Rain and storm for Cpt

Status
Not open for further replies.
hese are the temperatures that are showing variances of 25C to 36C on a specific day on different weather websites.
Gordon_R said:
Many forecasts nowadays are for the Cape Town CBD (or Waterfront), not the airport weather station. There are several computer models, and some are sufficiently detailed that they can differentiate between the two, but you may not know which until you look at the track record. I always look at the SAWS aviation forecast, since that is for a specific location (and I can do my own interpretations from that).

There are significant technical challenges in extrapolating from one of the (several) gridded 3D weather model to a specific location, and each prediction you see will have different biases. In addition there are often random variations (hard to get below 2-3C) depending on wind direction during a fickle weather event.

Edit: You get what you pay for (on the internet or TV), all of the forecasts you see are 'free' and designed for a large group of people. There are meteorologists who can produce a custom output just for you, and it may be slightly more accurate overall, but not worthwhile unless you have thousands of bucks at stake. In principle every forecast should show a range of probabilities, but this is hard to convey, and might be even more confusing.

An overview of the topic can be found in this article: https://en.wikipedia.org/wiki/Model_output_statistics
Click to expand...
A lot of interesting information, that still doesn't answer my basic question:

"temperatures that are showing variances of 25C to 36C on a specific day on different weather websites. "

That is a 11C variation that surely cannot be explained away by the info provided above?
Anyway, I'm sure the flamers are hovering, so I'll let it go.

Thanks @Gordon_R for responding in such detail. I will give SAWS a go.
 
MightyQuin said:
A lot of interesting information, that still doesn't answer my basic question:

"temperatures that are showing variances of 25C to 36C on a specific day on different weather websites. "

That is a 11C variation that surely cannot be explained away by the info provided above?
Anyway, I'm sure the flamers are hovering, so I'll let it go.

Thanks @Gordon_R for responding in such detail. I will give SAWS a go.
Click to expand...

The range in predicted temperatures today (Wed 24th) is indeed real, due to a combination of the factors: geographic differences, and random variations in wind direction.

I don't have a particularly detailed chart, but this one gives some idea of the extremes:

t14[20181024_12Z_24Z].gif
 
Gordon_R said:
The range in predicted temperatures today (Wed 24th) is indeed real, due to a combination of the factors: geographic differences, and random variations in wind direction.

I don't have a particularly detailed chart, but this one gives some idea of the extremes:

View attachment 568581
Click to expand...
Maybe I wasn't clear enough:
That range of temperatures I have mentioned a few times are on different websites for the same location: Cape Town CBD.
 
MightyQuin said:
Maybe I wasn't clear enough:
That range of temperatures I have mentioned a few times are on different websites for the same location: Cape Town CBD.
Click to expand...
I think you missed this part from Gordon_R;

You get what you pay for (on the internet or TV), all of the forecasts you see are 'free' and designed for a large group of people.
Click to expand...

All these free and cheap apps get their data from different sources, models and different stations. Some with low resolution covering a large area, which explains the inconsistencies in forecasting for the same location.
 
MightyQuin said:
Maybe I wasn't clear enough:
That range of temperatures I have mentioned a few times are on different websites for the same location: Cape Town CBD.
Click to expand...

It is quite possible that some of the forecasts are wrong, but today (Wed 24th) really is going to be fickle in terms of wind direction and temperature. It is statistically impossible to be right every time, and it is much easier to make predictions when conditions are consistent.
 
At the risk of getting too technical: One of the major issues with today's outlook (wed 24th) is that the air temperature profile with height is inverted (it is currently warmer at the top of Table Mountain than it is at sea-level). This is because the air is not uniformly mixed, but instead there are layers of cold and warm air moving in opposite directions. Depending on the exact wind profile and surface location, this can produce wildly different results.

The following SAWS aviation chart for latt 34S long 18E (offshore of Hout Bay) gives some idea of the issues. By contrast, the chart for 19E (near Jonkershoek) is completely different, due to the lack of cold maritime air.

P.S. It is currently too dry, but coastal fog forms under these kind of conditions (with large contrasts in temperature).

xl-3418[20181024_00Z_24Z_large].gif
 
@Mighty Quinn, which are the 5 apps you looked at?
Have you found out which base model each use and do they all update their predictions at the same time based on the same model runs?
You might be quite surprised to find out how "old" (up to 72 hours old) the data is that some of these apps use.
That generally (for me anyway) explains the wildly different predictions we see being offered.
 
The geographical coverage of yr.no seems to be very good (although no single model is always perfect). Have a look at the hourly predictions for these 3 stations today:
https://www.yr.no/place/South_Africa/Western_Cape/Sea_Point/hour_by_hour.html
https://www.yr.no/place/South_Africa/Western_Cape/Cape_Town/hour_by_hour.html
https://www.yr.no/place/South_Africa/Western_Cape/Cape_Town_International_Airport/hour_by_hour.html

Sea Point shows temperatures actually dropping between 10am and midday, as the wind changes. The range of maximum temperatures (over a range of 20km) is almost exactly the same as your original comment (25C to 33C). Further inland is even hotter!

HeftyCrab said:
I mainly use

https://www.windy.com/-33.932/18.412?-33.937,18.412,15
Click to expand...

LOL, that windy.com animation uses enough processing to overheat my CPU fan!
 
Exactly correct. Being pedantic about precise values is an exercise in futility. Not even the peak temp times agree. Weather predictions are there as a guide only. Unless you are in an industry where accuracy is essential such as aviation, there is absolutely no value in trying to be precise.

As the old Afrikaans idiom goes (I hope I have got it right):

"As die son skyn, vat jou jas. As did reen, maak wat jy will"
 
Temperatures seem to be tapering off slightly from Monday but we still need to get through the weekend. 35 on Saturday and 34 on Sunday. Sheesh...
 
Jonkershoek via yr.no has rain for coming MON_pm/TUE/WED/THU/FRI_am - fingers X'ed it does not change!
 
air said:
Jonkershoek via yr.no has rain for coming MON_pm/TUE/WED/THU/FRI_am - fingers X'ed it does not change!
Click to expand...

Several recent model runs show a significant change in the weather starting on Mon 29th. Sharply lower temperatures (10-15C cooler) will see an end to the current heatwave, though rainfall predictions are very uncertain at this stage..
 
I was toying with the idea of doing a trail run in Paarl on Saturday morning. Yeah, that's not going to happen!
 
Just when you think winter has gone for good, the forecasts throw up something freaky! A few recent model runs show light snow over the high ground of the SW-Cape on Tues 30th or Wed 31st:
https://www.mountain-forecast.com/peaks/Groot-Wellington-Sneeukop/forecasts/1703
https://www.yr.no/place/South_Africa/Western_Cape/Matroosberg~3364263/long.html

Even if the snow does not occur, the cold and wet conditions will mark a huge contrast with the current heatwave. Some places could see a 20C drop in daytime temperatures (mid 30s to mid teens) between Sun 28th and Tues 30th:
https://www.yr.no/place/South_Africa/Western_Cape/Paarl/long.html
 
Last edited:
Status
Not open for further replies.
Back
Top
Sign up to the MyBroadband newsletter
X