April 15-17, 2007 Tax Day/Patriot’s Day Storm

(Click on thumbnails for larger images)

For the fourth consecutive month, a huge storm impacted our area right in the middle of month. This trend had started with the Martin Luther King Storm, continuing with the Valentine’s Day Storm, and the St. Patrick’s Day Storm.

Antecedent conditions prior to the storm were not the usual teleconnections that one might expect prior to and during the development of nor’easter, a stronger one than any during the winter months. The NAO was positive (as it had been much of the previous winter), but turned abruptly negative after to storm exited eastern North America (Fig. 1a). The PNA was slightly positive (Fig. 1b) and MJO was apparently “inactive” (Fig. 1c). Taking a look at the North American hemisphere, the polar jet stream was displaced a little further south than usual over North America and more importantly, was highly amplified with several cutoff lows found across the North Atlantic. Further east, a high amplitude ridge was seen over Europe.

Figure 1.  North Atlantic Oscillation and Pacific North American Oscillation January to present, and 7, 10 and 14 day forecasts, and Madden Julian Oscillation for march through 24 April.

A strong Pacific storm with lots of strong wind slammed into the Pacific coast and migrated across the southern plains. This storm wreaked havoc in many parts of the country, bringing unusually late heavy snow to the plains and tornadoes further south and east. As energy from this storm phased in the jet stream downstream of it, this system also became strongly cutoff from the main west to east flow, while the surface cyclone tapped copious from moisture off the Gulf Stream and south Atlantic Ocean, tracked up the eastern seaboard, and eventually was “captured” by the upper level feature near Long Island. The evolution from the southern plains through the northeastern U.S. is evident in the visible and water vapor satellite imagery (Figs. 2a-f)

a) b) c)

d) e) f)
Figure 2.  Satellite water vapor imagery with lightning overlay from a) 2330 UTC 14 April, b) 1730 UTC 15 April, c) 1100 UTC 16 April, visible satellite imagery with lightning overlay from d) 1730 UTC 15 April, e) 1130 UTC 16 April and f) visible satellite imagery with mean sea level pressure overlay at 1800 UTC 16 April.  Note the clearing area at the center of the strong surface low (@977 hPa) over southern NY.

The storm proved to be the most complex storm of the season. With enough cold air in place to produce dynamical cooling and thermodynamical cooling due to melting snowflakes, not only were we looking at a critical rain/snow elevation dependent event, but enough of a warm air intrusion wedged between about 6-8 thousand feet to further complicate the thermal profile. As result we had plenty of sleet and even a little freezing rain.

Models wavered just enough with the track of the storm to make forecasters realize that once again, the Capital District would be right on the fence between heavy rain and heavy snow (possibly mixed with sleet and even freezing rain). The overnight runs up to the actual event, tended to track the storm a little further east, increasing the risk of heavy snow right into the Capital District, while the daytime runs trended further west and warmer.

Ensemble members did consistently forecast huge anomalies in the surface to H5 wind fields, up to 5 deviations from normal, from the east, veering to southerly direction. Also, the ensembles showed U wind anomalies of 3 to 4 standard devations below normal at the 250 hPa level. The evolution of ensemble and HPC guidance is described below.  Ensemble guidance is composed of the Medium Range Ensemble Forecast (MREF) and Short Range Ensemble Forecast (SREF).  Other guidance that will be shown is from the Global Forecast System (GFS), North American Mesoscale (NAM), Hydrometeorological Prediction Center (HPC), and Plumes.  Guidance from 12 April will be described in Figs. 3-10, guidance from 13 April will be described in Figs. 11-16, and 15 April guidance in Figs. 17-22.

Many statewide conference calls were conducted on Friday through the weekend discussing the complexities of the storm and how just a slight shift in the track, precipitation intensity and just a degree shift in the thermal profile would ultimately dictated whether rain, or a mix of snow and sleet would fall. 

