The Weather Outlook

Remove ads from site

Cumbrian Snowman
04 February 2021 09:26:42
Thanks for that update johncs2016

I have looked at the weatherobs site - found it to be very very slow, it gave me an idea of temps in Cumbria from the main Met Office Sites but I could have used Wunderground and used the many more private stations to build up a picture of whats going on.

From a personal point of view I need access to

Carlisle, Spadeadam, Shap, Great Dun Fell, Bridgefoot, Keswick, Levens, Warcop and Walney

I have my own website so could use the Met Office data point service to drag the data into some readable format, but I dont know how to do it or where to start. If anyone has any code or wants to have a go on my behalf that would be great !!


johncs2016
04 February 2021 10:28:26

Thanks for that update johncs2016

I have looked at the weatherobs site - found it to be very very slow, it gave me an idea of temps in Cumbria from the main Met Office Sites but I could have used Wunderground and used the many more private stations to build up a picture of whats going on.

From a personal point of view I need access to
Carlisle, Spadeadam, Shap, Great Dun Fell, Bridgefoot, Keswick, Levens, Warcop and Walney

I have my own website so could use the Met Office data point service to drag the data into some readable format, but I dont know how to do it or where to start. If anyone has any code or wants to have a go on my behalf that would be great !!

Originally Posted by: Cumbrian Snowman 

For a while, I had forgotten all about Wunderground and I can remember that it was Arty who gives the daily updates on the CET threads, who first pointed me towards that site some time ago. That is meant for dealing with local personal weather stations such as what you have mentioned, and therefore doesn't give any data for the major Met Office stations such as Edinburgh Gogarbank. This has a downside to it because data from small individual weather stations is probably not likely to be accurate as the data for those major Met Office stations, although this data should be good enough to give a general picture of what is happening.

Although Wundergound doesn't give data for the major Met Office stations though, it is another place to get data for Edinburgh Airport although by default, it always reports temperatures in °F instead of °C, so I'm always having to change that in the settings for that site. If I was an actual member of that site though (which would only really be worthwhile if you had your station for which you wanted its data to be publicly available, you would probably be able to set that as a default in your account settings.

 


The north of Edinburgh, usually always missing out on snow events which occur not just within the rest of Scotland or the UK, but also within the rest of Edinburgh.

Crepuscular Ray
04 February 2021 13:25:15
Thanks for all this John. I'd missed this thread somehow. I'll experiment with the links above.
Jerry

Edinburgh, in the frost hollow below Blackford Hill

johncs2016
04 February 2021 14:20:15

Thanks for all this John. I'd missed this thread somehow. I'll experiment with the links above.

Originally Posted by: Crepuscular Ray 

With all of this in mind, it's a good job tat there is sill the website of your own local station at Swanston in the south of Edinburgh. Although I live on the other side of Edinburgh from there, I have a really good view of the Pentland Hills from my south facing kitchen window and the Swanston website is quite good for telling us what the snow level is for the Pentland Hills if this lying snow doesn't extend down into that particular station itself. It's just a shame though, that there aren't more sites which are like that.

 


The north of Edinburgh, usually always missing out on snow events which occur not just within the rest of Scotland or the UK, but also within the rest of Edinburgh.

08 February 2021 20:59:55

Thanks for that update johncs2016

I have looked at the weatherobs site - found it to be very very slow, it gave me an idea of temps in Cumbria from the main Met Office Sites but I could have used Wunderground and used the many more private stations to build up a picture of whats going on.

From a personal point of view I need access to
Carlisle, Spadeadam, Shap, Great Dun Fell, Bridgefoot, Keswick, Levens, Warcop and Walney

I have my own website so could use the Met Office data point service to drag the data into some readable format, but I dont know how to do it or where to start. If anyone has any code or wants to have a go on my behalf that would be great !!

Originally Posted by: Cumbrian Snowman 

I have had a bit of a play with the weatherobs site and although it can be a bit slow, once you get the hang of using it I found the data easy and quick to access. It provides raw data in the same format as the weathercast site used to (go to the dashboard for each site) and does include the 99xxx stations. It goes back a full 7 days which is slightly further than weathercast which was about 6 days. No max or mins which is a shame as that is the missing piece when it comes to the 99xxx stations. But at least it does provide an acceptable alternative to the now defunct weathercast site.

johncs2016
09 February 2021 01:17:04

 

I have had a bit of a play with the weatherobs site and although it can be a bit slow, once you get the hang of using it I found the data easy and quick to access. It provides raw data in the same format as the weathercast site used to (go to the dashboard for each site) and does include the 99xxx stations. It goes back a full 7 days which is slightly further than weathercast which was about 6 days. No max or mins which is a shame as that is the missing piece when it comes to the 99xxx stations. But at least it does provide an acceptable alternative to the now defunct weathercast site.

Originally Posted by: Global Warming 

One thing which I have discovered about weatherobs.com though is that just like a number of other sites, this doesn't actually show all of the available data in decoded form With the botanic gardens in Edinburgh for example (which is a 99xxx station, its WMO index is 99180 in this case), the minimum temperatures are shown every day at 6z in amongst the decoded data (you have to go to the extra decoded data tab to be able to see that). This decoded data doesn't show any maximum temperatures but if I look closely at the raw SYNOP data, I have noticed that this raw data does indeed, include the maximum temperatures every day at 18z.

As an example of that, I discovered from this data that yesterday's maximum temperature was 2.7°C which looks to be about right as yesterday's highest hourly reading was 2.6°C and the actual maximum temperatures are always the same as or higher than the highest hourly reading. There is also one other subtle difference in how rainfall is reported because unlike with the main 03xxx stations (such as Edinburgh Gogarbank), there is no separate entry in the raw SYNOP data for the 24 hour rainfall totals, As it turns out that, this doesn't actually matter since the rainfall totals can still be easily worked out by adding up the relevant 12 hour totals which are released twice a day at 06z and 18z.

Unlike with the main 03xxx stations though, the SYNOP data for the 99xxx stations doesn't actually include everything which has actually been recorded, and sunshine totals is one example of that. THat is why weatherobs.com has to be used in addition to the Met Office WOW site where I can get that other information (although there is often a delay with that being updated). When I do that though, I then still end up with all of the same data which I have been able to get for Edinburgh Gogarbank from a single source and this is all that I need.

However, it would be nice if we could get data from weatherobs.com for up to a time period like a week, but where we are able to choose any such period back to a certain date in the past instead of that always being just for the last week. If we were able to do that, I would then be able to catch up with all of the past data for the botanic gardens in Edinburgh in my various spreadsheets but without such a feature in place, this isn't something which I am able to do just now.

