Crowdsourced Meteorological Data to Supplement Limited Official Sources: A Survey and Case Study of Precipitation Monitoring in Guangzhou, China

Yu Yu aNational Meteorological Information Centre, China Meteorological Administration, Beijing, China

Search for other papers by Yu Yu in
Current site
Google Scholar
PubMed
Close
https://orcid.org/0000-0002-2455-840X
,
Lei Cao aNational Meteorological Information Centre, China Meteorological Administration, Beijing, China

Search for other papers by Lei Cao in
Current site
Google Scholar
PubMed
Close
,
Zhihua Ren aNational Meteorological Information Centre, China Meteorological Administration, Beijing, China

Search for other papers by Zhihua Ren in
Current site
Google Scholar
PubMed
Close
,
Yan Xu aNational Meteorological Information Centre, China Meteorological Administration, Beijing, China

Search for other papers by Yan Xu in
Current site
Google Scholar
PubMed
Close
,
Wei Feng aNational Meteorological Information Centre, China Meteorological Administration, Beijing, China

Search for other papers by Wei Feng in
Current site
Google Scholar
PubMed
Close
, and
Licheng Zhao aNational Meteorological Information Centre, China Meteorological Administration, Beijing, China

Search for other papers by Licheng Zhao in
Current site
Google Scholar
PubMed
Close
Open access

Abstract

Crowdsourced meteorological data may provide a useful supplement to operational observations. However, the willingness of various parties to share their data remains unclear. Here, a survey on data applications was carried out to investigate the willingness to participate in crowdsourcing observations. Of the 21 responses, 71% expressed difficulty in meeting the requirement of data services using only their own observations and revealed that they would be willing to exchange data with other parties under some framework; moreover, 90% expressed a willingness to participate in crowdsourcing observations. The findings suggest that in a way the social foundation of crowdsourcing has been established in China. Additionally, a case study on precipitation monitoring was performed in Guangzhou, the capital city of Guangdong Province, South China. Three sources of hourly measurements were combined after data quality control and calibration and interpolated over Guangzhou (gridded precipitation was based on combined data, and it is referred to as the COM grid). Subsequently, the COM grid was compared with the grid data based only on observations from the China Meteorological Administration using three indices, namely, cumulative precipitation, precipitation intensity, and heavy rain hours. The results indicate that requirement for more observations could benefit from crowdsourced data, especially on uneven terrain and in regions covered by sparse surface stations.

© 2024 American Meteorological Society. This published article is licensed under the terms of the default AMS reuse license. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Publisher’s Note: This article was revised on 14 March 2024 to designate it as open access.

Corresponding author: Lei Cao, caolei@cma.gov.cn

Abstract

Crowdsourced meteorological data may provide a useful supplement to operational observations. However, the willingness of various parties to share their data remains unclear. Here, a survey on data applications was carried out to investigate the willingness to participate in crowdsourcing observations. Of the 21 responses, 71% expressed difficulty in meeting the requirement of data services using only their own observations and revealed that they would be willing to exchange data with other parties under some framework; moreover, 90% expressed a willingness to participate in crowdsourcing observations. The findings suggest that in a way the social foundation of crowdsourcing has been established in China. Additionally, a case study on precipitation monitoring was performed in Guangzhou, the capital city of Guangdong Province, South China. Three sources of hourly measurements were combined after data quality control and calibration and interpolated over Guangzhou (gridded precipitation was based on combined data, and it is referred to as the COM grid). Subsequently, the COM grid was compared with the grid data based only on observations from the China Meteorological Administration using three indices, namely, cumulative precipitation, precipitation intensity, and heavy rain hours. The results indicate that requirement for more observations could benefit from crowdsourced data, especially on uneven terrain and in regions covered by sparse surface stations.

© 2024 American Meteorological Society. This published article is licensed under the terms of the default AMS reuse license. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Publisher’s Note: This article was revised on 14 March 2024 to designate it as open access.

Corresponding author: Lei Cao, caolei@cma.gov.cn

1. Introduction

Timely and fine spatial and temporal weather monitoring has been in high demand in recent years based on the rapid development of meteorological services, such as weather analysis and forecasting or disaster mitigation (Mahoney et al. 2010). Crowdsourcing, which refers to the idea of outsourcing data collection to the broader community, is now an established technique for collecting mass data in various scientific disciplines (Chapman et al. 2017). The increased demand for real-time, high-spatiotemporal-resolution meteorological observations indicates that crowdsourcing has been introduced to meteorological applications at the right time. For example, a mature mechanism has been established in Austria’s National Weather Service to collect information from trained voluntary storm spotters and make it available after thorough automatic quality management to forecasters who are responsible for severe weather warnings (Krennert et al. 2018a). Ten other national meteorological and hydrological services across Europe have established official collaborations with voluntary spotter groups (Krennert et al. 2018b). However, Muller et al. (2015) suggested that the approaches for data gathering have expanded from a large number of people to a range of public sensors, which are typically connected via the internet. With the advances and proliferation of the “Internet of Things,” citizen (or personal) weather stations (CWSs), cheap sensors, and smart devices have become the primary form of crowdsourcing devices, which actively perceive and extract data and have played significant roles in research on urban effects, meteorological disaster monitoring, and forecast verification, where their potential has been demonstrated (Masson et al. 2020; Zhu et al. 2020; Hintz et al. 2021; Giazzi et al. 2022).

For instance, Netatmo weather stations, which can be easily configured and operated through “smartphones” or tablets, were chosen to observe local environmental elements such as air temperature, air pressure, precipitation, and wind (e.g., Fenner et al. 2017; de Vos et al. 2017; Droste et al. 2020). The urban heat island (UHI) effect, which is one of the greatest environmental problems associated with urban growth examined in urban climate research, was studied in several megacities, such as Berlin, Germany; London, United Kingdom; and Vienna, Austria (Chapman et al. 2017; Fenner et al. 2017; Feichtinger et al. 2020; Brousse et al. 2022); based on the CWSs’ data.

As data quality is a major concern when using crowdsourced data (Chapman et al. 2017), several studies have focused on data quality control (QC) and quality assessment (QA). After applying a series of QC and bias correction protocols to air temperature observations (e.g., Meier et al. 2017; Napoly et al. 2018), suspicious data were identified, and uncertain data were filtered out. Other original studies on QC have been applied to precipitation and wind speed observations provided by CWSs. For example, de Vos et al. (2017, 2019) developed a real-time QC algorithm that could detect typical errors in precipitation measurements using spatial consistency checks. Chen et al. (2021b) performed a four-stage QC scheme on the wind speed data from CWSs and found that the crowdsourced data were comparable with official station data after QC. However, owing to the inherent properties of CWS observations, a key challenge remains in data QA and application; thus, QC is an ongoing priority for further research (Chapman et al. 2017; Meier et al. 2017).

To observe spatially fine-scale interactions and complex links between cities and their atmosphere, special observation networks are required (Muller et al. 2013; Meier et al. 2015; Caluwaerts et al. 2020). One application for air temperature observations with fine temporal (hourly) and spatial scales, which is also applicable for UHIs, was based on smartphones, and it leveraged the high density of smartphones in urban areas, especially in megacities. Droste et al. (2017) developed an algorithm based on the study of Overeem et al. (2013), which converts measurements of smartphone battery temperature into air temperature using a heat transfer model. Cabrera et al. (2021) also established that temperature observations from a professional weather station can be replicated by certain smartphone temperature observations after applying bias correction based on wind speed and solar radiation.

A different surface-observation network design utilized vehicles and leveraged the large number of vehicles and high-density of spatial coverage. Mahoney et al. (2010) described a scenario in which different environmental parameters were directly captured or deduced and then sent to high-speed data hubs from sensors and safety systems onboard millions of private automobiles. After the concept “vehicles as weather sensors” (Mahoney and O’Sullivan 2013) was proposed, several research projects were conducted on data QA, evaluations and applications in weather nowcasting (Riede et al. 2019; Bell et al. 2022; Marquès et al. 2022). Trüb et al. (2018) also presented a cost-efficient proposal that used crowdsourced air traffic control data to obtain air temperature, pressure, and wind measurements in the upper air on a large scale.

Other than smart devices, manually crowdsourced meteorological observations remain an important data source. Community Collaborative Rain, Hail and Snow Network (CoCoRaHS) and Meteorological Phenomena Identification Near the Ground (mPING) are successful citizen science initiatives in the United States that contribute to weather monitoring and forecasting. CoCoRaHS, a network of volunteers who measure precipitation in their local areas using standardized rain gauges, has been highly successful in providing valuable data for rainfall analysis and expanding the observation network (Reges et al. 2016). mPING is a mobile app that allows users to report various weather phenomena, especially severe weather events, directly to the National Weather Service (Elmore et al. 2014). These real-time data help meteorologists to validate and improve weather forecasting and warnings. However, volunteer-based data collection is associated with potential biases and other limitations.