12 April 2007 Guidance


Figure 3.  Mean Sea Level Pressure (MSLP) from a) Medium Range Ensemble Forecast initialized 1200 UTC 12 April valid 0000 UTC 16 April, and b) NAM initialized 1200 UTC 12 April valid 0000 UTC 16 April.  Note the consistency from the MREF and the NAM, and both indicating a significant storm with MSLP forecasted to be 4 to 5 SD below normal.


Figure 4.  Spread of 500 hPa heights, mean and anomalies for a) 12Z 12 April MREF valid 12Z 16 April, and b) 15Z 12 April SREF valid 03Z UTC 16 April.  Note the 500 hPa low is forecasted to be 2 to 3 SD below normal.  No clear statement can be made about position of the 500 hPa low since the valid times are 9 hours apart.


Figure 5.  Spread of 850 hPa temperatures, mean and anomalies for a) 15Z 12 April SREF valid 03Z 16 April, b) 12Z 12 April MREF valid 00Z 16 April, and c) 12Z 12 April MREF valid 12Z 16 April.  Note the relatively large spread in each of the sources of guidance, suggesting considerable uncertainty.  Although the mean from each is fairly consistent across the interior northeastern U.S., there is more disagreement and uncertainty across southern and eastern New England.  Temperatures at 850 hPa were cold enough to support snow over the interior northeastern U.S. if surface temperatures were expected to be cold enough.

Figure 6.  Wind barbs, U and V wind anomalies (shaded) at 850 hPa for a) 12Z 12 April GFS valid 00Z 16 April, and b) 12Z 12 April NAM valid 00Z 16 April.  Note the U anomalies in both models are more than 5.5 SD below normal, but the centers of the mean are in considerably different positions, with the NAM further south.  Also note the V wind anomalies in both models show 4 to 5 SD above normal, but again, the centers of the mean are displaced.  The extreme U and V wind anomalies suggest very anomalously high winds, moisture advection, convergence and frontogenesis on the southern and eastern peripheries of this evolving system.

a)b) c)

Figure 7.  Wind barbs, U and V wind anomalies (shaded) at 850 hPa for a) 12Z 12 April MREF valid 00Z 16 April, b) 15Z 12 April SREF valid 03Z 16 April, and c) 12Z 12 April MREF valid 12Z 16 April.  Note the mean anomalies in the MREF and SREF are considerably less than the operation GFS and NAM from figure 6.  The averaging of the 15 members of each ensemble masks some of the extremes in the operational models.  Still, a U wind anomaly in the ensemble mean of 4 SD or less suggests a potential historical storm, if the upper level anomaly threshold is also met (will be evaluated later).  Also note that in the MREF, the extreme U wind anomalies expand with time.  U wind anomalies of such extreme values, and such agreement between operational models and ensembles at 84 hours is EXTREMELY unusual and provides a high level of confidence of a big event with an unusually long lead time.


Figure 8.  Wind barbs, U and V wind anomalies (shaded) at 250 hPa for a) 12Z 12 April MREF valid 12Z 16 April, b) 15Z 12 April SREF valid 03Z 16 April, and c) 12Z 12 April GFS valid 03Z 16 April.  Note the differences in U wind anomalies between the MREF and SREF, with the MREF suggesting 2 to 3 SD below normal, and the SREF 1 to 2 SD below normal.  The operational GFS suggests a very small area between -2.5 and 3 SD below normal, which is the threshold for a slow-moving storm, capable for producing historical precipitation amounts.

Figure 9.  Probability for 1.00” in 36 hours from the 12Z 12 April MREF valid 18Z 15 April through 06Z 17 April.  Note the high probability of 1.00” or more across most of eastern NY and New England.  This is an unusually high probability for this much precipitation forecasted 3 or more days in the future.



Figure 10.  Plume diagrams from 15Z 12 April SREF for a) Albany, NY and b) Islip, NY, and 12Z 12 April MREF for c) Albany, NY and d) Islip, NY.  Note the large spread beyond 15 April for both Albany and Islip.  Also note the mixed precipitation type at Albany, but with significant snow potential.  Finally, there is clustering in the MREF with nearly 2.00” or more for both Albany and Islip.

