Scan barcode
A review by mburnamfink
The Weather Machine: A Journey Inside the Forecast by Andrew Blum
5.0
Blum has a talent for capturing the social aspect of sociotechnical systems, in this case the weather forecast that is an entirely unremarkable part of our lives. Weather forecasting has improved immensely, even in a few decades, and this book is how those forecasts are made worldwide.
The first forecasts were local, limited by what the metrologist can see and measure with their own eyes and perhaps a barometer. But weather systems span continents, and tomorrow's rain is often already pouring some miles to the west. In the 19th century, metrologists, with a surprising contribution from Norwegians, began a process of systematic observations reported by telegraph and telephone. While physicists had dreams of being able to calculate the weather, including an immense cathedral amphitheater of 65000 human calculators, the partial differential equations were far beyond abilities at the time. Instead, forecasters relied of crude pattern recognition, matching the fronts and winds they saw with with historical weather to provide a guess which was moderately skillful, one more precise than the simple average for that time and date.
Weather is continuous, but observations are only collected at discrete stations and times. Good data was a matter of massive international cooperation, since only a handful of nations were large enough to meaningfully encompass weather systems. The Cold War, with its advances in rocketry and computers, brought weather prediction into its modern form. Blum sees the launch of a NASA satellite with a billion dollar sensor for measuring surface moisture, and then goes to the European Centre for Medium-Range Weather Forecasts, a supercomputer site that gulps in torrents of data from satellites and ground stations and has incrementally produced a model accurate to five to eight days out. You might remember the Euro from hurricane season, when different models try to figure out where the big storm is going to make landfall.
Weather is international science diplomacy, but all is not well. While the traditional system was egalitarian, there's now a split between major powers capable of launching satellites and running supercomputers, and the minor nations still dependent on freely available forecast data. Private companies pose another threat, including US giant The Weather Company, which supplies forecasts to consumers, and a potential future of big data from millions of small sensors, collected and collated by a tech titan.
Blum keeps it light, focusing on the people and places of weather forecasting, so while a nerd may want a more in depth book, this is one that ordinary people can finish.
The first forecasts were local, limited by what the metrologist can see and measure with their own eyes and perhaps a barometer. But weather systems span continents, and tomorrow's rain is often already pouring some miles to the west. In the 19th century, metrologists, with a surprising contribution from Norwegians, began a process of systematic observations reported by telegraph and telephone. While physicists had dreams of being able to calculate the weather, including an immense cathedral amphitheater of 65000 human calculators, the partial differential equations were far beyond abilities at the time. Instead, forecasters relied of crude pattern recognition, matching the fronts and winds they saw with with historical weather to provide a guess which was moderately skillful, one more precise than the simple average for that time and date.
Weather is continuous, but observations are only collected at discrete stations and times. Good data was a matter of massive international cooperation, since only a handful of nations were large enough to meaningfully encompass weather systems. The Cold War, with its advances in rocketry and computers, brought weather prediction into its modern form. Blum sees the launch of a NASA satellite with a billion dollar sensor for measuring surface moisture, and then goes to the European Centre for Medium-Range Weather Forecasts, a supercomputer site that gulps in torrents of data from satellites and ground stations and has incrementally produced a model accurate to five to eight days out. You might remember the Euro from hurricane season, when different models try to figure out where the big storm is going to make landfall.
Weather is international science diplomacy, but all is not well. While the traditional system was egalitarian, there's now a split between major powers capable of launching satellites and running supercomputers, and the minor nations still dependent on freely available forecast data. Private companies pose another threat, including US giant The Weather Company, which supplies forecasts to consumers, and a potential future of big data from millions of small sensors, collected and collated by a tech titan.
Blum keeps it light, focusing on the people and places of weather forecasting, so while a nerd may want a more in depth book, this is one that ordinary people can finish.