In China, considering the surface weather observation as an example, there are approximately 2440 national-level stations distributed in the mainland region supplemented by more than 65 000 unmanned regional stations. Despite the size of this network, the distribution remains uneven, especially over western China (e.g., Shen et al. 2018). The China Meteorological Administration (CMA) has noticed that observations of the atmosphere and weather conducted by other ministries and commissions of China, including commercial companies, have boomed in recent years. Consequently, the CMA is expected to capitalize on the “crowdsourcing” concept to supplement meteorological observations based on social resources, which are not limited to only individuals or citizens. The CMA has released guideline documents intended to improve crowdsourcing data collection and application (e.g., China Meteorology Administration 2018), and these official guidelines are designed to promote the establishment of a basic system that combines multiparty participation from the public, enterprises, and social institutions, among others. However, we have not yet obtained a clear understanding of whether these parties are willing to participate in crowdsourcing observations and share their data freely.

To bridge this knowledge gap, a questionnaire survey was conducted by the National Meteorological Information Centre (NMIC) of CMA and distributed through email to approximately 30 enterprises and social institutions selected from NMIC’s cooperative users after considering their business size. As a preliminary attempt in China, businesses and institutions were deemed supplementary and potential crowdsourcing observation data sources at this initial stage, and the questionnaire was mainly designed for these organizations. Last, 21 responses were received. To reveal the differences in precipitation monitoring before and after including crowdsourcing data, a case study was also designed. The evaluation results shed positive light on crowdsourcing development in China.

The major results of the survey are summarized in section 2. In section 3, we compared precipitation measurements from two datasets: a merged dataset consisting of data from three sources [the Ministry of Water Resources (MWR), a commercial company, and the CMA] and a dataset from the CMA only. All measurements were taken in the summer of 2022 in Guangzhou City, located in South China. Observations from the MWR are technically not crowdsourced data (e.g., Garcia-Marti et al. 2022), but they serve as a convenient proxy for crowdsourced data because they are not collected by the CMA. The data from two sources other than the CMA were used to identify the effect of involving more surface stations on rainfall monitoring, and they were also employed to test our data processing procedures, including QC, calibration, and assessment, which are required prior to data utilization. Challenges and recommendations concerning crowdsourced data reliability, and incentives for crowdsourcing are discussed in section 4. The major points of this study are summarized in section 5.

2. Survey and results

The questions from our survey can be grouped into three themes. Group A consists of property and attribute of the respondents and contains only question A.1: “Which of the following categories do you fall under: company, research institute (college), government department, or volunteer observer?” Group B consists of data management, application, and service and contains two questions:

  • B.1—Do you have your own data center?

  • B.2—Why do you collect these data; to meet your own requirements, for social or business services, or uncertain?

Group C consists of data exchange and participation in crowdsourcing observations and contains three questions:

  • C.1—Do you have or anticipate requiring other data sources of data, right now or in the future, in addition to your own data?

  • C.2—Are you willing to exchange or share your own data directly with other parties in certain circumstances?

  • C.3—Are you willing to join crowdsourcing observational programs initiated and organized by the CMA?

The feedback collected has been summarized as a graphic overview in Fig. 1. The total number of valid responses was 21.

Fig. 1.
Fig. 1.

Graphic overview of replies to the questionnaire collected from 21 valid responses. The legends for each result are shown above the graphics.

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0065.1

a. Properties and attributes of the respondents

Because there was no response from volunteer or amateur observers (see results for question A.1), the following statistics are based on the answers from organizations. The information collected indicated that the CWS network of China is in the early developmental stage. Observations carried out by government departments should ideally not be attributed to crowdsources. However, to expand the scope of our survey, they were retained and one piece of feedback was obtained in this category. Eighteen responses were obtained from companies that focused on information or data services, thus accounting for the ∼86% of respondents.

b. Data management, application, and services

A total of 13 of the 21 surveyed organizations (62%) had built their own data centers, which can be inferred as having used certain proven techniques for data management. Seventeen respondents also offered additional information on the methods by which the observation data or data files were stored, while 11 of them used a database including relational and unstructured databases for local management and 3 uploaded their data to cloud storage, whereas some used removable hard disks or arrays of disks to store or back up data. Among them, several organizations used two of the aforementioned techniques.

Regarding the usage of their observations, five of the surveyed organizations (24%) employed the data to meet their own requirements, while two organizations carried out observations only to support their business services. Two-thirds of them applied their data to both research and day-to-day operational work.

c. Data exchange and participation in crowdsourcing observations

A total of 18 of the 21 respondents (86%) agreed that more categories of data were needed or would be needed in the future. Most of these positive answers were obtained from companies, highlighting the importance of data in enterprise development.

Regarding willingness to share data directly with other parties through service purchases or data exchanges, only one organization declined while the other 17 demonstrated willingness to engage. Combining the responses to questions C.1 and C.2 showed that 15 of the respondents (71%) had a clear need for data from other sources and were willing to exchange data with other parties under some framework. We also consulted with these companies about the volume of their annual observed data and found that more than one-half were above 1 TB. Thus, a significantly large potential volume of data is available for exchange.

In response to the last question, 19 organizations (90%) were willing to join crowdsourcing observational programs initiated and organized by the CMA. This higher positive answer rate implies that participants recognize the authority from national organizations and associate these authorities with the avoidance of potential risks.

In conclusion, it is conceivable that crowdsourcing of meteorological data may expand in China, with solid foundations for data management and data services, and a strong willingness to exchange data under structured conditions being demonstrated.

3. Case study

a. Data

Guangzhou, with more than 18 million inhabitants, is the capital city of the Guangdong Province and is located along the coast of South China. The elevation of the city generally increases from southwest to northeast, with a coastal alluvial plain in its southern part and a hilly region in its northeastern part. This area commonly experiences heavy rainfall every warm season, mainly related to low-level onshore winds (Zhang and Chen 2018; Wang et al. 2021) and the subdaily precipitation has regional and diurnal characteristics (Li et al. 2019, 2021).

Three different sources of hourly precipitation data during summer (June–August) in 2022 were employed in this study. The measurements archived at the NMIC, which were obtained from national- and regional-level surface stations of the CMA, constitute the main part of the data (referred to as CMA stations and CMA data). The second data source was the surface stations of the MWR (referred to as MWR stations and MWR data), whose real-time observations were transferred to the CMA under a previously agreed upon data exchange protocol. CMA and MWR rainfall data were gathered using tipping-bucket rain gauges with resolutions of 0.1 and 0.5 mm, respectively. The remainder of the data were collected using a type of piezoelectric precipitation sensor, with a precision of 0.01 mm and maximum range of 200 mm h−1 (referred to as Pie stations and Pie data). This instrument was independently developed by a commercial meteorological and agricultural instrument manufacturer, who also participated in this survey.

Several studies have focused on QC of crowdsourced measurements of precipitation, such as those by de Vos et al. (2017) and Chen et al. (2021a). In our study, the hourly precipitation underwent a plausible value check, time consistency check, and spatial consistency check, intended to screen out erroneous data. This procedure was developed based on the QC scheme for real-time automatic weather station hourly precipitation data (version 1.2) from the NMIC. Details can be found in appendix A. First, the QC scheme was applied to the CMA data, and then the QC-accepted measurements were employed as reference data in the spatial consistency checks for the MWR and Pie data. Stations with more than 10% erroneous or missing data were excluded. After QC, ∼10% of the original stations were omitted, and the stations that fulfilled the criteria are shown in Fig. 2. Notably, all Pie stations were located in Guangzhou city, and the CMA and MWR stations outside Guangzhou were included to facilitate more accurate interpolation at the city border. Further, the CMA stations north of 23.4°N within Guangzhou were relatively underrepresented, and the MWR and Pie stations provided good compensation.

Fig. 2.
Fig. 2.

Distribution of surface stations that passed QC. Blue dots, purple triangles, and red squares represent CMA stations, MWR stations, and Pie stations, respectively. Dark-orange lines outline different districts within Guangzhou City, and the white part indicates the coastal area. White solid lines divide the map into a 0.05° × 0.05° grid.

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0065.1

Crowdsourced data users typically face uncertain data quality and reliability. Therefore, a comparison of these data with those of authoritative departments is an optimal choice (Zheng et al. 2018; Coney et al. 2022). CMA data were well qualified and have been employed as a reference in many studies, such as evaluations of satellite precipitation products and analyses of daily precipitation (e.g., Shen et al. 2010; Yu et al. 2020). Consequently, the MWR and Pie data were compared and calibrated with the CMA data after QC. Details of this procedure can be found in appendix B.

