Precipitation

Unusually large amounts of rain deluged parts of Burlington, Camden and Ocean Counties in southern New Jersey from the morning of July 12 through the early morning hours of July 13, 2004. The heavy rainfall was caused by a warm front approaching from the southwest that stalled over the area as a low pressure system developed along the front. The system received upper-level support from the northwest. (Anthony Gigi, National Weather Service, oral commun., July 14, 2004). Doppler radar estimates of total rainfall for the 24-hour period ending 0700 hours EDT on July 13 were from 8-12 inches over central Burlington, western Ocean, and eastern Camden Counties. More than 11 inches of rain was recorded by rain gages in Pemberton and Tabernacle Townships in Burlington County. The rain gage at Tabernacle recorded 13.2 inches, and the rain gage at the U.S. Geological Survey (USGS) gaging station on Greenwood Branch at New Lisbon (01466900) in Pemberton Township, N.J., recorded 11.3 inches. Both rain gages are located in the Rancocas Creek basin; a tributary to the Delaware River.

The most intense period of rainfall observed at the USGS rain gage was from 1730 to 2130 hours on July 12. The gage recorded 7.88 inches during that 4-hour period. The most intense hourly rainfall was observed during the 1-hour period between 1930 and 2030 hours, when 3.51 inches was recorded.

Peak Flows and Stages

Flood peaks were the highest ever recorded at six USGS stream gaging stations in Burlington and Camden Counties N.J. The stream gages are located on South Branch Rancocas Creek, McDonalds Branch, Greenwood Branch, and North Branch Rancocas Creek in the Rancocas Creek basin and at gaging stations on the Cooper River and South Branch Pennsauken Creek. The Rancocas Creek, Pennsauken Creek, and Cooper River are tributaries to the Delaware River.

The flood peak on the South Branch Rancocas Creek at Vincentown significantly exceeded the 100-year recurrence interval flood. Flood frequency statistics at this site are based on peak flow data from 1961 through 1990. The flood peak exceeded the upper end of the current stage/discharge relation. A discharge measurement made just after the stream peaked measured 4,100 ft³/s (cubic feet per second). As a result of the measurement and the peak stage recorded, the peak flow was estimated to be 4,200 ft³/s. A total of four discharge measurements were made during the highest stages of the flood. Further analysis of the current stage/discharge relation and the new streamflow measurements is needed to more accurately compute the peak discharge. The peak stage recorded was 12.34 ft, 4.36 ft above the previous record of 7.98 ft set on August 28, 1971.

The flood peak on the Cooper River at Haddonfield significantly exceeded the 100-year recurrence interval flood. Flood frequency statistics at this site are based on peak flow data from 1964 through 1990. This flood peak exceeded the upper end of the current stage/discharge relation. As a result, the peak discharge was estimated to be 5,000 ft³/s on the basis of the shape of the hydrograph below the value of 3,000 ft³/s. Additional computations and analyses are needed to more accurately compute the peak discharge. A discharge measurement was made on the receding limb of the hydrograph. The peak gage height was recorded at 6.27 ft, 0.81 ft above the previous record of 5.46 ft set on August 28, 1971.

The flood peak on the Greenwood Branch at New Lisbon exceeded the previously recorded peak of record. The peak gage height recorded for this flood was 8.91 ft, 1.13 ft above the previous record of 7.78 ft set on May 11, 1998. The flood peak exceeded the upper end of the current stage-discharge relation. Four discharge measurements made at stages higher than any stages previously recorded (>8.00 ft) were used to extend the stage-discharge relation. Peak flow was estimated to be 1,280 ft³/s. This is 30 ft³/s less than the peak flow of record on May 11, 1998. A large tree and other debris caused an artificially high stage for the flow measured. Additional computations and analyses are needed to more accurately compute the peak discharge. Because the gaging station was installed in 1998, the record of instantaneous peak flows at this site is not long enough to compute flood frequency statistics.

The flood peak on the South Branch Pennsauken Creek at Cherry Hill approximately equaled the 100-year recurrence interval flood. The peak discharge was approximately 1,560 ft³/s, 60 ft³/s higher than the previous peak flow of record. The peak stage recorded was 11.75 ft, 0.12 ft higher than the previous record of 11.63 ft set on July 14, 1994. One discharge measurement was made at this site on July 13, 2004.

The flood peak on the North Branch Rancocas Creek at Pemberton equaled approximately the 90-year recurrence interval flood. The peak discharge was 2,050 ft³/s, 320 ft³/s higher then the previous peak flow of record. The peak gage height recorded was 4.20 ft, 0.03 ft less than the previous record of 4.23 ft set on August 21, 1939. Two discharge measurements were made at this site during the flood.

The flood peak on the McDonalds Branch in Lebanon State Forest tied the historical peak set on August 25, 1958. The peak stage recorded was 2.33 ft, and the peak flow was approximately 35 ft³/s. The event is approximately a 50-year recurrence interval flood.

The only other gaging station in southern New Jersey to record a peak greater than a 2-year recurrence interval flood during the July 12-14, 2004, storm is the station at Batsto River at Batsto New Jersey. The peak stage was 7.15 ft and the peak flow was approximately 1,030 ft³/s. This event is approximately a 10-year recurrence interval flood.

Response to the event

The USGS New Jersey District hydrologic data assessment program responded to flood conditions beginning at 0530 hours on July 13, 2004. Real-time data from USGS gaging stations in southern New Jersey played a significant role in planning the first line of response to the storm. The real-time data helped the USGS personnel to focus efforts on the watersheds experiencing the most flooding. Plans were made to send crews to the areas hardest hit.

Five crews of hydrologic technicians and hydrologists visited the stream gaging stations and crest-stage gages in the area affected by the flooding. Eighteen discharge measurements were made on July 13-14, 2004. Many measurements were made at stages that had never been measured before. All stream gaging stations and crest-stage gages in the affected areas and surrounding areas were inspected, even though no flow measurements were made. High-water marks were flagged at some sites. Photographs were taken of the flooded areas at the South Branch Rancocas Creek at Vincentown gaging station.

The New Jersey District worked with the New Jersey Department of Environmental Protection (NJDEP) Bureau of Dam Safety and the National Weather Service River Forecast Center. Requests for discharges measured during the flooding, information on record high stages and flows, and assistance with timing of peaks between sites were answered. Information on peak stages and flows and river conditions were provided to the Associated Press. The Federal Emergency Management Agency (FEMA) was contacted.

The NJDEP Bureau of Dam Safety reported 13 dams that failed in the Rancocas Creek basin during this flood. Two of the dam failures were upstream from the gaging station on the South Branch Rancocas Creek at Vincentown. The other 11 dams that failed are upstream from two discontinued crest-stage gages on the Southwest Branch Rancocas Creek. High water marks have been flagged to record the record breaking flooding in the basin.

No damage to USGS gaging stations or other USGS property was observed.

Table 1. Summary of historical flood peaks and peaks from the July 12-14, 2004, flood event at USGS sites