 


The north of Edinburgh, usually always missing out on snow events which occur not just within the rest of Scotland or the UK, but also within the rest of Edinburgh.

09 February 2021 04:57:35

 

One thing which I have discovered about weatherobs.com though is that just like a number of other sites, this doesn't actually show all of the available data in decoded form With the botanic gardens in Edinburgh for example (which is a 99xxx station, its WMO index is 99180 in this case), the minimum temperatures are shown every day at 6z in amongst the decoded data (you have to go to the extra decoded data tab to be able to see that). This decoded data doesn't show any maximum temperatures but if I look closely at the raw SYNOP data, I have noticed that this raw data does indeed, include the maximum temperatures every day at 18z.

As an example of that, I discovered from this data that yesterday's maximum temperature was 2.7°C which looks to be about right as yesterday's highest hourly reading was 2.6°C and the actual maximum temperatures are always the same as or higher than the highest hourly reading. There is also one other subtle difference in how rainfall is reported because unlike with the main 03xxx stations (such as Edinburgh Gogarbank), there is no separate entry in the raw SYNOP data for the 24 hour rainfall totals, As it turns out that, this doesn't actually matter since the rainfall totals can still be easily worked out by adding up the relevant 12 hour totals which are released twice a day at 06z and 18z.

Unlike with the main 03xxx stations though, the SYNOP data for the 99xxx stations doesn't actually include everything which has actually been recorded, and sunshine totals is one example of that. THat is why weatherobs.com has to be used in addition to the Met Office WOW site where I can get that other information (although there is often a delay with that being updated). When I do that though, I then still end up with all of the same data which I have been able to get for Edinburgh Gogarbank from a single source and this is all that I need.

However, it would be nice if we could get data from weatherobs.com for up to a time period like a week, but where we are able to choose any such period back to a certain date in the past instead of that always being just for the last week. If we were able to do that, I would then be able to catch up with all of the past data for the botanic gardens in Edinburgh in my various spreadsheets but without such a feature in place, this isn't something which I am able to do just now.

 

Originally Posted by: johncs2016 

Thanks I will take a look at that further. At first glance it looked like the maximum and minimum figures I could see were just taken from the hourly data rather than actual maximums and minimums. Given the 03xxx stations are SYNOP stations then it would make sense that maximums are available at 18z and minimums at 06z.

To your last point about historic data this is now freely available for all Met Office stations from the CEDA website. The data currently goes up to July 2020 and includes data since each individual station was opened. This is an excellent resource and should I think give you what you are looking for. You have to register and get a login to access the data but this is quick and easy to do. The datalink page is here

http://data.ceda.ac.uk/badc/ukmo-midas-open/data/uk-daily-temperature-obs

johncs2016
09 February 2021 05:41:44

 

Thanks I will take a look at that further. At first glance it looked like the maximum and minimum figures I could see were just taken from the hourly data rather than actual maximums and minimums. Given the 03xxx stations are SYNOP stations then it would make sense that maximums are available at 18z and minimums at 06z.

To your last point about historic data this is now freely available for all Met Office stations from the CEDA website. The data currently goes up to July 2020 and includes data since each individual station was opened. This is an excellent resource and should I think give you what you are looking for. You have to register and get a login to access the data but this is quick and easy to do. The datalink page is here

http://data.ceda.ac.uk/badc/ukmo-midas-open/data/uk-daily-temperature-obs

Originally Posted by: Global Warming 

.. and that is the big deal breaker for me in this case.

I have looked at the registration form for this and everything there seems to assume that you are studying for some form of degree (which I'm not) and asks questions such as where you are applying for such a degree and even who your supervisor is for that. This website says that you can access certain datasets without a login as long as this isn't being done through their FTP servers and yet as soon as I click on any links to even begin that process, I am asked to login before I can go any further, so that is obviously one big fat lie.

Furthermore, there is the question of privacy because even if I was actually studying for a degree, that is information which is private and confidential to me which means that I might not necessarily want to be giving out such information to ant third party organisation, even if the purpose of that is to get a login for a certain site, and even if this is a major company or organisation such as the Met Office.

That is why I will still continue to refuse to sign up for any sort of accounts which I need to login to, if all I am doing to trying to get access to data which I believe, should be freely available without the need for any registrations or logins. That is why I mentioned that it would be good if I could get a week's worth of data from weatherobs.com for any other period of time other than just the last week because even in this day and age, I strongly believe that it should be possible to do this (as it is with the 03xxx stations at sites such as meteomanz or ogimet) without any registration or login being required for that.

 


The north of Edinburgh, usually always missing out on snow events which occur not just within the rest of Scotland or the UK, but also within the rest of Edinburgh.

Rob K
11 February 2021 11:32:25

 

I have had a bit of a play with the weatherobs site and although it can be a bit slow, once you get the hang of using it I found the data easy and quick to access. It provides raw data in the same format as the weathercast site used to (go to the dashboard for each site) and does include the 99xxx stations. It goes back a full 7 days which is slightly further than weathercast which was about 6 days. No max or mins which is a shame as that is the missing piece when it comes to the 99xxx stations. But at least it does provide an acceptable alternative to the now defunct weathercast site.

Originally Posted by: Global Warming 

Weatherobs does show the max and min as taken from the 0600 and 1800 synop. You have to click on the station and then on the pop-up you click at the bottom where it says "Click to load dashboard for this site" (this doesn't show up on mobile for me, so use a desktop).

This then gives you three tabs: "Raw reports", "Decoded - Main" and "Decoded - Extra". The "Decoded - Extra" tab gives you the max and min at 12 hourly intervals in the table.

 

 

There are still some stations missing from this map, though. I wonder if there is a wider problem with availability? For example I found this site which looked perfect as it has an extensive list of dropdowns, but many of the sites only have data up to the end of January 2021, eg Braemar: Starlings Roost Weather (asuscomm.com)


Yateley, NE Hampshire, 73m asl

"But who wants to be foretold the weather? It is bad enough when it comes, without our having the misery of knowing about it beforehand." — Jerome K. Jerome

johncs2016
11 February 2021 12:09:47

 

Weatherobs does show the max and min as taken from the 0600 and 1800 synop. You have to click on the station and then on the pop-up you click at the bottom where it says "Click to load dashboard for this site" (this doesn't show up on mobile for me, so use a desktop).

This then gives you three tabs: "Raw reports", "Decoded - Main" and "Decoded - Extra". The "Decoded - Extra" tab gives you the max and min at 12 hourly intervals in the table.

