Evaluation of uWRF Performance and Modeling Guidance Based on WUDAPT and NUDAPT UCP Datasets for Hong Kong
Other outputs


Full Text

Other information
AbstractThe urban Weather Research and Forecasting Model (uWRF) (WRF BEP/BEM, Salamanca et al., 2010) has been widely used to study the urban boundary layer physics of several major cities. However, its performance in modeling Hong Kong and the Pearl River Delta (PRD) region has received less attention. One reason is the lack of a complete regional dataset of urban canopy parameters (UCPs), i.e. the National Urban Data and Access Portal Tool (NUDAPT) (Ching et al., 2009). The World Urban Database and Access Portal Tools (WUDAPT) (Ching et al., 2014) approach provides an alternative estimation of building morphology dimensions based on satellite retrieved local climate zones (LCZs) (Stewart et al., 2012). Such an approach provides a simple open source means of generating the input data required for uWRF modeling. The implementation of WUDAPT in uWRF simulation involves uncertainties that arise from various sources, as compared with the NUDAPT approach in which more accurate building data are used.

These uncertainties include the following.
(1) The supervised classification of different LCZs in the WUDAPT approach by satellite images introduces uncertainty in identifying the correct LCZs.
(2) Given that the current WUDAPT approach based on LCZs requires selecting UCPs from a range of values associated with each LCZ, the choices may not accurately represent local conditions, thus introducing some degree of uncertainty into the modeling results.
(3) As the WUDAPT approach divides urban grids into classes (LCZs), compared with the continuous UCPs of the NUDAPT approach, this discretization introduces another
source of uncertainty.
(4) The subsampling of WUDAPT and NUDAPT datasets into the uWRF’s domain also generates uncertainty.

The quantification of these sources of uncertainty can
improve the understanding of the efficacy of the WUDAPT model compared with the results from using NUDAPT based UCP values. Some uncertainties could be reduced.
Identifying the source of relatively large but reducible uncertainty could provide a framework and guidance for other regions, thus contributing to progress in the development of next generation WUDAPT levels 1 and 2 datasets, which are more accurate than the current WUDAPT level 0.

In this study, different methods of WUDAPT level 0 preprocessing methods are carried out to isolate the source of uncertainties (1) to (4) above. The corresponding WRF result is compared with NUDAPT, which acts as the baseline “reference”; the extent of deviation of the WUDAPT cases vis à vis the NUDAPT case provides the means to quantify different sources of uncertainty.

After quantifying those different sources of uncertainty, we offer guidance for implementing WUDAPT in uWRF to possibly minimize uncertainty. Then we evaluate the improvement in the performance of uWRF with NUDAPT, along with the corresponding and suggested WUDAPT approach as input, for comparison with the traditional Noah bulk scheme and local and nonlocal planetary boundary layer (PBL) schemes, by considering surface observation station data.
All Author(s) ListMichael Mau Fung Wong, Jimmy Chi Hung Fung, Jason Ching, Peter Pak Shing Yeung, Jason Wai Po Tse, Ran Wang, Meng Cai, Chao Ren
Title of PublicationUrban Climate News (Issue No. 71)
Year2019
Month3
PublisherInternational Association for Urban Climate
Pages19 - 23
LanguagesEnglish-United States

Last updated on 2019-01-04 at 15:37