithm for cloud detection on ground-based color images. Journal of Atmospheric and Oceanic Technology, 28(10): 1286-1296
  34. 34.
    Li X. 2014. Characterization, controlling, and reduction of uncertainties in the modeling and observation of land-surface systems. Science China Earth Sciences, 57(1): 80-87
  35. 35.
    Li X, Liu S M, Liu Q H, Xiao Q, Che T, Ma M G, Jin R, Ran Y H, Wen J G, Xu Z W and Li Z Y. 2023. Heihe remote sensing experiments: retrospect and prospect. National Remote Sensing Bulletin, 27(2): 224-248
  36. 36.
    Lin X W. 2018. Land Surface Albedo Remote Sensing Products Validation over Rugged Terrain. Beijing: University of Chinese Academy of Sciences (Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences)
  37. 37.
    Liu F and Guo J W. 2013. Study on quality control approach for Heihe wireless sensor network observation data. Remote Sensing Technology and Application, 28(2): 252-257
  38. 38.
    Liu Q H, Wen J G, Zhou X, Zhao J, Li Z Y, Li X, Ma M G, Wang W Z, Liao X H, Liu S M, Fan W J, Xiao Q, Zhong B, Li J, Xin X Z, Li L, Jia L, Gao Z H, Jin J D, Liang S, Xin J, Liao C J and Wu Y R. 2023. Technique system of remote sensing product generation and validation of GF common products. National Remote Sensing Bulletin, 27(3): 544-562
  39. 39.
    Liu X N and Ren Z H. 2005. Progress in quality control of surface meteorological data. Meteorological Science and Technology, 33(3): 199-203
  40. 40.
    Lu B B, Ge Y, Qin K and Dong G P. 2025. Theoretical and technical frameworks for geographically weighed models. National Remote Sensing Bulletin, 29(3): 596-609
  41. 41.
    Lu G Y and Wong D W. 2008. An adaptive inverse-distance weighting spatial interpolation technique. Computers and Geosciences, 34(9): 1044-1055
  42. 42.
    Ma J. 2021. Retrieval and Validation of Long-Term Land Surface Temperature from Satellite Remote Sensing. Chengdu: University of Electronic Science and Technology of China (马晋. 2021. 长时序地表温度遥感反演与真实性检验方法研究. 成都: 电子科技大学) [DOI: 10.27005/d.cnki.gdzku.2021.005062]
  43. 43.
    Ma M G. 2010. Quality control and evaluation for the ground observation data of the watershed allied telemetry experimental research. Remote Sensing Technology and Application, 25(6): 766-771
  44. 44.
    Miatselskaya N S, Bril A I, Chaikovsky A P, Yukhymchuk Y Y, Milinevski G P and Simon A A. 2022. Optimal interpolation of aeronet radiometric network observations for the evaluation of the aerosol optical thickness distribution in the Eastern European region. Journal of Applied Spectroscopy, 89(2): 296-302
  45. 45.
    Nappo C J, Caneill J Y, Furman R W, Gifford F A, Kaimal J C, Kramer M L, Lockhart T J, Pendergast M M, Pielke R A, Randerson D, Shreffler J H and Wyngaard J C. 1982. The workshop on the representativeness of meteorological observations, June 1981, Boulder, Colo. Bulletin of the American Meteorological Society, 63(7): 761-764
  46. 46.
    Pandžić K, Kobold M, Oskoruš D, Biondić B, Biondić R, Bonacci O, Likso T and Curić O. 2020. Standard normal homogeneity test as a tool to detect change points in climate-related river discharge variation: case study of the Kupa River Basin. Hydrological Sciences Journal, 65(2): 227-241
  47. 47.
    Papachristopoulou K, Fountoulakis I, Bais A F, Psiloglou B E, Papadimitriou N, Raptis I P, Kazantzidis A, Kontoes C, Hatzaki M and Kazadzis S. 2024. Effects of clouds and aerosols on downwelling surface solar irradiance nowcasting and short-term forecasting. Atmospheric Measurement Techniques, 17(7): 1851-1877
  48. 48.
    Parsons M A, Brodzik M J and Rutter N J. 2004. Data management for the Cold Land Processes Experiment: improving hydrological science. Hydrological Processes, 18(18): 3637-3653
  49. 49.