 

 

There are still some stations missing from this map, though. I wonder if there is a wider problem with availability? For example I found this site which looked perfect as it has an extensive list of dropdowns, but many of the sites only have data up to the end of January 2021, eg Braemar: Starlings Roost Weather (asuscomm.com)

Originally Posted by: Rob K 

What's interesting about weatherobs.com though, is that I am finding stations which I didn't know that even existed. As an example of that, it shows that there is a station to the north of the botanic gardens in Edinburgh. By looking more closely, this station has the name of SHIP AMOUK 36, and appears to be a ship which is currently moored at Leith Docks.

This isn't a major station and doesn't have a WMO number, but it does give out SYNOP codes every hour, although this is done in a slightly different way from the other stations which I know of because those are automated land stations where the SYNOP codes always start with AAXX, whereas this ship is obviously a marine station whose SYNOP codes always start with BBXX.

From that, I can get the temperature and pressure for each hour which in terms of the temperatures are often enough to show what it is like right on the coast itself, in comparison with locations which are just slightly further inland such as where I am.


The north of Edinburgh, usually always missing out on snow events which occur not just within the rest of Scotland or the UK, but also within the rest of Edinburgh.

johncs2016
13 February 2021 06:54:28

I can now give an update on what is happening with SEPA as regards to getting rainfall data from there which will be very valuable for a lot of my fellow Scottish members of this forum who have used the SEPA site to obtain rainfall data for their local stations in the past.

First of all, I must add that certain things on that site have not changed. If you try to get the rainfall data from there in the old way, you will still get that same old Service Unavailable page. However, I have been informed on this forum that the nature of the recent cyber attack means that certain new systems might need to be developed and from what I can see, it looks as though this is what is going to happen as regards to being able to get any rainfall data.

If you go to the SEPA homepage and then click on the link which says "rainfall" under the section which says "water" (don't click on the link below that for rainfall data, as that will just take you the usual service unavailable page), you will be taken to another page which tells you all about what is happening as regards to the cyber attack and the current inability to get any rainfall data from there.

On the right-hand side of that page, you will see two different links which are listed under related items. The one which says "rainfall data" is just the old link which will once again, just take you to the service not available page. However, there is another link which says "Citizen Science Portal - Rainfall Observers" and that is where you want to go to as it looks as this is going to become the new way of getting rainfall data from SEPA in the (hopefully) not too distant future.

Just in case you have trouble getting to that section of the SEPA site, here is the direct link to it. It looks as though this site is going to operate in a similar manner to how wundergound operates for all types of weather observations. Anyone who has their own weather station (and I know that there are a number of people on this forum who do) will no doubt, also have their own way of recording rainfall and if your station is located somewhere in Scotland, you will be able to register there to enter your own rainfall data and have that available on-line.

However, SEPA was also previously good at giving me that important data for the major stations such as Edinburgh Gogarbank and the botanic gardens in Edinburgh, so I am hoping that this will continue to be the case under the new system once that is fully up and running. I haven't tested that part of it though, as I don't have my own weather station or rain gauge, but you are free to test that and come back to this thread on that matter if you so wish.

As for getting data from there, you won't need to register for that and so, it will just be a case of clicking on where it says "view data". As I write, there isn't actually any data available just yet and it may well be a while before there is. However, there is already a clickable map there and when that data does eventually become available,you will then be able to select the station from there, for where you are trying to get data for.

Given the nature of the recent cyber attack, this will obviously all have to be done with security in mind and so, I would probably expect this new system to be harder to hack or scrape than the old system, but it is nice to see that things are now happening on that front and I am now looking forward to seeing that new system fully up and running, whenever that eventually happens.

 


The north of Edinburgh, usually always missing out on snow events which occur not just within the rest of Scotland or the UK, but also within the rest of Edinburgh.

johncs2016
12 September 2026 13:30:55
I know that this is a very old thread dating all the way back to 2021 but I have a very important update on the availability of online past weather data which I believe would be much better dealt with by bringing this particular thread back to life, rather than starting a brand new thread from scratch on which will basically just be the same subject.

The reason why I started this particular thread is because back in 2021, we lost what was a very good site in weathercast.co.uk which led to quite a good discussion on here about the fact that it was becoming more difficult as a result, to get available data for stations such as Kew Gardens in London or the botanic gardens in Edinburgh which had (and still has) a WMO number which starts with 99 rather than 03.

Over the years, I have continued to be able to get such data from weatherobs.com but even that site only gives you nothing more than the last week's worth of data which is very difficult to scrape by using a python script (I was eventually able to overcome that, but that was a very difficult task to achieve).

However, I am seeing signs now that we are starting to turn the corner in regards to that data being harder to get because with recent developments, I am starting to find that this is no longer the case now to the same extent. That all started when I noticed just recently that the site at roostweather.com which has been previously mentioned by others on this thread, now gives us quite extensive data for stations such as Kew Gardens or the botanic gardens in Edinburgh. For the botanic gardens in Edinburgh, that data still only goes back to 2011 and since that particular station has been around since around 1956, you are still only getting a small fraction of all of the available data from there even with that. Nevertheless, even that is still a vast improvement on the 7 days worth of data which you get for there from weatherobs.com, and is therefore very clearly a big step in the right direction.

However, I have just discovered one other thing which I believe to be so important that I thought that it was enough for this thread to be revived as a result. When you go onto ogimet.com, you can do a search for a particular station if you know at least part of its name. I did that search entering EDINBURGH as the search term which gave me a list of stations here in Edinburgh. To my astonishment, I discovered that the botanic gardens in Edinburgh was actually on that list despite it not having a WMO number which began with 03. What I discovered though, is that every major station has what is known as WIGOS ID which consists of 4 different parts to it. The first and third part of that is always '0' with the second part of that being a number which identifies which country that particular station is in. For here in the UK, that number is always 826 and this means that the WIGOS-ID for all of the major SYNOP stations in the UK is always "0-826-0-" followed by a number which uniquely identifies the station itself.

For Edinburgh Gogarbank, the WIGOS-ID is "0-826-0-88" whilst the WIGOS-ID for the botanic gardens in Edinburgh is "0-826-89". As another example, the WIGOS-ID for Kew Gardens in London is "0-826-0-127". It turns out in fact that instead of entering the WMO Code when you are looking up data on ogimet.com, you can enter the WIGOS-ID. When I did that for Edinburgh Gogarbank, I still got exactly the same data as I get whenever I'm entering its WMO code of 03166. In addition to that though, I then entered the WIGOS ID for the botanic gardens and to my astonishment, I discovered that I was actually able to get the data for there as a result. There wasn't as much data for there as there was for Gogarbank but the important thing here nevertheless, is that there was some data there and that I'm no longer limited to just a week's worth of that data.

