NE.Wx Snowfall Forecast Contest

Complete Results @ http://www.newx-forecasts.com/

Follow the left-panel link from '26^th Annual Snow Storm Contest > Verified Forecasts > ’15-FEB-25’ to see the complete forecasters’ verification table by station.

In the table ...

Yellow cells indicate the best score in the category.

Forecast STP cells: yellow if within +/- 5% of observed STP.

Blue (Red) cells indicate the 1^st (4^th) quartile

SUMSQ: sum of square errors (measure of forecast accuracy accounting for magnitude and distribution of snowfall)

STP: storm total precipitation

TAE: total absolute error
AAE: average absolute error

Final Standings and Measures of Skill - all Forecasters

---

Station by Station Comparison of Top 4 Forecasts and Observed Storm-total Snowfall (STP)

---

Perfect Forecasts (Batting Average - Forecast Stations with No Error)

Best Station Forecasts (Batting Average - Forecast Stations with Lowest Absolute Error)

Best Station Forecast Busts (Batting Average - Forecast Stations with Highest Absolute Error)

Skill v. NWS ER WFOs

- SKILL: positive (negative) skill value indicates a forecast/s improvement (degradation) relative to the current NWS forecast at entry deadline.

- R^2: how well the forecast model 'explained' the variability of the observed snowfall (i.e., an R^2 of 0.710 means the forecast model explained 71% of the observed snowfall's variability).

- BIAS: arithmetic difference between the average Forecast snowfall and the average Observed snowfall (averageForecasts – averageObserved).

---
Consensus … Extrema … and NWS ER WFO Forecasts with Observed Snowfall

Consensus forecast best @ CAR …CON …ORH …JFK

NWS forecast best @ BOS …ORH …BDR …ALB …ISP …JFK

MAX forecast best @ BOS

MAX forecast less than observed @

MIN forecasts best @ BGR …ABE

MIN forecasts more than observed @

* all evaluated subjectively

---

Total Absolute Error and SUMSQ Error

Strong correlation (R = 0.971) between SUMSQ and TAE Z-scores

---

Forecast by Observed Snowfall Scatterplots (Top 4 Forecasts)

- A Forecast’s dashed BLUE trend line above (below) solid RED Observed snowfall line ==> over (under) forecast

- R^2: how well the forecast model 'explained' the variability of the observed snowfall (i.e., an R^2 of 0.710 means the forecast model explained 71% of the observed snowfall's variability).

---

Storm-total Snowfall Verification

Verification of storm-total snowfalls based on reporting from CDUS41 (CLI) ... CXUS51 (CF6) ... METARs ... and PNS bulletins issued by NWS.

Good coverage and reporting.

---

HYA

STP derived by an inverse distance-weighting technique using vicinity STP reports from Barnstable county carried in the BOXPNS bulletin as inputs.

---
Snow-to-liquid ratios (SLR) less than 8:1 are not reported for stations with measurable snowfall b/c significant liquid and / or freezing precipitation likely occurred during the verification period.

The 'TOT SLR' field is a quantity-weighted AVERAGE of those stations with at least an 8:1 SLR.

---
Stations observing >= Trace: 25 (93%)

Stations observing > Trace: 18 (67%)

Given a station had measurable snowfall; stations observing at least:

4" - 10 (37%)

6" - 1 (4%)

8" - 1 (4%)

10" - 1 (4%)

MAX snow melt-water (minimum SLR 8:1)

BDR - 0.58'

ORH - 0.51"

PVD - 0.46"

MAX precipitation (frozen + freezing + liquid)

JFK - 0.72"

JFK - 0.69"

EWR - 0.64"

---

New daily snowfall record(s)
16-FEB-25

CAR - 9" (8.4"; 2021)

---

Daily snowfall data table

ORANGE cells: new daily record
GREY cells: data source(s) - PNS and / or METARs or inverse distance weighting
Trace amounts (displayed as 0.05") not included in STP

---

A map of the united states

AI-generated content may be incorrect.