13 April 2007 Guidance

Figure 11.  PMSL spread and anomalies (color shaded) valid 12Z 16 April from a) 12Z 13 April MREF and b) 15Z 13 April SREF.  Note the differences in location of the 4 to 5 SD below normal anomaly near Long Island.  However, the forecasted central pressure continues to be in good agreement.


Figure 12.  Temperature spread, mean and anomlies (color shaded) for 850 hPa valid 12Z 16 April from a) 12Z 13 April MREF and b) 15Z 13 April SREF.  Note the warmer mean temperatures in western New England and eastern NY in the MREF, a subtle but important trend.



Figure 13.  Wind barbs and U and V wind anomalies (color shaded) for 850 hPa valid 12Z 16 April from a)00Z 13 April MREF, b) 09Z 13 April SREF, c) 12Z 13 April MREF and d) 15Z 13 April SREF.  Note the consistency in the ensembles for 5 to 6 standard deviation below normal over New England, an EXTREMELY rare value for both ensembles at any forecast time, but especially for a 3 or more day forecast.  Note also the increasing values for the V wind anomalies toward 4 SD above normal.  It is also very rare to see such strong U and V wind anomalies associated with the same storm.

a)b) c)

Figure 14.  Wind barbs and U and V wind anomalies (color shaded) for 250 hPa valid 12Z 16 April from a) 00Z 13 April MREF,  b) 12Z 13 April MREF, and c) 09Z 13 April.  Note the MREF consistently suggested stronger U wind anomalies at 250 hPa with values between 3 and 4 standard deviations below normal.  These values correspond to some of the longest duration storms of the past.



Figure 15.  Probability for 2.00” of rain in 36 hours (color shaded) and spread from a) 00Z 13 April MREF valid 12Z 15 April through 00Z 17 April,  b) 12Z 13 April MREF valid 12Z 15 April through 00Z 17 April, c) 09Z 13 April SREF valid 09Z 15 April through 21Z 16 April and d) 15Z 13 April SREF valid 12Z 15 April through 00Z 17 April.  Note the very large area of high probability of 2.00” or more in the MREF for consecutive runs, while the SREF had less  probabilities.  The spread in both MREF runs seemed smaller than the more spread in the SREF.



Figure 16.  Plume diagrams for Albany, NY from a) 00Z 13 April MREF, b) 12Z 13 April MREF, c) 09Z 13 April SREF and d) 15Z 13 April SREF.  Note the MREF and SREF plumes are converging on 2.00”+ with most members suggesting a significant period of snow as precipitation type.

April 15 Guidance – The last guidance prior to the onset of the storm


Figure 17.  PMSL spread and anomalies (color shaded) valid 12Z 16 April from a) 00Z 15 April MREF and b) 03Z 15 April SREF.  Note the continued subtle differences in location of the center of the mean surface low pressure, but the anomaly is still 4 to 5 SD below normal.  This was a consistent signal since the 12 April model/ensemble runs.


Figure 18.  Precipitable water (PWAT) spread, mean and anomalies (color shaded) from 00Z 15 April MREF valid a) 00Z 16 April and b) 12Z 16 April.  Note the anomalous PWAT anomalies of 4 to 5 SD above normal into southern New England decreasing to 2 to 3 SD above normal as the axis shifted into northeastern New England.  These high PWATs signaled well above normal moisture associated with the storm, which can alert forecasters to flood potential.


Figure 19.  Temperature spread, mean and anomalies (color shaded) for 850 hPa valid 12Z 16 April from a) 00Z 15 April MREF and b) 03Z 15 April SREF.  Note the significant change in the mean 850 hPa temperatures in both ensembles.  Both ensembles indicated much warmer mean temperatures in western New England and eastern NY, likely as a result of the strong easterly winds that were forecasted from the surface through at least 850 hPa, and the lack of a source of cold air due to a lack of a strong surface high to the north.  This resulted in high confidence that any snow or mixed precipitation would change to a cold rain over much of our forecast area between 00Z and 12Z 16 April.

a)b) c)

Figure 20.  Wind barbs, U and V wind anomalies (shaded) at 850 hPa for a) 00Z 15 April MREF valid 00Z 16 April, b) 03Z 15 April SREF valid 06Z 16 April, and c) 00Z 15 April MREF valid 12Z 16 April.  Note the consistency between the MREF and SREF through the forecast periods.  The U and V wind anomalies exceeding |5| SD at the same time in 2 sets of ensembles, in the same storm are EXTREMELY rare, and suggests a particularly historic storm.