    Pastorello G, Agarwal D, Papale D, Samak T, Trotta C, Ribeca A, Poindexter C, Faybishenko B, Gunter D, Hollowgrass R and Canfora E. 2014. Observational data patterns for time series data quality assessment//2014 IEEE 10th International Conference on e-Science. Sao Paulo: IEEE: 271-278
  50. 50.
    Pastorello G, Gunter D, Chu H S, Christianson D, Trotta C, Canfora E, Faybishenko B, Cheah Y W, Beekwilder N, Chan S, Dengel S, Keenan T, O’Brien F, Elbashandy A, Poindexter C, Humphrey M, Papale D and Agarwal D. 2017. Hunting data rogues at scale: data quality control for observational data in research infrastructures//2017 IEEE 13th International Conference on e-Science (e-Science). Auckland: IEEE: 446-447
  51. 51.
    Pastorello G, Trotta C, Canfora E, Chu H S, Christianson D, Cheah Y W, Poindexter C, Chen J Q, Elbashandy A, Humphrey M, Isaac P, Polidori D, Reichstein M, Ribeca A, van Ingen C, Vuichard N, Zhang L M, Amiro B, Ammann C, Arain M A, Ardö J, Arkebauer T, Arndt S K, Arriga N, Aubinet M, Aurela M, Baldocchi D, Barr A, Beamesderfer E, Marchesini L B, Bergeron O, Beringer J, Bernhofer C, Berveiller D, Billesbach D, Black T A, Blanken P D, Bohrer G, Boike J, Bolstad P V, Bonal D, Bonnefond J M, Bowling D R, Bracho R, Brodeur J, Brümmer C, Buchmann N, Burban B, Burns S P, Buysse P, Cale P, Cavagna M, Cellier P, Chen S P, Chini I, Christensen T R, Cleverly J, Collalti A, Consalvo C, Cook B D, Cook D, Coursolle C, Cremonese E, Curtis P S, D’Andrea E, da Rocha H, Dai X Q, Davis K J, De Cinti B, de Grandcourt A, De Ligne A, De Oliveira R C, Delpierre N, Desai A R, Di Bella C M, di Tommasi P, Dolman H, Domingo F, Dong G, Dore S, Duce P, Dufrêne E, Dunn A, Dušek J, Eamus D, Eichelmann U, ElKhidir H A M, Eugster W, Ewenz C M, Ewers B, Famulari D, Fares S, Feigenwinter I, Feitz A, Fensholt R, Filippa G, Fischer M, Frank J, Galvagno M, Gharun M, Gianelle D, Gielen B, Gioli B, Gitelson A, Goded I, Goeckede M, Goldstein A H, Gough C M, Goulden M L, Graf A, Griebel A, Gruening C, Grünwald T, Hammerle A, Han S J, Han X G, Hansen B U, Hanson C, Hatakka J, He Y T, Hehn M, Heinesch B, Hinko-Najera N, Hörtnagl L, Hutley L, Ibrom A, Ikawa H, Jackowicz-Korczynski M, Janouš D, Jans W, Jassal R, Jiang S C, Kato T, Khomik M, Klatt J, Knohl A, Knox S, Kobayashi H, Koerber G, Kolle O, Kosugi Y, Kotani A, Kowalski A, Kruijt B, Kurbatova J, Kutsch W L, Kwon H, Launiainen S, Laurila T, Law B, Leuning R, Li Y N, Liddell M, Limousin J M, Lion M, Liska A J, Lohila A, López-Ballesteros A, López-Blanco E, Loubet B, Loustau D, Lucas-Moffat A, Lüers J, Ma S Y, Macfarlane C, Magliulo V, Maier R, Mammarella I, Manca G, Marcolla B, Margolis H A, Marras S, Massman W, Mastepanov M, Matamala R, Matthes J H, Mazzenga F, McCaughey H, McHugh I, McMillan A M S, Merbold L, Meyer W, Meyers T, Miller S D, Minerbi S, Moderow U, Monson R K, Montagnani L, Moore C E, Moors E, Moreaux V, Moureaux C, Munger J W, Nakai T, Neirynck J, Nesic Z, Nicolini G, Noormets A, Northwood M, Nosetto M, Nouvellon Y, Novick K, Oechel W, Olesen J E, Ourcival J M, Papuga S A, Parmentier F J, Paul-Limoges E, Pavelka M, Peichl M, Pendall E, Phillips R P, Pilegaard K, Pirk N, Posse G, Powell T, Prasse H, Prober S M, Rambal S, Rannik Ü, Raz-Yaseef N, Rebmann C, Reed D, de Dios V