That to me is quite a big breakthrough because this shows that you can now get data for stations such as the botanic gardens in Edinburgh or Kew Gardens on ogimet.net just in the same way that I can for stations such as Gogarbank even if that station doesn't have a WMO code that starts with 03, as long as you can find its WIGOS-ID. I then tried that at meteomanz.com and couldn't actually get any data from there for the botanic gardens in Edinburgh, but the record for that particular does exist on meteomanz.com, so my guess is that I might only be a matter of time before we start to get some data on that site as well for those particular stations.

As for ogimet.com, I found that with the data for the botanic gardens in Edinburgh, I could also link back to the raw SYNOP/BUFR data for that in the same way that I could for Gogarbank and for both of those stations, it turns out that the raw SYNOP/BUFR data is actually very similar with the only difference being that for the SYNOP data, '00000' is shown instead of any WMO code with those positions where there isn't the equivalent data being taken up by a series of forward slashes. Nevertheless, that should still be easy enough to decode with my python scripts so with that, I am hoping that I will be able to make more use of that by reporting more data on here for the botanic gardens as a result, especially as it is my actual local station here in Edinburgh.


The north of Edinburgh, usually always missing out on snow events which occur not just within the rest of Scotland or the UK, but also within the rest of Edinburgh.

johncs2016
11 October 2026 04:44:31
As a further update, I was recently pondering over the question of how the Met Office knows what the 1991-2020 (and even earlier) averages are for Edinburgh Gogarbank even though that particular station didn't even exist before 1998 which is a whole 7 years after the beginning of that 30 year time frame. Just recently though, the Met Office published a video on their main YouTube channel about something called the HadUK grid and it turns out that this is actually what is used to determine those averages.

I got that answer when I posed the same question to Google's AI bot which is accessible at google.com. With that, there is the 2km grid which is capable of tracking back over a period of time and give an accurate picture of what the observations in a particular area would be if there was an actual weather station there. That in turn allowed the Met Office to be able to determine what the conditions were like in the Gogarbank area before its weather station was even setup, thus allowing them to be able to come up with those averages.

It turns out in fact, that this can be utilised to give similar averages for any area of the UK and in my case, I was even able to build my own 1991-2020 averages for my actual location here in NW Edinburgh by creating a python script for that which downloads data from open-meteo.com and then uses that to build those average. It is possible to directly download HadUK grid data from the Met Office site but that is rather bulky and difficult to navigate. At the other site at open-meteo.com, I can access data from Copernicus's ERA-5 land dataset which is similar in nature (if not the same) as the Met Office's HadUK grid data and I have now created a python script for doing that.

Developing such a script used to be a very time consuming process which required you to understand quite a lot of detailed maths and physics, and a lot of in depth programming concepts. I am now approaching retirement ago and my brain doesn't quite work as well as it used to, so it's a lot more difficult now for me to be able to deal with those concepts than it used to be. However, AI is now changing that. In today's world, there are a lot of time pressures involved which mean that in some situations, there might not be the time to waste on such time consuming processes, especially given that a lot of projects might have to be up and running very quickly.

Because of that, even professional computer programmers now collaborate with each other and use AI to help them to develop their projects. That is what I have started to do and by doing so, I am now able to produce these python scripts in only a small fraction of the time that it would have taken had I just tried to work on that on my own. That in turn makes quite a huge difference.

As for my own 1991-2020 averages, I found them to be very interesting. Since I live in the same city as Gogarbank, my averages are similar but they are not exactly the same because of the fact that I live in a different area. I live in NW Edinburgh but Gogarbank is actually in SW Edinburgh which technically makes it a part of the south of Edinburgh rather than the north of Edinburgh. If you study the data from there more closely, what you actually find is that as a result, the conditions there are much more likely to be more like what they are at a station such as Swanston, that they are at the botanic gardens in Edinburgh which is closer to the coast. Indeed, that was shown really well in 2019 when the temperature there rose to over 30°C, but only reached 26°C at the botanic gardens due to some slight cooling from the North Sea.

However, the botanic gardens are also very close to the city centre and that North Sea effect actually grows even stronger as you move even closer to the coast from there, yet there are no actual weather stations in that vicinity which can pick that up. As a result, it is possible for snow to be lying at the botanic gardens in the winter, with my area being completely snow free at the same time. That is something which these localised averages pick up really well, and I'm glad I've been able to find them.

This is good not just for long term averages though because it turns out that I can also utilise that to give a fairly accurate picture of what my own data would be like in terms of the actual current conditions if I actually had my own weather station to be able to record that. This then actually allows me to report the conditions specifically in my area without any need for an actual weather station. This will help me even more when I travel down to the Borders because there are no nearby major Met Office stations and yet, this system would probably still be able to give an accurate picture of the actual conditions there.

Furthermore, this also means that I would no longer have the same issues with missing data as I would for recording data from any station as this system automatically fills in those gaps. Another example of that comes with rainfall because I have lost count of the number of times when I have had a lot of rain here, only for very little of that to actually be picked up at any of the nearby major stations. This system actually picks that up really well though and with that, I can get a more accurate picture of what those rainfall totals actually are in this particular area.

With all of that in mind, I am now in the process of giving my own python script a complete overhaul and replacing the system which I have been using up until now, with a system which give reports for my actual area where I live, rather than for a station such as Gogarbank. That was a system which required me to understand more complex concepts such as dealing with raw BUFR/SYNOP data and looking back in it, it turns out that my old system was a lot more complicated than it actually needed to be.

On the other hand, the system which I am now creating to replace that will be a lot simpler and easier to use. As an example of what this system provides, I noticed that according to SEPA, there been 11.6 mm of rain during this month so far at the botanic gardens but only 10.6 mm at Gogarbank. Meanwhile, my new system is telling me that where I live, there has been a total of 12.4 mm of rain during this month so far. As for the temperatures, It is currently 8.2°C at Gogarbank and 7.4°C at the botanic gardens. Meanwhile, my new system is telling me that it is currently 8.8°C. This nicely shows that it is a west to east variation which is overriding everything in that regard just now. From that, I am seeing that this is a system which I am very happy with and so, I'm looking forward to bringing that into use in my future reports on here on the relevant threads for that.


The north of Edinburgh, usually always missing out on snow events which occur not just within the rest of Scotland or the UK, but also within the rest of Edinburgh.

Roger Parsons
11 October 2026 05:18:50
Thanks John. Intriguing piece. 
RogerP

West Lindsey district of Lincolnshire

Everything taken together, here in Lincolnshire are more good things than man could have had the conscience to ask.

