Is Flooding Getting Worse in New Jersey?

A Historical, Empirical Analysis of Temporal Trends in Flood Frequency and Magnitude

Kirk R. Barrett (1) and Eric Slaff

 Passaic River Institute, Montclair State University, 1 Normal Ave., Montclair, NJ 07043

(1) 973-655-7117, kirk.barrett@montclair.edu, www.csam.montclair.edu/pri

Flooding is a severe and reoccurring problem throughout nearly all of New Jersey.   Furthermore, it seems a popular perception that flooding is getting worse – more frequent floods and/or more extreme floods.  Certainly, urbanization has occurred and it is well known to increase both rate and volume of runoff.   However, the link between the increase in impervious surface and the magnitude and frequency of floods is not as obvious as it may seem.  First, increase runoff should be mitigated somewhat by stormwater management systems installed during development, most commonly detention basins.  Also, during extremely long and large precipitation events, pervious surfaces become more and more saturated and their behavior approaches that of impervious surfaces. Therefore, replacement of pervious with impervious surface may not cause a large increase on the largest floods. 

The project includes examination of temporal changes in the “channel-forming flow” (aka “dominant discharge” or “bankfull flow”), which is approximated as the 2-year flow.  Changes to the channel-forming flow can be linked to streambank erosion, which in turn causes several serious problems:  smothering benthic habitats, increasing turbidity, filling in lake and reservoirs, destroying threaten riparian habitats, undermining structures and endangering human safety. 

We are determining how flow associated with the 2, 5, 10, 20, 50 and 100 year return periods has changed over time for ~60 stream flow gauges in New Jersey with at least 50 years of record, starting with ~20 gages in the Passaic Basin.  Using the computer program HEC-FFA, we analyze the peak annual flow series of running 30-year blocks at successive 5 year intervals.  

Statistical tests will be applied at each gage to determine if there are statistically significant correlations between increases in flood flows with time, and, if so, the rate of increase will be computed.  We will aggregate the data and report, for each return period, the percent of gauges that showed a statistically significant correlation.  This should give a good indication of whether flows of each given return period are showing increases statewide. 

An analysis of a single gage should give a good indication of whether non-extreme floods (2 to 10 years) are indeed increasing.  However, the inherent randomness and infrequency of extreme events confounds identification of trends in large return-period flows (50-100 years) at a single gauge. Fortunately, the large number of gauges distributed throughout a large geographic area (all of NJ) helps overcome this problem  -- we can examine if large return-period flows are increasing in New Jersey as a whole by examining all of the gages together.   To do this, we will first determine, for each gauge, the return period of each year’s recorded peak flow based on the gauge’s entire record.   Then, for each year, we will determine the percentage of gages whose peak flow was greater than or equal to the 100-yr, 50-yr, 20-yr, 10-yr, 5-yr and 2-yr flows.  We will then plot these percentages versus time to see if the percentage of gages exceeding a particular return period is increasing with time.