Long Island and Mid-Atlantic Record Deluge Pattern and Analysis

1. 
   Introduction
   

The larger scale pattern, with a north-south frontal and moisture boundary (Fig. 2) had many of the characteristics often associated with a Maddox Synoptic rainfall event type (Maddox et al. 1979). It will be shown that as the wave moved up the frontal boundary, strong easterly flow developed north of the cyclone. On the time scale of hours, this event transitioned into a Maddox Frontal system, with stronger easterly flow. Heavy rainfall events which transition from synoptic to frontal types can produce record rainfall. For example, the record breaking rainfall 29-30 March event in southeastern New England (Grumm 2011) was associated with this transition.

Events of this type are difficult to forecast with numerical guidance due to grid resolution issues and the need in models with grid spacing over about 11 km (Kain et al. 2008; Weisman et al. 1997) to use parameterization schemes to deal with instability and the implied convective rainfall. The modeling gray area is horizontal resolutions of 5-11km. Convection allowing models; with horizontal resolutions generally below about 5km; should perform better and produce higher quantitative precipitation amounts than coarser models. However, these systems have limitations and often fail to predict mesoscale heavy rainfall events or have both temporal and spatial errors. These uncertainties and limitations require human interventions related to flash flooding and heavy rainfall forecasts and warnings. Situational awareness in these events requires knowledge of patterns in which heavy rain may occur and the ability to discern a potential ordinary rainfall event from a potential record event.

The use of standardized anomalies (SA: Hart and Grumm 1991; Grumm and Hart 1991; Graham and Grumm 2010) applied to re-analysis data and forecast guidance (Junker et al. 2009) has been demonstrated to have value in identifying potential record events. Du et al (2014) demonstrated how this concept could be applied to ensembles to produce a simple extreme forecast index. Du et al. (2014) used anomaly-based forecast concepts as tools to identify a potential record rainfall event near Beijing, China. Another iteration of anomaly-based forecast and situational awareness, leveraging the forecast probability distribution function (PDF) was demonstrated by Graham et al (2013). Using the full PDF, the extreme events can be placed into percentiles and in some events, the forecasts may be of the record event relative to the 30-year climatological PDF.

This paper will show the pattern associated with the historic Long Island rainfall event and the record rainfall in Maryland and New Jersey. The focus is on the pattern and the value of the pattern and anomalies in improving forecaster situational awareness for this and similar extreme events, which typically occur with short, if not minimal predictability horizons. A companion paper will demonstrate the limitations of NWP in predicting events of this nature.

2. 
   Methods and Data
   

The stage-IV rainfall data was retrieved in gridded form from NCEP. Unless specified otherwise, all plan-view images show accumulations using the 6-hourly data files. The point diagrams use the hourly files for comparison to the point observations. METAR sites were used to get point data at select locations. These point data tended to be higher than those in the gridded datasets.

Rainfall return periods in 1, 3, 6 and 12 hour intervals were examined using the U.S. Department of Commerce Technical Paper No. 40 and the NOAA Atlas 14 Precipitation Frequency Data. For select points such as Islip, Baltimore and Portland, the data were collected to compare to METAR data. These data were also examined using the National Severe Storms Laboratory (NSSL) Flooded Locations and Simulated Hydrographs Project (FLASH) webpage. These data are available in near real-time providing return periods based on rainfall estimates. The return periods at the FLASH site are computed relative to the NOAA-14 Precipitation Atlas data.

3. 
   Pattern and Standardized Anomalies
   

The developing trough (Fig. 4) pulled a plume of deep moisture (Fig. 2) into the region. Though not shown, PW values over 65 mm and 2-3 were present over the southwestern Atlantic as the trough deepened and the surface waves developed along the frontal boundary (Fig.3). The deep moisture plume was pulled into the Mid-Atlantic region and southern New England. Strong southerly flow at 850 hPa was present (Fig. 5). The north-south PW plume and strong southerly flow was indicative of a Maddox Synoptic pattern. But ahead of the waves along the frontal boundary, easterly flow developed. The 850 hPa winds and 850 hPa u-wind anomalies indicated +2 to +3 u-wind anomalies in Maryland and New Jersey at 1800 UTC 12 August and 0000 UTC 13 August respectively. The low-level easterly wind anomalies peaked at -4 below normal over Long Island at 0600 UTC 13 August then quickly lifted northward into southern New England, where additional heavy rainfall was observed.

Another contributing factor in this event may have been the strong frontal system and the injection of unseasonably cold air into the system from the west (Fig. 6). The 850 hPa temperature anomalies were -2 to -4 below normal in the Ohio Valley (Fig. 6e) while a narrow ribbon of +1 above normal air was pulled into the coastal regions ahead of the front. These data imply strong baroclinicity in the larger scale pattern. Backward trajectories from a point near Islip, New York implied that flow ahead of the frontal boundary ingested air from the southwestern Atlantic (Fig. 7) and the air at 1500 meters originated near the 65 mm PW area off the southeastern United States.

The heavy rainfall and the extreme rainfall over portions of Long Island were likely related to the enhanced low-level flow and a coastal front. The latter feature is beyond the scope of this study to examine. This case shows the value of anomaly-based analysis to identify patterns of potentially high impact weather events.