William Cobbett, in his Rural Rides - c.1830

lanky
11 October 2026 09:47:55

As a further update, I was recently pondering over the question of how the Met Office knows what the 1991-2020 (and even earlier) averages are for Edinburgh Gogarbank even though that particular station didn't even exist before 1998 which is a whole 7 years after the beginning of that 30 year time frame. Just recently though, the Met Office published a video on their main YouTube channel about something called the HadUK grid and it turns out that this is actually what is used to determine those averages.

I got that answer when I posed the same question to Google's AI bot which is accessible at google.com. With that, there is the 2km grid which is capable of tracking back over a period of time and give an accurate picture of what the observations in a particular area would be if there was an actual weather station there. That in turn allowed the Met Office to be able to determine what the conditions were like in the Gogarbank area before its weather station was even setup, thus allowing them to be able to come up with those averages.

It turns out in fact, that this can be utilised to give similar averages for any area of the UK and in my case, I was even able to build my own 1991-2020 averages for my actual location here in NW Edinburgh by creating a python script for that which downloads data from open-meteo.com and then uses that to build those average. It is possible to directly download HadUK grid data from the Met Office site but that is rather bulky and difficult to navigate. At the other site at open-meteo.com, I can access data from Copernicus's ERA-5 land dataset which is similar in nature (if not the same) as the Met Office's HadUK grid data and I have now created a python script for doing that.

Developing such a script used to be a very time consuming process which required you to understand quite a lot of detailed maths and physics, and a lot of in depth programming concepts. I am now approaching retirement ago and my brain doesn't quite work as well as it used to, so it's a lot more difficult now for me to be able to deal with those concepts than it used to be. However, AI is now changing that. In today's world, there are a lot of time pressures involved which mean that in some situations, there might not be the time to waste on such time consuming processes, especially given that a lot of projects might have to be up and running very quickly.

Because of that, even professional computer programmers now collaborate with each other and use AI to help them to develop their projects. That is what I have started to do and by doing so, I am now able to produce these python scripts in only a small fraction of the time that it would have taken had I just tried to work on that on my own. That in turn makes quite a huge difference.

As for my own 1991-2020 averages, I found them to be very interesting. Since I live in the same city as Gogarbank, my averages are similar but they are not exactly the same because of the fact that I live in a different area. I live in NW Edinburgh but Gogarbank is actually in SW Edinburgh which technically makes it a part of the south of Edinburgh rather than the north of Edinburgh. If you study the data from there more closely, what you actually find is that as a result, the conditions there are much more likely to be more like what they are at a station such as Swanston, that they are at the botanic gardens in Edinburgh which is closer to the coast. Indeed, that was shown really well in 2019 when the temperature there rose to over 30°C, but only reached 26°C at the botanic gardens due to some slight cooling from the North Sea.

However, the botanic gardens are also very close to the city centre and that North Sea effect actually grows even stronger as you move even closer to the coast from there, yet there are no actual weather stations in that vicinity which can pick that up. As a result, it is possible for snow to be lying at the botanic gardens in the winter, with my area being completely snow free at the same time. That is something which these localised averages pick up really well, and I'm glad I've been able to find them.

This is good not just for long term averages though because it turns out that I can also utilise that to give a fairly accurate picture of what my own data would be like in terms of the actual current conditions if I actually had my own weather station to be able to record that. This then actually allows me to report the conditions specifically in my area without any need for an actual weather station. This will help me even more when I travel down to the Borders because there are no nearby major Met Office stations and yet, this system would probably still be able to give an accurate picture of the actual conditions there.

Furthermore, this also means that I would no longer have the same issues with missing data as I would for recording data from any station as this system automatically fills in those gaps. Another example of that comes with rainfall because I have lost count of the number of times when I have had a lot of rain here, only for very little of that to actually be picked up at any of the nearby major stations. This system actually picks that up really well though and with that, I can get a more accurate picture of what those rainfall totals actually are in this particular area.

With all of that in mind, I am now in the process of giving my own python script a complete overhaul and replacing the system which I have been using up until now, with a system which give reports for my actual area where I live, rather than for a station such as Gogarbank. That was a system which required me to understand more complex concepts such as dealing with raw BUFR/SYNOP data and looking back in it, it turns out that my old system was a lot more complicated than it actually needed to be.

On the other hand, the system which I am now creating to replace that will be a lot simpler and easier to use. As an example of what this system provides, I noticed that according to SEPA, there been 11.6 mm of rain during this month so far at the botanic gardens but only 10.6 mm at Gogarbank. Meanwhile, my new system is telling me that where I live, there has been a total of 12.4 mm of rain during this month so far. As for the temperatures, It is currently 8.2°C at Gogarbank and 7.4°C at the botanic gardens. Meanwhile, my new system is telling me that it is currently 8.8°C. This nicely shows that it is a west to east variation which is overriding everything in that regard just now. From that, I am seeing that this is a system which I am very happy with and so, I'm looking forward to bringing that into use in my future reports on here on the relevant threads for that.

Originally Posted by: johncs2016 

I have been downloading and analysing the HadUKGrid files of max/min temperature and rainfall for some time now. I use them in conjunction with the Met Office MIDAS historical Station Data to construct a temperature and rainfall daily history for my local area (Kew, London) and also to look at the details of temperature and rainfall data for particular notable weather events.

The files are in NetCDF (.nc) format covering Rainfall, Max/min Temp, MSLP, sunshine hours, vapour pressure, relative humidity, air frosts, ground frosts for 1,5,12,25 and 60 km grid squares over 1 day, 1 month, 1 season or 1 year (not for every measurement type)

However they are only updated annually in July up to then end of the previous year so they are no good for (say) analysing this years exceptional summer weather.

https://data.ceda.ac.uk/badc/ukmo-hadobs/data/insitu/MOHC/HadOBS/HadUK-Grid/v1.3.2.ceda 

My understanding is that they use all official Met Office recording stations in the UK daily data and pass this through an extrapolation/interpolation process to get the data down to 1km grid squares for each day then average/accumulate this to get up to the higher levels of time period or grid size.

NetCDF data is not too easy to analyse so I use a routine I purchased from a company called AgriMetSoft to extract the data over whatever time period and latitue/longitude range  from the correct file or set of files and present the extracted data in the form of an Excel spreedsheet with the dimensions of latitude and longitude with repeating groups for each time unit (e.g. 30 lat/lon matrices for each day of December 2010)

I then use conditional formatting by range and colour to develop maps of local events such as the one below

UserPostedImage

It is a great pity that these files only relate to dates up to 18 months or more before the current time. You first email on data availability is much more relevant to analysing more recent events (and I am still trying to get to grips with some of the comments you made about ogimet.com)  As you pointed out in that email, it is quite hard to get hold of local data for recent weather data


Martin

Richmond, Surrey

johncs2016
11 October 2026 10:41:40

I have been downloading and analysing the HadUKGrid files of max/min temperature and rainfall for some time now. I use them in conjunction with the Met Office MIDAS historical Station Data to construct a temperature and rainfall daily history for my local area (Kew, London) and also to look at the details of temperature and rainfall data for particular notable weather events.

