Remnants of Dennis
06-07 September 1999
by
Richard H. Grumm
And
David Ondrejik
National Weather Service
State College PA
Draft 10 Sept 1999
INTRODUCTION
The remnants of hurricane Dennis moved over central Pennsylvania during the afternoon and evening hours of 06 September. Early bands produced some severe weather over southwestern sections of the Commonwealth. The features associated with these bands were similar to those associates with hurricane Fran three years earlier.
The most significant problem was with heavy rains. Between 0300 and 0700 UTC, locally heavy rains pounded portions of the Susquehanna Valley region of central Pennsylvania. Despite the magnitude of the flooding, no deaths were reported.
This page will slowly grow to show the forecast products and nowcast products, which were of use during this heavy rainfall event.
METHOD
All data were collected and archived locally. The RIDDS tape made during the event was used to evaluate the rainfall and severe weather. Using WATADS, the WSR-88D tropical Z-R and default Z-R were used to create imagery. Operationally, the tropical Z-R was employed during this event. It will be shown that radar estimate were 2-3 inches higher, and closer to reality using the tropical Z-R relationship relative to the default Z-R relationship. Please note the scale changes in some figures to accent the locally heavy rainfall amounts.
The default Z-R is :
R = 300 Z1.4
and the tropical Z-R is:
R = 250 Z1.2
Where R is the rainfall rate and Z is the reflectivity. The tropical Z-R should do better with a smaller drop distribution and with low topped convection. In this case, the convection was not very deep and the tropical air mass probably provided a drop distribution favoring the tropical Z-R.
As the
Other data used include Eta grib files, surface data, and local mm5 model runs.
A quick look at all images can be obtained here.
OVERVIEW
Under development. Overall, it appeared the surface low tracked over central Pennsylvania.
The band during the early afternoon brought some locally strong and damaging winds to southwestern Pennsylvania. No images are available yet.
The character of the heavy rainbands can be seen in the 0416 UTC base reflectivity image. During this time and for the next 3 hours, a north-south oriented band set up over east central Pennsylvania. This band and cells moving along it produced heavy rains to the area.
Some radar imagery showing the weak mesoscale circulation's along the bands between 1800 and 1900 are available to view below:
Base reflectivity at 1819 Base velocity at 1819
Base reflectivity at 1824 Base velocity at 1819 Base reflectivity at 1839 Base velocity at 1819
A cross section through the storm over southern Cambria County was cut to show the
reflectivity and velocity data at 1819 UTC. The reflectivity data showed the narrow high intensity reflectivity core, quite low compared to most severe weather events, and a rapid tilt of the cloud top material to the north, toward the RDA. The storm relative velocity data showed a strong outbound area near the updraft core with strong inbounds aloft, blowing the cloud top material toward the RDA. The inbound/outbound couplet at near 45 miles in the image at times produced mesoscale circulations in the WSR-88D algorithms. This storm was one of two, which passed over this particular region.Similar bands and circulation's appeared with Fran in 1996. These too, produced severe weather and one confirmed and one unconfirmed tornado.
The storm total rainfall estimated by using the default Z-R relationship is shown in the STP valid
at 1143 UTC. The maximum rainfall estimates were around 5.1 inches of rain. There was a significantly larger area of 4.4-5.1 inches of estimated rainfall.The hourly and three hourly precipitation products (OHP and THP) showed that the period of heavy rains was confined to between 0300 and 0700 UTC. The
0405 UTC OHP product showed that up to 1.5 inches of rain was estimated to have fallen over the region where the flood would occur. The corresponding 3-hour rainfall product showed that nearly 2.4 inches of rain had fallen in the past three hours, suggesting that the last hour accounted for 60-70% of the total rainfall in the area.During the next hour, the
0501 UTC OHP showed about another inch fell over the area, most of it to the north of the heavier rain from the previous hour. Despite this, the 0501 UTC THP showed about 2.5 inches. The rain continued into the neat hour as seen in the 0600 UTC OHP and STP images. Rain continued to fall over the area between 0600 and 0700 UTC as seen in the 0700 UTC OHP and STP images. By 0800 UTC, the rain had shifted north of the flood area as seen in the 0800 OHP showing the rainfall from 0700-0800 UTC.The focus of the heavy rain, with nearly 1-inch per hour rates was confined to portions of eastern Union and northeastern Snyder Counties.
