Going Negative - Another Look

Last season/s five-part series 'Going Negative' took a look at the conventional wisdom (CW)...from the perspective of NEWxSFC snow storms...about the 'need' for the NAO to be negative or the PNA to be positive or both to get good snows in the Northeast and mid-Atlantic. The main conclusions were 1) there was no preference in sign for PNA...2) AO was generally negative...and 3) contest snow storms were more likely when the NAO was positive. The data set was limited by the number of years the contests have been held and the number of storms where PNA...NAO...and AO data were readily available.

The question has come up again recently...as it does in one form or another every year...about the teleconnection CW. Near-record and / or well-above period-of-record normal (PORN) snows were observed in the Northeast last month during 'unfavorable' conditions (-PNA...+NAO)...which suggests that region/s snows are not dependent on the CW.

So..if the CW does not apply to the Northeast...then does it apply to the mid-Atlantic or is it more a myth?

The Synoptiscope® has been in VCP 31 since JAN 02 and indications are it will stay that way for @ least another week. We/ll take another...broader look at the 'Big 3' teleconnection indices while we wait for the next snow storm to see if the CW can stand the light of data in the mid-Atlantic.

Daily snow fall data (1950 - 2007) from U of Utah and daily normalized indices for PNA...NAO...and AO will be analyzed to find the probabilities for +4" and 12/24hr warning criteria events for several contest stations in the M-A. UofU data has missing values for portions of '96 and '97...which is unfortunate b/c of the bonus snows observed in JAN '96...but it/s the best available free data.

First up: Philly!

Interim Standings

Click to enlarge
Shanabe takes possession of 1st place after four snowstorms with a best '3 out of 4' average SUMSQ Z-Score of -1.225...having moved up from 4th place in the previous interim standings. He was propelled into the lead by his run-away 1st place finish in Contest #4.

Donsutherland1 (-1.182) and TQ (-0.589) drop one click to second and third place...respectively.

Herb @MAWS (-0.443) holds on to 4th place. Donald Rosenfeld (-0.430) made the largest move by advancing from 9th to 5th.

SUMSQ errors are normalized with a 'Z-score' for each contest snow storm...then averaged to compute the standings. Forecasters who have entered at least 2/3 of all snow storm contests are included in the standings.

If a forecaster has participated in more than two-thirds of all snow storm contests...then Z-scores from their 'best two-thirds' forecasts are used to calculate interim and final standings. Same idea as dropping the lowest quiz score before computing the final grade.

Contest # 4 - Final Results

NEWxSFC/s perennial bridesmaid Shanabe (aka SYRMAX) finally made it to the alter of Contest #4 and crushed the field in doing so by verifying with a SUMSQ error score 47% better than the average error. His winning forecast accounted for 95% of the observed snowfall/s variability.

Shanabe hit the trifecta finishing first in Total Absolute Error (11.6") and Average Absolute Error (0.48"). His STP forecast error of 8.26" was good for second place.

Congratulations...Shanabe!

Full forecast verifications and contest summary @ the web site.

Contest # 4 - Teleconnections

Sharp drop in AO ahead of the storm associated with big push of arctic air into CONUS/ eastern half. PNA swung the distance of one STD from negative to positive along with a near-normal NAO moving into higher territory.

Favorable PNA and AO conditions produced MAX snowfall @ contest stations that were the lowest of the four snow storms to date.

Contest # 4 - Verification

Preliminary storm-total snowfalls from CDUS41 and CF6 bulletins for Tuesday and Wednesday.

BGR amount is suspect given PNS reports of 9 - 10" SE of the station on the coast and 10 - 12" 7 - 10 miles to the N.

New daily records
None

Final results and storm summary on Friday.
Please report errors in 'Remarks.'

Interim Standings

After three snow storms...donsutherland1 is in 1st Place with a hefty lead.

Click to enlarge
SUMSQ errors are normalized with a 'Z-score' for each snow storm contest...then averaged to compute the standings. Forecasters who have entered at least 2/3 of all snow storm contests are included in the standings.

If a forecaster has participated in more than two-thirds of all snow storm contests...then Z-scores from their 'best two-thirds' forecasts are used to calculate interim and final standings. Same idea as dropping the lowest quiz score before computing the final grade.

Contest # 3 - Final Results

Chief Forecaster Donsutherland1 finished in 1st place for contest snow storm #3. This is his second 1st place finish this year.

No only did he have the lowest SUMSQ error...he hit the trifecta with top honors for Total Absolute Error (30% improvement over Consensus) and Average Absolute Error (1.16").

Donsutherland1/s STP forecast for 60" had an error of 0.95" (total observed snowfall @ all stations less total forecast snowfall @ all stations). The R-square statistic of 0.837 shows his forecast accounted for 84% of the observed snowfall/s variability.

Congratulations...Don!

Full forecast verification and contest summary @ the web site.

Mild...Then Wild

Despite some disagreement among the ENSM members...teleconnection indice forecasts suggest movement toward a colder...stormier planetary flow regime for the EC around mid-JAN following a brief spell of mild wx.

Today/s GooFuS depicts a series of SE PAC cyclones coming ashore then migrating NE through the central CONUS...blocked from coming too far E by the stubborn SE heat ridge through D+8. The extended range progs create a hi-amplitude ridge W that allows intense arctic anticyclones back into the Lower 48.

When the cold returns...so do the EC storms.

Snowfall Summary - DEC '07

Bonus snows were observed over much of the Northeast in December...with some stations getting up to three and half times their period-of-record normal (PORN) amounts.

BOS snowfall of 27.7" was 360% PORN (7.7") and amounts to 68% of D-J-F-M PORN (41").
CON snowfall of 44.5" is 78% of their D-J-F-M PORN (57.3"). DEC normally contributes 24% of the CON/s season-total snowfall and 19% @ BOS.

Most stations...where DEC snowfall was well in excess of PORN...also observed below normal temperatures. BGR had the greatest departure of -4°F. The exception was ACY where DEC/s snowfall of 4.4" was 183% of PORN and temperatures were 1.7°F above normal.

RED ==> Top 25%
BLUE ==> Bottom 25%
White ==> everywhere in between

Contest # 3 - Verification

Preliminary storm-total snowfalls from CDUS41 and CF6 bulletins for Sunday and Monday.

SN-to-H2O ratio for MDT is estimated.

Minor discrepancy between amounts reported by 12/30 CDUS41 (0.9") and F6 (0.3") @ PWM.

CAR amount is suspect given its 10:1 ratio and the vertical temperature profile during the event; however..VCNTY PNS reports suggest 3.4" is likely correct.

New daily records
12/31
BGR - 8.2" (6.4"; 1972)
CON - 10.1" (6"; 1879)
PWM - 6.6" (4.3"; 1893)
ALB - 7.9" (6.4"; ?)

Final results and storm summary on Wednesday.
Please report errors in 'Remarks.'