Saturday, March 31, 2007

6th Annual Season-total Snowfall Contest Comes to A Close


Today marks the end of the season for the NEWxSFC/s 6th Annual Season-total Snowfall Forecast Contest.

Summary results will be posted on the web blog. The full results...analysis...and verification will be posted on Sunday, 1 April 2007 at the web site and the NE.Wx Google Group.

Wednesday, March 21, 2007

Interim Standings


After four storms...


Click image to enlarge.

First Place: don sutherland1 with an average SUMSQ Z-score of -0.883
Second Place: herb@maws -0.705
Third Place: Raven -0.675

The second interim summary finds donsutherland1 holding onto 1st place. Herb@maws moves into 2nd from 3rd. Raven jumps from 5th to 3rd.

A data table with the complete interim standing statistics...including Sum Squared Error (SUMSQ)...Storm Total Precipitation (STP)...Total Absolute Error (TAE)...Average Absolute Error (AAE)..and R-Squared (RSQ) at the web site.

To qualify for ranking in the interim standings...forecasters must have entered at least 2/3 of all contests to date. If a forecaster has entered all four contests...only the best three forecasts are used to calculate an average normalized Z-score.

The chart shows the distribution of forecaster SUMSQ Z-Scores (plum) and Total Absolute Error Z-Scores(cyan). Lower (higher) Z-Scores indicate better (worse) forecasts compared to all other forecasts made for each storm.

A Z-Score of 0 means the forecaster's error was equal to the average of all forecast errors. A Z-Score of -1 (+1) means the forecaster's error was 1 standard deviation below (above) the average of all forecast errors.

Monday, March 19, 2007

Contest # 4 - Results



Full forecaster verification table and contest results at the web site.

Forecasters: 11
9 veterans
2 rookies

Station forecasts: 235
Average stations per forecaster: 21
Stations with observed snowfall: 23

Storm-total precipitation (STP), all stations
Forecast
Minimum: 67.5" (Donald Rosenfeld)
Consensus Median: 132.25"
Maximum: 184.5" (mitchel volk)

Observed: 146.45"
Max single station: 16.9" ORH
Shutout: RIC...ORF...and RDU

***************************************************
Storm #4 - Results Summary

1st Place - donsutherland1
SUMSQ Error: 128.5"
SUMSQ - Z: -0.799
STP: 15.05 (4th)
TAE: 38.95" (1st)
AAE: 1.69" (1st)

2nd Place - ilibov
SUMSQ Error: 129.1"
SUMSQ - Z: -0.796
STP: 7.75" (2nd)
TAE: 44.25" (2nd)
AAE: 1.92" (2nd)

3rd Place - herb@maws
SUMSQ Error: 163.6"
SUMSQ - Z: -0.615
STP: 22.15" (5th)
TAE: 48.45" (6th)
AAE: 2.11" (5th)

Honorable Mention - bruced
SUMSQ Error: 167.7"
SUMSQ - Z: -0.593
STP: 0.20" (1st)
TAE: 44.5" (3rd)
AAE: 1.93" (3rd)

Congratulations to the Winners!

SUMSQ: sum of square errors
STP: storm total precipitation
TAE: total absolute error
AAE: average absolute error

Sunday, March 18, 2007

1/4SM +SN SNINCR 02/14 P0007


METARs taken Friday evening and into the early morning hours of the Saturday at KALB...records the occurrence of two inch per hour snowfall rates at 05z and again at 07z.

The 'dP' notation at the end of Albany/s observations is the one-hour pressure change. Falling surface pressure means vertical motion in the column above the station is rising. The quicker the pressure falls...the greater the upward vertical motion (UVM); therefore...more snow.

METAR KALB 170151Z 01012G17KT 1/4SM SN FZFG OVC002 M07/M09 A3007 RMK AO2 SFC VIS 1/2 SLP187 SNINCR 01/07 P0004 [...]
METAR KALB 170251Z 01010KT 1/4SM SN FZFG OVC002 M07/M08 A3001 RMK AO2 SFC VIS 1/2 SLP166 SNINCR 01/08 P0006 [...] dP = 2.1
METAR KALB 170351Z 01012KT 1/4SM SN FZFG OVC002 M07/M08 A2994 RMK AO2 SLP144 SNINCR 1/9 P0005 [...] dP = 2.2
METAR KALB 170451Z 01010KT 1/4SM +SN FZFG OVC002 M06/M08 A2989 RMK AO2 SLP124 SNINCR 2/11 P0006 [...] dP = 2
METAR KALB 170551Z 36012G18KT 1/4SM +SN FZFG OVC002 M06/M07 A2984 RMK AO2 SLP110 SNINCR 1/12 P0006 60031 [...] dP = 1.4
METAR KALB 170651Z 36011KT 1/4SM +SN FZFG OVC002 M06/M07 A2981 RMK AO2 SLP100 SNINCR 02/14 P0007 [...] dP = 1
METAR KALB 170751Z 36012KT 1/4SM +SN FZFG OVC002 M06/M07 A2978 RMK AO2 SLP087 SNINCR 01/014 P0003 [...] dP = 1.3

In six hours...the snow cover at KALB increased from 7" to 14".

