Typhoon Gaemi (WP052024) is making landfall on Taiwan over the next 24 hours. It is an intense typhoon, estimated winds by the US Joint Typhoon Warning Center of 125 knots. Although other forecast agencies such as the official China Meteorological Agency are a bit lower, the higher number is supported by some satellite analysis.
In any event, the impacts are likely to be catastrophic:
On this track and intensity, the eyewall will sweep over the densely populated northern shore of the island, including Taipei. With nearly 16 million people estimated within the hurricane wind swath, impacts will be significant even with preparation and a good emergency management system. Economic impacts for this storm could be incredible. It’s hard to imagine this will be less than $50 Billion, and some of the computer model estimates are topping out at over $100 Billion dollars! So even the low end impact is an eye watering 6.3% of GDP (six times worse than the worst storm to ever hit the US) upwards of 16.4% of Taiwan’s GDP of $791 Billion USD. Of course behind these numbers are people – as noted, a lot of them. So the human toll is likely to be quite high.
The storm is projected to make landfall again Thursday on the mainland, in Fujian province, inflicting another $5 to $7 Billion USD in impacts on upwards of another 5 to 10 million people.
So how are these impact estimates made? As you might imagine, it’s a complex process. The first step is to create a model of the earth – bathymetry (water depth), topography, and land cover. This is used to simulate the event itself (hurricane, earthquake, explosion of some kind) and the forces it will inflict on infrastructure and people, forces like wind pressure, ground motion, wave height, flood water depth and flow rates, etc.
The next piece you need is a model of infrastructure, the construction characteristics and value of that infrastructure and population. This is created from a combination of economic data from the various countries, population estimates from both satellite and local sources, and analysis of satellite data to determine the type and distribution of infrastructure. This is all combined in to an exposure data base that shows the type and density of stuff on the ground that can be damaged:
The final step in the calculation is to apply the forces computed from the hazard model (wind pressure, etc.) on the infrastructure. In the TAOS(tm) modeling system, the model (estimate) of stuff on the ground is typically aggregated in to 0.5km grid squares (about 1/3 of a mile across), although for specialized purposes it can be as find as a 30 meter (~98 foot) grid. Each type of infrastructure or land cover class (depending on the type of simulation) has a unique “damage curve” for each hazard that estimates how bad the damage is depending on the hazard.
The above curves are for wind; there are similar curves (equations, actually) for water, ground motion, and so forth. Although I rarely discuss it here because it is too depressing, we also have vulnerability equations for people to estimate injuries and deaths.
Because no method is perfect, we have multiple models (estimates) for both the storm itself and the infrastructure on the ground. We can also do the estimates for each forecast agency, leading to literally hundreds of simulations of the storm every six hours as we update the tracks. For example, here are the estimates from the 5am forecasts based on the GFS model, the US Joint Typhoon Warning Center, and the Japan Meteorological Agency:
When I create the estimates for the graphics, I summarize all of this and use our award-winning “plain English” impact maps to get a coherent picture of the potential impact of the storm on people and economies.
So, hopefully that gives you an insight in to how this works. It’s an incredibly complex process.
If you like this kind of commentary and analysis, you can subscribe (and donate to keep things going!) by visiting this link.