It will be interesting to see your weather observations 
Using windy pro app, 3 of the 9 models say rain in the morning ECMWF model, the one I am using says dry with rain in the afternoon
I am certain I will need to tweak the weather models used for input. 
There has been a few occasions recently when the Met Office rain radar has shown light drizzle falling here, but none was detected. Humidity at ground level was quite low (advancing warm front) so I believe that the drizzle was evaporating before it reached the ground.
..thatās what I do.
Having lived for on 80 years, you get a feel for what the weather will be; donāt need an app!
Geoff vk3sq
That approach probably works quite reliably in a continental climate like VK. Big, slow moving weather systems and repeatable weather patterns given known wind directions. But in a mountainous maritime environment such as ZL or MM things vary so much faster and over short distances. One work trip last year we pulled out under cloudless skies from back-country camps (rightly) on a forecast of 70-150mm over 24 hrs across the various models. The actual rainfall was 750mm at the nearest weather station (Roaring Billy) in that 24 hrs. Yet 30km away at Makarora <6mm fell.
Likewise next week I need to decide whether to get my crew of 4 to drive 600km up the West Coast, then chopper them into alpine camps for 8 days of work. Given that level of outlay and commitment, Iāll take as many weather models from as many sources as I can get, please! SOTA is slightly less committing than that, itās only my diesel and my time that gets wasted, but Iām still happy to see more data rather than less. I can then make my own determinations based on personal knowledge of how weather behaves in that specific area / terrain, given the various predictions as a starting point.
Yes, I agree about the weather in large continental countries. They usually have more kinds and extremes of weather over the year but small islands (ZL, UK) surrounded by seas and oceans often have more variation hour-to-hour.
Mark Twain famously lampooned New England weather for its extreme unpredictability but when I lived there in the 1980s the variation during the day wasnāt so much as in the UK. Even small variations in the position of the jet stream w.r.t the UK (e.g. to the N, to the S or over it) make short-term changes to wind direction, temperature, and rainfall.
On extremes, from my point of view, the UK has (most years) four proper seasons [despite cynics with confirmation bias who remember a year with no proper summer]. IMO, the spring and autumn (fall) in N.E. lasts only a few weeks going from deep winter to deep summer and back again. Our next-door neighbo(u)rs from the Mid-West say it was even more extreme back home.
I have made several changes.
The weather model has now been changed to the newer ECMWF AI model as this offers an hourly 9km resolution.
All altitude interpolations now come from the actual summit height.
A key on the page would be useful as I donāt get what some of the data is telling me.
Can the actual temperature be made more prominent?
As a perfect illustration of the changeability of the weather in this part of the world, 10 minutes ago before pegging out a full load of washing on the line outside, I stepped out to check the weather: blue sky and large cumulus clouds - great. But before I had even finished hanging the clothes out a medium to dark grey cloud drifted over dumping some drizzle.
Itās why I always carry my KX2 or FT817 in a waterproof case or bag should the weather decide to be fickle.
Thanks for the feedback Richard, I will swap the temperature to the conventional view.
I have also noticed gust (g) are shifting the data row, I will try and fix that today. Well when I say I will fix it, I will have a word with my programmer Claude He is much quicker than me.
Out of curiosity what device are you using?
Iām waiting for the last line of that songā¦
āin short, weather will be normal for the time of yearā.
It feels like a long time since the weather has been normal for the time of year. At least 3 years, anyway. Maybe the shift to El NiƱo will put the seasons back where they belong, but Iām not holding my breath.
Chrome, Windows 11, Desktop PC
Large cumulus (cumulus congestus?) that early is a warning that things may go downhill rapidly, its frequently an early stage in the formation of cumulonimbus, and the rapid unplugging of antennas!
Thanks was just wondering incase it was phone or similar.
Should all be fixed now
I have dropped a legend in at the bottom of the page for now.
I will tidy the legend up with a html link later
Funny you should mention that Brian. As I write a biblical-strength hailstorm has just started. Not good for drying the washing (my wife and I dashed to bring it inside) nor for 2m Yagi-based dxāing this lunchtime.
Great work on the app. Paul, and for being responsive to feedback.
I love it, except for not being allowed out tomorrow because the app. suggests I postpone, on what is pretty reasonable weather for a 1157m (3800ā) Cairngorm mountain in Spring.
With your āGo or No Goā warning at the bottom, you are taking some responsibility away from the activator and placing it on your own shoulders.
Iāve bookmarked it and will use it. Displays well on an android phone.
Thanks for the explanation. It also helps to spot the large temperature is the result with wind chill added which is useful, Iām just use to thinking the other way - show me the temp and Iāll apply windchill myself.
My view is that will be better. 
More feedback, it may depend on eyes but I donāt find the font and colours that readable on a monitor. I havenāt tried it on a phone. It may just be my preferences for a more standard/boring font etc.
That seems low, very low. 1987 was just yesterday in my mind but desktop computers were still somewhat nascent and ātoy-likeā. Big desktop systems would be 68020+68881 which were about 5MIPS and 250kFlops. The 80386/387 combination was about 1 year old and had rapidly gained traction with similar performance. However, Intel was still forcing segmented architecture programming on us at that time 
The real performance machines were still ābig-ironā mainframes. But the writing was on the wall for them as desktop micros were destroying the mini-computer sales and it was only a matter of time. Still Control Data, IBM, Cray and Amdahl were selling high performance machines. The Met Office own computer history shows they replaced their 1972 IBM 360/195 (fast for its time) with a Control Data Cyber 205 in 1982 which means by 1987 it would be well bedded in and being worked hard. The Cyber 205 was a Fortran/Assember monster and even the base model could process 800 million 32 bit vector operations/sec. Bigger models had multiple vector stream processors. Or 400 million 64 bit operations. That would typically be for multiply and add combinations. Vector processing is designed for processing massive single dimension arrays of data which is a typical workload in simulation / numerical computation problems. In addition the Cyber205 could perform 50MIPS of scalar data processing in parallel with the 400/800milion vector operations.
So their big computer could run at 400MFLOPs and 50MIPS. Thatās 10x faster than the fastest desktop computer standard performance and 1666x faster on floating point workloads. In 1987 that was still kick-ass performance though NEC and Cray were producing faster vector machines. However, the cost of installing new mainframes and then porting over all your custom software to new architectures meant you couldnāt update to faster hardware quickly.
For comparison I have a 2022 vintage Intel i7 12700 CPU in my desktop (30x30x9cm box) which can run at 765 gigaFLOPS compared to the Cyberās 400 megaFLOPS, or 1900x faster. And a 12700 is 3 generations off the leading edge Intel/AMD CPUS.
Me too, itās now change to a white background
It looks quite different now than it did yesterday, including using a different weather model.
And the powers that be want to ban clothes driers! 