SN:H2O ranged between 17 and 33 during the period. This suggests the presence of stellar dendrites in the crystal factory. It/s important to note the SNINCR report rounds the snow cover value to the nearest whole integer...so the SN:H2O conversion is not precise.


ALB launched a special rawindsonde flight at 06z on the 17th...right at the height of the storm.


Two interesting features are immediately evident. The depth of the region inside the cloud where the -12°C and -18°C temperature range is most favorable for dendrite growth and the deep layered veering wind profile indicating warm air advection (WAA) which is producing strong UVM in the 850-700 mb layer.

KALB was also in the right entrance region of a moderately strong upper level jet max...which contributed additional UVM through its transverse ageostrophic circulation.


Having a look inside the storm near its peak intensity with a well-timed balloon flight reveals important details that can be applied when similar features are indicated on forecast soundings.

Contest # 4 - Verification



Preliminary storm-total snowfall based on final CDUS41 bulletins from Friday and Saturday.

Report any errors in 'Remarks' along with a link to the correct data. Final results and contest summary will be posted tomorrow evening.

Rank Station Snow
7 CAR 9.30
15 BGR 3.80
11 PWM 7.00
6 CON 9.50
2 BTV 13.30
10 BOS 8.10
19 HYA 2.50
1 ORH 16.90
14 PVD 3.95
12 BDR 6.05
4 BDL 10.20
3 ALB 13.20
8 BGM 8.80
16 ISP 3.35
18 JFK 2.75
9 ABE 8.50
5 MDT 9.55
17 PHL 3.05
23 ACY 0.05
13 EWR 6.00
21 BWI 0.65
20 IAD 1.25
22 DCA 0.15
23 SBY 0.05
25 RIC 0.00
25 ORF 0.00
25 RDU 0.00
TOT 147.95

Saturday, March 17, 2007

Contest # 4 - Teleconnections


So much for the conventional wisdom about needing +PNA...-AO...or -NAO to get good snows in the NE CONUS.



AO for this historic snowstorm...where six daily snowfall records were broken...is more than 2 standard deviations above normal. NAO is +1.5 STD and PNA is normal.

Everything else being equal...all you really need is amplification of the long wave in a ridge-W and trof-E flow regime. Shorten the wavelength and Sutcliff/s self-development process kicks in...which takes care of the rest.


Click image to animate.

Contest # 4 - Day One Leader Board


Based on Friday/s CDUS41 and Saturday/s PNS bulletins.

herb@maws
ilibov
MatthewRydzik
Raven
bruced
shanabe
TQ
donsutherland1
emoran
Donald Rosenfeld
mitchel volk

Friday, March 16, 2007

Snowless in Tokyo


Tokyo has experienced it first winter without snow since 1876. The Japan MET Agency...as do other MET services world-wide...defines winter as December...January...and February.

Although the wait here on the EC wasn/t quite as long...Tokyo observed its first snow...and latest snowfall EVER... on Friday morning. The snowfall reportedly was not enough to cover the ground.

Contest # 4 - The Forecasts



Clearly an interior NE event where the consensus +SN axis runs from BTV - CAR - CON -ORH - BDL - ABE - BTV with a max of 13" at ALB.

Forecasters: 11
Rookies: 2
Veterans: 9

Minimum STP: 67.5" (Donald Rosenfeld)
Maximum STP: 184.5 (mitchel volk)
Average STP: 135.8"
Median STP: 132.3"
10th percentile: 111.2"
90th percentile: 171.5"

All individual forecasts have been posted to the NEWxSFC web site. Follow the link to 'Contest # 4 - The Forecasts' The forecasts are ranked from minimum to maximum storm total snowfall (STP) for all stations.

Verification snowfall reports will be posted Sunday morning. They are considered preliminary and subject to change / challenge for 24 hours...if updated information becomes available. Final results will be posted late Monday evening.