Figure 21.  Probability for 2.00” of rain in 36 hours (color shaded) and spread from a) 00Z 15 April MREF valid 06Z 15 April through 18Z 16 April, and b) 03Z 15 April SREF valid 03Z 15 April through 15Z 16 April.  Note the large areal extent of the high probabilities, but also notice the MREF suggesting an axis of high probabilities through central and western NY, unlike the SREF with lower probabilities over the same region.

Figure 22.  HPC WWD forecast graphic of day 1+2+3 snowfall, issued at0617Z 15 April.  Note the southern edge of the accumulating snow was displaced southward again, with Winter Storm Warning criteria snows forecasted for the Capital District, and over 2’ of snow in the Adirondacks.  Significant changes to the snow forecasts within our forecast areas contributed to high levels of uncertainty in precipitation type and location of precipitation transition zones.

Ultimately, forecasters correctly identified that the heaviest rain and flooding issues would be south and east of Albany, while the heaviest snow would generally fall across the Adirondacks. We also correctly forecast the strongest winds would be over the higher terrain, especially southern Vermont and generally fall just short of warning criteria. The initial timing of the storm was forecast fairly well.

However, there were aspects of the storm that were not forecast as well. The timing of the heaviest snow in the valleys came much earlier than originally thought. Model data from WFOWRF did not support a change to snow in the Capital District until Sunday night, when in fact, melting of snow flakes cooled the column down enough to turn our precipitation to all snow midday Sunday. Despite a strong April sun angle, the snow fell hard enough to even coat some roadways before switching back to sleet and then rain by nightfall. Snow and sleet accumulations of one to four inches were common in the capital district. After that, no more snow fell in the Capital District and points south.

By far the biggest forecast challenge was with the QPF. Once again, virtually all model data missed the QPF by a magnitude of up to two fold. The thinking by HPC, our forecast staff and strongly supported by all American model data was a 2 to 4 inch would fall to the southeast of Albany, with 1 to 2 inches elsewhere, much of that falling as up to a foot or more of heavy wet snow to the northwest of Albany. Right in the Capital District, 1 to about 4 inches of snowfall was correctly forecast to fall, while just miles to our west, more than a foot buried the Helderbergs. Once again, the QPF was seriously under forecast. This seems to be a chronic problem of late, with many of our past flood events (and there have been a lot of them in the past several years) reflecting the same disturbing trend. Convective and tropical feeds of moisture most likely were not taken account much of the guidance. Rainfall from the storm turned out to be 5 to 8 inches to our southeast, 3 to 5 inches locally, and even 2 to 3 inches as far northwest as our southern Adirondacks. This may have been the greatest April rainstorm of all time, and April 15th was officially the rainiest April day ever in Albany’s history. This QPF had huge implications in that flooding was more widespread that initially expected, and often a category higher than initially thought.

Downsloping did not turn out to be as much of an issue as we thought for the entire storm. However, downsloping was viewed Sun night into early morning on the radar over southern Washington county, and Bennington county.

Once again, heavy snow fell across the elevated range of northwest Saratoga County, and was the only place a winter storm was missed. In fact, up to 13 inches of snow fell in those parts, not too far behind the Adirondacks! We did catch the winter storm warning with some lead time in southern Vermont Western Albany and Schenectady counties.

Also, this storm broke the all-time lowest barometric pressure ever for April, 28.84 inches. Of course non of the models forecast this storm to get this low, but several of the GFS members and the ECMWF did suggest the sea level pressure would get to at least 969 hPa just off the New Jersey coastline Sunday night-Monday morning. This became apparent during the mid week period preceding the storm.

