23 June Flash Flooding in Herkimer County
(click on images to enlarge)
Upper-level ridging along with a retreating surface high pressure system off the eastern seaboard allowed for a southerly/southwesterly return flow to transport warm and humid conditions across the forecast area...setting the stage for afternoon showers and thunderstorms. These showers/thunderstorms were not initially expected to contain a widespread, or even scattered, severe threat but rather resemble more of the typical summertime pulse-popcorn thunderstorms. As the event grew closer, SPC first highlighted the forecast area in a "See Text" by the 0600 UTC Day 1 outlook for 5% hail and wind, saying that isolated to scattered thunderstorms would develop and merge into clusters/bowing segments with the potential for damaging winds/large hail.
It should be noted the "See Text" emphasis was for northern NY and Northern New England at 0600 UTC. The 5% probabilities did cover the northern half of the forecast area (Capital Region North), but through collaboration with neighboring WFO's and local analysis the severe threat was deemed minimal, especially in the ALY CWA. It was mentioned that an isolated/rogue severe thunderstorm would be possible (NOT great enough to put in the HWO). The midnight crew updated the HWO to emphasis that any thunderstorms that did form along the diffuse frontal boundary would be yield locally heavy rainfall that could produce ponding of water on roadways...and poor drainage flooding in low-lying areas. To pick where an isolated flash flood would occur was difficult due to drier weather over the past week, and the "noise" from the various storm-scale 3-km columnar REF simulations. The bottom line...heavy rain was emphasized in the morning AFD...HWO...and gridded forecast products with the thunderstorms.
Our forecast area was then upgraded to a "Slight Risk" with the 1630 UTC Day 1 outlook for 15% wind as SPC again highlighted the potential for thunderstorm clusters capable of producing damaging winds. As the afternoon progressed, large CAPE developed (2000-3000+J/kg) as temperatures soared into the upper 80s/lower 90s with moisture advection bringing sultry dew points up into the upper 60s to near 70F. The high temperature at KALB was actually 91, with one observation indicating a dew point of 70. PWAT values of 1.5-1.9 inches (roughly 2 SD above normal) indicated that the unstable air mass had plenty of water to ring out. A high freezing level (~13-14kft) also gave a hint that thunderstorms that would develop would be capable of highly efficient warm rain processes, given the high PW, high FZL, and tall, thin CAPE profile (via 12z KALB sounding). Although the thermodynamic environment was primed for heavy rain, a weak shear environment (25 knots at best) and weak midlevel lapse rates (generally less than 6.0 C/km) hampered the severe threat in our CWA (better shear/dynamics occurred in BTV/GYX/northern BOX CWAs).
It was a bit surprising SPC upgraded the entire northern half of the forecast area to a Slight Risk with the lack of shear/weak lapse rates, etc. It was impressive the advection of the 70F dew point air in from the west/southwest. Dew points were much higher over the Capital Region than the mid-Hudson Valley. A question to answer is...despite the abundance of CAPE in place, why the convection weaken with the fine line outflow boundary ? It fire a bit north of the Capital Region into southern VT and then weaken. I suspect or hypothesize that the CAPE was not "thick" but packed down low in the profiles, and the outflow boundary actually dried out the higher surface dew point air, and the CAPE shriveling with the loss of the daytime heating.
Showers and thunderstorms had already begun to develop during the early afternoon hours across portions of upstate New York, Vermont, and other portions of New England. Initially, there were two main areas of convection that formed in west-central New York along a theta-e gradient and lake breeze boundary and northeast in far upstate New York along a secondary theta-e gradient. The storms that were north of the forecast area and that moved into BTV's CWA had several SVRs issued as slightly better shear helped organize storms enough to produce large hail and damaging winds. Much of the forecast area did not see convection initially as there was no trigger.
There was a diffuse frontal boundary that was the main focus all day. It was extremely difficult to analyze the front in terms of winds/dew points, etc. This was the focusing mechanism for convection during the morning hours...as the forecast area got into a robust warm sector. I agree multiple theta-e gradients were around, and perhaps a lake breeze interacted or phased with the frontal boundary. The extension of the showers and thunderstorms into northern VT and northern New England was due to the front and a mid-level vort max. A weak surface trough also may have moved across central NY in the afternoon (straddled the border on NY-PA. This was evident in the surface wind field which acted as a focusing mechanism for showers and thunderstorms across central and eastern NY. The 21Z WPC surface analysis analyzed this surface trough.
The cluster of storms in west-central New York formed into a quasi-linear feature oriented southwest/northeast and progressed into southern Herkimer County. With a mean wind of only 12 kts from 266 degrees (12z KALB sounding), these thunderstorms were fairly slow-movers but more importantly, began to train over the same areas - areas that were still saturated from previous rainfall events over the last several weeks. An FLS was initially issued in anticipation of this training nature of the storms. Within an hour though, the 3 hour FFG for basins in southern Herkimer county was exceeded prompting the issuance of a FFW for portions of southern Herkimer county (from FFMP).
