weather


We wrote about this sometime back, that too referring to the Economist. Seems that Kenneth Cukier and Abu Saeed Khan are interested in the same kinds of things. But earlier, the talk was about reporting rain. Now it’s about predicting, which is way more interesting: Though it is useful to know how much rain is falling right now, forecasting is even better. Telecoms data promise to make this easier as well.
The government predicted rainfall more than 150 mm on the 25th of May. Over 500 mm of rain fell. Technically, they were not wrong (550 mm is within the range of “more than 150 mm”), but obviously, forecasts like this might as well not be made. [an error was corrected in the above para] But it is wrong to condemn the Met Department which operates even without Doppler radar, though they have been talking about it since 2012. But as discussed below, Doppler radar is old and can only tell about large rain drops.
It’s a little odd to use a concept like vertical integration but that seems to best explain what IBM is doing. IBM calls what Watson does “cognitive computing,” heralding an age of machines that supposedly think. What the company has not figured out is how to make this into an engine of growth. The tech giant has had years of shrinking revenue, but says its investments in Watson will take time to bear fruit. It will be picking up more talent in the deal.
Just today, my friend Abu Saeed Khan was telling me this was something the mobile operators of Bangladesh could easily do. And here, the Economist carries a story about how the Dutch have done it. Each technique has its upsides and downsides. Radar and satellites can cover swathes of land, yet they lack detail. Gauges are much more accurate, but the price of that accuracy is spotty coverage.
Yesterday, I was talking with an Indian colleague who was involved in improving the Indian weather information system based on INSAT while working for the Indian Space Research Organization (ISRO). The trigger had been the devastating cyclone that hit Andhra Pradesh in 1977. This was also related to initiating my interest in disaster early warning because that cyclone was supposed to hit the East Coast of Sri Lanka, but veered away at the last minute. I remember tracking news of its journey while working at the Sri Lanka Broadcasting Corporation. One year later the cyclone did not change track and we lost over 900 people on the East Coast.