4. 
   Rainfall and Rainfall Return Periods
   
   
   
   

1. 
   Rainfall Patterns and Amounts
   

These rapid accumulations in short 3-6 hour windows produced remarkable rates which accounted for 80 to 95% of the event total estimated rainfall (Fig. 1). The point data from the stage-IV data (Fig. 10) and observational data showed the period of extreme rainfall at Islip and Baltimore (BWI). The rainfall near BWI was very light from 0000 UTC then peaked dramatically from 1500 to 1800 UTC. The intense rainfall was in a 3 hour window.

At Islip the gridded data implied the rain began at 0300 UTC and picked up in intensity at 0600-0700 UTC when over 9 mm rain fell. The gridded data then showed 5 hours of hourly rainfall in excess of 18 mm and a peak value of 69.13 mm. Point data showed even higher amounts, illustrating some of the limitations of gridded rainfall estimates. Summaries of the rainfall totals in the Mid-Atlantic region, Long Island and southern New England, and New England are provided in Tables 1-3. Many locations had over 4 inches (100mm) rainfall.

2. 
   Rainfall Return Periods
   

The rainfall return period data for the Baltimore-Washington International Airport (KBWI: Table 5) shows the observed 1- and 3-hour peak observed rainfall compared to the NOAA Atlas 14 data. The overall rates and accumulations were not as extreme as those observed at Islip. However the site did exceed the 100-year return period for 6-hour rainfall amounts. The Plainview data near the peak of the event is shown in Fig. 12. The plainview data show the focus of the heaviest rainfall south of the city of Baltimore and another area of heavy rainfall to the south near Washington, DC.

Similar data is shown for the rainfall and return periods for Portland, ME in Table 6. These data show that the 6-hour rates in Portland were just under those observed during a 500-year event and the 3-hour rates exceeded the 500 year return period. The planview data near the peak of the event in the Portland area is shown in Fig. 13.

A summary for each METAR sites’ hourly rainfall is provided in the Appendix.

5. 
   Summary
   

The pattern of the rainfall shows how mesoscale this event was in both time and space. In most areas, the heaviest rain fell in a 3-6 hour period. Across Long Island, where rainfall rates were historically high, only a narrow band received the extreme rainfall rates, and areas on the warm side of the boundary received significantly lower rainfall amounts. Deeper into the cold air, the rainfall amounts fell off dramatically.

From a short-term forecast and warning perspective, this case demonstrates the value of knowing the patterns in which heavy rains can occur. This case also shows the critical input rainfall return period data can play in the near-term forecast process. At both the synoptic and mesoscale, climate-based forecasts are useful and effective forecast tools.

This case shows the value of anomaly-based analysis to identify patterns of potentially high impact weather events. When applied to forecast output from numerical weather prediction models and ensemble forecast systems, these same concepts can be applied to anomaly-based forecasts of high impact weather events. Probabilistic information relative to the forecast and climatological PDF likely add significant value to this process.

The return period data Islip, Baltimore, and Portland show the value of return periods in the forecast and warning process. Previous work on NOAA-14 data suggests rainfall rates in the 10 to 50-year interval produce nuisance floods. As the rates approach the 100-year event, the floods become significant and rainfall rates exceeding 100-year and 500-year events tend to be record-breaking and historic events. These data need to be better incorporated into the flood warning process in real-time.

6. 
   Acknowledgements
   

7. 
   References
   

Associated Press, 2014: Flooding-Metro NYC Story. And similar stories.

Fritsch,J.M and L.A. Giordano, 1991: Strong Tornadoes and Flash-Flood-Producing Rainstorms during the Warm Season in the Mid-Atlantic Region, Wea. Forecasting,6,437-455.

Du, J, R.H. Grum, and G. Deng, 2014: Ensemble anomaly forecasting approach to predicting extreme weather demonstrated by extremely heavy rainfall even in Beijing. Chin., J. Atmos, Sci,38(4): 685-699 (in Chinese)

Graham, R.A, T. Alcott, N. Hosenfeld, and R. Grumm, 2013: Anticipating a rare event utilizing forecast anomalies and a situational awareness display: The Western Region US Storms of 18-23 January 2010, BAMS,2013.

Graham, R A. and R. H. Grumm, 2010: Utilizing Normalized Anomalies to Assess Synoptic-Scale Weather Events in the Western United States. Wea. Forecasting, 25, 428-445

Grumm, R.H. and R. Hart. 2001: Standardized Anomalies Applied to Significant Cold Season Weather Events: Preliminary Findings. Wea. and Fore., 16,736–754.

Hart, R. E., and R. H. Grumm, 2001: Using normalized climatological anomalies to rank synoptic scale events objectively. Mon. Wea. Rev., 129, 2426–2442.

Kain,J.S, S.J. Weiss, D. R. Bright, M. E. Baldwin, J J. Levit, G W. Carbin, C S. Schwartz, M L. Weisman, K K. Droegemeier, D B. Weber, and K W. Thomas, 2008: Some Practical Considerations Regarding Horizontal Resolution in the First Generation of Operational Convection-Allowing NWP. Wea. Forecasting, 23, 931–952.

Maddox,R.A., C.F Chappell, and L.R. Hoxit. 1979: Synoptic and meso-alpha aspects of flash flood events. Bull. Amer. Meteor. Soc., 60, 115-123.

Weisman,M.L, W. C. Skamarock, and J. B. Klemp, 1997: The Resolution Dependence of Explicitly Modeled Convective Systems. Mon. Wea. Rev., 125, 527–548.