Fairhaven, The River

Good Hurricane Forecasts Considered Harmful

Good hurricane forecasts may have reached the point of doing more harm than good. Since 1970 the American Meterological Society has editorialized against the bad hurricane policies of the US. In particular, the coastal development planning and regulation emphasizes evacuation and encourages people to live in areas at risk. Evacuation is a brittle protection. Failures are complete, not partial. The recommended policies are mitigations that reduce risk and provide partial protection even when the hurricane overwhelms the protections.

This is a subject that simple linear text handles poorly. So I'm experimenting with a diagram of the interactions.

(My first experiment was with SVG. That doesn't work so well with Typepad. It puts in a nice link that you can follow, so if your browser supports SVG you then get to see it. This diagram does OK as an image, so it's in as an image with thumbnail.)

GPS and COSMIC

COSMIC is proceeding into operation. It is a really neat idea that has gone from crazy notion to operational over the past decade. The basic idea is that one satellite (COSMIC) watches another satellite (GPS) pass from above the Earth horizon to below. With very precise and accurate receivers, a lot of computer power, and good atmospheric models you can derive the atmospheric temperature and humidity profile from the radio distortions. This animation illustrates the result.

GPS was chosen as the signal source because the GPS signals are extremely precise, well controlled, and at predictable locations. The first few experiments were piggy backed on other satellites to save money. There was a lot of doubt whether it would actually work in practice.

Funding has come from Taiwan and the United States. This is of special interest to Taiwan because it addresses the huge data gaps in the pacific. The GPS/COSMIC soundings are less detailed than those from weather balloons, but there are only a few islands in the Pacific to take balloon soundings. If the theory worked, COSMIC would capture 5-15,000 soundings per day, spread uniformly around the globe. That would improve pacific forecasts substantially.

It worked in the experiments and a low cost constellation of 6 COSMICs are now in orbit. They are working out the problems and adjusting firmware and ground data issues. The logs at COSMIC give a hint at the reality of running satellites. They are not launch and forget.

As they come online there will be a series of experimental analyses and trials of different methods for incorporating the data into routine weather prediction. It always takes some research to figure out how to incorporate a new data source with random measurement times and locations, and with new and different accuracy and precision characteristics. In a few years the Pacific, Antarctica, and other data poor areas will see better weather forecasts.

The COSMIC sites also talk about climate, but that is mostly for PR and funding purposes. There is government funding to be found in climate work. The big win will be in weather forecasting. The climate analysis is just a side show. The largest climate impact might be from improved fuel use due to better forecasts that enable better routing decisions by ships and airplanes. It might be less property destruction due to smaller forecast errors on typhoons.