College of Engineering | Department of Atmospheric Science | Apply to CSU
Disclaimer | Equal Opportunity Statement | Privacy | Search CSU
words

Locally Extreme Precipitation Forecasts

UNDER CONSTRUCTION!

Welcome to the webpage for extreme precipitation forecasting with machine learning. This page contains real time CONUS-wide probabilistic forecasts for excessive rainfall, analogous to Weather Prediction Center's Excessive Rainfall Outlooks, based on Average Recurrence Interval exceedances. Several model formulations are currently running in real-time. 2017 models for Days 2-4 predict Stage IV exceedances of 1- and 10-year ARIs. 2018 models generate a single probabilistic field, incorporating 1-year 24-hour ARI exceedances of CCPA, MRMS, or Stage IV QPE in addition to Flash Flood Local Storm Reports into a single field, and also uses a larger 40 km neighborhood radius. Day 1 models running in real-time are proof-of-concept models developed based on historical runs of the convection-allowing NSSL-WRF model, otherwise using an identically defined predictand and same underlying formulation as Day 2/3 models. Forecasts are made for the 24-hour Day 1 period, as well as products specific to the 18-00 and 00-06 UTC periods. GEFS-based models are currently running separetely off of both the GEFS/R model output and operational GEFS output ('-O' suffix).

2017 model development and performance characteristics are described in a two-part publication in Monthly Weather Review: Part I and Part II. Methodology of 2018 models and other pending developments will be described in forthcoming publications. Methodology is currently under development and subject to change.

DISCLAIMER:These forecasts are provided for informational purposes only; the creator assumes no responsibility for the forecasts or actions of others derived therefrom. This disclaimer applies to all forecasts provided on this page or any other forecast product presented elsewhere on this website.

Please send any comments, feedback, or suggstions to: gherman at atmos dot colostate dot edu

Currently Supported Forecast Lead Times:

- F12-F36

- F18-F24

- F24-F30

- F36-F60

- F60-F84

- F84-F108

Initialization Date:

Year:
Month:
Day:


Hover mouse over an hour to see forecasts for forecasts for precipitation accumulations beginning at the given start time (in hours since initialization).

2018 Models 1218243636-O6060-O

2017 Models 3636-O6060-O84