To obtain gridded data, Shepard’s angular and distance weighting algorithm (Shepard 1968; New et al. 2000) was applied in this study to interpolate hourly measurements of precipitation from surface stations. This method has been widely used to develop the global precipitation datasets from the Global Precipitation Climatology Centre, Deutscher Wetterdienst (Becker et al. 2013; Schneider et al. 2017). The grid length was set to 0.04° to enable almost all of the grids covering the study area to find at least one station within 0.04° of the grid center.

b. Method

We set the gridded precipitation based on the CMA data (referred to as the CMA grid) as a reference, and compared the gridded precipitation based on the combined data (referred to as the COM grid) to the CMA grid. Three indices, the cumulative precipitation (CP), precipitation intensity (PI), and heavy rain hours (HRH), were calculated grid by grid and were employed to assess the effects of including crowdsourced data on precipitation monitoring. CP is the sum of hourly data indicating no less than 0.3 mm of effective precipitation (considering the precision of each type of measurement) during the whole period. PI is the hourly mean value of effective precipitation. It is crucial to monitor subdaily heavy rain (Wu et al. 2019) for the cities’ sustainable management; therefore, in this study, we employed HRH to analyze the precipitation events that may trigger disasters. The peak-over-threshold nth-percentile method is typically applied to identify heavy rainfall (Jones et al. 1999; Wang et al. 2021). The grid data indicate that ∼85% of rainfall events produced less than 8 mm precipitation. Therefore, we set the heavy rain threshold to 8 mm and HRH to the sum of rain hours for which the hourly precipitation was greater than or equal to this threshold. An additional metric, the critical success index (CSI), was used to develop for HRH and can be expressed as follows:
CSI=HH+M+F,
where H, M, and F represent hit (CMA grid ≥ 8 mm and COM grid ≥ 8 mm), miss (CMA grid ≥ 8 mm and COM grid < 8 mm), and false (CMA grid < 8 mm and COM grid ≥ 8 mm), respectively.

c. Results

The CP distribution of the CMA grid from June to August in 2022 is shown in Fig. 3a, indicating that the total amount increased from south to north, and the maximum exceeded 1300 mm, in an area mainly characterized by hilly terrain. Although the CP of the COM grid (Fig. 3b) showed a similar pattern to that of the CMA grid, and the differences between these two grid datasets can be observed clearly in Fig. 3c, with the major differences located in the western and northeastern parts of Guangzhou. The CP of the COM grid was lower than that of the CMA grid with a maximum bias of −214.5 mm at the western corner, where the distributions of CMA stations were sparse. Hence, the observations employed in the CMA data interpolation were performed farther away from the grid center than those used in COM data interpolation, which may cause discrepancies between these two analyses. As for the hilly area of Guangzhou, the difference in CP also shows negative values over this region, but large positive values up to 192.7 mm were found immediately adjacent to this area. The statistics of CP differences based on the division of the grid’s altitude are presented in Table 1. The results show that the mean absolute differences and standard deviation of absolute differences increased when the altitude increased, indicating that uneven terrain may strengthen uncertainty in precipitation analyses.

Fig. 3.
Fig. 3.

Distribution of (a) CP of CMA grid, (b) CP of COM grid, (c) CP of COM grid minus CMA grid, (d) PI of CMA grid, (e) PI of COM grid, and (f) PI of COM grid minus CMA grid during June–August in 2022.

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0065.1

Table 1.

Statistics of CP differences based on the division of grid’s altitude.

Table 1.

Figure 3d shows the PI of the CMA grid with a high-value center (up to 4.75 mm h−1) in the middle-west of the city. The PI of the COM grid had a high value center at the same location, but it also showed relatively higher values in the northwest and lower parts of the north. Figure 3f reveals that the PI of the COM grid is higher than that of the CMA grid in almost the entire northern region.

A comparison of Fig. 3c with Fig. 3f shows opposite trends in the western and northern parts of Guangzhou. Theoretically, high CP should be accompanied by heavy PI. We suppose that the uneven distribution of the observational network was the main factor underlying these differences. The percentage of precipitation in different ranges was calculated based on grid data, and only the grids with COM stations but no CMA stations were counted. Figure 4 shows that the percentage of COM grid data ranging from [0.3, 0.6) to [2.0, 4.0) was lower than that of CMA grid data while that ranging from [1.0, 2.0) to ≥8.0 was greater than that of CMA data, indicating that the COM network observed relatively heavier rainfall in some places.

Fig. 4.
Fig. 4.

Percentage distribution of precipitation in different ranges counted on the grids with COM stations in them but without CMA stations.

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0065.1

The distribution of HRHs on the CMA grid is shown in Fig. 5a, which indicates almost the entire area north of 23.5°N had more than 32 HRHs. Similar to the difference in PI, Fig. 5c shows that the HRH of the COM grid was greater than that of the CMA grid in the northwest and lower parts of the north, which also verifies the result in Fig. 4. The CSI shown in Fig. 5d is greater than 0.9 in almost the southern part of Guangzhou, indicating that the HRHs of the COM grid matched well with the CAM grid. Lower values were found in the north, and particularly in the northwest.

Fig. 5.
Fig. 5.

Distribution of (a) HRHs of CMA grid, (b) HRHs of COM grid, (c) HRHs of COM grid minus CMA grid, and (d) CSI of HRHs during June–August in 2022.

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0065.1

d. Implications for surface observations

In this section, the intention is not to show which grid (COM or CMA) better reflects the real situation but to reveal their differences. From the analysis, it is apparent that bias was introduced into the monitoring of cumulative amount, intensity, and heavy rainfall aspects over the region sparsely covered by surface stations, especially on uneven terrain. Other observation techniques, such as weather radar, may be relatively more mature technical methods for precipitation monitoring; however, they may not be representative of rainfall at ground level (de Vos et al. 2017). Hence, surface observations are still the most important and reliable source for precipitation monitoring. Introducing high-quality crowdsourcing data as a supplement would be of great benefit.

4. Discussion

The survey results revealed that most organizations are willing to exchange their own meteorological observations and participate in crowdsourcing, especially under the organization of a national authority. In addition, data mining has not been thoroughly conducted, and it has not yet generated sufficient economic benefits. It is, therefore, necessary to develop a crowdsourcing observation system with wide participation to maximize the value of data already available in China. In doing so, the following challenges will likely need to be addressed:

  1. Observational specifications and data reliability. As more observations are incorporated, erroneous or unreasonable data may be included. At present, there are still some unavoidable problems, such as whether (i) the instruments meet observation standards, (ii) the instruments are regularly calibrated and maintained by their owners or managers, (iii) data collection and the environment of the observation site meet the criteria, and (iv) the observers may lack technical training and guidance that could be offered by official authorities. Data quality cannot be guaranteed until these considerations are addressed.

  2. Incentives. Because of the lack of a good profit or a mutual benefit model, the actual ratio of companies and institutions participating in the crowdsourcing observations was extremely low and the scale was not adequate. Moreover, due to insufficient funding, unclear data policies, and other reasons, enterprises, academic institutions, and even private citizens do not have adequate incentives to enter this field.

  3. Data security and privacy protection. Considering national security and personal privacy, not all meteorological data or information are appropriate for direct release to the public or even for cross-border transfer; therefore, data management should be strengthened, which will require the formulation of appropriate policies. It is essential to categorize data to establish which types can be freely used and which can only be used under specific conditions or after relevant processing. It is also necessary to clarify the responsibilities and obligations of data users.

Garcia-Marti et al. (2022) provided six recommendations for crowdsourced data applications that emphasize cooperation in sharing data and outcomes between different national meteorological services. For China, we propose the establishment of a unified crowdsource data platform at the CMA for data collection, storage, secure management, and usage to fully exploit crowdsourced data now and in the future. Moreover, we hope that these suggestions will provide a reference for other countries to carry out similar work. Some related works include the following:

  1. Strengthening the national-level design, planning, and formulation of standards relating to each aspect of the data flow. The CMA and its subordinate provincial administrations should provide observational training and guidance. Most crowdsourcing entities in China at present are companies and organizations; thus, it will be easier to handle data quality assurance than it would be for data from weather observation amateurs. To lead the development of crowdsourcing and strengthen cooperation, a national-level manager is required to follow national and international developments of crowdsourced data application.

