Historical 29-31 March 2010 Rain Storm

(Click on images for larger displays)

 

On 29-31 March 2010, a historical rainstorm affected areas from the mid-Atlantic states through eastern NY and southern and eastern New England. This storm was anticipated days in advance, and the extreme magnitude of the potential flooding was communicated to the user community up to 2 days in advance. Locally heavy rains from several systems in the previous 2 weeks resulted in already swollen rivers across New England and Eastern New York, so any additional rain from the 29-31 March system was expected to initiate a new period of flooding.

 

This storm followed an evolution outlined in recent research on extreme flood events in the northeastern U.S., including the 14-16 April 2007 storm. However, even when regions are warned of a potential catastrophe, there is only so much communities can do to prepare for such an event. Such was the case in southern New England, where extreme flooding was observed and people just had to respond to the extremely hazardous conditions in as safe a manner as possible, and then begin the slow recovery process. The key is to make sure people do make the best decisions before, during and after the event, as these decisions will often determine how much risk people will take, and affects the chances for individuals to become victims of the hazards.

 

Selected guidance shows that there were some uncertainties 3-5 days prior to the onset, but within 2 days of the onset, there were very clear signals for an extreme rainfall and flooding event. The biggest uncertainty was determining the area of highest impact. Comparing guidance with the observed rainfall, it can be seen that even within hours prior to the onset, all sources of guidance were far from perfect in precisely defining the area of greatest impact.

 

The storm occurred in 2 phases, a warm frontal southerly low-level flow phase, and an Atlantic flow upper deformation zone phase. Each phase was resolved in the guidance in the short and near term, and both phases were predicted to produce very heavy rains. The predicted total rainfall in guidance was unusually extreme, and very rare to see the predictions that were seen in sources of guidance, contributing to the high confidence in an extremely heavy rainfall event.

 

The heaviest rain and the catastrophic flooding occurred in RI and eastern CT, which was slightly east of where all sources of guidance and ensembles were suggesting. However, all of eastern NY and New England were affected by very heavy rain that eventually contributed to varying degrees of flooding due to preceding wet conditions, even though some areas were outside the areas of maximum predicted rainfall. Observed rainfall amounts were in some cases 1.5 to 2 times the predicted rainfall in guidance, but forecasting techniques from recent research highlighted the strong possibility for rainfall totals near what was observed.

 

Long Range Guidance

 

 

Above loops from left to right: Four-panel of 500 hPa heights and vorticity (color shaded) from the 12Z 25 March GFS, ECMWF, GGEM and GFSEnsemble; four-panel of MSLP from the same guidance sources, four-panel of 850 hPa and 925 hPa temperatures from the same guidance sources, and four-panel of 48 hour precipitation, total accumulated precipitation in the GGEM, and probability of 1 inch in 24 hours from the GFSEnsemble.

 

Notice that the GFS was suggesting a very strong northern stream upper low dropping out of Canada in phase 2 while the other sources of guidance suggested more of a steady-state southern stream upper low. The GFSEnsemble depicted an open trough, suggesting the operational GFS was an outlier.

 

The presence of a strong upper low was also evident in the MSLP fields, based on the evolution and track of the secondary low, but the spread in the GFSEnsemble illustrates the large spread in all GFSEnsemble members. The thermal fields at 850 hPa and 925 hPa were consistent with the upper-level and surface feature evolutions. The precipitation fields showed that the ECMWF and GGEM were predicting significant rains over eastern New England, while the GFS and GFSEnsemble were not predicting such extreme rains, and not in such a well-defined region.

 

Short Range Guidance

 

 

Above from left to right: 84 hour forecast of 500 hPa heights from the 12Z 27 March and valid at 00Z 31 March from the GFS, ECMWF, GGEM and NAM; MSLP from the same guidance sources, four-panel of 850 hPa and 925 hPa temperatures from the same guidance sources, and four-panel of total accumulated precipitation in same guidance sources.

 

Notice the consistency between all sources of guidance, all converging on a dominant southern stream system affecting the northeastern U.S. Details on the exact track of surface and upper features are still not clear, but there is enough of a broad consensus in all model/ensemble guidance, including storm total rainfall, to increase forecaster confidence in a significant precipitation event for southern and eastern NY and New England. Note the guidance sources were suggesting the axis of heaviest precipitation was over southern NY and southwestern, central and northeastern New England.

 

 

Above from left to right: Phase 1 winds and anomalies (color shaded) at 850 hPa from the 12Z 28 March NAM, GFS, 15Z SREF and GEFS, valid 06Z 29 March.