R, Restrepo-Coupe N, Reverter B R, Roland M, Sabbatini S, Sachs T, Saleska S R, Sánchez-Cañete E P, Sanchez-Mejia Z M, Schmid H P, Schmidt M, Schneider K, Schrader F, Schroder I, Scott R L, Sedlák P, Serrano-Ortíz P, Shao C L, Shi P L, Shironya I, Siebicke L, Šigut L, Silberstein R, Sirca C, Spano D, Steinbrecher R, Stevens R M, Sturtevant C, Suyker A, Tagesson T, Takanashi S, Tang Y H, Tapper N, Thom J, Tomassucci M, Tuovinen J P, Urbanski S, Valentini R, van der Molen M, van Gorsel E, van Huissteden K, Varlagin A, Verfaillie J, Vesala T, Vincke C, Vitale D, Vygodskaya N, Walker J P, Walter-Shea E, Wang H M, Weber R, Westermann S, Wille C, Wofsy S, Wohlfahrt G, Wolf S, Woodgate W, Li Y L, Zampedri R, Zhang J H, Zhou G Y, Zona D, Agarwal D, Biraud S, Torn M and Papale D. 2020. The FLUXNET2015 dataset and the ONEFlux processing pipeline for eddy covariance data. Scientific Data, 7(1): 225
  52. 52.
    Peterson T C and Vose R S. 1997. An overview of the Global Historical Climatology Network temperature database. Bulletin of the American Meteorological Society, 78(12): 2837-2850
  53. 53.
    Ren Z H, Xu S, Sun H N and Zhang Q. 2006. Quality check and analysis on the historic global surface synoptic observations. Journal of Applied Meteorological Science, 17(4): 412-420
  54. 54.
    Ren Z H, Zhang Z F, Sun C, Liu Y M, Li J, Ju X H, Zhao Y F, Li Z P, Zhang W, Li H K, Zeng X J, Ren X W, Liu Y and Wang H J. 2015. Development of three-step quality control system of real-time observation data from AWS in China. Meteorological Monthly, 41(10): 1268-1277
  55. 55.
    Román M O, Schaaf C B, Woodcock C E, Strahler A H, Yang X Y, Braswell R H, Curtis P S, Davis K J, Dragoni D, Goulden M L, Gu L H, Hollinger D Y, Kolb T E, Meyers T P, Munger J W, Privette J L, Richardson A D, Wilson T B and Wofsy S C. 2009. The MODIS (Collection V005) BRDF/albedo product: assessment of spatial representativeness over forested landscapes. Remote Sensing of Environment, 113(11): 2476-2498
  56. 56.
    Schaaf C B, Gao F, Strahler A H, Lucht W, Li X W, Tsang T, Strugnell N C, Zhang X Y, Jin Y F, Muller J P, Lewis P, Barnsley M, Hobson P, Disney M, Roberts G, Dunderdale M, Doll C, d’Entremont R P, Hu B X, Liang S L, Privette J L and Roy D. 2002. First operational BRDF, albedo nadir reflectance products from MODIS. Remote Sensing of Environment, 83(1/2): 135-148
  57. 57.
    Schimel D, Pavlick R, Fisher J B, Asner G P, Saatchi S, Townsend P, Miller C, Frankenberg C, Hibbard K and Cox P. 2015. Observing terrestrial ecosystems and the carbon cycle from space. Global Change Biology, 21(5): 1762-1776
  58. 58.
    Sha Y K, Gagne II D J, West G and Stull R. 2021. Deep-learning-based precipitation observation quality control. Journal of Atmospheric and Oceanic Technology, 38(5): 1075-1091
  59. 59.
    Sharp R A, Mudford O C and Elliffe D. 2015. A data-based method for selecting parameters of momentary time sampling to provide representative data. European Journal of Behavior Analysis, 16(2): 279-294
  60. 60.
    State Administration for Market Regulation and Standardization Administration of the People’s Republic of China. 2018. GB/T 36344-2018 Information technology—Evaluation indicators for data quality. Beijing: Standards Press of China
  61. 61.