The STP product using the tropical Z-R relationship showed a local maximum of 9.5 inches of rain in Union County. Overall, in the area of locally heavy rains associated with the rainbands on East Side of the track of Dennis, the tropical Z-R produced significantly more rainfall than the default Z-R relationship. Interestingly, the high-resolution data from the archive II produced about 0.7 inches more rainfall than the operationally available STP.
For comparison purposes, the 0405, 0501, 0600, 0700, and 0800 UTC OHP's were produced along with matching THP's. As shown earlier, the heaviest rain began after 0300 UTC with around 3 inches of estimated rainfall in the 0405 UTC OHP (nearly double that in the default OHP). The rainfall decreased in intensity after 0400 UTC as seen in the 0501 UTC OHP. Rainfall picked up in intensity after 0500 UTC as seen in the 0600 UTC OHP where nearly 2.5 inches of rain was estimated over northern Northumberland County. During the next hour, the rain band shifted west and northward producing around 2.5 inches of rain over eastern Union County northward into Lycoming County as shown in the 0700 UTC OHP. The rainfall rapidly diminished after 0700 UTC.
The 0405 UTC THP showed that nearly 4 inches of rain was estimated to have fallen over northern Dauphin County between 0100 and 0400 UTC. Over Union County, the OHP estimate accounted for most of the rainfall in through 0400 UTC. The heavy rains were moving northward with time up the Susquehanna Valley. The 0501 UTC THP showed the continued northward progression of the axis of heavy rainfall. Note that these estimates were 150% of those estimated using the default Z-R relationship. The 0600 and 0700 UTC THP's continued to show the trend of the northward progression of the rainband. Note that the rainfall color scales were changed in these latter products to show the estimates of over 5 inches of rain in the 3-hour periods from both 0300-0600 and 0400-0700 over portions of the middle Susquehanna Valley. These three-hour estimates were nearly equivalent to the STP produced using the default Z-R relationship!
The most difficult part of the model forecasts was the track of the remnant surface low. The heaviest rain was forecast by all models to remain to the east of the surface low track. Based on the source of warm moist air off the Atlantic Ocean, this was a likely scenario. Unfortunately, the models had difficulty forecasting the track of the surface low as seen in the Eta surface pressure forecasts, valid at 07/0000 UTC. The forecasts are from the 05/0000, 05/1200, 06/0000, and 06/1200 UTC model runs. Note the large variation in the forecast track of the surface low. The corresponding precipitation forecasts (mm) are shown in the next image.
The next forecast cycle offered some additional help, just prior to the onset of the heavy rains as shown in the surface forecast valid at 07/0600 UTC. The corresponding precipitation plots area also shown. Note the general convergence of the heavy rainfall forecasts over central Pennsylvania from the 06/0000 UTC through 07/0000 UTC Eta forecasts.
CONCLUSIONS
The remnant of Dennis brought locally heavy rains to central Pennsylvania during the evening hours of 06 September 1999. The initial bands brought heavy rains and some isolated severe weather. These bands were oriented from SE to NW across the region. Later, as the surface low entered the State, a more N-S band produced locally heavy rains and flash flooding in portions of the lower and middle Susquehanna Valley's. The most severe flooding was in the latter region, where over 8 inches of rainfall was observed.
Similar to hurricane Fran, three years earlier, the tropical Z-R was used operational during the event. This lead to reasonable rainfall estimates relative to observations. The default Z-R relationship grossly underestimated the areas of heavy rainfall. This case represents the second tropical system for which a comparison of the default and tropical Z-R relationships could be accomplished. In both instances, the tropical Z-R produced estimates, which were more reliable and probably played a critical role in the successful issuance of Flash Flood Warnings by the National Weather Service in State College.