The files are in NetCDF (.nc) format covering Rainfall, Max/min Temp, MSLP, sunshine hours, vapour pressure, relative humidity, air frosts, ground frosts for 1,5,12,25 and 60 km grid squares over 1 day, 1 month, 1 season or 1 year (not for every measurement type)

However they are only updated annually in July up to then end of the previous year so they are no good for (say) analysing this years exceptional summer weather.

https://data.ceda.ac.uk/badc/ukmo-hadobs/data/insitu/MOHC/HadOBS/HadUK-Grid/v1.3.2.ceda 

My understanding is that they use all official Met Office recording stations in the UK daily data and pass this through an extrapolation/interpolation process to get the data down to 1km grid squares for each day then average/accumulate this to get up to the higher levels of time period or grid size.

NetCDF data is not too easy to analyse so I use a routine I purchased from a company called AgriMetSoft to extract the data over whatever time period and latitue/longitude range  from the correct file or set of files and present the extracted data in the form of an Excel spreedsheet with the dimensions of latitude and longitude with repeating groups for each time unit (e.g. 30 lat/lon matrices for each day of December 2010)

I then use conditional formatting by range and colour to develop maps of local events such as the one below

UserPostedImage

It is a great pity that these files only relate to dates up to 18 months or more before the current time. You first email on data availability is much more relevant to analysing more recent events (and I am still trying to get to grips with some of the comments you made about ogimet.com)  As you pointed out in that email, it is quite hard to get hold of local data for recent weather data

Originally Posted by: lanky 

Just recently, I tried downloading data in netcdf format directly from Copernicus after signing up for an API key and that appeared to be successful. There are a number of tools which can be used in python which can then process those files but as you have said, these files are very difficult to process. When my script was running, it was appearing as though the download was successful but when I was actually trying to process them, I was getting a file format error.

 

After querying an AI bot about that, the answer which I got was that although the download appeared to be successful, there is a possibly that this might not actually have been the case, resulting in what was actually a corrupted download. In that scenario, it is possible that what I might actually have got is an html file disguised as an nc file. The nc format is a binary format and the real issue with that is only one bit needs to fail to download properly in order for the whole thing to crash.

That is why I am much happier after changing my focus by getting that data from open-meteo.com. With that, it is possible to get that data in plain JSON format which then completely avoids that issue. Furthermore, if I chose the Met Office as the model on open-meteo.com, I can actually get that same Met Office HadUK 2km grid data which you can get directly from the Met Office and I find it much easier to process from there.

On another note, I have also noticed that SEPA have revamped their rainfall data pages so that they are now using their new timeseries api for that instead of the previous api which only gave you a much more limited amount of data. The page for each station has also been revamped and also lists the catchment area which that station is in. It turns out as well that these are the same catchment names which are used in their weekly water scarcity status which means that I can get that station's rainfall data and its current water scarcity status without leaving the SEPA site.


The north of Edinburgh, usually always missing out on snow events which occur not just within the rest of Scotland or the UK, but also within the rest of Edinburgh.

lanky
11 October 2026 14:05:45

Just recently, I tried downloading data in netcdf format directly from Copernicus after signing up for an API key and that appeared to be successful. There are a number of tools which can be used in python which can then process those files but as you have said, these files are very difficult to process. When my script was running, it was appearing as though the download was successful but when I was actually trying to process them, I was getting a file format error.

 

After querying an AI bot about that, the answer which I got was that although the download appeared to be successful, there is a possibly that this might not actually have been the case, resulting in what was actually a corrupted download. In that scenario, it is possible that what I might actually have got is an html file disguised as an nc file. The nc format is a binary format and the real issue with that is only one bit needs to fail to download properly in order for the whole thing to crash.

That is why I am much happier after changing my focus by getting that data from open-meteo.com. With that, it is possible to get that data in plain JSON format which then completely avoids that issue. Furthermore, if I chose the Met Office as the model on open-meteo.com, I can actually get that same Met Office HadUK 2km grid data which you can get directly from the Met Office and I find it much easier to process from there.

On another note, I have also noticed that SEPA have revamped their rainfall data pages so that they are now using their new timeseries api for that instead of the previous api which only gave you a much more limited amount of data. The page for each station has also been revamped and also lists the catchment area which that station is in. It turns out as well that these are the same catchment names which are used in their weekly water scarcity status which means that I can get that station's rainfall data and its current water scarcity status without leaving the SEPA site.

Originally Posted by: johncs2016 

Thanks for the tip off about HadUK Grid. 

I had assumed that the CEDA data I referenced was the most recent available (currently up to the end of 2025).

I had not spotted that the Met Office also have much more recent versions of a lot of it  all the way up to End September 2026 (currently) outside of CEDA

I will update my own databases with more recent data now I know where to find it !


Martin

Richmond, Surrey

johncs2016
13 October 2026 08:58:25
As I'm continuing to develop my new data collection system, I've just made another discovery and it was my thinking about the data for the botanic gardens in Edinburgh. As I've mentioned on countless occasions, the data for there comes in SYNOP/BUFR format just as it does for Gogarbank. However, Gogarbank is on the WMO list (its WMO number is 03166) whereas the botanic gardens is not (its WMO number is 99180 and although that is also shown in any BUFR data for there, any SYNOP data from there just shows '00000' where the WMO number would normally be shown (the same is true at other similar stations such as Kew Gardens in London).

This means that the info which you get from this raw data for the botanic gardens is a lot more limited than the info which you get from that data for stations such as Gogarbank which are on the WMO list. An example of that shows up really well when it comes to sunshine totals because for a station such as Gogarbank, it is easy to get that info from the raw SYNOP/BUFR data. As I understand it, sunshine totals are also recorded at the botanic gardens but that is not included in the raw SYNOP/BUFR data for there.

For one thing, the 1991-2020 averages for the botanic gardens include figures for sunshine totals. This doesn't prove that this is something which is recorded there these days, but it does show that sunshine totals have been recorded there before. A while ago, it used to be possible to get that information from the Met Office's WOW site and I had even created a python script for doing that, but the Met Office are now in the process of retiring that site and that is going to make it more difficult to get that data in the longer data.