The first report of flash flooding came 5 minutes later, and several more reports came in after that. Herkimer county EM director Bob Vandawalker called the office around 530pm stating that there was "incredible flash flooding throughout all of southern Herkimer county. This is the worst I've seen it since the summer of 2006." The towns of Middleville and Herkimer were the hardest hit, with several road washouts, reports of debris in roadways, and homes being flooded. Little Falls also experienced flash flooding with roads being washed out. Homes in Kast Bridge were flooded as well as roads being closed. Multiple roads were also washed out in Fairfield. Based on the reports and the analysis during the event via FFMP/GoogleMaps/etc, excessive runoff, including from the West Canada Creek, contributed significantly to the flash flooding.
Someone at 0902 pm in the WKTV blog reported 2.4" of rain in Schuyler with the event. Other ground truth reports will have to wait until the COOP and CoCoRahs data comes in tomorrow morning.
Several SPSs and Bullet NOWs were also issued during this event to highlight stronger storms that were capable of producing small hail, frequent lightning, gusty winds, and torrential rainfall. An additional FLS was issued for extreme northeastern Warren County and extreme northern Washington County as FFMP had indicated basins reaching or exceeding FFG there as well.
An impressive outflow boundary sagged east from the convection over the Mohawk Valley and Adirondack Region. This boundary moved south of the Capital Region during the late afternoon, but did not fire any convection, despite steep low-level lapse rates, and midlevel lapse rates of 6.5C/km and the abundance of CAPE. See the 00Z 24 JUN 2013 KALB sounding I put in the post mortem folder. It looks like the low-levels mixed out, and perhaps significant drying from the outflow boundary mixed convective possibilities. Monty put a nice loop of the event in the Post Mortem folder.
2. What was learned from this event:
The legacy FFMP was used after the recent issues experienced with the QPE version of FFMP during the Schoharie Floods. FFMP proved to be a valuable tool as the Downstream Basin Trace tool was utilized to determine where the water would flow relative to I90 in Herkimer County when coordinating with NYSOEM and NYDOT.
While the legacy FFMP was mainly utilized, I also did check Dual-Pol STA, instantaneous rate product, and one-hour products and the HPE FFMP. It's important to use all products in conjunction with each other - and make assessments as you go along. Just because on product is bad during one event is not enough evidence that it will not work in the next one.
The polygon created was drawn to the threat area - with keeping in mind terrain and where the water will flow (towards the Mohawk River). In other events, people have a tendency to keep the county-based warning era alive (which is now over 5+ years ago) and draw entire counties/zones to the polygon without thinking of the consequences. With WEA alerts going for flash flooding, it's important to consider your polygon, HOWEVER, always keeping in mind where water flows (as flash flooding can occur downstream from areas that receive heavy rain as well), but not warning unnecessary areas just because Itís the same county.
Similar to the Schoharie flooding event, excessive runoff from already saturated soils played a crucial role in the flash flooding. While local streams and creeks exacerbated the flooding, a lot of this was contributed by the excessive runoff.
1-hr and 3-hr Zonal FFG values were 2" and 2.5" respectively for southern Herkimer Co. I am not sure the soils were as saturated as indicated. Some of these locations have not had rain in the past week or so (June 13-14th last time robust rain). Will checked the % soil moisture graphs too. High PWAT air and training cells in the favored location with weak steering flow look like probable causes for the flash flooding.
Although the flash flooding was significant, a Flash Flood Emergency was not issued since evacuations were not ordered and that there were no reports of injuries/fatalities. Instead, enhanced wording was used in the FFS statements (see DSS section below). Eastern Region was not notified as well since their criteria was not met.
Flash Flood Climatology was useful again with Herkimer CTY being the hot spot for Flash Flooding. The high dew point air, high FZL's, anomalous PWAT, weak shear, training thunderstorms were all keys for the Flash Flood!
STA and STP did well. I think the STA showed a significant improvement from the poor performance from the stratiform rain events.
Once again the meso-scale models proved disappointing in that they failed to indicate where convection would be focused and therefore missed the rainfall maximum in both Herkimer and northern Washington counties.
Again...finding where convection will initiate in a "Pulse" environment is like finding a needle in a haystack. I think the storm-scale models have a long way to come yet. The NAM12 at 00Z and 06Z, and the 00Z CAN REG kept showing a well-defined "frontal/convective band" traversing the central portion of the forecast area in the late pm and early evening. This did not happen at all! It crumbled and weakened with no impact in the Capital Region and points just south.
Above:† Graphics showing the anomalously high soil moisture in NY and New England.
Above:† Graphics showing prior rainfall and antecedent moisture over the northeastern U.S.
Above:† Sounding from Albany, NY (KALB) at 1200 UTC 23 June.
Above:† Soundings from 0000 UTC 24 June at (left to right) Albany, NY (KALB), Buffalo, NY (KBUF), Maniwaki, Canada (KWMW) and Upton, NY (KOKX).
Above:† Loop of the High Resolution Weather Research Forecast model (WRF) simulated forecast radar reflectivity.
Above:† Loop of KENX Base Reflectivity.
Above:† KENX rainfall estimates from the legacy Storm Total Precipitation and the dual pol Storm Total Accumulation.
Above:† Flash Flood Monitoring Program showing the excessive rainfall in parts of southern Herkimer County.
Above:† AHPS displays of the water rises at Little Falls and Kast Bridge.
Above:† Flooding in Middleville.
Above:† Flooding in Herkimer.
Above:† Flooding around Schuyler and Fairfield.