NOAA predictions: Grib vs. Website??

[alavarre]

Delighted with the program, thank you.

Would enjoy hearing of others' experience in predicting snowfall:

Are the precipitation levels actually in centimeters per hour, not the millimeters per hour shown in the sidebar display?

=====

Right now zyGrib is showing between 0.7-2.5 MM/h of precipitation (presumably snow) for 41° 25' 13"N 70° 56' 20" for the next fifteen hours, based on 20140215_210922_.grb.bz2, which I assume is a NOAA-GFS product.

+ Weather map has precipitation selected

+ The "Snow Depth" box in the "Weather data" panel shows 19 CM after the storm passes at 10AM local on Sunday 2/16. This is curiously close to my calculations below in MILLIMETERS (NOT centimeters)

+ For each three hour interval of the storm I read the "Precipitation" box (MM/H) in the "Weather data" panel, multiply the average over the interval by three for the three hour time interval and add it to the previous total, then do the unit conversion. This gives a maximum accumulation of 0.97 INCHES:

Time MM/h Interval Total MM Inches
13:00 0.02
16:00 1.65 2.505 2.51 0.10
19:00 2.17 5.73 8.24 0.32
22:00 2.34 6.765 15.00 0.59
01:00 1.86 6.3 21.30 0.84
04:00 0.13 2.985 24.29 0.96
07:00 0.06 0.285 24.57 0.97
10:00 0 0.09 24.66 0.97
13:00 0 0 24.66 0.97

However, the NOAA website:

http://forecast.weather.gov/MapClick.ph ... mnyrapSeRt

says between four to eleven inches. That is at least an order of magnitude difference...

Now, if the sidebar is actually CM/H, rather than MM/H then the answer is 9.7 (not 0.97) inches, which matches the website prediction...

What am I missing?

Kind regards (stay wahm in New England), Andy

[jza]

Hi,

Unit is mm/h, as mentioned.

The displayed value is the average over a mesh of 0.5x0.5 degrees and for a period of 3 hours.
You can not directly compare such a value with a very local data.

[alavarre]

>You can not directly compare such a value with a very local data.

Mmm.

Well, then that particular field doesn't do much good...

What many interests people is how much snow do they have to shovel in the morning...

So:

Start with current accumulation (e.g., 0)
+
mm/h x hours(3) = total accumulation during the three hour period
= total accumulation at end of three hours.

Repeat for the next periods...

Which is what my algorithm does.

but in that manner the answer is off by a factor of ten.

Perhaps you might explain how to get current accumulation, ignoring the melting in the intervening time.

Thanks again, excellent app.

Kind regards, Andy

[alavarre]

Bonjour Jacques...

Thank you again for this excellent program.

Again on the subject of snow: How might I correct this?

The current grib for New England does not give any Snow Depth values for the coastal region around Newport, RI:

https://www.dropbox.com/s/xyj5uw4ao4whn ... h.png?dl=0

(the img button above cannot read the link)

I have selected in NOAA-GFS Standard the parameters:
Total precipitation
Snow (snowfall possible)
Snow (depth)
As seen in the image the data exist further inland but not in this coastal region.

Thank you again.

Best regards, Andy

[DomH]

Again on the subject of snow: How might I correct this?

I have selected in NOAA-GFS Standard the parameters:
Total precipitation
Snow (snowfall possible)
Snow (depth)
Best regards, Andy

Hello,

I do not think the GFS Model to be the best if you want precise info at a very local level.

A short and simple explanation on the different types of models : Openskiron : http://openskiron.org/en/why-are-regional-models-needed, but of course you find a lot of very good infos about models on the web (given by scientists and pros of weather forecast).