  2. Performing data transactions, paid data services, or data exchanges through legal mechanisms. Projects funded by CMA should be established to provide examples and guidance in the development of crowdsourced data sharing and application development. In addition, social institutions could be encouraged to actively conduct observations and join crowdsourcing programs. It is possible to identify a feasible process to periodically assess their work, including the data, and provide commendations or rewards to those who have made outstanding contributions.

  3. An essential aspect would be establishing policies for data security, by establishing the demand, considering the actual situation, and incorporating advice from the data user side during the policy development process. It is also necessary to formulate proposals for the definition of observation data ownership, as well as measures for user privacy protection.

  4. Exploring data applications and mining high-value-added data. Crowdsourcing data should be applied in operational works and has potential utility in services for the public and outside the CMA. Thus, users should be engaged to develop data applications together.

The abovementioned aspects are reflections on the development of crowdsourcing data platforms, as well as some pertinent concerns. With the expansion of crowdsourcing observations in the future, data quantities will become massive and the categories and quality of data will vary widely, raising further questions related to practical application.

We analyzed the differences in precipitation monitoring before and after including crowdsourcing data in Guangzhou and tested the data processing procedures, including data quality control, calibration, and assessment, which were important before data application. The evaluation result on crowdsourcing development in China was very encouraging, hence we will continue with it by attracting more participating companies. The research and analysis presented in this paper did not include CWS observations, which differs from overseas studies. This difference may be related to the current status and development of CWSs in China. In the future, with the economic development and implementation of related policies, it is possible that there will be a growth in CWSs observations in China, implying that the encouragement of amateurs to actively participate will provide a specific set of problems.

5. Conclusions

This study demonstrated that meteorological data collected by official observation networks can be effectively broadened by incorporating crowdsourced data from enterprises and institutions. Our survey also revealed that a certain number of organizations in China are willing to participate in crowdsourced data exchanges based on their own needs, although the approaches to be implemented for participation have not been fully defined. Because of their own internal requirement, most of these parties already have solid foundations for data management and data services. An increase in crowdsourced data use in China in the future is anticipated. As crowdsourcing is in a nascent phase in China, we further suggest that a unified crowdsourcing data platform be constructed at the CMA to support further work related to securing the data life cycle, including collection, storage, secure management, sharing, and application.

In this study, we employed hourly precipitation data from three sources at surface stations to verify our crowdsourced data analysis procedure, including quality control, data assessment, and calibration. Guangzhou was chosen as a test site to investigate the differences in precipitation monitoring, with inclusion of additional stations. The results demonstrate that additional observations enhance accuracy on uneven terrain and in areas with limited surface station coverage, suggesting that this may indicate crowdsourcing to have developmental potential.

Acknowledgments.

This study was supported by the National Key Project of the Ministry of Science and Technology of China (Grant 2022YFC3701205), and the Specialized Research Funding Program from the CMA Institute for Development and Program Design (Grant ZCYJ2022009). The authors thank the owners of the Pie stations and the Ministry of Water Resources for offering hourly precipitation data. We also thank the 21 organizations that participated in our survey. We thank the three anonymous reviewers and the editor for their thoughtful and constructive comments on our paper.

Data availability statement.

Because of the data policy of the CMA, supporting data cannot be made openly available except for gridded hourly precipitation. Further information about the gridded data for access are available on request from Dr. Yu Yu (yuyu@cma.gov.cn).

APPENDIX A

Data Quality Control

The QC procedure was employed and modified from the QC scheme for real-time automatic weather station hourly precipitation data (version 1.2), which was developed by the NMIC. After QC, every hourly precipitation (phour) was given a QC code of the value from 0, 2, or 8, which implies right, error, or missing. First, all the hourly data were given an initial QC code of 0, except those with missing data, which were flagged with 8. Subsequently, three steps of the QC procedure, plausible value check, time consistency check, and spatial consistency check, were applied successively.

a. Plausible value check

Any phour larger than 150 mm was flagged with 2 and selected for manual verification, as it was extremely unusual for such great rainfall to occur in mainland China.

b. Time consistency check

The aim of the check was to detect unrealistic repeating values or dead band data, which were usually caused by blocked sensor. All continuous phour that fulfilled any of the following criteria were flagged with 2. These criteria were set based on an analysis of historical precipitation data from automatic weather stations of the CMA, which were also applied in our operational data QC:

  1. phour lies in the range of [0.1, 0.5 mm], and repeats for no less than 14 h;

  2. phour lies in the range of (0.5, 1.0 mm], and repeats for no less than 9 h;

  3. phour is greater than 1.0 mm and repeats for no less than 6 h.

c. Spatial consistency check

As it is difficult to identify the veracity of a measurement by employing only hourly data, this step was included to find the unreasonably heavy or low precipitation or unrealistic long-lasting precipitation by comparing with adjacent stations during that period.

1) Adjacent station selections

The focal station was set as a center, and the search range for adjacent stations was gradually enlarged from 30 to 50 km, until at least 20 stations could be found. If there were more neighbor stations, the closest 20 stations were selected. Then, extremely inconsistent precipitation at the focal station were verified based on adjacent stations.

2) 12-h cumulative precipitation check

The sliding measurements for 12 consecutive hours were cumulated station by station. If the cumulative precipitation of the focal station was greater than 15 mm, while during the same period the cumulative values from the neighboring stations were all 0 mm, the 12 consecutive phour were flagged as errors. This step is used to screen out excessive precipitation at a single station.

3) 10-day cumulative precipitation check

A sliding window of 10 days was selected, and the cumulative precipitation (p10-day) and hours of precipitation of no less than 0.1 mm (h10-day) were calculated respectively on the focal station and its adjacent stations. Then, the mean and standard deviation (std) of the adjacent stations’ p10-day and h10-day were calculated. If the focal station fulfills any items of the following list 1–4, its 10-day phour is flagged as an error:

  1. focal stations’ p10-day > mean(adjacent stations’ p10-day) + std(adjacent stations’ p10-day) × 7,

  2. focal stations’ h10-day > mean(adjacent stations’ h10-day) + std(adjacent stations’ h10-day) × 7,

  3. focal stations’ p10-day < mean(adjacent stations’ p10-day) − std(adjacent stations’ p10-day) × 7, or

  4. focal stations’ h10-day < mean(adjacent stations’ h10-day) − std(adjacent stations’ h10-day) × 7.

The multiplier factor of 7 was employed to identify significant outliers.

The data in Tables A1 and A2 present two examples of erroneous observations detected at focal stations when applying the “10-day cumulative precipitation check.” In Table A1, the nonzero rainfall hour value at the focal station was 59 h, which was almost 2 times the adjacent stations’ value (between 32 and 36 h). Precipitation fell continuously from the 10th hour to the 65th hour at the focal station, while similar rainfall was not observed at the adjacent stations given in Table A1. In Table A2, the total precipitation of the focal station was 45.3 mm, which was much less than that of the adjacent stations at 225.9 mm (calculated from adjacent station 3) and 237.4 mm (calculated from adjacent station 4).

Table A1.

Example of unrealistic long-lasting precipitation detected by the “10-day cumulative precipitation check.” The consecutive hourly precipitation in millimeters at the focal station and its five adjacent stations is shown (precipitation of 0.0 mm before and after the series was not given).

Table A1.
Table A2.

As in Table A1, but for unreasonably low precipitation.

Table A2.

APPENDIX B

Data Comparison and Calibration

MWR and Pie data were compared with the CMA data on stations that were within a range of 500 m (this distance was chosen to include an adequate number of stations for comparison). Consequently, there were 28 pairs of MWR data versus CMA data and 11 pairs of Pie data versus CMA data. Only hourly measurements greater than 0 mm at the same time from the two parties were selected.

A comparison of the MWR and CMA data is shown in Fig. B1. The fitted line can be expressed as y = 0.98x, which is very close to the 1:1 line. In addition, the standard deviation of their differences, that is, 1.24 mm, was not significant. There is an obvious systematic deviation between the Pie and CMA data, which can be seen in Fig. B2. The fitted line was y = 0.8x.

Fig. B1.
Fig. B1.

Comparison of hourly precipitation from MWR stations and CMA stations: std is standard deviation, cc is correlation coefficient, and n is sample number.

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0065.1

Fig. B2.
Fig. B2.

As in Fig. B1, but comparing Pie stations and CMA stations.

Citation: Weather, Climate, and Society 16, 1; 10.1175/WCAS-D-23-0065.1

According to the above assessment, the MWR data had no distinctive deviation from the CMA data and were maintained at the original values. The Pie data were multiplied by 0.8 as calibrated to the CMA data.

