Improving Tropical Cyclone Intensity Forecasts by Assimilating Ocean Surface Drifter paths with altimeter sea level in a Coupled Atmosphere-Ocean Forecast System

James Carton - (University of Maryland, College Park)

Co-Investigator(s):

Luyu Sun (University of Maryland, College Park)
Stephen Penny (University Of Colorado, Boulder)

Collaborator(s):

Avichal Mehra (National Weather Service, College Park)
Travis Sluka (University Corporation For Atmospheric Research (UCAR))

Abstract:

Tropical cyclones and hurricanes are a hugely destructive risk for coastal communities. This is a proposal to improve the accuracy of tropical cyclone forecast intensities by including sea level observations and near-surface drifter path observations as constraints in a fully operational coupled (atmosphere-ocean) data assimilation forecast system. By building on the new NOAA coupled forecast system this project also serves to advance the use of satellite altimetry in operational weather prediction.

Increases atmospheric model resolution over the past decade have led to improvements in hurricane track forecasts, but have not led to corresponding improvements in intensity forecasts (e.g., Goni et al. 2017; Zhang and Weng 2015). Several recent theoretical and idealized studies suggest that improvements in intensity will require switching to a fully coupled forecast modeling framework as well as adopting a data assimilation system for initial conditions that is also fully coupled (Zhang and Emanuel, 2018, Li and Toumi 2018). Li and Toumi, in particular, show promising results by including HF RADAR currents to set initial conditions for the ocean, but such observations are necessarily sparse and also inaccurate in high wind conditions. Recent work by Sun and Penny (2019) and Sun (doctoral dissertation, 2019) have developed the methodology to assimilate surface drifter paths as well as sea level anomaly as constraints (along with atmospheric variables) in a real-time 1/24-deg coupled forecast system (1m nearsurface).

We propose to build on the results of Sun and Penny (2019) and Sun (2019) to examine how much we can improve historical hurricane forecasts by including information from surface drifter paths and satellite sea level anomaly. The model we will use, in coordination with our NOAA collaborator Dr. Avichal Mehra, is the new NOAA Unified Forecast System (UFS) which couples the FV3 atmospheric dynamical core with a high resolution implementation of ocean model HYCOM and Wavewatch III for the Gulf. The coupled data assimilation we propose to use will leverage complementary projects being conducted at the NOAA Physical Sciences Laboratory by co-PI Penny and collaborators that support development of the “Local Ensemble Transform Kalman Filter (LETKF) in the Hurricane Analysis and Forecast System (HAFS)”. The time periods of particular interest for this proposal are those spanned by recent Gulf surface drifter deployments such as the Grand LAgrangian Deployment (GLAD), and the Global Drifter Program (GDP). Additional oceanographic data includes available L2 altimetry, L3 SST, and hydrographic profiles. Implementation of the Joint Effort for Data assimilation Integration (JEDI) software framework will be greatly aided by the involvement of our JCSDA Collaborator Travis Sluka.

In the previously-supported OSTST project Sun and Penny completed development of the Lagrangian Data Assimilation (LaDA) system using LETKF to directly assimilate surface drifter paths. This proposal extends this work by adding the other components of a coupled data assimilation system in order to move towards an operational fully coupled hurricane coupled forecast system.

Supported by NOAA