Saturday, November 5, 2011

Maps of Thai Floods and Inside the War Room

In addition to the usual official sources of flood forecasts and the webpage by Google (mentioned previously) there are several new (to me at least) great map-based interfaces to Thai flood information. If these are new, it would be interesting to know if, in this age of the Internet and data sharing, things like this would pop up in other countries like the US or Australia. In my own experience, there has been a lot of interest in realtime flood mapping and prediction from a research-applications sense, but it's rare to see this information being so well shared and repackaged.

Much of the official english-language resources (e.g. summary bulletins) are available at this page: However, a picture is worth a thousand words and the people at our hotel in Bangkok were mostly glued to Thaiflood, refreshing a couple times an hour:

For finer spatial resolution the Thailand Flood Monitoring System has some comparisons of the current situation to a past few years of floods (this is just flood monitoring and historical reference, not forecasts)

The Flood Monitoring System is a clearing house for satellite data, the office of which is shown below (and there's a gallery of photos of the office in action)

The GIS "War Room"

This "Longdo" page has more flood monitoring, including realtime video cameras for various roads:
Longdo also has overlays of flood direction

This United Nations sponsored webpage UNITAR-UNOSAT also allows people to upload Geotagged digital photos and combines it with various layers of information, such as flood extent:
Red color is flooded areas, green circles are locations/directions of photos
Again, all this is just flood monitoring, but for maps of forecasts, the Bangkok Post recently posted an image from the Disaster Warning Centre, describing a "Worst-case scenario". This somewhat resembles Bangkok's land elevation map, the medium blue areas in the south center area being the lower elevations  (and hence some of the deeper flooding).

A phrase like the "worst case scenario" is a double-edged sword. This is the kind of information that decision-makers crave and request from forecasters. But what is the chance of this worst case scenario happening? 1 in 3? 1 in 10? 1 in 100? History is filled with examples of where the actual outcome was above the worst case scenario and decisionmakers ended being very resentful of this. Of course, forecasters could instead conjure up a wildly high scenario (e.g. take the above map and multiply it by 3), but this leads to wasteful overplanning and its own form of resentment. 


  1. It seems to me that "worst case scenario" is inconsistent with a continuous probabilistic forecast.

  2. Daniel,

    You're right. A good forecast would include a complete set of statements like, say, "the river has:

    1 in 2 odds of getting above 3 meters
    1 in 10 odds of getting above 4 meters
    1 in 100 odds of 5 meters"

    and so on.

    Things get really crazy on the very rare end of things. What is the water level for 1 in a trillion odds? Maybe that's the case for when an asteroid hit the ocean, causing an inconceivable wall of water to rip around the planet.

    On the other hand, imagine if you had side pains and you went to a doctor. It's likely indigestion but could be something more serious, like a problem with your appendix. What might the doctor say is the "worst case scenario"? You'd need surgery and a few days recovery?

    However, an even worse but very outlandish scenario would be that you've been infected by a mutant space virus and now all humanity is going to die. Literally, it's true. Most people wouldn't want to go to a doctor with such a wild imagination though.

    Thanks for the comments!