Urban climate monitoring network design: Existing issues and a cluster-based solution

Chen, Xuan; Yang, Jiachuan

doi:10.1016/j.buildenv.2022.108959

Cited by 8 publications

(6 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The station density and selection strategy has an impact on the interpolation performance [52,53] . The kmeans clustering strategies have been used for the optimization of weather station density and location [19,54] . The k-means strategy clusters data by separating all data points into a certain number ( ) of k groups or components with similar characteristics based on minimal within-cluster variances (squared Euclidean distances) which are calculated based on the data matrix representing each data point's characteristic.…”

Section: Selecting Weather Stationsmentioning

confidence: 99%

“…However, previous studies only considered coordinates or local climate characteristics in clustering [19,54] . Local land cover is not involved in station location selection, while it is believed to have a signi cant impact on local air temperature and should be considered when deploying weather stations [20] .…”

Section: Selecting Weather Stationsmentioning

confidence: 99%

“…Conventional meteorological observational networks were designed for the detection of synoptic atmospheric conditions, and thus are sparsely distributed and cannot capture the full picture of intra-urban climate differences [18] . Besides, monitoring networks currently in operation in most cities are designed based on administrative districts and the population [19] . Thus, it may not be the ideal strategy to obtain representative observations at urban sites.…”

Section: Introductionmentioning

confidence: 99%

“…Despite its great potential, rigorous data quality assessment is required, and it is still a major challenge to bene t from such novel datasets [21,22] . Some cities have started to implement dense monitoring networks to observe the intra-city variation of urban climate [19] . A dense monitoring network has been built and put into use in Guangzhou, China, with 321 automatic weather stations (AWS) during our study period, which has a mean sample spacing of 5 km (approximately 25 km 2 /sample).…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Machine learning-assisted mapping of city-scale air temperature: Using sparse meteorological data for urban climate modeling and adaptation

Ding

Zhao

Fan

et al. 2023

Building and Environment

View full text Add to dashboard Cite

Section: Selecting Weather Stationsmentioning

confidence: 99%

Section: Selecting Weather Stationsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Machine learning-assisted mapping of city-scale air temperature: Using sparse meteorological data for urban climate modeling and adaptation

Ding

Zhao

Fan

et al. 2023

Building and Environment

View full text Add to dashboard Cite

Section: Selecting Weather Stationsmentioning

confidence: 99%

Machine learning-assisted mapping of city-scale air temperature: Using sparse meteorological data for urban climate modeling and adaptation

Ding

Zhao

Fan

et al. 2023

Preprint

View full text Add to dashboard Cite

The availability of high spatio-temporal resolution of urban air temperature is paramount for understanding urban heat island (UHI) and developing effective mitigation strategies, in particular for local-scale mitigations. Obtaining high spatial resolution of air temperature at city-scale is challenging as the quantity of weather stations is often limited in cities, particularly in those less developed ones. In this study, based on the existing weather station network in Guangzhou city, China, we compare eight different air temperature interpolation models and select one with the best performance to interpolate city-scale air temperature. The training and validation of the models are performed using observatory meteorological data of 321 weather stations in Guangzhou. Deep learning-derived land cover information and social-economic data are encoded to be used as explanatory variables. The regression kriging combined with multiple linear regression is found to result in the best performance, with an average root mean squared error (RMSE) of 0.92℃ and a coefficient of determination (R2) of 0.959. Furthermore, the quantities and locations of current weather stations can be optimized by the proposed model. Guided by the k-means clustering alongside the information of geocoordinates and land cover, the number of current weather stations in Guangzhou can be reduced by 50% (i.e., 160 weather stations) while retaining the model performance. This study proposes and demonstrates an effective model for obtaining city-scale air temperature at high spatio-temporal resolution with data from sparse weather stations, which is much needed for cities which want to enhance their city-scale air temperature mapping by complementing new weather stations to their existing weather station network.

show abstract

Low-cost urban carbon monitoring network and implications for china: a comprehensive review

Jiang,

Han,

Zalhaf

et al. 2023

Environ Sci Pollut Res

View full text Add to dashboard Cite

Urban climate monitoring network design: Existing issues and a cluster-based solution

Cited by 8 publications

References 46 publications

Machine learning-assisted mapping of city-scale air temperature: Using sparse meteorological data for urban climate modeling and adaptation

Machine learning-assisted mapping of city-scale air temperature: Using sparse meteorological data for urban climate modeling and adaptation

Machine learning-assisted mapping of city-scale air temperature: Using sparse meteorological data for urban climate modeling and adaptation

Low-cost urban carbon monitoring network and implications for china: a comprehensive review

Contact Info

Product

Resources

About