[cosc-grad-students-list] RESEARCH PROPOSAL SEMINAR NOTICE - Abdullah Bdeir

[Eulenfeld, Menda]

RESEARCH PROPOSAL SEMINAR NOTICE
GEOSPATIAL COMPUTER SCIENCE PROGRAM
SCHOOL OF ENGINEERING AND COMPUTING SCIENCES
TEXAS A&M UNIVERSITY-CORPUS CHRISTI
SUBJECT:Pixel-wise Fog Forecasting Using HRRR and METAR Data over Texas:
Spatiotemporal Deep Learning Framework
SPEAKER: Abdullah Bdeir
DATE: March 23, 2026
TIME: 02:00 p.m.
PLACE: Island Hall 156

ABSTRACT

Fog is a hazardous weather phenomenon that significantly disrupts aviation, marine transportation, and roadway safety, particularly in coastal regions such as southern Texas. Accurate fog forecasting remains challenging due to the complex atmospheric processes involved and the sparse distribution of surface observations. This dissertation proposes a pixel-wise fog forecasting framework that integrates meteorological variables from the High-Resolution Rapid Refresh (HRRR) numerical weather prediction model with surface observations from METAR stations derived from ASOS/AWOS networks.
The proposed framework addresses the mismatch between point-based observations and grid-based numerical models by labeling HRRR grid cells using METAR-reported visibility conditions at station locations. To overcome the limited spatial coverage of observations, a semi-supervised learning strategy is employed. METAR observations provide strong labels at station-overlapping grid cells, while heuristic fog classifications derived from HRRR atmospheric variables generate weak pseudo-labels across non-observed grid cells. A weighted loss function combines both supervision sources, enabling machine learning models to learn spatially consistent fog patterns and generate continuous fog probability maps across the study domain. The research is organized into three objectives: (1) evaluating HRRR visibility as a proxy for observed fog, (2) developing spatially explicit Deep learning models for pixel-wise fog prediction, and (3) comparing the proposed framework with operational forecasting systems such as HRRRv4 and HRRR-Cast, HREF and SREF. The results aim to improve fog prediction skill and enhance early warning capabilities for transportation safety and emergency response.



Best,
Abdullah Bdeir
Graduate Research Assistant
iCore Lab
Conrad Blucher Institute for Surveying and Science
Texas A&M University - Corpus Christi
6300 Ocean Drive
Corpus Christi, TX 78412
abdeir at islander.tamucc.edu<mailto:abdeir at islander.tamucc.edu>

──────────
Menda Eulenfeld is inviting you to a scheduled Zoom meeting.
Join Zoom Meeting
https://tamucc.zoom.us/j/95510672188?pwd=jGVSPn3QKkfBURx1X2Giargfstgr62.1

Meeting ID: 955 1067 2188
Passcode: 540092

---

One tap mobile
+13462487799,,95510672188#,,,,*540092# US (Houston)
+16694449171,,95510672188#,,,,*540092# US

---

Join by SIP
• 95510672188 at zoomcrc.com
Passcode: 540092

Join instructions
https://tamucc.zoom.us/meetings/95510672188/invitations?signature=2yVlqcWNHTjICSLrq8_WpmGEdL_koWP060fJR_hnOGg


──────────
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.tamucc.edu/pipermail/cosc-grad-students-list/attachments/20260323/12da2c3d/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/calendar
Size: 10571 bytes
Desc: not available
URL: <http://listserv.tamucc.edu/pipermail/cosc-grad-students-list/attachments/20260323/12da2c3d/attachment-0001.ics>