A RIVER AND FLASH FLOOD CLIMATOLOGY OF SOUTHERN NEW ENGLAND:
RESULTS FROM 1994-2000

David R. Vallee and Joseph DelliCarpini
NOAA/National Weather Service Forecast Office
Taunton, Massachusetts 02780

 

1. INTRODUCTION
   The National Weather Service in Taunton, MA assumed its Hydrologic Service Area (HSA) in November 1994, extending from the lower Connecticut River Valley eastward to coastal Massachusetts. This study presents a climatology of all flood and flash flood events during the period 1994-2000. Events are separated into two distinct categories. These categories are Flood/Flash Flood, which includes flooding that occurs within twelve hours of a heavy rain event, but not strictly limited to true flash flooding associated with short duration heavy rainfall; and River Flood, which is defined as all flooding associated with the larger main stem rivers and their largest tributaries. Flood/Flash Flood events are calculated on a county by county basis, while main stem river flooding is calculated on a forecast point specific basis for 34 forecast locations on 19 major rivers in the region.

2. HYDROLOGIC SERVICE AREA
   The National Weather Service Forecast Office, in Taunton, MA, has a very complex HSA. Geographically, it extends from the eastern slopes of the Berkshires and Monadnocks of western Massachusetts and southwest New Hampshire, to the coastal plain of southeast New England including Cape Cod, Martha’s Vineyard and Nantucket Island. Small stream response times range from a matter of hours in the hilly terrain of the northwest to days along the coastal tributaries. The HSA includes the Connecticut and Merrimack Rivers, two of the largest river valleys in New England. In addition, the HSA includes three highly urbanized, hydrologically fast responding metropolitan regions including the Hartford, CT, Providence, RI and Boston, MA metropolitan areas, all with their own unique modes of urban flood control.

3. DATA SOURCES
 
The National Climatic Data Center (NCDC) publishes a monthly publication entitled Storm Data. This publication contains a chronological listing by state of occurrences of storms and unusual weather phenomena. Included in the publication are reports pertaining to these categories of Flood/Flash Flood and River Flood events. The database which is used for the Storm Data publication draws upon a variety of sources to determine the occurrence of flooding. These include, trained NWS Spotters, local Amateur Radio Spotters, newspaper clippings and United States Geological Survey (USGS) stream gaging stations for location where official flood stages have been determined.

4. METHODOLOGY
   For this study, a Flood/Flash Flood event was defined as the occurrence of flooding associated with heavy rainfall, melting snow, or a combination of events leading to the inundation of normally dry areas. This included both inundation along known waterways, including some gaged small streams, as well as significant urban flooding. River Flooding is defined as flooding which occurs along a major river or large tributary at a known gaged and forecasted location in the HSA. Official flood stages have been determined for each location and are used as the reference point at which flooding begins along a corresponding reach of the river.

   For Flood/Flash Flood events, any occurrence of flooding or the first observed rise above flood stage at a gaged small stream location was considered one event for the county. Multiple occurrences of observed flooding from the same city or along the same small stream were not included as a new event. For River Flood events, the first observed rise above flood stage at the forecast point along the river was considered one event. Multiple crests were ignored unless a corresponding fall below flood stage occurred.

5. FLOOD/FLASH FLOOD CLIMATOLOGY
   The following subsections provide a summary of yearly and monthly Flood/Flash Flood distributions. The HSA includes the counties of Hartford, Tolland and Windham in northern Connecticut, all of Massachusetts, with the exception of Berkshire County, all of Rhode Island and Cheshire and Hillsborough counties of southwest New Hampshire.

5.1 Yearly Distribution
   Flood/Flash Flood episodes varied considerably from year to year. Note in Figure 1, the dramatic periods of quiet versus extremely active years. There were 154 Flood/Flash Flood events, from 1994 through 2000. Two major flood episodes affected the HSA. These included the devastating floods of October 1996, resulting from a developing coastal storm tapping tropical moisture from offshore Hurricane Lili. An equally devastating urban flood event occurred in June 1998 and was associated with a slow moving frontal boundary in which training of convection was the catalyst.

wpe1.jpg (17945 bytes)
Figure 1. Yearly distribution of Flood/Flash Flood events for 1994-2000.

5.2 Monthly Distribution

Flood/Flash Flood can occur any month of the year, as evident in Figure 2. Clearly, the October 1996 and June 1998 events stand out dramatically. Events are distributed throughout the year. Maxima occur in the later spring and early summer related to convective activity, and in late summer and early fall, related to the effects of tropical cyclones (Vallee, 2000).

wpe52.jpg (20618 bytes)
Figure 2. Monthly distribution of Flood/Flash Flood events for 1994-2000.

5.3 Hourly Distribution

Figure 3 provides the hourly distribution of events, divided into 6 hour periods. A maxima of activity does appear to occur in the late afternoon through the overnight hours.
wpe2.jpg (14461 bytes)
Figure 3. Time distribution of Flood/Flash Flood events for 1994-2000.

6. RIVER FLOOD CLIMATOLOGY
The following section provides a summary of yearly and monthly distribution of River Flooding. The HSA includes the following major rivers:
  In the Connecticut Basin: Connecticut, Chicopee, Westfield,             Farmington
  In the Merrimack Basin: Merrimack, Piscataquog, Souhegan,           Nashua, North Nashua,     Squannacook, Assabet, Sudbury,           Concord, and Shawsheen
  In southeast New England: Shetucket, Blackstone, Charles and           Neponset

6.1 Yearly Distribution
River Flood episodes vary considerably from year to year. There were 100 River Flood events from 1994 through 2000. Note in Figure 4, the dramatic periods of quiet versus extremely active years as was the case for Flood/Flash Flood events. The 1996 year was hydrologically active, with record breaking winter snows followed by multiple River Flood episodes through late spring. This was followed by Tropical Storms Bertha and Edouard, and ended with the devastating floods in October.

wpe54.jpg (21145 bytes)

Figure 4. Yearly distribution of River Flood events for 1994-2000.

6.2 Monthly Distribution

River Flood events can occur in any month of the year, as evident in Figure 5, but there are two more distinct periods of activity. These periods are related to the late winter/spring snowmelt season and the late summer and early fall associated with the affects of nearby tropical cyclones. A third maxima is also noted in the early winter, associated with extratropical cyclones producing heavy liquid precipitation.

wpe55.jpg (20201 bytes)

Figure 5. Monthly distribution of River Flood events for 1994-2000.

6.3 Hourly Distribution

Figure 6 provides the hourly distribution of River Flood events. The hour of occurrence is defined as that time at which a forecast point location exceeds its official flood stage. River Flood events appear to be rather evenly distributed in time.

wpe56.jpg (20094 bytes)
Figure 6. Hourly distribution of River Flood events from 1994-2000.

7. CONCLUSION

After reviewing Flood/Flash Flood events and River Flood events since 1994, several trends were noted for both categories. The frequency of Flood/Flash Flood and River Flood events vary considerably from year to year. Flood/Flash Flood events tend to occur most frequently in the late spring and early summer, with a second maxima in the late summer and early fall. River Flood events also show maxima in these time periods. Flood/Flash Flood events were more common during the late afternoon and evening time frame, while River Flood events showed little preference in time of occurrence.

8. REFERENCES

National Climatic Data Center (NCDC), 1994-2000: Storm Data, Asheville, North Carolina.

Vallee, D. R., 2000: A Centennial Review of Land Falling Tropical Cyclones in Southern New England. Preprints, Twenty-forth Conference on Hurricanes and Tropical Meteorology (Fort Lauderdale), American Meteorological Society, Boston, 551-552.