• Hello, please take a minute to check out our awesome content, contributed by the wonderful members of our community. We hope you'll add your own thoughts and opinions by making a free account!

Misc General Banter Thread

To set the tone for December: Post your best Saturday football snow photos.....so far here's Alma vs Wisconsin-Platteville.
 
FortMillSCWx said:
To set the tone for December: Post your best Saturday football snow photos.....so far here's Alma vs Wisconsin-Platteville.
Click to expand...

3522d78c66c9a961bb95534cffafadbf.jpg



Sent from my iPhone using Tapatalk
 
Waking up to 38 and rain this morning. 34 and rain for Tuesday. Winter in the Carolina’s is here!! lol. 😂 Hopefully we step down into glory through the month. ❄️🌨️
 
Rain Cold said:
I can't believe I'm the first person to post this. Y'all need to wake up in here! Winter is coming.

View attachment 177257
Click to expand...
Im sure its been covered before in here but I haven't seen it and have to ask. Why are there differences in the graphics on the models depending on which site you look at? I know the precip maps are not what you should be looking at but more so the atmospheric conditions.
 
jeremyt said:
Im sure its been covered before in here but I haven't seen it and have to ask. Why are there differences in the graphics on the models depending on which site you look at? I know the precip maps are not what you should be looking at but more so the atmospheric conditions.
Click to expand...

So on a technical side, the data is all the same inside the grib files. IE the raw data inside the grids (the km thing), like little squares.

When you manipulate the grid or generate graphics from the grib data onto the maps, the maps can appear differently depending on how much resolution their actual plot is.

For example, one site uses a low resolution usa map, it'll appear more blockish.

The color tables matter a lot too. Think of it like a radar app such as GRLEVELX. With a software like that, you can get finer and finer smoothing with more color steps in the color table. In other words, if your color table has 4 colors it in, you're going to see sharper cut offs. But if someone uses a color table that has say, 15 colors and they progressively step up in shade.. like light light blue > light blue > medium blue > dark blue etc, you will see better smoothing.

When it comes to precipitation type, some people lazily generate them and just go off the actual model's offered parameters in the grib file. Some models will for example, have a "categorial freezing rain" type of table. When models have this, scripts can be written to actually take a look at that table along with the actual precipitation and decide if it should be printed out as ice or not.

You won't find many of these types of maps, because the more advanced the scripting, the less friendly it is for both a programmer/map site, and more computational power it would require over the whole dataset. Therefore you end up with lots of maps showing ZR and sleet as snow on the snowfall maps. This is because they have not gone through the process of actually separating out the precipitation types with logical scripts. Again, sometimes it's laziness, sometimes the model's grib file doesn't provide it, and other times, its for computational reasons.


A good example is this: take a look at the instantweathermaps site. It's one of the absolute fastest websites, along with StormVISTA. The maps are low resolution and not very detailed. So while we get an initial idea of the overall forecast, we wait for the higher resolution maps that have better smoothing/graphics quality. Those two sites are meant to be as fast as possible, while offering lower quality, but they're generally using the same exact base model data.
 
Last edited:
weather nerd said:
So on a technical side, the data is all the same inside the grib files. IE the raw data inside the grids (the km thing), like little squares.

When you manipulate the grid or generate graphics from the grib data onto the maps, the maps can appear differently depending on how much resolution their actual plot is.

For example, one site uses a low resolution usa map, it'll appear more blockish.

The color tables matter a lot too. Think of it like a radar app such as GRLEVELX. With a software like that, you can get finer and finer smoothing with more color steps in the color table. In other words, if your color table has 4 colors it in, you're going to see sharper cut offs. But if someone uses a color table that has say, 15 colors and they progressively step up in shade.. like light light blue > light blue > medium blue > dark blue etc, you will see better smoothing.

When it comes to precipitation type, some people lazily generate them and just go off the actual model's offered parameters in the grib file. Some models will for example, have a "categorial freezing rain" type of table. When models have this, scripts can be written to actually take a look at that table along with the actual precipitation and decide if it should be printed out as ice or not.

You won't find many of these types of maps, because the more advanced the scripting, the less friendly it is for both a programmer/map site, and more computational power it would require over the whole dataset. Therefore you end up with lots of maps showing ZR and sleet as snow on the snowfall maps. This is because they have not gone through the process of actually separating out the precipitation types with logical scripts. Again, sometimes it's laziness, sometimes the model's grib file doesn't provide it, and other times, its for computational reasons.


A good example is this: take a look at the instantweathermaps site. It's one of the absolute fastest websites, along with StormVISTA. The maps are low resolution and horrible. So while we get an initial idea of the overall forecast, we wait for the higher resolution maps that have better smoothing/graphics quality. Those two sites are meant to be as fast as possible, while offering lower quality.
Click to expand...
Thanks, that makes sense!
 
jeremyt said:
Thanks, that makes sense!
Click to expand...
No worries.

I'd also like to note, this is the biggest reason why a site like ours doesn't have weather model data.
- It costs a lot of bandwidth to download the model files.
- We need a lot of computational power to process it.
- We then would need to offer the images, costing even more bandwidth

While there are some services/models like NCEP that you can "zoom in" and just request what you're interested in from the model files, it's still a lot of back and forth with data. For the model data providers that don't include this tech, we see grib files get very large into the multiples of gigabytes for each forecast hour.

Now considering there are a lot of models these days, you can see how the cost would exponentially increase when each of these models are producing 4+ forecast cycles per day. That's not even counting the storage space needed.
 
jeremyt said:
Im sure its been covered before in here but I haven't seen it and have to ask. Why are there differences in the graphics on the models depending on which site you look at? I know the precip maps are not what you should be looking at but more so the atmospheric conditions.
Click to expand...
Give me an example of what you mean.
 
Rain Cold said:
Give me an example of what you mean.
Click to expand...

Same time frame on the 6z GFS on Pivotal, compared to the 6z GFS you posted from TT. Pivotal barely had snow into Western NC but TT had it through most of SC. Again I know that looking at the atmospheric conditions on the models is what matters but always wondered why the sites showed differently. What weather nerd said made sense to me.
1764531099366.png
 
You must log in or register to reply here.
Back
Top