REFERENCES

  • Becker, A., P. Finger, A. Meyer-Christoffer, B. Rudolf, K. Schamm, U. Schneider, and M. Ziese, 2013: A description of the global land-surface precipitation data products of the Global Precipitation Climatology Centre with sample applications including centennial (trend) analysis from 1901–present. Earth Syst. Sci. Data, 5, 7199, https://doi.org/10.5194/essd-5-71-2013.

    • Search Google Scholar
    • Export Citation
  • Bell, Z., S. L. Dance, and J. A. Waller, 2022: Exploring the characteristics of a vehicle-based temperature dataset for kilometer-scale data assimilation. Meteor. Appl., 29, e2058, https://doi.org/10.1002/met.2058.

    • Search Google Scholar
    • Export Citation
  • Brousse, O., C. Simpson, N. Walker, D. Fenner, F. Meier, J. Taylor, and C. Heaviside, 2022: Evidence of horizontal urban heat advection in London using six years of data from a citizen weather station network. Environ. Res. Lett., 17, 044041, https://doi.org/10.1088/1748-9326/ac5c0f.

    • Search Google Scholar
    • Export Citation
  • Cabrera, A. N., A. Droste, B. G. Heusinkveld, and G.-J. Steeneveld, 2021: The potential of a smartphone as an urban weather station—An exploratory analysis. Front. Environ. Sci., 9, 673937, https://doi.org/10.3389/fenvs.2021.673937.

    • Search Google Scholar
    • Export Citation
  • Caluwaerts, S., and Coauthors, 2020: The urban climate of Ghent, Belgium: A case study combining a high-accuracy monitoring network with numerical simulations. Urban Climate, 31, 100565, https://doi.org/10.1016/j.uclim.2019.100565.

    • Search Google Scholar
    • Export Citation
  • Chapman, L., C. Bell, and S. Bell, 2017: Can the crowdsourcing data paradigm take atmospheric science to a new level? A case study of the urban heat island of London quantified using Netatmo weather stations. Int. J. Climatol., 37, 35973605, https://doi.org/10.1002/joc.4940.

    • Search Google Scholar
    • Export Citation
  • Chen, A. B., M. Behl, and J. L. Goodall, 2021a: Assessing the trustworthiness of crowdsourced rainfall networks: A reputation system approach. Water Resour. Res., 57, e2021WR029721, https://doi.org/10.1029/2021WR029721.

    • Search Google Scholar
    • Export Citation
  • Chen, J., K. Saunders, and K. Whan, 2021b: Quality control and bias adjustment of crowdsourced wind speed observations. Quart. J. Roy. Meteor. Soc., 147, 36473664, https://doi.org/10.1002/qj.4146.

    • Search Google Scholar
    • Export Citation
  • China Meteorology Administration, 2018: CMA released letter No. 80: Guiding opinions on the development of meteorological crowdsourcing observation (in Chinese). CMA, 6 pp.

  • Coney, J., B. Pickering, D. Dufton, M. Lukach, B. Brooks, and R. R. Neely III, 2022: How useful are crowdsourced air temperature observations? An assessment of Netatmo stations and quality control schemes over the United Kingdom. Meteor. Appl., 29, e2075, https://doi.org/10.1002/met.2075.

    • Search Google Scholar
    • Export Citation
  • de Vos, L., H. Leijnse, A. Overeem, and R. Uijlenhoet, 2017: The potential of urban rainfall monitoring with crowdsourced automatic weather stations in Amsterdam. Hydrol. Earth Syst. Sci., 21, 765777, https://doi.org/10.5194/hess-21-765-2017.

    • Search Google Scholar
    • Export Citation
  • de Vos, L., H. Leijnse, A. Overeem, and R. Uijlenhoet, 2019: Quality control for crowdsourced personal weather stations to enable operational rainfall monitoring. Geophys. Res. Lett., 46, 88208829, https://doi.org/10.1029/2019GL083731.

    • Search Google Scholar
    • Export Citation
  • Droste, A. M., J. J. Pape, A. Overeem, H. Leijnse, G. J. Steeneveld, A. J. Van Delden, and R. Uijlenhoet, 2017: Crowdsourcing urban air temperatures through smartphone battery temperatures in São Paulo, Brazil. J. Atmos. Oceanic Technol., 34, 18531866, https://doi.org/10.1175/JTECH-D-16-0150.1.

    • Search Google Scholar
    • Export Citation
  • Droste, A. M., B. G. Heusinkveld, D. Fenner, and G.-J. Steeneveld, 2020: Assessing the potential and application of crowdsourced urban wind data. Quart. J. Roy. Meteor. Soc., 146, 26712688, https://doi.org/10.1002/qj.3811.

    • Search Google Scholar
    • Export Citation
  • Elmore, K. L., Z. L. Flamig, V. Lakshmanan, B. T. Kaney, V. Farmer, H. D. Reeves, and L. P. Rothfusz, 2014: MPING: Crowd-sourcing weather reports for research. Bull. Amer. Meteor. Soc., 95, 13351342, https://doi.org/10.1175/BAMS-D-13-00014.1.

    • Search Google Scholar
    • Export Citation
  • Feichtinger, M., R. de Wit, G. Goldenits, T. Kolejka, B. Hollósi, M. Žuvela-Aloise, and J. Feigl, 2020: Case-study of neighborhood-scale summertime urban air temperature for the City of Vienna using crowd-sourced data. Urban Climate, 32, 100597, https://doi.org/10.1016/j.uclim.2020.100597.

    • Search Google Scholar
    • Export Citation
  • Fenner, D., F. Meier, B. Bechtel, M. Otto, and D. Scherer, 2017: Intra and inter “local climate zone” variability of air temperature as observed by crowdsourced citizen weather stations in Berlin, Germany. Meteor. Z., 26, 525547, https://doi.org/10.1127/metz/2017/0861.

    • Search Google Scholar
    • Export Citation
  • Garcia-Marti, I., A. Overeem, J. W. Noteboom, L. de Vos, M. de Haij, and K. Whan, 2022: From proof-of-concept to proof-of-value: Approaching third-party data to operational workflows of national meteorological services. Int. J. Climatol., 43, 275292, https://doi.org/10.1002/joc.7757.

    • Search Google Scholar
    • Export Citation
  • Giazzi, M., and Coauthors, 2022: Meteonetwork: An open crowdsourced weather data system. Atmosphere, 13, 928, https://doi.org/10.3390/atmos13060928.

    • Search Google Scholar
    • Export Citation
  • Hintz, K. S., C. McNicholas, R. Randriamampianina, H. T. P. Williams, B. Macpherson, M. Mittermaier, J. Onvlee-Hooimeijer, and B. Szintai, 2021: Crowd-sourced observations for short-range numerical weather prediction: Report from EWGLAM/SRNWP meeting 2019. Atmos. Sci. Lett., 22, e1031, https://doi.org/10.1002/asl.1031.

    • Search Google Scholar
    • Export Citation
  • Jones, P. D., E. B. Horton, C. K. Folland, M. Hulme, D. E. Parker, and T. A. Basnett, 1999: The use of indices to identify changes in climatic extremes. Climatic Change, 42, 131149, https://doi.org/10.1023/A:1005468316392.

    • Search Google Scholar
    • Export Citation
  • Krennert, T., R. Kaltenberger, G. Pistotnik, A. M. Holzer, F. Zeiler, and M. Stampfl, 2018a: Trusted spotter network Austria–A new standard to utilize crowdsourced weather and impact observations. Adv. Sci. Res., 15, 7780, https://doi.org/10.5194/asr-15-77-2018.

    • Search Google Scholar
    • Export Citation
  • Krennert, T., G. Pistotnik, R. Kaltenberger, and C. Csekits, 2018b: Crowdsourcing of weather observations at national meteorological and hydrological services in Europe. Adv. Sci. Res., 15, 7176, https://doi.org/10.5194/asr-15-71-2018.

    • Search Google Scholar
    • Export Citation
  • Li, J., N. Li, and R. Yu, 2019: Regional differences in hourly precipitation characteristics along the western coast of South China. J. Appl. Meteor. Climatol., 58, 27172732, https://doi.org/10.1175/JAMC-D-19-0150.1.

    • Search Google Scholar
    • Export Citation
  • Li, Z., J. C. L. Chan, K. Zhao, and X. Chen, 2021: Impacts of urban expansion on the diurnal variations of summer monsoon precipitation over the South China coast. J. Geophys. Res. Atmos., 126, e2021JD035318, https://doi.org/10.1029/2021JD035318.

    • Search Google Scholar
    • Export Citation
  • Mahoney, B., S. Drobot, P. Pisano, B. McKeever, and J. O’Sullivan, 2010: Vehicles as mobile weather observation systems. Bull. Amer. Meteor. Soc., 91, 11791182, https://doi.org/10.1175/2010BAMS2954.1.