Verification – What happened


Figure 23.  Initialized PMSL spread and anomalies (color shaded) from a) 06Z 16 April MREF and b) 09Z 15 April SREF.  Note the anomalies in both ensembles are more than 5 SD below normal, and the SREF suggests deeper low pressure.  This was still not deep enough as the surface pressure dropped to around 969 hPa.


Figure 24.  Initialized 850 hPa wind barbs and anomalies (color shaded) from a) 06Z 16 April MREF and b) 09Z 15 April SREF.  Note the extreme U and V wind anomalies did occur as forecasted.  Interior and northern Vermont experienced some of the worst wind damage seen in years.

Figure 25.  Initialized 250 hPa wind barbs and anomalies (color shaded) from the 06Z 16 April MREF.  Note the 250 hPa wind anomalies were 3 to 4 SD below normal, as was forecasted, suggesting an unusually prolonged event since the storm was very much cut off from the upper-level steering flow.


c) d)

Figure 26.  Upper air soundings valid a) 12Z 15 April at Albany, NY, and 00Z 16 April from b) Albany, NY (ALY), c) Buffalo, NY (BUF) and d) Manawaki Canada (WMW).  Note that Albany was virtually all below freezing at 12Z 15 April, supporting snow as a precipitation type, then went near isothermal by 00Z 16 April with a small layer just above freezing, suggesting mixed precipitation.  Also note that Buffalo and Manawaki were both well below freezing throughout the soundings.


Figure 27.  Upper air soundings valid 06Z 16 April from a) Albany, NY (ALY) and b) Upton, NY (OKX).  Note the nearly isothermal sounding at Albany, with slightly warmer layers at 900 hPa and 700 hPa suggesting mixed precipitation continuing.  Also note the deep layer above freezing at Upton, and the very strong veering winds at low levels.  Damaging winds and significant coastal flooding  occurred along the Long Island and New England coast.


Figure 28.  Upper air soundings at 12Z 16 April from a) Albany, NY (ALY) and b) Chatham, MA (CHH).  Note both soundings have a deep layer above freezing supporting rain as a precipitation type.  Note the veering winds at Chatham, reflecting the damaging winds and coastal flooding that was occurring along the New England coast.

Figure 29.  Base reflectivity from KENX at 1036Z 16 April, with PMSL overlay.  Note the heaviest precipitation over southern NY through southern and eastern New England, and another axis through central NY, similar to what the 00Z 15 April MREF suggested in the 36 hour probability of 2.00”.  Also note the comma-shaped area of precipitation circling around the 969 hPa surface low centered around New York City.

Figure 30.  One-day accumulated precipitation valid 12Z 16 April.  Note the 5-8” maxima from the DELMARVA through NJ, southern NY, southern and eastern New England.

What was learned from this event:

This was an extremely complex storm from a forecasting standpoint. Confidence was high for a significant, perhaps historical storm 3 or more days prior to the onset.  The MREF, SREF, and operational models were in good agreement for a deep surface low pressure center, cut-off low at 500 hPa, and wind anomalies at 850 hPa and 250 hPa pointing to a prolonged event with extreme precipitation potential.  Guidance agreement of this type, 3 or more days prior to the storm, was reminiscent of the Valentine’s Day Storm.  However, the combination of anomalies at such extreme values for MSLP, 500 hPa heights, 850 hPa winds (U and V), and 250 hPa winds were unprecedented since the implementation of the SREF and MREF about 5 years ago, and rivaled the Ash Wednesday storm of March 1962 in overall magnitude.

