How we set up our school weather station

view of weather station enclosure looking west: click for close-up

In 2011 Simon Collins, Geography & ICT Departments, put in a bid to RGS PFA for the funding of a wireless automated weather station (AWS) to run an RGS weather station and website and feed good quality data to weather websites (WOW, Weather Underground, NOAA etc).

(Update: in April 2016 we set up another AWS (robust single unit Vantage Vue with Connect 3G data transfers to web)

Some of the aims of this project are as follows:

  • to provide the school and wider community with weather information and educational data (for Geography, Maths, ICT etc);
  • stimulate student interest in weather and engage student involvement in a weather club to collect data (for manual data collection and calibration purposes) and perhaps provide amateur “forecasts” for school events, display summary data and raise the profile of weather interest in the school (assemblies, Open Day displays etc?)
  • over time build-up and back-up good quality weather data about RGS weather;
  • log data over forthcoming years and months which will stimulate increasing interest as time goes by.
  • provide an online LIVE weather data summary for RGS;
  • register the site on WOW, Weather Underground and NOAA.
  • provide future proof back up of data
  • allow data to be used for GCSE and A level geography fieldwork to supplement primary data.

Read on to find out how it was set-up and the technical and siting problems overcome along the way.


March: £800 funding agreed for AWS purchase, with help from Sean Davey and Martin Hallpike: shared cost with Geography department agreed with HOD: Malcolm Cline

April: Choice of weather station finalised after much research: Davis Vantage Pro2 wireless model.

Potential locations for weather station enclosure and anemometer mast investigated with Fred George and Carmel Grater (maintenance and groundstaff).

May: Purchase of VP2 from Prodata: £980.75 serial model shared cost with Geography department

Funding for security fencing (costly) and mast erection (for anemometer) agreed by bursar

Testing of wireless range from proposed locations on school roof for anemometer and ISS: used 2-way radios to establish wireless strength with console in Geography office (no signal) and then DofE office (D1) – good signal. Help from Peter Kline and Mark Hadley and Malcolm Cline.

PC requirements discussed with ITS: help from Graham Redfern, Keith MacPherson and Peter Townsend.

June: Fred George organises installation of anemometer on school roof: aerial company attach 5m aerial pole to top of lift shaft: good exposure, especially to SW winds; pole can be dismantled quite easily for maintenance of anemometer. Anemo is up; signal is good, real wind data streaming to console; main ISS unit still sits under the desk in office! Question to Prodata regarding difference between indoor and outdoor temperature readings (despite both being in same room at the time).

July 4 -5 Security fencing constructed for ISS: completion of weather station enclosure.

August: ISS mounted on tripod and installed in weather station enclosure.

Weatherlink software installed on pc: console communication established with pc: complete weather data display seen on pc monitor for the first time!

Some troubleshooting of console data logging: Prodata solution = switch it off and switch it on again!

Grass mowing inside enclosure discussed with grounds staff.

SAC develops weather blog & discusses build of live weather page with ITS using Cumulus software.

August 13: Cumulus software installed and set-up completed.  Exploration of webpage set-up through ftp and cumulus.  Peter Townsend supplies webspace through; ftp upload.

September: RGS live weather goes online!  Web page installed and uploaded by SAC.

Questions and troubleshooting

Here is a chronological list of the snags (very few), questions and emails sent during the weather station set-up process.  John Dann of Prodata sent back excellent replies immediately in every case.

Question 1: Which AWS is most suitable for a school weather station?

14 May 2012

Many thanks for your enquiry about our range of Davis weather stations. A catalogue and price list together with a copy of our introductory booklet to choosing a weather station are in the post to you. Once you’ve had a chance to browse through the information pack, please don’t hesitate to get back to us if you need any further information. In the meantime, I include below a brief summary introduction to Davis stations and to Prodata Weather Systems.


Other than the Davis printed catalogue and our introductory booklet, the summary information we have available is, to a large extent, already posted on our web site. Possibly you have already found the site, but in case you may have missed some of the details please see the following specific pages: introduces the Davis Vantage Pro 2 range of weather stations, which includes both cabled and wireless models; describes the new Davis Vantage Vuestations; explains how best to mount the different weather sensors for optimum observational accuracy; summarises the extensive benefits available from connecting a Davis weather station to a PC;

The most popular model for typical UK customers is the Vantage Pro 2 wireless model, which is a fully integrated weather station able to measure temperature, wind speed and direction, pressure, humidity and rainfall. Sunshine intensity, UV, leaf wetness, soil moisture and additional temperature sensors are available as options.

The separate Weatherlink package allows any Davis station to be connected to a PC and consists of a hardware module that can store weather data when the PC is not switched on, plus software to generate an impressive display of weather data. The Weatherlinksoftware also includes the capability to upload weather data frequently and automatically to a web site in a wide range of text and graphical formats.We’re also pleased to announce that the new Davis Vantage Vue stations are now available, although in limited supply as yet.