    • Search Google Scholar
    • Export Citation
  • Mahoney, W. P., III, and J. M. O’Sullivan, 2013: Realizing the potential of vehicle-based observations. Bull. Amer. Meteor. Soc., 94, 10071018, https://doi.org/10.1175/BAMS-D-12-00044.1.

    • Search Google Scholar
    • Export Citation
  • Marquès, E., V. Masson, P. Naveau, O. Mestre, V. Dubreuil, and Y. Richard, 2022: Urban heat island estimation from crowdsensing thermometers embedded in personal cars. Bull. Amer. Meteor. Soc., 103, E1098E1113, https://doi.org/10.1175/BAMS-D-21-0174.1.

    • Search Google Scholar
    • Export Citation
  • Masson, V., A. Lemonsu, J. Hidalgo, and J. Voogt, 2020: Urban climates and climate change. Annu. Rev. Environ. Resour., 45, 411444, https://doi.org/10.1146/annurev-environ-012320-083623.

    • Search Google Scholar
    • Export Citation
  • Meier, F., D. Fenner, T. Grassmann, B. Jänicke, M. Otto, and D. Scherer, 2015: Challenges and benefits from crowdsourced atmospheric data for urban climate research using Berlin, Germany, as testbed. Ninth Int. Conf. on Urban Climate/12th Symp. on the Urban Environment, Toulouse, France, Météo-France, http://www.meteo.fr/icuc9/LongAbstracts/nomtm6-2-6171335_a.pdf.

  • Meier, F., D. Fenner, T. Grassmann, M. Otto, and D. Scherer, 2017: Crowdsourcing air temperature from citizen weather stations for urban climate research. Urban Climate, 19, 170191, https://doi.org/10.1016/j.uclim.2017.01.006.

    • Search Google Scholar
    • Export Citation
  • Muller, C. L., L. Chapman, C. S. B. Grimmond, D. T. Young, and X. Cai, 2013: Sensors and the city: A review of urban meteorological networks. Int. J. Climatol., 33, 15851600, https://doi.org/10.1002/joc.3678.

    • Search Google Scholar
    • Export Citation
  • Muller, C. L., L. Chapman, S. Johnston, C. Kidd, S. Illingworth, G. Foody, A. Overeem, and R. R. Leigh, 2015: Crowdsourcing for climate and atmospheric sciences: Current status and future potential. Int. J. Climatol., 35, 31853203, https://doi.org/10.1002/joc.4210.

    • Search Google Scholar
    • Export Citation
  • Napoly, A., T. Grassmann, F. Meier, and D. Fenner, 2018: Development and application of a statistically-based quality control for crowdsourced air temperature data. Front. Earth Sci., 6, 118, https://doi.org/10.3389/feart.2018.00118.

    • Search Google Scholar
    • Export Citation
  • New, M., M. Hulme, and P. Jones, 2000: Representing twentieth-century space–time climate variability. Part II: Development of 1901–96 monthly grids of terrestrial surface climate. J. Climate, 13, 22172238, https://doi.org/10.1175/1520-0442(2000)013<2217:RTCSTC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Overeem, A., J. C. R. Robinson, H. Leijnse, G. J. Steeneveld, B. K. P. Horn, and R. Uijlenhoet, 2013: Crowdsourcing urban air temperatures from smartphone battery temperatures. Geophys. Res. Lett., 40, 40814085, https://doi.org/10.1002/grl.50786.

    • Search Google Scholar
    • Export Citation
  • Reges, H. W., N. Doesken, J. Turner, N. Newman, A. Bergantino, and Z. Schwalbe, 2016: CoCoRaHS: The evolution and accomplishments of a volunteer rain gauge network. Bull. Amer. Meteor. Soc., 97, 18311846, https://doi.org/10.1175/BAMS-D-14-00213.1.

    • Search Google Scholar
    • Export Citation
  • Riede, H., and Coauthors, 2019: Passenger car data—A new source of real-time weather information for nowcasting, forecasting, and road safety. Third European Nowcasting Conf., Madrid, Spain, AEMET, https://edrop.zamg.ac.at/owncloud/index.php/s/9tiKLz9JFisqPCP/download.

  • Schneider, U., P. Finger, A. Meyer-Christoffer, E. Rustemeier, M. Ziese, and A. Becker, 2017: Evaluating the hydrological cycle over land using the newly-corrected precipitation climatology from the Global Precipitation Climatology Centre (GPCC). Atmosphere, 8, 52, https://doi.org/10.3390/atmos8030052.

    • Search Google Scholar
    • Export Citation
  • Shen, Y., A. Xiong, Y. Wang, and P. Xie, 2010: Performance of high‐resolution satellite precipitation products over China. J. Geophys. Res., 115, D02114, https://doi.org/10.1029/2009JD012097.

    • Search Google Scholar
    • Export Citation
  • Shen, Y., Z. Hong, Y. Pan, J. Yu, and L. Maguire, 2018: China’s 1 km merged gauge, radar and satellite experimental precipitation dataset. Remote Sens., 10, 264, https://doi.org/10.3390/rs10020264.

    • Search Google Scholar
    • Export Citation
  • Shepard, D., 1968: A two-dimensional interpolation function for irregularly spaced data. Proc. 23rd ACM Nat. Conf., Princeton, NJ, Association for Computing Machinery, 517–524, https://doi.org/10.1145/800186.810616.

  • Trüb, R., D. Moser, M. Schäfer, R. Pinheiro, and V. Lenders, 2018: Monitoring meteorological parameters with crowdsourced air traffic control data. Proc. 17th ACM/IEEE Int. Conf. on Information Processing in Sensor Networks, Porto, Portugal, Institute of Electrical and Electronics Engineers, 25–36, https://doi.org/10.1109/IPSN.2018.00010.

  • Wang, C., K. Zhao, A. Huang, X. Chen, and X. Rao, 2021: The crucial role of synoptic pattern in determining the spatial distribution and diurnal cycle of heavy rainfall over the South China coast. J. Climate, 34, 24412458, https://doi.org/10.1175/JCLI-D-20-0274.1.

    • Search Google Scholar
    • Export Citation
  • Wu, M., Y. Luo, F. Chen, and W. K. Wong, 2019: Observed link of extreme hourly precipitation changes to urbanization over coastal South China. J. Appl. Meteor. Climatol., 58, 17991819, https://doi.org/10.1175/JAMC-D-18-0284.1.

    • Search Google Scholar
    • Export Citation
  • Yu, Y., U. Schneider, S. Yang, A. Becker, and Z. Ren, 2020: Evaluating the GPCC full data daily analysis version 2018 through ETCCDI indices and comparison with station observations over mainland of China. Theor. Appl. Climatol., 142, 835845, https://doi.org/10.1007/s00704-020-03352-8.

    • Search Google Scholar
    • Export Citation
  • Zhang, J., and Y. Chen, 2018: Typhoon Nida’s rainfall characteristics in Guangzhou city based on Doppler radar estimation. IOP Conf. Ser. Earth Environ. Sci., 169, 012002, https://doi.org/10.1088/1755-1315/169/1/012002.

    • Search Google Scholar
    • Export Citation
  • Zheng, F., and Coauthors, 2018: Crowdsourcing methods for data collection in geophysics: State of the art, issues, and future directions. Rev. Geophys., 56, 698740, https://doi.org/10.1029/2018RG000616.

    • Search Google Scholar
    • Export Citation
  • Zhu, Y., S. Zhang, Y. Li, H. Lu, K. Shi, and Z. Niu, 2020: Social weather: A review of crowdsourcing-assisted meteorological knowledge services through social cyberspace. Geosci. Data J., 7, 6179, https://doi.org/10.1002/gdj3.85.

    • Search Google Scholar
    • Export Citation
Save
  • Becker, A., P. Finger, A. Meyer-Christoffer, B. Rudolf, K. Schamm, U. Schneider, and M. Ziese, 2013: A description of the global land-surface precipitation data products of the Global Precipitation Climatology Centre with sample applications including centennial (trend) analysis from 1901–present. Earth Syst. Sci. Data, 5, 7199, https://doi.org/10.5194/essd-5-71-2013.

    • Search Google Scholar
    • Export Citation
  • Bell, Z., S. L. Dance, and J. A. Waller, 2022: Exploring the characteristics of a vehicle-based temperature dataset for kilometer-scale data assimilation. Meteor. Appl., 29, e2058, https://doi.org/10.1002/met.2058.