One of the most difficult aspects of forecasting this storm was precipitation types across the region.  The most significant complicating factor in dealing with the rain/snow issues on Sunday was that it was mid-April, and forecasting snow/rain is usually tied to diurnal variation (and elevation too). However, this was not the case diurnally, as locally in Albany the precipitation actually began as rain early Sunday morning, then changed to snow later in the morning, and eventually changed to a mix of sleet and rain late in the afternoon. After sunset, all rain fell at Albany. Early ensemble guidance 3 to 4 days prior to the storm suggested snow potential through Sunday 15 April over most of western New England and eastern NY.  However, WFO forecasters across the region noted the lack of a strong surface high anchored in Canada, so there was no source of anomalously cold air.  Consequently, forecasters trended warmer than guidance and favored primarily elevation dependent snows.  HPC did not quite resolve the localized nature of the heavy snow in their WWD graphics, and if taken literally, valley locations throughout the Capital District would have had close to winter storm warning criteria.  HPC cannot be expected to resolve local terrain effects well in their snow forecasts near precipitation transition zones IN ANY WINTER STORM SCENARIO. 

By 24 to 48 hours prior to the storm onset, ensemble and operational guidance resolved the strong easterly jet at 850 hPa that warmed low level temperatures in all areas except the Adirondacks.  Forecasters were correct to trend warmer for temperatures, especially in valley locations.  Hence, pattern recognition identified the correct temperature and precipitation trends before the guidance did.  Still, forecasting snow amounts at various elevations was very difficult, as some snowfall amounts within a county (including Albany and Saratoga) ranged from around an inch in valley areas to over a foot in the higher elevations.

So what happened here? There was likely a significant amount of dynamic cooling, as well as cooling due to melting snow. There was likely not a lot of evaporation, as the low levels became saturated rather quickly early Sunday morning. The 06Z ALB sounding depicted the freezing level to be around 2700 feet, with exclusively below freezing temps above this layer, and unanimously above freezing temps below. This explains why the precipitation started as rain in Albany then quickly went to snow. Later in the day, a pronounced warm layer aloft moved overhead from the southeast, which aided in the transition to sleet/rain. The NAM/GFS had a difficult time forecasting precipitation types, and the forecast soundings/precipitation type algorithms were not accurate. The higher elevations did receive a significant amount of snow before the changeover to rain, making snowfall forecasts very difficult before the storm commenced. Remembering some of the complexities during this storm will hopefully help us to better identify potential significant features in future early/late season Nor'Easters involving multiple precipitation types.

We also got into more widespread flooding and bigger trouble in the central and northern portions of the HAS, when we broke into the warm air. All the sludge (sleet and snow) melted contributing to the runoff causing flooding in the upper Hudson, Capital Region, Mohawk Valley, etc. The models did not forecast ALY to get to 50 degrees ! There was evidence of the warm air occlusion over the lower Hudson Valley around 10Z. However, it looked like this mild/warm air would be pinched off to the E/SE with maybe KPOU getting into it.

Try to split Saratoga County up by elevations higher than 1000 feet, NOT by a north/south split which is somewhat useless.

Utilize the satellite imagery for the trajectory of moisture during the event. The models didn't handle the QPF from the decaying convection and/or tropical connection in the Monday morning portion of the event. There was not much of a consensus on the areal extent of the high QPF potential prior to the storm.  However, the MREF guidance did show likely to categorical chances for 2” of liquid in an axis through the western Catskills and into the Adirondacks, a feature that SREFs and NAM-based products did not show.  There was a strong consensus 3 or more days prior to the storm, that 2” or more liquid was likely from near the NY/New England border through New England.  Plumes throughout the region suggested a widespread area of 2” liquid in interior parts of the northeast U.S., to around 3” liquid over southern NY and southern New England. Observed liquid equivalents and rainfalls were generally double the wettest ensemble members in the plumes.  Widespread categorical chances for 2” of liquid equivalent almost always result in some observed flooding within the region.

When 850 hPa U and V wind anomalies exceed |4| standard deviations at the same time, seriously consider doubling the QPF, especially where convective (upright or SLANTWISE) elements are possible.  If the 250 hPa wind anomalies exceed -3 SD, this suggests a particularly slow-moving system cut off from the upper steering flow, so QPF can be tripled or more in locations where convective elements (upright or SLANTWISE) are expected.  Also, when the U wind anomaly exceeds -4 SD, strongly consider increasing the QPF along the east facing slopes of our Catskills, Litchfield Hills, Berkshires and southern Greens. More research is needed in this field, a LOT more.