Wednesday, March 14, 2007

Contest # 4 - Call for Forecasts


Latest SR NWP guidance suggests quite a few M-A and NE stations in play Friday and Saturday when a moisture rich...late season nor'easter makes its turn up the coast. The storm should be a good opportunity for another round of snowfall forecasts.

Entry Deadline: Thursday...15 March 2007 @ 10:30 PM EDT.

Verification begins 12:01 AM EDT Friday...16 March 2007 and ends 11:59 PM EDT Saturday...17 March 2007.

Your forecast must be entered via the Contest/s web site. Follow the link to ‘Enter Storm Forecast.’

Forecasters will need to register once before entering...even if they were registered last year. Registration is simple...requiring only a user name and password. If you provide a valid e-mail address...a copy of your forecast will be sent to you immediately after your entry is submitted. Please ensure your browser is enabled to accept first-party cookies.

All forecasts will be posted to the NE_Wx Google Group by the Contest Administrator before 11 PM EDT Thursday...15 March 2007 and to the Contest web site by Friday evening.

Contest updates are posted on the Contest/s web log.

The NE.Wx Snowfall Forecast Contest (NEWxSFC) is a multi-month event that continues into late March or early April. In general...contests are held whenever a decent... synoptic-scale storm rears its head and threatens at least a half-dozen forecast stations with more than nuisance snowfall amounts. Forecasters are called to post their 'storm total' snowfall predictions...on deadline...for 27 NWS / FAA observing stations scattered about New England and the Mid-Atlantic regions.

You can find the Contest Rules and additional information about Error Scoring... current monthly snowfall climatology from NCDC...daily CPC teleconnection indices... daily NESDIS N-Hemi snow cover...and NWS Daily Climate Bulletins (CDUS41) at the contest/s web site.

Tuesday, March 13, 2007

Contest # 4 - Take Two


MR model solutions continue trending toward a possible contest storm later in the week. Heavy rain at coastal stations with good snows inland.


Click to animate.

Upper trof sharpens on Saturday as it approaches the EC then acquires a negative tilt with the surface LOW wraping up over Newfoundland and a broad polar HIGH settling over the M-A.

Saturday, March 10, 2007

Coastal Teaser #6


MR has certainly had its way with NEWxSFC forecasters this year. A frustrating parade of one fantasy coastal teaser after another. The best storm has always been just a week away...so it comes as no surprise that the models are singing their siren song for our unrequited affections yet again.


ECMWF...GooFuS...and even the Canadians are currently trending toward a sharper EC trof scenario at the coming week/s end. The EC D+6 prog has a secondary coastal LOW around the 40°N / 70°W 'benchmark' and an arctic HIGH to its N and W...which could offer a late season opportunity for another snowfall forecast contest.

Thursday, March 8, 2007

Interim Standings


After three contests comes the first opportunity to post interim standings. Forecasters must have entered at least two of the three contests held so far this season to be included. If a forecaster entered all three contests...scores from the two best forecasts are used to calculated an average normalized SUMSQ score.

First Place: Donsutherland1 with an average SUMSQ Z-Score of -0.925.
Second Place: Bubbler86 -0.882
Third Place: Herb@MAWS -0.751

A data table with the complete interim standing statistics...including Sum Squared Error (SUMSQ)...Storm Total Precipitation (STP)...Total Absolute Error (TAE)...Average Absolute Error (AAE)..and R-Squared (RSQ) at the web site.


Click on image to enlarge.

There appears to be a tie for 3rd but Herb@MAWS leads TQ by 0.00007.

The chart shows the distribution of forecaster SUMSQ Z-Scores (plum) and Total Absolute Error Z-Scores(cyan). Lower (higher) Z-Scores indicate better (worse) forecasts compared to all other forecasts made for each storm.

A Z-Score of 0 means the forecaster's error was equal to the average of all forecast errors. A Z-Score of -1 (+1) means the forecaster's error was 1 standard deviation below (above) the average of all forecast errors.

Monday, March 5, 2007

Contest # 4?


(Updated)

Not impressed by tonight/s 00z f-soundings from the Eta / NAM. This run doles out saturated columns to ACY and BWI for all of ~3 hours. Columns at the other stations under the clipper look too dry.

The combination of a short duration event and limited moisture doesn/t usually make for a good forecasting event.

Update
Turned out to be not much of an event afterall...especially given the advanced billing from private...public...amd hobby forecasters.

Station / SN / SN:H2O
MDT 2.0 - 12
ACY 1.8 - 15
IAD 1.4 - 18
PHL 1.4 - 16
ABE 1.2 - 11
EWR 1.0 - 10
DCA 1.0 - 9
JFK 0.8 - 16
BWI 0.8 - 10
BDR 0.7 - 14
ISP 0.5 - 10
BGM 0.2
BDL T

Some of the snowfall totals seem low considering the SN:H2O. 10:1
ratios from clipper-type snows are highly suspect.