    Steinacker R. 2023. Mean value splines and their use for climatological time series. International Journal of Climatology, 43(9): 4326-4336
  62. 62.
    Su L, Gao C C, Ren X L, Zhang F Y, Cao S S, Zhang S Q, Chen T D, Liu M Q, Ni B C and Liu M. 2022. Understanding the spatial representativeness of air quality monitoring network and its application to PM2.5 in the mainland China. Geoscience Frontiers, 13(3): 101370
  63. 63.
    Toreti A, Bratu M, Müller E, van der Grijn G, Oliveira I, Zampieri M and Ceglar A. 2022. Advancing near-real-time quality controls of meteorological observations. Bulletin of the American Meteorological Society, 103(4): E1078-E1087
  64. 64.
    Toreti A, Kuglitsch F G, Xoplaki E, Della-Marta P M, Aguilar E, Prohom M and Luterbacher J. 2011. A note on the use of the standard normal homogeneity test to detect inhomogeneities in climatic time series. International Journal of Climatology, 31(4): 630-632
  65. 65.
    Wan G N, Yang M X, Liu Z C, Wang X J and Liang X W. 2017. The precipitation variations in the Qinghai-Xizang (Tibetan) Plateau during 1961-2015. Atmosphere, 8(5): 80
  66. 66.
    Wang C, Caja J and Gómez E. 2018. Comparison of methods for outlier identification in surface characterization. Measurement, 117: 312-325
  67. 67.
    Wang J F, Stein A, Gao B B and Ge Y. 2012. A review of spatial sampling. Spatial Statistics, 2: 1-14
  68. 68.
    Wang Q. 2021. Progress of environmental remote sensing monitoring technology in China and some related frontier issues. National Remote Sensing Bulletin, 25(1): 25-36
  69. 69.
    Wang S K, Zhang Z Y, Wang P G and Tian Y R. 2021. Failure warning of gearbox for wind turbine based on 3σ-median criterion and NSET. Energy Reports, 7: 1182-1197
  70. 70.
    Wang X L. 2008. Penalized maximal F test for detecting undocumented mean shift without trend change. Journal of Atmospheric and Oceanic Technology, 25(3): 368-384
  71. 71.
    Wang X L, Wen Q H and Wu Y H. 2007. Penalized maximal t test for detecting undocumented mean change in climate data series. Journal of Applied Meteorology and Climatology, 46(6): 916-931
  72. 72.
    Wang Z S, Schaaf C B, Strahler A H, Chopping M J, Román M O, Shuai Y M, Woodcock C E, Hollinger D Y and Fitzjarrald D R. 2014. Evaluation of MODIS albedo product (MCD43A) over grassland, agriculture and forest surface types during dormant and snow-covered periods. Remote Sensing of Environment, 140: 60-77
  73. 73.
    Wen J G, Liu Q H, Li Z Y, Li X, Liu S M, Xiao Q, Gao Z H, Ma M G, Che T, Liu L Y, Fang H L, Yan G J, Ge Y, Chen E X, Zhang Y, Ma L L, Wu X D and Chen X. 2023a. A review of the development of remote sensing field experiments and product validation in China. National Remote Sensing Bulletin, 27(3): 573-583
  74. 74.
    Wen J G, Wu X D, Wang J P, Tang R Q, Ma D J, Zeng Q C, Gong B C and Xiao Q. 2022. Characterizing the effect of spatial heterogeneity and the deployment of sampled plots on the uncertainty of ground “truth” on a coarse grid scale: case study for Near-Infrared (NIR) surface reflectance. Journal of Geophysical Research: Atmospheres, 127(11): e2022JD036779
  75. 75.
    Wen J G, Xiao Q, Zhong S Y, Tang Y, Chen X, Wei Q F, Wu X D, Lin X W, Ouyang X Y, You D Q and Liu Q H. 2023b. Technology system for product validation and algorithm test of GF common products and an application example. National Remote Sensing Bulletin, 27(3): 780-789
  76. 76.
    Wu X D, Wen J G, Xiao Q, Lin X W, You D Q, Yin G F and Liu Q H. 2025. Theory and method of ground observation on satellite pixel scales: taking the numerical land surface parameters as an example. National Remote Sensing Bulletin, 29(5): 1227-1244
  77. 77.
