Many weather events are signalled by changes in sky brightness.When the sky darkens quickly, people know to take cover from impending rain. Today, it is easy to make cheap devices to measure sky brightness, and this website explains how a network of such devices could be built and networked by schoolchildren. This opens up many educational opportunities for students with a wide range of interests, and students would be motivated by hearing of their work in weather reports.
See also: methods and results.
A darkening sky is a good indicator of the onset of bad weather. On an unsettled day, we keep an eye on the sky, to decide whether to run to the shelter before a downpour. Although weather forecasts do very well at predicting weather variations over days and hundreds of kilometers, people have still rely on the sky for quick, localized forecasts.
A limitation of this observer-based forecasting is that even small-scale weather patterns are not entirely localized. Events such as fronts and thunderstorms tend to sweep across regions, so that one neighborhood could warn another of impending bad weather.
The idea of this project is to develop a network of skylight sensors that would mimic the human eye, testing for darkening of the sky hour by hour, or even minute by minute. The network would need to cover populated areas in a reasonable data density, say one sensor per square kilometre or so. The national weather service could do that in a professional way, probably at great expense for equipment, networking, and maintenance.
But there is another way–put the sensors in schools. Let the students build the sensors, network them, write the code to integrate them, and perform regular calibration and other maintenance. Each of the steps would provide students with fun, and good learning experiences. Building the sensors would help them to learn about the physics of light sensing and about the electronics of data logging. Networking and coding provide the chance to learn about computing hardware and software. And the maintenance would do two important things. First, it would continually reinforce the lessons about physics and electronics, spreading the learning across the entire class, so that it does not end at the construction phase. Second, it would underline for the students the importance of assessing and documenting data quality, a topic that occupies scientists for much of their time, but that is seldom discussed in textbooks.
The sensors could be built for under 100 dollars each, putting them within the reach of many schools. The networking would be free, since the sensors could be connected to existing networked computers. And the data compilation and presentation on the web would be free, with work done by students themselves. Schools are spaced at a distance of a few kilometers in towns and cities, so the data sampling density would be very high. (There is no reason that students would not put these sensors on their homes, either, yielding even higher data densities.)
Beyond some initial proof-of-concept work on the electronics and on computer security, there would be no need for wide organization of the project. The students could run it, themselves. One model would be a single hosting computer, perhaps at Environment Canada. But another model, probably of more interest to the students, would be a distributed model, using peer-to-peer methods.
As to the name of the project, and a logo to distinguish it, and an aesthetic theme for the project website, these are things best decided by students, and provide avenues for participation of students whose interests focus on communication skills.
This project may be interesting to students even at the very early stages, and should garner more and more attention as widening data streams suggest widening application.