    • Search Google Scholar
    • Export Citation
  • Brousse, O., C. Simpson, N. Walker, D. Fenner, F. Meier, J. Taylor, and C. Heaviside, 2022: Evidence of horizontal urban heat advection in London using six years of data from a citizen weather station network. Environ. Res. Lett., 17, 044041, https://doi.org/10.1088/1748-9326/ac5c0f.

    • Search Google Scholar
    • Export Citation
  • Cabrera, A. N., A. Droste, B. G. Heusinkveld, and G.-J. Steeneveld, 2021: The potential of a smartphone as an urban weather station—An exploratory analysis. Front. Environ. Sci., 9, 673937, https://doi.org/10.3389/fenvs.2021.673937.

    • Search Google Scholar
    • Export Citation
  • Caluwaerts, S., and Coauthors, 2020: The urban climate of Ghent, Belgium: A case study combining a high-accuracy monitoring network with numerical simulations. Urban Climate, 31, 100565, https://doi.org/10.1016/j.uclim.2019.100565.

    • Search Google Scholar
    • Export Citation
  • Chapman, L., C. Bell, and S. Bell, 2017: Can the crowdsourcing data paradigm take atmospheric science to a new level? A case study of the urban heat island of London quantified using Netatmo weather stations. Int. J. Climatol., 37, 35973605, https://doi.org/10.1002/joc.4940.

    • Search Google Scholar
    • Export Citation
  • Chen, A. B., M. Behl, and J. L. Goodall, 2021a: Assessing the trustworthiness of crowdsourced rainfall networks: A reputation system approach. Water Resour. Res., 57, e2021WR029721, https://doi.org/10.1029/2021WR029721.

    • Search Google Scholar
    • Export Citation
  • Chen, J., K. Saunders, and K. Whan, 2021b: Quality control and bias adjustment of crowdsourced wind speed observations. Quart. J. Roy. Meteor. Soc., 147, 36473664, https://doi.org/10.1002/qj.4146.

    • Search Google Scholar
    • Export Citation
  • China Meteorology Administration, 2018: CMA released letter No. 80: Guiding opinions on the development of meteorological crowdsourcing observation (in Chinese). CMA, 6 pp.

  • Coney, J., B. Pickering, D. Dufton, M. Lukach, B. Brooks, and R. R. Neely III, 2022: How useful are crowdsourced air temperature observations? An assessment of Netatmo stations and quality control schemes over the United Kingdom. Meteor. Appl., 29, e2075, https://doi.org/10.1002/met.2075.

    • Search Google Scholar
    • Export Citation
  • de Vos, L., H. Leijnse, A. Overeem, and R. Uijlenhoet, 2017: The potential of urban rainfall monitoring with crowdsourced automatic weather stations in Amsterdam. Hydrol. Earth Syst. Sci., 21, 765777, https://doi.org/10.5194/hess-21-765-2017.

    • Search Google Scholar
    • Export Citation
  • de Vos, L., H. Leijnse, A. Overeem, and R. Uijlenhoet, 2019: Quality control for crowdsourced personal weather stations to enable operational rainfall monitoring. Geophys. Res. Lett., 46, 88208829, https://doi.org/10.1029/2019GL083731.

    • Search Google Scholar
    • Export Citation
  • Droste, A. M., J. J. Pape, A. Overeem, H. Leijnse, G. J. Steeneveld, A. J. Van Delden, and R. Uijlenhoet, 2017: Crowdsourcing urban air temperatures through smartphone battery temperatures in São Paulo, Brazil. J. Atmos. Oceanic Technol., 34, 18531866, https://doi.org/10.1175/JTECH-D-16-0150.1.

    • Search Google Scholar
    • Export Citation
  • Droste, A. M., B. G. Heusinkveld, D. Fenner, and G.-J. Steeneveld, 2020: Assessing the potential and application of crowdsourced urban wind data. Quart. J. Roy. Meteor. Soc., 146, 26712688, https://doi.org/10.1002/qj.3811.

    • Search Google Scholar
    • Export Citation
  • Elmore, K. L., Z. L. Flamig, V. Lakshmanan, B. T. Kaney, V. Farmer, H. D. Reeves, and L. P. Rothfusz, 2014: MPING: Crowd-sourcing weather reports for research. Bull. Amer. Meteor. Soc., 95, 13351342, https://doi.org/10.1175/BAMS-D-13-00014.1.

    • Search Google Scholar
    • Export Citation
  • Feichtinger, M., R. de Wit, G. Goldenits, T. Kolejka, B. Hollósi, M. Žuvela-Aloise, and J. Feigl, 2020: Case-study of neighborhood-scale summertime urban air temperature for the City of Vienna using crowd-sourced data. Urban Climate, 32, 100597, https://doi.org/10.1016/j.uclim.2020.100597.

    • Search Google Scholar
    • Export Citation
  • Fenner, D., F. Meier, B. Bechtel, M. Otto, and D. Scherer, 2017: Intra and inter “local climate zone” variability of air temperature as observed by crowdsourced citizen weather stations in Berlin, Germany. Meteor. Z., 26, 525547, https://doi.org/10.1127/metz/2017/0861.

    • Search Google Scholar
    • Export Citation
  • Garcia-Marti, I., A. Overeem, J. W. Noteboom, L. de Vos, M. de Haij, and K. Whan, 2022: From proof-of-concept to proof-of-value: Approaching third-party data to operational workflows of national meteorological services. Int. J. Climatol., 43, 275292, https://doi.org/10.1002/joc.7757.

    • Search Google Scholar
    • Export Citation
  • Giazzi, M., and Coauthors, 2022: Meteonetwork: An open crowdsourced weather data system. Atmosphere, 13, 928, https://doi.org/10.3390/atmos13060928.

    • Search Google Scholar
    • Export Citation
  • Hintz, K. S., C. McNicholas, R. Randriamampianina, H. T. P. Williams, B. Macpherson, M. Mittermaier, J. Onvlee-Hooimeijer, and B. Szintai, 2021: Crowd-sourced observations for short-range numerical weather prediction: Report from EWGLAM/SRNWP meeting 2019. Atmos. Sci. Lett., 22, e1031, https://doi.org/10.1002/asl.1031.

    • Search Google Scholar
    • Export Citation
  • Jones, P. D., E. B. Horton, C. K. Folland, M. Hulme, D. E. Parker, and T. A. Basnett, 1999: The use of indices to identify changes in climatic extremes. Climatic Change, 42, 131149, https://doi.org/10.1023/A:1005468316392.

    • Search Google Scholar
    • Export Citation
  • Krennert, T., R. Kaltenberger, G. Pistotnik, A. M. Holzer, F. Zeiler, and M. Stampfl, 2018a: Trusted spotter network Austria–A new standard to utilize crowdsourced weather and impact observations. Adv. Sci. Res., 15, 7780, https://doi.org/10.5194/asr-15-77-2018.

    • Search Google Scholar
    • Export Citation
  • Krennert, T., G. Pistotnik, R. Kaltenberger, and C. Csekits, 2018b: Crowdsourcing of weather observations at national meteorological and hydrological services in Europe. Adv. Sci. Res., 15, 7176, https://doi.org/10.5194/asr-15-71-2018.

    • Search Google Scholar
    • Export Citation
  • Li, J., N. Li, and R. Yu, 2019: Regional differences in hourly precipitation characteristics along the western coast of South China. J. Appl. Meteor. Climatol., 58, 27172732, https://doi.org/10.1175/JAMC-D-19-0150.1.

    • Search Google Scholar
    • Export Citation
  • Li, Z., J. C. L. Chan, K. Zhao, and X. Chen, 2021: Impacts of urban expansion on the diurnal variations of summer monsoon precipitation over the South China coast. J. Geophys. Res. Atmos., 126, e2021JD035318, https://doi.org/10.1029/2021JD035318.

    • Search Google Scholar
    • Export Citation
  • Mahoney, B., S. Drobot, P. Pisano, B. McKeever, and J. O’Sullivan, 2010: Vehicles as mobile weather observation systems. Bull. Amer. Meteor. Soc., 91, 11791182, https://doi.org/10.1175/2010BAMS2954.1.

    • Search Google Scholar
    • Export Citation
  • Mahoney, W. P., III, and J. M. O’Sullivan, 2013: Realizing the potential of vehicle-based observations. Bull. Amer. Meteor. Soc., 94, 10071018, https://doi.org/10.1175/BAMS-D-12-00044.1.

    • Search Google Scholar
    • Export Citation
  • Marquès, E., V. Masson, P. Naveau, O. Mestre, V. Dubreuil, and Y. Richard, 2022: Urban heat island estimation from crowdsensing thermometers embedded in personal cars. Bull. Amer. Meteor. Soc., 103, E1098E1113, https://doi.org/10.1175/BAMS-D-21-0174.1.

