Dr Jack forecast high too low

It seems that for many months Dr Jack's forecast temperature is off by -4° F consistently, sometimes I've see it too low by 7°.

Has anyone else seen this bias?

Tomorrow it is forecast 3.7° lower than the other commercial forecasts
(69.3 vs 73). I've seen this happen in the observed high temperature pretty often, not just the forecast. 4+ degrees is enough to throw off a lift and cloudbase prediction pretty far, and I've missed some good days because of this.

Has anyone else seen this?
I thought if there was widespread issues with this I could provide better feedback to Jack.

I wonder if the many updates to the RUC and NAM models have created some issues with the data or computations.


DrJack (www.drjack.info/BLIP) miniBLIPSPOT
Lat,Lon= 32.79,-84.44 (429,144)
Feb 19 - NAM
---------------------------------------
VALID.TIME 18z 21z VALID.TIME
ForecastPd 24h 27h ForecastPd
---------------------------------------
W* 430 334 W*
Sfc.Heating 217 98 Sfc.Heating
Hcrit 4153 3641 Hcrit
BL Depth 5213 5178 BL Depth
BL Top 5908 5873 BL Top
Hgt.Variab. 753 497 Hgt.Variab.
B/S Ratio 6 6 B/S Ratio
BL Wind 13 10 BL Wind
Direction 327 317 Direction
Wind Shear 19 11 Wind Shear
Max.Converg -5 1 Max.Converg
CUpot 2142 1319 CUpot
CUbase 3766 4554 CUbase
ODpot 957 953 ODpot
ODbase 4951 4920 ODbase
maxRH 88 90 maxRH
CAPE 20 20 CAPE
Temp@2m 67.2 69.3 Temp@2m

On Saturday, February 18, 2017 at 9:05:38 PM UTC-5, wrote:
It seems that for many months Dr Jack's forecast temperature is off by -4° F consistently,

Well one good work around is to use the Hcrit for the height, then add to it any additional height shown in the Thermal Height Variability, parameter. Just by coincidence(?) it shows what would happen if 4 degrees hotter.


Thermal Height Variability
This parameter measures the atmospheric stability above the BL and thereby indicates the variability of the BL top (TI=0) height which can arise from (1) actual variations in surface temperature over the region encompassing a model grid cell due to surface changes, etc., (2) variations in actual surface elevation which are omitted by the smoothed topography over a model grid cell (since surface elevation changes are effectively changes in surface temperature), or (3) error in the model's surface temperature prediction. The value given is the expected height change which would be produced by a surface change of 4 degF, but is usually best evaluated in a relative sense. (Numerically this parameter is calculated as the difference between the TI=+4 and TI=0 heights - strictly speaking this only give the effect of an increase in surface temperature, since the effect of a surface temperature decrease cannot be easily estimated.) Weak stability above the BL top gives large variability values which are often good for soaring, since thermal heights due to small sub-grid-scale variations can then be much higher than the predicted average BL height. However, high variability values can also be accompanied by soaring conditions being much poorer than those predicted if actual surface temperatures are much cooler than those predicted by the model. In short, this parameter represents the uncertainty of the predicted BL height. [see the BL Variability diagram].

http://www.drjack.info/BLIP/INFO/DOC/blvariability.jpg

Did you notice the same in the NAM model as well? I usually don't use blipspot.
one thing I noticed since last year with the NAM model is that the cloud base prediction is persistently lower than actual. It used to be more accurate.

Ramy

On Sunday, February 19, 2017 at 10:44:50 PM UTC-5, Ramy wrote:
Did you notice the same in the NAM model as well? I usually don't use blipspot.
one thing I noticed since last year with the NAM model is that the cloud base prediction is persistently lower than actual. It used to be more accurate.

Ramy

FWIW it seems to be the same issue in XCSkies too. Apparently the model
has a bias these days. Apparently it was a pretty good soaring day yesterday
around here (5K AGL), even though the blipmap forecasted weak, low, ripped up lift.

On Monday, February 20, 2017 at 8:15:18 AM UTC-8, wrote:
On Sunday, February 19, 2017 at 10:44:50 PM UTC-5, Ramy wrote:
Did you notice the same in the NAM model as well? I usually don't use blipspot.
one thing I noticed since last year with the NAM model is that the cloud base prediction is persistently lower than actual. It used to be more accurate.

Ramy

FWIW it seems to be the same issue in XCSkies too. Apparently the model
has a bias these days. Apparently it was a pretty good soaring day yesterday
around here (5K AGL), even though the blipmap forecasted weak, low, ripped up lift.

Matt, how did that compare to the RASP?
Jim

On Monday, February 20, 2017 at 10:15:18 AM UTC-6, wrote:
On Sunday, February 19, 2017 at 10:44:50 PM UTC-5, Ramy wrote:
Did you notice the same in the NAM model as well? I usually don't use blipspot.
one thing I noticed since last year with the NAM model is that the cloud base prediction is persistently lower than actual. It used to be more accurate.

Ramy

FWIW it seems to be the same issue in XCSkies too. Apparently the model
has a bias these days. Apparently it was a pretty good soaring day yesterday
around here (5K AGL), even though the blipmap forecasted weak, low, ripped up lift.

a nice reminder that sometimes you just have to take a tow and give it a try

Like I always answer to the other club members, when they ask what's the weather going to be like today? "Good as long as I get to fly"