    Wu X D, Wen J G, Xiao Q, Liu Q, Peng J J, Dou B C, Li X H, You D Q, Tang Y and Liu Q H. 2016. Coarse scale in situ albedo observations over heterogeneous snow-free land surfaces and validation strategy: a case of MODIS albedo products preliminary validation over northern China. Remote Sensing of Environment, 184: 25-39
  78. 78.
    Wu X D, Wen J G, Xiao Q, You D Q, Gong B C, Wang J P, Ma D J and Lin X W. 2022. Spatial heterogeneity of albedo at subpixel satellite scales and its effect in validation: airborne remote sensing results from HiWATER. IEEE Transactions on Geoscience and Remote Sensing, 60: 4407114
  79. 79.
    Wu X D, Wen J G, Xiao Q, You D Q, Lin X W, Wu S B and Zhong S Y. 2019. Impacts and contributors of representativeness errors of in situ albedo measurements for the validation of remote sensing products. IEEE Transactions on Geoscience and Remote Sensing, 57(12): 9740-9755
  80. 80.
    Wu X D, Wen J G, Xiao Q, You D Q, Wang J P, Ma D J and Lin X W. 2021. A multiscale nested sampling method for representative albedo observations at various pixel scales. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 14: 8193-8207
  81. 81.
    Wu X D, Xiao Q, Wen J G and You D Q. 2019. Advances and challenges in the validation of remote sensing albedo products. Journal of Remote Sensing (in Chinese), 23(1): 11-23
  82. 82.
    Xiao J J, Li Z Q, Guo F F, Yao Y P, Ma H, Wang K and Wen Q P. 2018. Construction and analysis of annual precipitation series from 1901 to 2017 in Zhejiang province. Climate Change Research, 14(6): 553-561
  83. 83.
    Xu B D. 2018. Leaf Area Index Retrieval and Product Validation over Heterogeneous Land Surfaces. Beijing: University of Chinese Academy of Sciences (Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences)
  84. 84.
    Xu B D, Li J, Liu Q H, Huete A R, Yu Q, Zeng Y L, Yin G F, Zhao J and Yang L. 2016. Evaluating spatial representativeness of station observations for remotely sensed leaf area index products. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 9(7): 3267-3282
  85. 85.
    Xu B D, Li J, Liu Q H, Xin X Z, Zeng Y L and Yin G F. 2015. Review of methods for evaluating representativeness of ground station observations. Journal of Remote Sensing (in Chinese), 19(5): 703-718
  86. 86.
    Xu C, Wang W, Hu Y J and Liu Y W. 2024. Evaluation of ERA5, ERA5-Land, GLDAS-2.1, and GLEAM potential evapotranspiration data over mainland China. Journal of Hydrology: Regional Studies, 51: 101651
  87. 87.
    Yang H Y, Zhao X Y and Wang L. 2023. Review of data normalization methods. Computer Engineering and Applications, 59(3): 13-22
  88. 88.
    You D Q, Wen J G, Tang Y, Liu Q, Zhong S Y, Han Y, Gong B C, Zhong B, Wu S L and Liu Q H. 2023. The GF land surface albedo product based on the high-spatial-and-temporal-resolution BRDF priori-knowledge and its preliminary validation. National Remote Sensing Bulletin, 27(3): 738-747
  89. 89.
    Zeng Y W and Gong J Y. 2004. Implementing technique of spatial data quality control and evaluation. Geomatics and Information Science of Wuhan University, 29(8): 686-690
  90. 90.
    Zhang R H, Tian J, Li Z L, Su H B, Chen S H and Tang X Z. 2010. Principles and methods for the validation of quantitative remote sensing products. Science China Earth Sciences, 53(5): 741-751
  91. 91.
    Zhang Y, Jia Z Z, Liu S M, Xu Z W, Xu T R, Yao Y J, Ma Y F, Song L S, Li X, Hu X, Wang Z Y, Guo Z X and Zhou J. 2020. Advances in validation of remotely sensed land surface evapotranspiration. Journal of Remote Sensing (in Chinese), 24(8): 975-999
  92. 92.
    Zhao H. 2015. Research on Quality Control Methods for Ground Observation Data and Its Application in Numerical Models. Nanjing: Nanjing University of Information Science and Technology

Lesen Sie die ganze Passage

The above content is generated by Large Model Translation. The translated content is for reference only. We do not assume any commercial or legal responsibilty for any consequences arising from the use of our website