Evaluation of Non-Gaussian Data Assimilation Behaviors in Tropical Storm Chris in 2006
Abstract
In 2006, tropical storm Chris developed off the west coast of Africa on July 26th then moved west towards the Lesser Antilles. The forecasts at the time predicted that Chris would go south to the Dominican Republic and hit Cuba or head north through the Bahamas and hit Florida, which could potentially cause substantial damage to these areas. However, this was not the case as Chris weakened into a tropical depression and had little impact. The question we pose is: "Why were the forecasts incorrect and why did the storm collapse?" Our working hypothesis on why the forecasts were incorrect has to do with the data assimilation aspect of the forecasts. Operational data assimilation is based on a Gaussian assumption, which we know is many times violated. Our hypothesis is that the moisture variable is lognormally distributed, and if that lognormality was considered, the data assimilation may have predicted the storm weakening. Our results will examine non-Gaussian behaviors in Chris and its supporting environment, and we plan to present preliminary results from a mixed Gaussian-lognormal data assimilation method using WRF.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2019
- Bibcode:
- 2019AGUFMNG21B0924G
- Keywords:
-
- 3315 Data assimilation;
- ATMOSPHERIC PROCESSES;
- 0555 Neural networks;
- fuzzy logic;
- machine learning;
- COMPUTATIONAL GEOPHYSICS;
- 1640 Remote sensing;
- GLOBAL CHANGE;
- 3275 Uncertainty quantification;
- MATHEMATICAL GEOPHYSICS