The Vue is a wireless-only station that shares its wireless technology and data logging features with the VP2 and aims to be a more compact and lower cost alternative to the VP2, while still offering the well-known Davis virtues of accuracy and reliability. The outside sensors are all located in a single non-separable sensor pack for ease of installation but please note that the Vue is designed as a non-expandable station. This will be perfectly acceptable to many customers but anyone seeking flexibility and expandability in the sensor set should look at the VP2 stations.

All Davis weather stations are supplied complete with everything needed to set up a basic weather monitoring station, but various accessories and additional installation options are also available.

The Vantage Pro 2 stations are available in cabled and wireless versions and include an integrated radiation shield. Standard list prices are £546.38 + VAT for the cabled system and £648.51 + VAT for the wireless system. The standard price for a VP weatherlink is £188.94 + VAT. Our standard shipping charge for next-day delivery is £15 + VAT.TheVue station has a suggested selling price of £285 + VAT + P&P, but we can usually offer at least a modest discount on this price.

Question 2: USB, serial or WLIP logging?

Sent: 21 May 2012 12:24

I’ve done more reading and I’d be very grateful if you could answer some more of my questions before I arrange pc specs and arrange a purchase. 

I’d like to prioritize setting up some webpages to display our weather data linked or embedded onto the school website and my own blog with the flexibility of possibly using alternative third party AWS software like weather display or cumulus in the future.  Would the WLIP allow this or, as I have read, is it dedicated to Weatherlink and rather tricky to feed into other software / webpages?  Would the standard USB or serial logger be better for this purpose of setting up our own webpage?  Am I correct in thinking we could subscribe to weatherlink if we wished to use that?  Wunderground is also available for uploading data whichever data logger is chosen: is this correct?

 The serial com datalogger is said to be more “stable” and has an option of a 12m cable extension.  I am tempted by both of these options … please can you confirm that the stability on the serial DB-9 connector is more stable than a USB or is there nothing much in it?  It makes a difference regarding the spec I ask for from our IT support.

 The choice between WLIP and standard datalogger seems a big decision at this stage. It seems that the standard data logger, though initially more time consuming to get on the web, allows for more future flexibility.  I am also considering requesting a quiet / low power PC which could “always be on” to reduce costs.  Is this a reasonable choice if we want flexibility of web publishing and live weather feeds onto the web?


WLIP is undoubtedly the simplest way of getting your weather data on to a web page and without requiring a PC to be running 24/7 locally. But if you definitely want the flexibility of being able to use e.g. WD or Cumulus then I would not recommend using the WLIP logger. Those are the basic facts – I can’t really make the decision for you, this is really down to how you weigh up ease-of use/simplicity/robustness against maximum flexibility.

Few users have trouble with the USB logger these days – mainly because of improvements to the USB drivers, but also because of increased awareness of the importance of keeping the USB logger away from any potential sources of interference. But the serial logger (usually plus a serial-to-USB adapter, though if you have a spare serial port on a PC then you can obviously use that instead) is still perhaps marginally more robust than the USB logger. It makes no difference to software compatibility whether you use serial or USB. The USB version can be slightly more confusing to use when setting up the interface but once set up there’s really no difference in use.

There is an option to upload to with the latest (v6.0) version of the Weatherlink software. The software has been released but we’re still waiting for UK pricing and how-to-order details – until we get these, the upload option cannot be activated. But I have been chasing these details today in fact and hopefully the information will come through to us shortly. (The delay is because this is the first time that Davis have launched a subscription service in the UK and so there are various international complications to be sorted out about the mechanics of both ordering and renewing the annual subscription  – it’s not just a pricing issue. Hopefully these issues will be resolved soon, but I’m not directly party to the negotiations.)

There are a variety of possible mechanisms for uploading to Wunderground, depending on which software you have in use. I’m sure you’ll be able to use one of these, but exactly which option will be available will obviously depend on your choice of software.

Question 3: weather station enclosure design and build:

From: Simon Collins []
Sent: 11 June 2012 09:54

We’ve tested wireless range and found locations for anemo on the roof and ISS on a suitable grassy patch in the school grounds.  I will need to put a fence around the VP2 weather station for security and protection; I am hoping to erect a fairly standard garden open wooden paneled fence arrangement giving the the ISS plenty of space to “breathe”.  Are there any fence designs which we should avoid or will a common sense approach regarding plenty of space, colour and open slatted design be enough to avoid temperature or rainfall errors?


I think common sense is the only easy guidance I can give and without knowing the size of the enclosure that you might be considering. Certainly the more open the fencing the better – personally I’d have said that wire mesh might be better than wooden fence panels but I appreciate that installation practicalities and aesthetics probably come into it too.

Also of course it depends on how large the enclosure might be – the closer the sensors are to the fence then the more influence the panel material/openness  might have. Too close to a thick trellis panel, for example, would start to have some noticeable effect on rainfall recording I would have thought, certainly with the wind from the nearest panel direction. But just by way of a benchmark, placing the VP2 ISS roughly in the centre of a 5-6m square metal chain-link enclosure ought to have fairly minimal effects on readings. (This all assumes that the anemometer is somewhere outside of the enclosure, which I understand to be the case, so I’m really just focusing on effects on temperature, rainfall etc.) Having an enclosure of about this minimum size (i.e. 5-6m square) would also give you a decent space to install additional sensors in future, should you wish to do so.

