Premium
High‐Resolution Simulation Study Exploring the Potential of Radars, Crowdsourced Personal Weather Stations, and Commercial Microwave Links to Monitor Small‐Scale Urban Rainfall
Author(s) -
Vos L. W.,
Raupach T. H.,
Leijnse H.,
Overeem A.,
Berne A.,
Uijlenhoet R.
Publication year - 2018
Publication title -
water resources research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.863
H-Index - 217
eISSN - 1944-7973
pISSN - 0043-1397
DOI - 10.1029/2018wr023393
Subject(s) - environmental science , weather radar , radar , meteorology , microwave , remote sensing , weather station , correlation coefficient , sampling (signal processing) , grid , computer science , statistics , geography , mathematics , telecommunications , geodesy , detector
Abstract Many applications in urban areas require high‐resolution rainfall measurements. Typical operational weather radars can provide rainfall intensities at 1‐km 2 grid cells every 5 min. Opportunistic sensing with commercial microwave links yields path‐averaged rainfall intensities (typically 0.1–10 km) within urban areas. Additionally, large amounts of urban in situ rainfall measurements from amateur weather observers are obtainable in real‐time. The accuracy of these three techniques is evaluated for an urban study area of 20 × 20 km, taking into account their respective network layouts and sampling characteristics. We use two simulated rainfall events described in terms of drop size distributions on a 100‐m grid and with a temporal resolution of 30 s. Accurate radar rainfall estimation with the Z ‐ R relationship relies heavily on an appropriate choice of parameters, and a dual‐polarization strategy is more suitable for higher intensities. Under ideal measurement conditions, the weather station network is the most promising, with a Pearson correlation coefficient above 0.86 and a relative bias below 4% for 100‐m rainfall estimates at 5‐min resolution. Microwave link rainfall observations contain the largest error, shown by a consistently larger coefficient of variation. The accuracy of all techniques improves when considering rainfall at larger scales, especially by increasing time intervals, with the strongest improvements found for microwave links for which errors are largely caused by their temporal sampling. Sparser networks are examined, showing that the decline in measurement accuracy only becomes significant when the link and station network density are reduced to less than half their levels in Amsterdam.