Evaluating Tropical Cyclone Intensity Forecasts from HAFS Using Satellite-Derived Ocean Profile and Heat Content Estimates

Accurately modeling and forecasting tropical cyclone (TC) intensity changes, especially from rapid intensification (RI) and rapid weakening (RW) events, is challenging. To date, much of the work to improve forecasts of RI/RW have focused on enhancing forecasting models’ representation of the atmosphere. However, both RI and RW events can be strongly influenced by the heat content and stratification of the upper ocean, particularly for larger and slower-moving TCs. As upper ocean heat content increases, TCs may become more dangerous and costly for coastal communities since this heat can increase TC intensity and precipitation amounts. Better information about upper ocean conditions can thus aid in the evaluation and potentially the improvement of RI/RW forecasts.

At the National Oceanic and Atmospheric Administration (NOAA), a new product is being developed to replace the current, operational NOAA National Environmental Satellite, Data, and Information Service (NESDIS) Satellite Ocean Heat Content Suite (SOHCS) product. The algorithm underlying this new product, known as the Next Generation Enterprise Ocean Heat Content (NGE OHC) algorithm, is an empirical parameterization that generates synthetic depth-resolved ocean temperature and salinity profiles. These daily profiles will initially be used operationally to evaluate ocean model initialization, and then ultimately to support the initialization of TC forecast models using data assimilation. Additional upper ocean metrics of interest will be calculated from these profiles, such as Tropical Cyclone Heat Potential (TCHP), mixed layer depth, and near-surface stratification strength, to improve statistical predictions of intensity (e.g., using Statistical Hurricane Intensity Prediction Scheme (SHIPS)) and to enable detailed analysis of the potential oceanic contributions to RI/RW. The NGE OHC method leverages known ocean dynamics from relatively sparse in situ data to project satellite data into the interior to directly estimate subsurface profiles at high vertical and horizontal resolution. In case studies from the 2020–2022 Atlantic hurricane seasons, daily oceanic conditions from the NGE OHC are retrospectively compared with in situ Argo float observations, model initializations from the Hurricane Analysis and Forecast System (HAFS version 1.0B), and the SOHCS product to illustrate examples of when the NGE OHC would have helped to improve the model’s initial ocean state in the Gulf of Mexico and Caribbean Sea. These comparisons show that many HAFS forecasts with intensity busts miss near-surface heat and barrier layers (which inhibit ocean mixing) found in observations and the NGE OHC synthetic profiles.