Improving hurricane heat content estimates

Hurricanes are amongst the most destructive natural disasters known to mankind. The primary energy source driving these storms is the latent heat release due to the condensation of water vapor, which ultimately comes from the ocean. While the Sea Surface Temperature (SST) has a direct correlation with wind speeds, the oceanic heat content is dependent on the upper ocean vertical structure. Understanding the impact of these factors in the mutual interaction of hurricane-ocean is critical to more accurately forecasting intensity change in land-falling hurricanes. Use of hurricane heat content derived from the satellite radar altimeter measurements of sea surface height has been shown to improve intensity prediction.

The general approach of estimating ocean heat content uses a two layer model representing the ocean with the anomalies from the altimeter. Although these estimates compare reasonably well with in situ measurements, recent studies show that there is scope for further improvement. Our objective is to develop a methodology to more accurately represent the upper ocean structure using in situ data. As part of a Global effort, more than 3000 profiling floats acquire temperature and salinity vertical structure over vast regions of world's ocean. Currently, over 260,000 individual profiles are available for comparison to altimetric heat content. Using these profiles of temperature and salinity data, continuous profiles of density are derived. Ocean heat content from these profiles are then estimated and compared to those derived from altimeter data. Using a procedure that conserves density in the vertical, the continuous density profiles are discretized in to multiple layers representative of the upper ocean for specified weeks. Statistical correlations are then derived between the altimetric sea surface height anomalies and thickness anomalies of these layers. Using these correlations, a higher resolution upper ocean structure is derived from the altimeter data. Withholding observations from one snapshot of data in the correlations and comparing the estimated ocean heat content with in situ values will allow us to quantify errors in this approach. Results from this approach in selected regions of the world's oceans will be presented in this paper.