-
- H24A-07: TxRain: A Bayesian framework for integrating historical observations and model projections to develop nonstationary IDF curves
-
NOLA CC
Author(s):Generic 'disconnected' Message
James Doss-Gollin, Rice University (First Author, Presenting Author)
Yuchen Lu, Rice University
John Nielsen-Gammon, Texas A&M University
William Baule, Michigan State University
Srikanth Koka, Texas Water Development Board
Statistical assessments of extreme rainfall, used to design critical infrastructure, price insurance, and map floodplains, have traditionally assumed a stable climate. As climate change invalidates this assumption, there is a clear need to develop 'nonstationary' models that account for changing storm characteristics. However, simply adding trends to these models is not a straightforward solution; it increases their complexity and can introduce significant uncertainty, especially when relying on limited historical data. This creates a dilemma: weather station records are accurate but often too short to reliably quantify long-term changes, while climate model simulations offer long-term projections but suffer from their own systemic biases. To resolve this, we developed a statistical model that integrates these two data sources. We apply this framework to study rainfall across Texas, leveraging a uniquely rich set of station observations. We discuss our method, findings, and implications for ongoing projects such as NOAA Atlas 15.
Scientific DisciplineNeighborhoodType
Enter Note
Go to previous page in this tab
Session