With my discovery of gridded data though, I have just discovered that even if there is no raw BUFR/SYNOP data coming in from a particular station, I can still get some form of data for there anyway from the gridded data. All I have to do for that is to plug in the exact latitude and longitude for the station in question and use that to get the gridded data for there from open-meteo.com. As an exercise, it would be great to be able to compare that data with what we get from any raw BUFR/SYNOP data coming from that particular as this would be a great test of just how accurate the gridded data actually is.

If anyone wishes to take up that challenge for their nearest station, they are more than welcome to do that. For me though, it is great that I can use this method to get any data for any station without having to rely on any raw BUFR/SYNOP/METAR data from that particular station, and that is I am liking this system quite a lot.


The north of Edinburgh, usually always missing out on snow events which occur not just within the rest of Scotland or the UK, but also within the rest of Edinburgh.

lanky
13 October 2026 09:44:22

As I'm continuing to develop my new data collection system, I've just made another discovery and it was my thinking about the data for the botanic gardens in Edinburgh. As I've mentioned on countless occasions, the data for there comes in SYNOP/BUFR format just as it does for Gogarbank. However, Gogarbank is on the WMO list (its WMO number is 03166) whereas the botanic gardens is not (its WMO number is 99180 and although that is also shown in any BUFR data for there, any SYNOP data from there just shows '00000' where the WMO number would normally be shown (the same is true at other similar stations such as Kew Gardens in London).

This means that the info which you get from this raw data for the botanic gardens is a lot more limited than the info which you get from that data for stations such as Gogarbank which are on the WMO list. An example of that shows up really well when it comes to sunshine totals because for a station such as Gogarbank, it is easy to get that info from the raw SYNOP/BUFR data. As I understand it, sunshine totals are also recorded at the botanic gardens but that is not included in the raw SYNOP/BUFR data for there.

For one thing, the 1991-2020 averages for the botanic gardens include figures for sunshine totals. This doesn't prove that this is something which is recorded there these days, but it does show that sunshine totals have been recorded there before. A while ago, it used to be possible to get that information from the Met Office's WOW site and I had even created a python script for doing that, but the Met Office are now in the process of retiring that site and that is going to make it more difficult to get that data in the longer data.

With my discovery of gridded data though, I have just discovered that even if there is no raw BUFR/SYNOP data coming in from a particular station, I can still get some form of data for there anyway from the gridded data. All I have to do for that is to plug in the exact latitude and longitude for the station in question and use that to get the gridded data for there from open-meteo.com. As an exercise, it would be great to be able to compare that data with what we get from any raw BUFR/SYNOP data coming from that particular as this would be a great test of just how accurate the gridded data actually is.

If anyone wishes to take up that challenge for their nearest station, they are more than welcome to do that. For me though, it is great that I can use this method to get any data for any station without having to rely on any raw BUFR/SYNOP/METAR data from that particular station, and that is I am liking this system quite a lot.

Originally Posted by: johncs2016 

There is also the MIDAS daily Met Office Station history data that might be of use to you.

I have used this a fair bit. It is in csv format so quite easy to access and analyse but the layout is a bit odd and I believe some of it is manually transcribed. Both your 2 Edinburgh sites are on there up to 2025. One from 1959 and 1 from 1998. They have a daily history of "actual" rather tan gridded local data but I do come across gaps from time to time

You need to register with MIDAS to get access but thati s quite straightforward

This is a url for one of the sites

https://data.ceda.ac.uk/badc/ukmo-midas-open/data/uk-daily-temperature-obs/dataset-version-201901/midlothian-in-lothian-region/19260_edinburgh-gogarbank/qc-version-1 


Martin

Richmond, Surrey

Rob K
13 October 2026 09:52:15
Re the HadUK grid I noticed that MetJam (Jamie McBean)'s list of hottest UK days for each year since 1875 also includes the maximum value from the HadUK grid for those days, since 1931: https://www.metjam.co.uk/blog/hottest-day-of-each-year-from-1875-onwards/

Generally it is pretty close to the actual recorded max, or slightly above (as you would expect if the hottest spot is missed by the recording network) but there are quite a few anomalies, eg 1942 the grid max was 32.09C but the Sprowston figure is 33.9C.


Yateley, NE Hampshire, 73m asl

"But who wants to be foretold the weather? It is bad enough when it comes, without our having the misery of knowing about it beforehand." — Jerome K. Jerome

johncs2016
13 October 2026 11:37:35

There is also the MIDAS daily Met Office Station history data that might be of use to you.

I have used this a fair bit. It is in csv format so quite easy to access and analyse but the layout is a bit odd and I believe some of it is manually transcribed. Both your 2 Edinburgh sites are on there up to 2025. One from 1959 and 1 from 1998. They have a daily history of "actual" rather tan gridded local data but I do come across gaps from time to time

You need to register with MIDAS to get access but thati s quite straightforward

This is a url for one of the sites

https://data.ceda.ac.uk/badc/ukmo-midas-open/data/uk-daily-temperature-obs/dataset-version-201901/midlothian-in-lothian-region/19260_edinburgh-gogarbank/qc-version-1 

Originally Posted by: lanky 

That is something which I had already known about but back in my earlier years on this forum, I can remember being a lot more finicky about having to deal with anything which required me to sign up for something in order to be able to access something. Back then, this even got to the point where I regarded that as deal breaker when someone on here mentioned that going down that road was a good idea which led me to not go down that road as I was refusing to sign up to anything.

A number of years then passed during which time, I was interested in what others had to say about that but wanted absolutely nothing to do with that myself. Nowadays though, I don't feel as bad about signing up to anything and I showed that by signing up for an API key to access the Copernicus ERA-5 land data. That has now given me the confidence to be able to sign up for other things as well. I may well as a result, have gone on to sign up for access to the MIDAS data in the end anyway but now that you mentioned that, I have decided to go ahead with that. As a result, I am now looking forward to exploring that much further.

EDIT:

Having signed up for access to the MIDAS data, I looked at your link which is for old data for up until 2019. At the top of that page though, there is a folder list and by navigating through that, I was able to find a much more recent version of the data which your link was pointing to, and which goes all the way up to this year.

What's interesting though is that I can get some old data from there for weather stations here in Edinburgh which are no longer in use, in addition to the three local stations which I have mentioned on so many occasions. With that in mind, Jerry from the south of Edinburgh might be interested to learn that there is actually a weather station right in his own neck of the woods on Blackford Hill in addition to the one at Swanston which he often mentions on here.

That station on Blackford Hill is even on the WMO list. It has a WMO number of 03173 and a WIGOS id of 0-20000-0-03173. According to ogimet.com, that station closed on 2 December 2015 but the ai bot on google.com has said that this is not actually the case. According to that, the introduction of the wigos system resulted in the Met Office making a decision to consolidate its station network here in Edinburgh with Gogarbank becoming the main station. Before then, SYNOP data (I don't think that BUFR data was even a thing back then) used to come from that station on Blackford Hill just as it does these days from Gogarbank and the botanic gardens, and that is why it is on the WMO list with an id of 03173.