    • Search Google Scholar
    • Export Citation
  • Masson, V., A. Lemonsu, J. Hidalgo, and J. Voogt, 2020: Urban climates and climate change. Annu. Rev. Environ. Resour., 45, 411444, https://doi.org/10.1146/annurev-environ-012320-083623.

    • Search Google Scholar
    • Export Citation
  • Meier, F., D. Fenner, T. Grassmann, B. Jänicke, M. Otto, and D. Scherer, 2015: Challenges and benefits from crowdsourced atmospheric data for urban climate research using Berlin, Germany, as testbed. Ninth Int. Conf. on Urban Climate/12th Symp. on the Urban Environment, Toulouse, France, Météo-France, http://www.meteo.fr/icuc9/LongAbstracts/nomtm6-2-6171335_a.pdf.

  • Meier, F., D. Fenner, T. Grassmann, M. Otto, and D. Scherer, 2017: Crowdsourcing air temperature from citizen weather stations for urban climate research. Urban Climate, 19, 170191, https://doi.org/10.1016/j.uclim.2017.01.006.

    • Search Google Scholar
    • Export Citation
  • Muller, C. L., L. Chapman, C. S. B. Grimmond, D. T. Young, and X. Cai, 2013: Sensors and the city: A review of urban meteorological networks. Int. J. Climatol., 33, 15851600, https://doi.org/10.1002/joc.3678.

    • Search Google Scholar
    • Export Citation
  • Muller, C. L., L. Chapman, S. Johnston, C. Kidd, S. Illingworth, G. Foody, A. Overeem, and R. R. Leigh, 2015: Crowdsourcing for climate and atmospheric sciences: Current status and future potential. Int. J. Climatol., 35, 31853203, https://doi.org/10.1002/joc.4210.

    • Search Google Scholar
    • Export Citation
  • Napoly, A., T. Grassmann, F. Meier, and D. Fenner, 2018: Development and application of a statistically-based quality control for crowdsourced air temperature data. Front. Earth Sci., 6, 118, https://doi.org/10.3389/feart.2018.00118.

    • Search Google Scholar
    • Export Citation
  • New, M., M. Hulme, and P. Jones, 2000: Representing twentieth-century space–time climate variability. Part II: Development of 1901–96 monthly grids of terrestrial surface climate. J. Climate, 13, 22172238, https://doi.org/10.1175/1520-0442(2000)013<2217:RTCSTC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Overeem, A., J. C. R. Robinson, H. Leijnse, G. J. Steeneveld, B. K. P. Horn, and R. Uijlenhoet, 2013: Crowdsourcing urban air temperatures from smartphone battery temperatures. Geophys. Res. Lett., 40, 40814085, https://doi.org/10.1002/grl.50786.

    • Search Google Scholar
    • Export Citation
  • Reges, H. W., N. Doesken, J. Turner, N. Newman, A. Bergantino, and Z. Schwalbe, 2016: CoCoRaHS: The evolution and accomplishments of a volunteer rain gauge network. Bull. Amer. Meteor. Soc., 97, 18311846, https://doi.org/10.1175/BAMS-D-14-00213.1.

    • Search Google Scholar
    • Export Citation
  • Riede, H., and Coauthors, 2019: Passenger car data—A new source of real-time weather information for nowcasting, forecasting, and road safety. Third European Nowcasting Conf., Madrid, Spain, AEMET, https://edrop.zamg.ac.at/owncloud/index.php/s/9tiKLz9JFisqPCP/download.

  • Schneider, U., P. Finger, A. Meyer-Christoffer, E. Rustemeier, M. Ziese, and A. Becker, 2017: Evaluating the hydrological cycle over land using the newly-corrected precipitation climatology from the Global Precipitation Climatology Centre (GPCC). Atmosphere, 8, 52, https://doi.org/10.3390/atmos8030052.

    • Search Google Scholar
    • Export Citation
  • Shen, Y., A. Xiong, Y. Wang, and P. Xie, 2010: Performance of high‐resolution satellite precipitation products over China. J. Geophys. Res., 115, D02114, https://doi.org/10.1029/2009JD012097.

    • Search Google Scholar
    • Export Citation
  • Shen, Y., Z. Hong, Y. Pan, J. Yu, and L. Maguire, 2018: China’s 1 km merged gauge, radar and satellite experimental precipitation dataset. Remote Sens., 10, 264, https://doi.org/10.3390/rs10020264.

    • Search Google Scholar
    • Export Citation
  • Shepard, D., 1968: A two-dimensional interpolation function for irregularly spaced data. Proc. 23rd ACM Nat. Conf., Princeton, NJ, Association for Computing Machinery, 517–524, https://doi.org/10.1145/800186.810616.

  • Trüb, R., D. Moser, M. Schäfer, R. Pinheiro, and V. Lenders, 2018: Monitoring meteorological parameters with crowdsourced air traffic control data. Proc. 17th ACM/IEEE Int. Conf. on Information Processing in Sensor Networks, Porto, Portugal, Institute of Electrical and Electronics Engineers, 25–36, https://doi.org/10.1109/IPSN.2018.00010.

  • Wang, C., K. Zhao, A. Huang, X. Chen, and X. Rao, 2021: The crucial role of synoptic pattern in determining the spatial distribution and diurnal cycle of heavy rainfall over the South China coast. J. Climate, 34, 24412458, https://doi.org/10.1175/JCLI-D-20-0274.1.

    • Search Google Scholar
    • Export Citation
  • Wu, M., Y. Luo, F. Chen, and W. K. Wong, 2019: Observed link of extreme hourly precipitation changes to urbanization over coastal South China. J. Appl. Meteor. Climatol., 58, 17991819, https://doi.org/10.1175/JAMC-D-18-0284.1.

    • Search Google Scholar
    • Export Citation
  • Yu, Y., U. Schneider, S. Yang, A. Becker, and Z. Ren, 2020: Evaluating the GPCC full data daily analysis version 2018 through ETCCDI indices and comparison with station observations over mainland of China. Theor. Appl. Climatol., 142, 835845, https://doi.org/10.1007/s00704-020-03352-8.

    • Search Google Scholar
    • Export Citation
  • Zhang, J., and Y. Chen, 2018: Typhoon Nida’s rainfall characteristics in Guangzhou city based on Doppler radar estimation. IOP Conf. Ser. Earth Environ. Sci., 169, 012002, https://doi.org/10.1088/1755-1315/169/1/012002.

    • Search Google Scholar
    • Export Citation
  • Zheng, F., and Coauthors, 2018: Crowdsourcing methods for data collection in geophysics: State of the art, issues, and future directions. Rev. Geophys., 56, 698740, https://doi.org/10.1029/2018RG000616.

    • Search Google Scholar
    • Export Citation
  • Zhu, Y., S. Zhang, Y. Li, H. Lu, K. Shi, and Z. Niu, 2020: Social weather: A review of crowdsourcing-assisted meteorological knowledge services through social cyberspace. Geosci. Data J., 7, 6179, https://doi.org/10.1002/gdj3.85.

    • Search Google Scholar
    • Export Citation
  • Fig. 1.

    Graphic overview of replies to the questionnaire collected from 21 valid responses. The legends for each result are shown above the graphics.

  • Fig. 2.

    Distribution of surface stations that passed QC. Blue dots, purple triangles, and red squares represent CMA stations, MWR stations, and Pie stations, respectively. Dark-orange lines outline different districts within Guangzhou City, and the white part indicates the coastal area. White solid lines divide the map into a 0.05° × 0.05° grid.

  • Fig. 3.

    Distribution of (a) CP of CMA grid, (b) CP of COM grid, (c) CP of COM grid minus CMA grid, (d) PI of CMA grid, (e) PI of COM grid, and (f) PI of COM grid minus CMA grid during June–August in 2022.

  • Fig. 4.

    Percentage distribution of precipitation in different ranges counted on the grids with COM stations in them but without CMA stations.

  • Fig. 5.

    Distribution of (a) HRHs of CMA grid, (b) HRHs of COM grid, (c) HRHs of COM grid minus CMA grid, and (d) CSI of HRHs during June–August in 2022.

  • Fig. B1.

    Comparison of hourly precipitation from MWR stations and CMA stations: std is standard deviation, cc is correlation coefficient, and n is sample number.

  • Fig. B2.

    As in Fig. B1, but comparing Pie stations and CMA stations.

All Time Past Year Past 30 Days
Abstract Views 150 150 0
Full Text Views 221 221 32
PDF Downloads 217 217 33