Integration of ground and satellite radar data for precipitation monitoring and estimation
Date Issued
May 2025
Author(s)
Advisor
Abstract
This work investigates the challenges and solutions associated with the use of ground-based X-band weather radars in Cyprus for quantitative precipitation estimation (QPE). The study addresses two primary limitations of these radars, i.e., attenuation and calibration uncertainties, by employing spaceborne observations from the Global Precipitation Measurement Mission (GPM). In the first phase, a constrained attenuation correction approach is applied, assuming the path integrated attenuation equals the difference between ground radar and GPM Dual-frequency Precipitation Radar (DPR) reflectivity. Various parameter combinations are tested, highlighting the need for event- and radar-specific adjustments. The second phase focuses on radar calibration using volume-matched GPM DPR Ku-band reflectivity. Multiple filtering and thresholding schemes are evaluated, with the most consistent schemes being combined to derive stable calibration periods and determine the final offset. Finally, a dual-stage machine learning framework is introduced to convert raw reflectivity into rainfall rate estimates. The first stage corrects the ground raw reflectivity using GPM DPR Ku-band volume-matched reflectivity, while the second stage estimates rainfall using ground truth from rain gauges. The performance of the latter is compared with traditional Z-R methods. This research presents the first comprehensive effort to validate, correct, and prepare the Cyprus radar network for operational use. The findings demonstrate the potential of integrating satellite data, ground-based observations, and machine learning techniques to enhance the reliability of ground-based radar QPE in Cyprus.
File(s)![Thumbnail Image]()
![Thumbnail Image]()
Name
Loulli_PHD_2025_abstract.pdf
Size
92.89 KB
Format
Adobe PDF
Checksum (MD5)
94e81f7b3d641326d50935edb06ecae3
Name
Loulli_PHD_2025.pdf
Size
25.34 MB
Format
Adobe PDF
Checksum (MD5)
af068239285967044e9438a4e1749192