When the Met Office consolidated its local station network, a decision was taken to stop transmitting SYNOP data from there but according to the Google AI, that date which was given as the date on which it closed, is actually just the date that it stopped transmitting SYNOP data. The Google AI bot claims that the station on Blackford Hill is still in operation today, and is managed by the Royal Observatory which is located there, having previously moved there from Calton Hill which is located just to the east of the city centre.


The north of Edinburgh, usually always missing out on snow events which occur not just within the rest of Scotland or the UK, but also within the rest of Edinburgh.

Viking3
13 October 2026 13:00:00

As I'm continuing to develop my new data collection system, I've just made another discovery and it was my thinking about the data for the botanic gardens in Edinburgh. As I've mentioned on countless occasions, the data for there comes in SYNOP/BUFR format just as it does for Gogarbank. However, Gogarbank is on the WMO list (its WMO number is 03166) whereas the botanic gardens is not (its WMO number is 99180 and although that is also shown in any BUFR data for there, any SYNOP data from there just shows '00000' where the WMO number would normally be shown (the same is true at other similar stations such as Kew Gardens in London).

This means that the info which you get from this raw data for the botanic gardens is a lot more limited than the info which you get from that data for stations such as Gogarbank which are on the WMO list. An example of that shows up really well when it comes to sunshine totals because for a station such as Gogarbank, it is easy to get that info from the raw SYNOP/BUFR data. As I understand it, sunshine totals are also recorded at the botanic gardens but that is not included in the raw SYNOP/BUFR data for there.

For one thing, the 1991-2020 averages for the botanic gardens include figures for sunshine totals. This doesn't prove that this is something which is recorded there these days, but it does show that sunshine totals have been recorded there before. A while ago, it used to be possible to get that information from the Met Office's WOW site and I had even created a python script for doing that, but the Met Office are now in the process of retiring that site and that is going to make it more difficult to get that data in the longer data.

With my discovery of gridded data though, I have just discovered that even if there is no raw BUFR/SYNOP data coming in from a particular station, I can still get some form of data for there anyway from the gridded data. All I have to do for that is to plug in the exact latitude and longitude for the station in question and use that to get the gridded data for there from open-meteo.com. As an exercise, it would be great to be able to compare that data with what we get from any raw BUFR/SYNOP data coming from that particular as this would be a great test of just how accurate the gridded data actually is.

If anyone wishes to take up that challenge for their nearest station, they are more than welcome to do that. For me though, it is great that I can use this method to get any data for any station without having to rely on any raw BUFR/SYNOP/METAR data from that particular station, and that is I am liking this system quite a lot.

Originally Posted by: johncs2016 

I don't know if you have explored the full BUFR data on Ogimet, but you can get the full text listing of each BUFR observation from the menus on the site.

On the left hand side - BUFR REPORTS - Get Bufr reports - Emisor Center - type in EGRR - select the hours you want (default is 24 hours but this presents a lot of reports so it's better to just specify a single hour) - press the button that says Ver

For the UK 99 stations you are looking for bulletins containing ISMN62 / ISIN62 / ISNN62. M is for main hours - 00/06/12/18, I is intermediate hours - 03 / 09 / 15 / 21 and N is non-intermediate hours - 01 / 02 / 04 / 05 etc

The file for 1200 GMT today is 20261013121735_ISMN62_EGRR_131200.bufr - click on this and you should see lots of subsets. Edinburgh Botanic Gardens is often subset 84 or subset 85. Click on the subset button and below you should see a detailed text listing of all the elements contained in the BUFR report. For the Botanic Gardens there is no information on sun and it appears that this is not something that has ever been recorded.

Edinburgh Gogarbank is contained in bulletins with ISMN01 / ISIN01 / ISNN01 and is usually subset 14. It contains all the observation details including sunshine recorded in minutes (53 in the case of 1200 GMT today) compared to the SYNOP code of 55308 which translates to 0.8 hours or 48 minutes.

I can't see a way of extracting the data other than by copy/paste as the data only opens up when you click on the subset button, but you may be able to work out how to do it.


Keith

Aboyne, Aberdeenshire

135m asl

johncs2016
13 October 2026 13:17:33

I don't know if you have explored the full BUFR data on Ogimet, but you can get the full text listing of each BUFR observation from the menus on the site.

On the left hand side - BUFR REPORTS - Get Bufr reports - Emisor Center - type in EGRR - select the hours you want (default is 24 hours but this presents a lot of reports so it's better to just specify a single hour) - press the button that says Ver

For the UK 99 stations you are looking for bulletins containing ISMN62 / ISIN62 / ISNN62. M is for main hours - 00/06/12/18, I is intermediate hours - 03 / 09 / 15 / 21 and N is non-intermediate hours - 01 / 02 / 04 / 05 etc

The file for 1200 GMT today is 20261013121735_ISMN62_EGRR_131200.bufr - click on this and you should see lots of subsets. Edinburgh Botanic Gardens is often subset 84 or subset 85. Click on the subset button and below you should see a detailed text listing of all the elements contained in the BUFR report. For the Botanic Gardens there is no information on sun and it appears that this is not something that has ever been recorded.

Edinburgh Gogarbank is contained in bulletins with ISMN01 / ISIN01 / ISNN01 and is usually subset 14. It contains all the observation details including sunshine recorded in minutes (53 in the case of 1200 GMT today) compared to the SYNOP code of 55308 which translates to 0.8 hours or 48 minutes.

I can't see a way of extracting the data other than by copy/paste as the data only opens up when you click on the subset button, but you may be able to work out how to do it.

Originally Posted by: Viking3 

That is exactly what the python scripts in my old system were using up until now. I hadn't got as far as extending that to include the data for the botanic gardens, but I had mentioned a few times recently that you can do that.

That said though, I'm glad that you posted that anyway as that will be a great help to any newcomers to this field who are reading this, and that at least saves me from doing that myself.

EDIT:

I forgot to add that for anyone trying to use a python script to get the data for a nearby station in their own area, there is a python package called pdbufr. It uses the ECMWF's eccodes python package which then does all of the heavy lifting to decode the relevant parts of these BUFR data files (with the eccodes package itself, that task is still possible but more difficult to achieve). Once that is done, it is then just a simple matter of getting that data into the right format for reporting it.


The north of Edinburgh, usually always missing out on snow events which occur not just within the rest of Scotland or the UK, but also within the rest of Edinburgh.

Remove ads from site