It’s the Physics (Dummy)—RA2 Contributions to Hurricane Modeling and Tropical Cyclone NWP (Invited Presentation)

I made the most consequential decision of my life in 1971 when I chose the ‘new guy’ as my undergraduate advisor. Dr. Anthes had recently joined the Penn State Meteorology Department after several years in Miami developing an asymmetric numerical hurricane model. From this first association with Dr. Anthes in 1971, to the present, I would be involved with tropical cyclones (TCs) and modeling at the main US TC forecast centers (JTWC and NHC) and most of the major NWP centers (FNMOC, NCEP, ECMWF and JMA) – an absolutely wonderful career thanks to Dr. Anthes and many of his earliest PhD students, especially Tom Warner and Jim Hoke.

My talk will briefly review TC NWP (1972-2017) and Dr. Anthes’ seminal contributions to what I consider the greatest success story in the history of NWP – the 85% reduction in the area defined by the day three mean TC position error. The only way this tremendous advance in TC forecasting was possible is with improved modeling of the tropics and this is precisely what we did at Penn State under Dr. Anthes’ leadership.

The basic principles of the Anthes-inspired TC modeling of the 1970s remain valid to this day, namely:

• Interactive physics – boundary layer and deep convection (and radiation and…). To forecast a TC, you must model the physics driving the system. This was a huge computational issue solved by Tom Warner – to get the maximum physics out of the tiny computers of the time (1/10th of an iPhone) – why I say: it’s the physics dummy
• Assimilate TC-critical observations – surface wind, rainfall and the TC vitals (position, intensity, wind radii)
• Use the model in the assimilation process – 4D Dynamic Initialization (a precursor to full 4DVAR)

Finally, on a personal note, I will comment on the limited progress since my 1978 M.S. thesis research in making the first TC NWP forecast with operational analyses in a ‘full-physics’ model (MM0) – vortex initialization is still a huge problem in the limited-area models (e.g., HWRF) and rainfall assimilation has yet to be implemented. Regardless, the vast improvement in TC forecasting would not have been possible without the pioneering work of Dr. Anthes and, perhaps more importantly, his inspiration and mentoring of subsequent generations of modelers.