Question 4: details about anemometer set-up

From: Simon Collins []
Sent: 12 June 2012 14:46

We are erecting a 5m mast on the school roof for the anemo on Tuesday.  Please can you advise on any precise fitting arrangements for the anemo? 

Specifically, does the anemo have to point True or Mag North?  Which part of it must be orientated to North?


From: Prodata Sales []

The anemometer needs to fit to a mast of 32-44mm outside diameter.

Strictly speaking, the anemometer should point to True North, but in the UK it really doesn’t make too much difference and in practice I suspect you’ll have trouble aligning the mast to better than about 10 degrees accuracy so it’s a fairly moot point. Also, wind direction is traditionally reported in the 16 compass sectors such as NNE, which cover 22.5 degree sectors so, again, there’s no real call for going to a lot of trouble in achieving the most precise alignment to True North. Looking from the pole out along the anemometer arm, the arm itself should point to North.

Question 5: Temperature difference between inside and outside:

Sent: 29 June 2012 16:57

Anemo is up; signal is good, data streaming to console.  The ISS is operational and fully functioning and sends data to the console but sadly the ISS still sits under my desk until the fencing is erected next week and I’ll set it up outside to start recording proper data.

Please can you help me with one strange thing…

The Temp Out always reads very differently from the Temp In and yet the ISS (which sends the Temp Out data) is in the same room only about a foot away from the console (which I suppose is responsible for the Temp In data).  The reading has been different since it was all switched on a few weeks ago. 

 For example, currently the current Temp Out reads 20C and Temp In reads 23C.  Surely they should read exactly the same if both sensors (in and out) are indoors and next to each other?

 Is this an error or have I missed something obvious here?


Is the backlight on for the LCD panel? This inevitably increases the temperature inside the console unit and hence the indicated inside temperature. If you need a more accurate inside temp then the best policy is to keep the backlight off until you wish to read the display and then turn it off again afterwards. If you need the display to be visible constantly then shone an external light on the display.

If the backlight isn’t on: There is unlikely to be any significant problem with the inside temp reading, but the explanation is a bit more involved. The main reason that the inside and outside temps often don’t agree with one another indoors is that room temperatures are typically uneven (e.g. one unit may be nearer the draught from a door or window or nearer to a radiator) and changing through the day and night. Then the response times of the sensors in the ISS and console can be quite different. The ISS is obviously meant for use outdoors where’s there’s typically at least a light breeze blowing to move air through the shield. Conversely, the console sensor is somewhere inside the unit and hence also has a relatively slow response time but with different characteristics to the ISS. Then of course each temp sensor has its own accuracy spec of about +/- 0.5degC and so the two readings can often disagree by up to 1degC and still be comfortably within spec.

We usually find that if you place the ISS and console literally side-by-side, with the console backlight off, in a room at reasonably constant temperature and blow a decent fan over both units to keep the air moving then after a few hours the two temps should agree reasonably within spec.

Question 6: Where is the weather data stored on weatherlink?

Sent: 09 August 2012 12:07
One question that stumps me despite using help files: where exactly is the weather data stored (in csv file format?)  in the weatherlink directory? Presumably this location is required to push data out to websites.


The long-term historical data is stored in binary files, not CSV, a new one started for every month with a filename based on the current month and year – so August 2012 is called 2012-08.wlk. You’ll find all these files in your main station folder, e.g. C:\Weatherlink\Reigate (if you’ve called your station Reigate or whatever name it has).

Full technical details of the file format are in the readme file which is usually access from the Windows Start button in the Weatherlink section.

NB There are also two text files called (from memory) downld02.txt and downld08.txt which contain the last complete single day and week’s worth of data. As you might imagine, these two files are constantly updated and overwritten with each new data update.

Question 7: What file format is used to back up weather station data?

9-Aug 2012

Weather data storage and back-up:

Regarding data back-up and the wider us of collected weather data (for example, students accessing stored data archives on the network / downloading from a website for projects using spreadsheets):

 Can the wlk. files simply be backed up by transferring them to a remote server or external hard drive – my IT support suggest writing a batch file to move them across once a week as back up? 

Also, is it normal to leave them backed up as wlk files or can I back them up as .csv files – which might be more flexible in use with other programs like excel? 

I would like to enable public / student access to historical data through files accessible on a website: .csv files might be the best option because they can be used with spreadsheets.


Yes, the wlk files are just standard files and can be backed up just like any other file, i.e. copying to somewhere else e.g. an external drive.

There is no automatic way I’m afraid to convert the binary wlk files to text files, but there is a way of doing it manually: If you go into Browse mode (using e.g. the Notepad-like icon on the toolbar) you’ll see a new Browse top-level menu item appear. On this menu there’s an Export Records option where you can select a time period to export and then a filename to export to. You could for example do this after the end of each month to create a separate monthly text file. The format of the text file will be tab-delimited text but this is usually fairly easily read by spreadsheet type program either directly or if you tell the import process that the delimiter character is tab.

One response to How we set up our school weather station


    This was very helpful,Thanks

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s