 

Notice the V wind anomalies above 3 SD with the core of the winds tracking right through eastern NY and New England. These winds typically transport anomalous moisture and contribute enhanced boundary layer convergence for upward motion.

 

 

Above from left to right: Phase 1 precipitable water (PWAT, color shaded) from the same guidance sources, but valid 12Z 29 March.

 

Notice the PWAT anomalies of 2-3 SD above normal, which in research correlates with extreme moisture associated with the system. Also notice the south/north orientation to the PWAT plume, with origins in the subtropics.

 

 

Above from left to right: Phase 2 winds and anomalies at 850 hPa (color shaded) from the same guidance sources, but valid 18Z 30 March.

 

Notice the U wind anomalies exceeding -4 SD, which research has shown, signals heavy snow or heavy rain, depending on what predominant precipitation type is expected.

 

Above from left to right: Phase 2 precipitable water (color shaded) from the same guidance sources, but valid 12Z 30 March.

 

Notice the PWAT anomalies exceeding 2 SD above normal, which in research correlates with extreme moisture associated with the system. Also notice the southeast/northwest orientation to the PWAT plume, indicating inclusion of Atlantic moisture into the system.

 

 

 

Above from left to right: Phase 2 winds and anomalies at 250 hPa (color shaded) from the same guidance sources, but valid 18Z 30 March, except 12Z 30 March in the SREF.

 

Note the 250 hPa U wind anomalies exceeded -3 SD, which research has shown indicated an upper-level system cut off from the steering flow, correlating to an unusually prolonged period of precipitation.

 

 

 

Above from left to right: Phase 2 15Z 27 March SREF loop of probability for 2.00 in 36 hours, 12Z 26 March GEFS loop of probability for 2.00 in 36 hours, 00Z 28 March GEFS loop of probability for 3.00 in 36 hours, and 12Z 28 March GEFS probability for 4.00 in 48 hours.

 

Notice the probabilities increasing in the guidance as the event drew nearer. Also note the area of highest probabilities narrowed to southern NY, and southwestern to northeastern New England. These probabilities were unusually high, and few if any cases of probabilities of 3.00 in 36 hours and 4.00 in 48 hours have ever been observed.

 

Above from left to right: Loop of SREF plumes for Albany, NY, and loop of GEFS plumes for Albany, NY.

 

Notice the increased agreement between the SREF and GEFS with time. Also note the lower end of the range increases significantly, and the clustering increases to above 1.50.

 

 

Above from left to right: Loop of SREF plumes for Bradley, CT, and loop of GEFS plumes for Bradley, CT.

 

Notice the increased agreement between the SREF and GEFS with time. Also note the lower end of the range increases significantly, and the clustering increases to above 2.50. It is extremely rare to see clustering of most of the members of any ensemble above 2.50, and in this case both ensembles were consistent. This was also consistent with the guidance showing the axis of maximum precipitation through CT.

 

 

Above from left to right: Loop of SREF plume for Islip, NY, and loop of GEFS plume for Keene, NH and 15Z 26 March SREF plume for Keene, NH.

 

Notice the increased agreement between the SREF and GEFS with time. Also note the lower end of the range increases significantly, and the clustering increases to above 2.50 at ISP. Similar signals were noted at EEN. Again, it is extremely rare to see clustering of most of the members of any ensemble above 2.50, and in this case both ensembles were consistent. This was also consistent with the guidance showing the axis of maximum precipitation from CT through central and northeastern New England

 

 

 

Above from left to right: Four-panel of total accumulation from the 12Z 28 March GFS, NAM, GGEM and ECMWF, and Probabilistic QPF for Litchfield County, CT, generated by NWS Albany, NY.

 

Note the near-term guidance was very consistent in the placement of the axis and amounts of maximum precipitation. The highest resolution model guidance depicted the highest rainfall amounts. Still, even with the driest guidance suggesting 4.66 of rain, it is very significant, and added to the high confidence flooding scenario. The consistent position of the heavy precipitation axis from southern and eastern NY through southwestern, central and northeastern New England also increased confidence in the areas most vulnerable to the worst potential flooding.

 

The Probabilistic QPF is showing unusually high probabilities for 1.00 and 2.00 for Litchfield County, CT. These probabilities are produced by combining probability of precipitation with QPF grids, generated by forecasters at NWS Albany, NY. It is different from the SREF or GEFS probabilities for 1.00 and 2.00 in that forecasters produce a deterministic probability of ≥0.01 of precipitation and a total QPF grid, then a statistical equation is used to produce the Probabilistic QPF output.

 

 

Above from left to right: Probability for 2.00 in 36 hours from the 03Z and 09Z 29 March SREF.

 

Notice the axis of highest probabilities shifted eastward into southern and southeastern New England. These probabilities were unusually high, and few if any cases of probabilities of 2.00 in 36 hours of the depicted areal extent have been observed. This subtle eastward shift to the axis of highest probabilities was very important in the near term forecast.

 

 

 

Above from left to right: Probability for 3.00 in 36 hours from the 03Z and 09Z 29 March SREF.

 

Notice the axis of highest probabilities shifted eastward into southern and southeastern New England. These probabilities were unusually high, and few if any cases of probabilities of 3.00 in 36 hours of the depicted areal extent have been observed. This subtle eastward shift to the axis of highest probabilities was very important in the near term forecast.

 

 

Above from left to right: Probability for 2.00 in 36 hours from the 00Z and 12Z 29 March GEFS.

 

Notice the axis of highest probabilities shifted eastward into southern and southeastern New England. These probabilities were unusually high, and few if any cases of probabilities of 2.00 in 36 hours of the depicted areal extent have been observed. This subtle eastward shift to the axis of highest probabilities was very important in the near term forecast.

 

 

Above from left to right: Probability for 3.00 in 36 hours from the 00Z and 12Z 29 March GEFS.

 

Notice the axis of highest probabilities shifted eastward into southern and central New England. The GFSEnsemble was suggesting more spread had compromised the probabilities for 3.00, and in fact, the areal extent decreased by the 12Z 29 March initialization. Still, these probabilities were unusually high, and few if any cases of probabilities of 3.00 in 36 hours of the depicted areal extent have been observed. The eastward shift to the axis of highest probabilities in the GEFS was less pronounced as was observed in the SREF contributing to inconsistency and uncertainty in the eastward shift to the axis of highest probabilities.

 

 

 

 

 

Above from left to right: Probability for 4.00 in 48 hours from the 03Z 29 March SREF and 00Z 29 March GEFS.

 

Notice the axis of highest probabilities remained over southwestern New England. The GFSEnsemble was suggesting more spread had compromised the probabilities for 4.00, and in fact, the areal extent decreased by the 12Z 29 March initialization. Still, these probabilities were unusually high, and few if any cases of probabilities of 4.00 in 48 hours of the depicted areal extent have been observed. There was no clear eastward shift in the axis of highest probabilities, providing some uncertainty and inconsistency when compared to the probabilities of lesser amounts.

 

 

 

Above from left to right: Phase 1 real-time initialized winds and anomalies at 850 hPa (color shaded) from the 09Z 29 March SREF, and 06Z 29 March GEFS, NAM and GFS.

 

Verified 850 hPa V wind anomalies exceeded 3.5 SD.

 

 

Above from left to right: Phase 2 real-time initialized winds and anomalies at 850 hPa (color shaded) from the 21Z 30 March SREF, and 12Z 30 March GEFS, 00Z 31 March NAM and 00Z March GFS.

 

Verified 850 hPa U wind anomalies exceeded -4 SD.

 

 

Above from left to right: Water vapor satellite imagery and lightning overlay for phase 1 and phase 2.

 

Notice the tropical connection throughout the entire storm, and the 2 phases, the Gulf/Tropical origins as phase 1, and the Atlantic origins as phase 2.

 

 

Above from left to right: Visible satellite imagery and lightning overlay for phase 1 and phase 2.

 

Notice the moisture flow and cloud movement throughout the entire storm, and the 2 phases, the Gulf/Tropical origins as phase 1, and the Atlantic origins as phase 2.

 

 

Above: Radar loop of the northeastern U.S. through phases 1 and 2 (courtesy of the College of Dupage).

 

Notice the moisture flow and cloud movement throughout the entire storm, and the 2 phases, the Gulf/Tropical origins as phase 1, and the Atlantic origins as phase 2. More importantly, notice that eastern CT and all of RI experienced the brunt of the prolonged period of heavy rain.

 

 

 

Above: Stage 4 storm total precipitation from the National Weather Service.

 

Notice the axis of maximum precipitation was east of where the guidance suggested, and the large area of 6-10 of rain over RI. As stated earlier, catastrophic flooding occurred in eastern Ct and most of RI.

 

Any comments, questions or suggestions, please mail to:Neil.Stuart@noaa.gov