7-8 September Remnants of tropical storm Lee flooding

(click on images to enlarge)


After a very brief respite from Irene, a slow moving cold front brought a round of thunderstorms on September 4th and 5th, some of which contained heavy rainfall, strong winds and one cell even spawned a tornado. This front then stalled along the eastern seaboard. At the same time, Tropical Storm Lee, which formed in the Gulf of Mexico, came inland before reaching hurricane strength. This storm pounded Louisiana including New Orleans with up to a foot of rain, then weakened to a tropical depression. However, it was far from done dealing additional damage from copious amounts of rain. The remnants of Lee effectively "hooked" onto the stationary front, slowly riding up along it as it continued to bring epic rainfall bands from the southern Appalachians, northward throughout most of our region.


Three to six inches of rain fell from the remnants of Lee from September 6th to 8th. Although the rain was not as continuous as Irene, and there were "breaks" between the bands of rainfall, saturated ground from Irene, and the thunderstorms prior to event, produced another round of minor to moderate flooding throughout our region. Many of the places that experienced horrific flooding from Irene, had to deal with more water with this storm. The rains also hampered on-going recovery from the aftermath of Irene.


Unlike Irene, which was well forecasted, certainly from a synoptic point of view, the remnants of Lee turned out to one of the most challenging warm weather events in recent memory. Initially some of the models including the European, caught on to the fact that a unseasonable low latitude upper air trough would develop over the Southeast Conus, which would help transport the circulation of Lee as well as its associated moisture, northward.


However, this feature was actually lost in the usually reliable European model as we headed to the weekend prior to the event. In fact, one run of the European forecasted a ridge over the southeast when in reality a trough persisted. This was the opposite of what happened on the European in 10-day forecast in late April when it incorrectly forecasted a huge trough in the east, while the operational GFS more correctly forecasted a huge ridge, which ended up being much closer to reality.


Later versions of the European insisted on having a cutoff low in the southeast, but keeping it bifurcated from the northern stream. This scenario assured the remnants of Lee would not make it into our region. Even a couple of days prior to the event, the European insisted the remnants from Lee would stay just to our south, but each successive run crept the moisture a bit further northward, but never caught on the amount of rain that ultimately fell.


The GFS operational model, while performing better than the European, was not consistent with the moisture. One run, (00z Sunday) was nearly identical to the European, keeping moisture to our south. This offered increased confidence that the rains from Lee would spare our region. Unfortunately, that was the wrong hat to hang onto, and later runs of the GFS quickly backed off this scenario. However, the GFS was also too light with the QPF, and several runs indicated the maximum QPF would end up in Western New York and western PA, about 200 miles west of where the maximum QPF ended up.


The North American Model (NAM) had its own solution. The first one that latched onto to Lee, incorrectly sent the storm back south into the Gulf of Mexico. The next ran the remnants of Lee coming up the coast. Then, the NAM has a few runs where it developed a significant mid latitude low pressure area that blasted our area with rain Monday night but then pushed the front and whatever was left of Lee, well out to sea. Our area did a fair amount of rain Monday night, but an organized low pressure area really did not develop. Of course, Lee's remnants did not go out to sea. Later NAM runs gradually backed off this idea, but not fully until right about as the event was unfolding. The 00z/06z NAM was the only model to hint of the last band of rain forming, back over eastern New York early on the 9th, bringing one more heavy rainfall to region that was already saturated.


The Canadian Meteorological Model (CMC) was the only operational model that brought Katia inland to near NYC. It did this one or two runs, which of course (and thankfully) did not prove correct. However, this model might have been the most consistent forecasting our region could get soaked with bands of Lee's rain with only a very small respite between the frontal thunderstorms and the more widespread rain courtesy of Lee behind the front. This scenario turned out to be more realistic.



Above: Four panel displays of precipitation forecasts initialized at 12Z 5 September (left), 12Z 6 September (center) and 12Z 7 September (right) from the NAM (upper left), GFS (upper right), ECMWF (lower left) and GFSEnsemble (lower right).



Above: Plume diagrams for Elmira, NY from the GEFS (left) and 2 runs of the SREF (center and right). Note the GEFS suggesting over 4 inches of rain in all but one member and the SREF showing a notable increase in values with later runs.



Above: Plume diagrams for State College, PA from the GEFS (left and center) and the SREF (right). Note the GEFS suggesting over 4 inches of rain in consecutive runs but one member and the SREF showing over 3 inches of rain.



Above: Plume diagrams for Dubois, PA from the GEFS (left), Binghamton, NY from the GEFS (center) and Williamsport, PA from the SREF (right). Note the GEFS suggesting over 6 inches of rain at Dubois, PA, over 3 inches at Binghamton, NY and the SREF showing mainly over 2 inches at Williamsport, PA.




Above: Loops of the HRRR forecasted radar reflectivity valid at 06z 8 September (left), and 12Z 8 September (right), showing the west-east oriented band of heavy rain developing along the old cold front over Long Island and southern New England merging with heavy rain tracking east. The consistency from run to run is noticeable and provided good guidance for anticipating flooding rains over southern NY and parts of New England.



Above: Loops of water vapor satellite imagery through the event.



Above: Visible satellite loop showing the narrow band of clouds representing the cold front tracking south of Long Island and New England.




Above: Visible satellite imagery loop with surface dew points and wind barbs overlayed. Note the surface dew point gradient over Long Island and southern New England providing low level focus for development of heavy rain.



Above: Stage 4 precipitation totals for 2 weeks encompassing Hurricane/Tropical Storm Irene and Tropical Storm Lee remnants. Note over 20 inches of rain in parts of the Catskills.