Contest # 3 - Final Results


Full forecast verification and results summary at the web site.

Follow the links to 'Forecasts - Contest #3' and 'Results - Contest #3.'

Forecasters: 10
9 veterans
1 rookie

Station forecasts: 123
Average stations per forecaster: 12
Stations with observed snowfall: 11

Storm-total precipitation (STP), all stations
Forecast
Minimum: 27" (mitchel volk)
Consensus Median: 50.9"
Maximum: 94.75" (Donald Rosenfeld)

Observed: 32.35"
Max single station: 9.5" CAR
Shutout: HYA...BDR...ISP...JFK...ABE...MDT...PHL...ACY...EWR...BWI...IAD...DCA...SBY...RIC...
ORF...and RDU

***************************************************
Storm #3 - Results Summary

Congratulations to the Winners!

1st Place - bubbler86
SUMSQ Error: 17.3
SUMSQ - Z: -1.164
STP: 3.65 (1st)
TAE: 10.5" (1st)
AAE: 0.91" (1st)

2nd Place - Raven
SUMSQ Error: 33.1"
SUMSQ - Z: -1.009
STP: 4.65" (2nd)
TAE: 15.75" (2nd)
AAE: 1.58" (2nd)

3rd Place - TQ
SUMSQ Error: 75.3"
SUMSQ - Z: -0.825
STP: 15.7" (4th)
TAE: 19.7" (4th)
AAE: 1.79" (4th)

Honorable Mention - mitchel volk
SUMSQ Error: 69.3"
SUMSQ - Z: -0.653
STP: 5.35" (3rd)
TAE: 19.15" (3rd)
AAE: 3.19" (8th)

SUMSQ: sum of square errors
STP: storm total precipitation
TAE: total absolute error
AAE: average absolute error

Sunday, March 4, 2007

One More?



Click image to animate.

The clipper poised to sweep across the M-A on Wednesday may produce enough snow to hold Contest #4. Time-height x-sects from the Eta / NAM / WRF-NMM / WTF suggest there could be 6...possibly 10 stations in play...although it remains to be seen just how many are likely to get more than nuisance amounts. Clippers are commonly good for a stripe of 2-4"...sometimes a bit more if it/s particularly vigorous.

Antecedent conditions look good with plenty-o-arctic air in place. Forecast temperatures aloft also appear favorable for the production of hi-fluff stellar dendrites in the crystal factory.


GooFuS forecast sounding for MDT.

Go / No GO decision coming tomorrow evening.

Contest # 3 - Verification



Preliminary storm-total snowfall based on final CDUS41 bulletins from Friday and Saturday.

Report any errors in 'Remarks' along with a link to the correct data. Final results and contest summary will be posted tomorrow evening.

1 CAR 9.50
2 BGR 6.60
5 PWM 2.50
3 CON 6.40
4 BTV 5.90
9 BOS 0.05
12 HYA 0.00
6 ORH 0.60
9 PVD 0.05
12 BDR 0.00
9 BDL 0.05
8 ALB 0.20
7 BGM 0.50
TOT 32.35

Saturday, March 3, 2007

Contest # 3 - Teleconnections



Event starts out with AO and NAO near normal...then both indices increase to about 0.5 standard deviations above their mean. PNA has been below normal since Contest #2.

Friday, March 2, 2007

Contest # 3 - Day One Leader Board


Based on Friday/s preliminary CDUS41 reports.

bubbler86
Raven
mitchel volk
herb@maws
TQ
ilibov
donsutherland1
DAROONEY
MatthewRydzik
Donald Rosenfeld

Thursday, March 1, 2007

Contest # 3 - The Forecasts


Forecasters: 10
Rookies: 1
Veterans: 9


Clearly a NNE event. The consensus +SN axis runs from BTV - CAR - PWM - CON - BTV with a max of 11" at BTV.

All individual forecasts have been posted to the NEWxSFC web site. Follow the link to 'Forecasts - Mar 2.' The forecasts are ranked from minimum to maximum storm total snowfall (STP) for all stations.

Minimum STP: 27" (mitchel volk)
Maximum STP: 94.75 (Donald Rosenfeld)
Average STP: 56.8"
Median STP: 50.9"
10th percentile: 35.1"
90th percentile: 81.9"

Verification snowfall reports will be posted Sunday morning. They are considered preliminary and subject to change / challenge for 24 hours. Final results will be posted late Monday evening.