Data journey

The process for making amateur weather observations has many similarities with official stations such as Weston Park. Typical differences include things such as the number of weather variables recorded, the type of instruments used, the site of the weather station, the frequency of observations and the format of the data.

Different types of amateur observer use different types of equipment to observe the weather. Here we focus on two different set-ups:

• A semi-professional weather station similar to the official stations.

• An experimental weather station built using a Raspberry Pi kit.

Collecting data – Professional quality

Edward Hanna is a Professor of Climate Change at the University of Sheffield. He has been actively observing the weather since he was a child when he began taking manual recordings using a thermometer. Climate change is a long term alteration in global weather patterns and average temperatures.

His temperature and pressure sensors are housed in a Stevenson Screen, a tipping bucket rain gauge measures rainfall, and an anemometer measures wind speed. This equipment has similar capabilities to that used by professional weather observers, but is more cost-effective for enthusiasts. In the past, Edward has also used the manual equivalents of this equipment, which many traditional observers still use.

Since he got his equipment setup and calibrated for optimum use, Edward’s weather station has run almost by itself with little input from him required. He only needs to do occasional checks to ensure that things are running smoothly and make sure that his equipment is maintained and functioning properly. The observing process for automated amateur stations is much less hands-on than it was in pre-automated times when the observer would take their own daily readings according to a prescribed routine.

Edward shares his data with a network called Climatological Observers Link (COL). COL provides detailed guidance on standards for observation routines and data formats in order to minimise variations in the data and make the data more accurate, reliable and reusable. Amateurs uploading to the network are encouraged to follow these guidelines. Edward takes great care to ensure that his data meets the standards, so that his data are high quality and can be reliably used by others when they are published in the monthly COL Bulletin.

Longer term, Edward is interested in observing trends in the weather over time, tying in with his professional interest in climate science. Over the years he has recorded his data in log books and computer data files, and archived it for future reference.

Collecting data – Raspberry Pi

A generation of new amateur observers is taking advantage of developments in cheap computing and electronics technologies such as Raspberry Pi and Arduino.

These observers use a variety of components to build their weather stations around these technologies. Some use out-of-the-box weather stations such as the AirPi that connect directly to the Raspberry Pi. Others, like us, have built their own weather stations using components sourced from a range of high street and online retailers.

We have published our own guide to building a simple Raspberry Pi weather station which transmits data to the Met Office WOW site. This makes a great project for schools and others who want to find out more about weather observing and computing.

The hobbyist we talked to was interested in weather observing from a technological point of view:

“I thought this was a good platform for further experimentation and growing my knowledge in electronics more, rather than in computing, because there’s very little you can do on a Raspberry Pi that you can’t do on a regular computer except for the electronics side and that’s what really caught my interest there as a nice way to go into electronics. And that got me then into, as an extension of electronics, the weather station which I thought – this matches very nicely with the other stuff that I’ve done.”

He had setup a full AirPi weather station, with a rain gauge add-on. His station generates observation data for a wide range of weather variables:

“So the AirPi project is essentially the collection of weather related sensors which you can mix and match as you choose. I got everything that was feasible to get in the UK so it has air pressure, temperature, relative humidity, light as in lux, then UV intensity and I have a rain gauge myself. So that wasn’t something the project actually recommended but because I have my masters stuff and PhD stuff relating to rain gauges I thought I must certainly add this and that was a nice meshing.”

His equipment was setup at home, in a setting that was less than optimal:

“I live in the basement. I’ve got a window which opens outdoors onto a patio which has a small table... which the rain gauge is sat on and there’s a nice wire trailing up over my window down onto my window sill where the Raspberry Pi and the rest of it sits. The rain gauge is the only bit outside... and so the Raspberry Pi then just sits on the window sill where it can get the light.”

His Pi weather station was fully automated. It generates data at very frequent intervals and writes directly to a CSV file. He then gets the data into standard units, and uses it to generate visual displays of the data for his own personal interest, using it as a sort of electronic barometer.

“I added CSV logging so I could more easily post analyse the data. All the sensors, the AirPi software provided no platform for calibration. So I added calibration, you know, for the different sensors... then CSV is very good for post analysis but I wanted a real time display as well. So I wrote a built in HTTP server, so you can go to a nice page and you click on the temperature reading and it gives you a graph for the last 24 hours. So you can quickly see what are the trends in temperature.”

Whilst he was not yet using his archive of data, he recognised possibilities for future analysis of the accumulated rain data that were tied to his professional interest in engineering.

“I suppose one of the things I’m most interested about it is the rainfall side, because that does tie in more to what I’ve done as part of my degree. Particularly what’s interesting is the shape of rainfall because from the design point of view, which is where I’ve mostly used rainfall data as part of my masters degree, rainfall is designed, and this seems to be part of essentially Gaussian distribution with different amounts of peakiness depending on whether you’re seeing a summer or winter storm, 90% profile, and the different shapes of storms.”

Sharing data

Our interviews and observations of websites about amateur observing demonstrated that sharing of amateur weather data takes place across a wide range of online spaces. Much of this sharing is done automatically using built-in software, although in some cases it is purpose built software code written by the observer.

As mentioned above, Edward shares the data produced by his weather station via the amateur observing network Climatological Observers Link (COL). COL collates weather observations from serious amateurs and its monthly bulletin is of sufficient quality to be archived in the Met Office Library.

“Well because – rather than just recording data and you know, the data not being useful for anything, it can be – in a small way it can help build up a picture of weather conditions across the UK, and how they vary. Get some historically archived data that it can be potentially used in research or even for teaching purposes on occasion.”

Other sites of data sharing we observed being used by amateur observers include Met Office WOW, Weatherlink and Wunderground. Similar to COL, these networks all provide detailed guidance for amateur observers on how to produce high quality, usable data.

As his data has some quality issues, our Raspberry Pi enthusiast preferred not to share his weather data at the moment, although he may do in the future:

“I have been considering doing that for the things which I know wouldn’t be affected by the sunlight so that’s particularly with the pressure and for the rainfall but also means I do have to write then the software model to do that. And it’s not hugely complex I just haven’t got into the right frame of mind where I’ll sit down and write this bit of software today. So I haven’t done it, but in the future I suppose I would be interested in doing that.”

Some amateur observers also post automated weather data to Twitter. These local amateur stations often have a sizeable Twitter following. The majority of tweets are simple weather observations that are generated automatically by the weather station equipment and sent out at regular daily intervals, with only occasional manual tweets made by the weather station owner.

This is a tweet from the Pi Weather account, which reports weather conditions from Clacton-on-Sea: