Description of Study Area

The study area is located within the New Jersey Atlantic Coastal Plain region and includes the drainage basins of four streams that discharge to the Barnegat Bay-Little Egg Harbor estuary—North Branch (NB) Metedeconk River (referred to as NB Metedeconk River), Toms River, Cedar Creek, and Westecunk Creek. The Barnegat Bay-Little Egg Harbor watershed encompasses more than 600 square miles (Kennish, 2001). The combined drainage area of the four study sub-basins is about 224 square miles (table 1; fig. 1) (U.S. Geological Survey, 2024).

Predominant land uses within the Barnegat Bay-Little Egg Harbor watershed vary regionally (fig. 2). The northeastern part of the watershed is mostly urban with both residential and nonresidential development and includes several major population centers. The southeasternmost part of the watershed contains several protected wildlife refuges and wildlife management areas and is less developed; however, the areas peripheral to this and proximal to Barnegat Bay and Little Egg Harbor (fig. 2) have undergone a substantial increase in development from 1986–2015. The complex of barrier islands on the eastern shore of the estuary is mostly urban, with the exception of Island Beach State Park. Much of the western portion of the watershed lies within the Pinelands Area (Hunchak-Kariouk and Nicholson, 2001; Kennish, 2001). This area is protected under the Pinelands Comprehensive Management Plan and is characterized by large tracts of forested land, wetlands, and some low-density development. The headwaters of the Toms River, Cedar Creek, and Westecunk Creek sub-basins originate in the Pinelands Area (fig. 1).

The percentage of land in each land-use category was computed for the four sub-basins using digital land-use/land-cover (LU/LC) datasets obtained from the New Jersey Department of Environmental Protection Bureau of Geographic Information Systems for years 1986, 1995, 2002, 2007, 2012, and 2015 (New Jersey Department of Environmental Protection Bureau of GIS, 1998, 2000a, 2010a, 2010b, 2015b, and 2019) and the U.S. Environmental Protection Agency Geographic Information Retrieval and Analysis System (GIRAS) for 1973 (U.S. Environmental Protection Agency, 2019). GIRAS percentages are included in table 2 as they represent the best available data for the period; however, GIRAS percentages were computed at a lower resolution than any of the other land-use datasets, so some differences in percent LU/LC were expected.

The four sub-basins differ in size and extent of development, and a summary of the LU/LC distributions in the four sub-basins are provided in table 2. The NB Metedeconk River sub-basin is the northernmost of the four study sub-basins and is the most heavily developed (fig. 2). Between 1986 and 2015, the percentage of urban land in this sub-basin increased from approximately 31.7 percent to 43.3 percent. During the same period, forested land cover decreased from 27.2 to 20.0 percent, and agriculture and wetland cover decreased by 1.5 and 2.9 percent, respectively (table 2).

The Toms River sub-basin is the largest of the four study sub-basins. More than 50 percent of the sub-basin lies within the Pinelands Area (fig. 1). Between 1986 and 2015, the percentage of urban land in this sub-basin increased from approximately 14.2 percent to 26.7 percent. During the same period, forested land cover decreased from 51.6 percent to 43.5 percent; and agriculture, barren land, and wetlands cover decreased by 2.0, 2.3, and 0.7 percent, respectively (table 2).

The Cedar Creek sub-basin is primarily undeveloped. Nearly 90 percent of the sub-basin lies within the Pinelands Area, and more than 70 percent of the sub-basin is forested. Between 1986 and 2015, there was little change in LU/LC distributions, with urban land cover increasing by 1.6 percent and barren and forested land decreasing by negligible amounts (table 2).

The southernmost sub-basin, Westecunk Creek, is the smallest of the four sub-basins and is the least developed. As of 2015, approximately 80 percent of the sub-basin was forested, and 0.3 percent was comprised of urban land cover. Between 1986 and 2015, there was little change in LU/LC distributions (table 2).

The shallow, unconfined Kirkwood-Cohansey aquifer system underlies most of the the Barnegat Bay-Little Egg Harbor watershed. The Cohansey Formation is composed primarily of medium- to coarse-grained sands, with localized occurrences of silt and clay lenses and gravel. The Kirkwood Formation is composed of fine- to medium-grained sands with clay and, to a lesser extent, coarse to fine gravelly sand (Canace and Sugarman, 2009). The sands and gravels of this aquifer system make it an important source of water for communities within the watershed. Groundwater discharges from the Kirkwood-Cohansey aquifer system to major streams in the watershed, including the Toms and Metedeconk Rivers, and to other smaller streams and tributaries, which eventually discharge to the Barnegat Bay-Little Egg Harbor estuary. Streams, wetlands, and other surface-water bodies are hydraulically connected with the Kirkwood-Cohansey aquifer system, as groundwater discharge from the aquifer system accounts for a high percentage (71–94 percent) of surface-water flow in the watershed (Watt and others, 1994; Nicholson and Watt, 1997; Gordon, 2003; Baker and others, 2014).

Previous Studies

In 2007, the USGS developed the Hydroecological Integrity Assessment Process (HIP) and associated tools in cooperation with the New Jersey Department of Environmental Protection (Kennen and others, 2007). The HIP is based on a large body of research that links hydrologic variability and aquatic-ecosystem integrity. This innovative work supports the idea that the full range of natural intra- and inter-annual variation of hydrological regimes and associated characteristics of magnitude, frequency, duration, timing, and rate of change are critical in sustaining the diversity and integrity of aquatic ecosystems (Poff and others, 1997). Furthermore, streamflow is strongly related to many physiochemical components of rivers, such as dissolved-oxygen concentration, channel geomorphology, and water temperature. Modifications in those attributes as a function of changes in the streamflow regime can limit the distribution, abundance, and diversity of many aquatic plant and animal species (Resh and others, 1988; Poff and others, 1997). The HIP was designed to assist decision makers (water-resource managers, planners, and regulatory agencies) with integrating ecological-flow assessments into broader management platforms.

Many components of HIP were directly incorporated into the ELOHA process (Poff and others, 2010) and this approach has been used as a guide for expanding the understanding of ecological-flow regimes and developing improved approaches for balancing human-water needs with ecosystem functions across the globe (Arthington and others, 2018). Additional studies in New Jersey using these methodologies included evaluating the effects of potential changes in streamflow regime on fish and aquatic-invertebrate assemblages in the New Jersey Pinelands Area (Kennen and Riskin, 2010), developing a method to support total maximum daily loads assessment using hydrologic alteration as a surrogate to address aquatic-life impairment in New Jersey streams (Kennen and others, 2013) and a study in the New Jersey Pinelands Area linking groundwater withdrawals to changes in aquatic assemblage response (Kennen and others, 2014). This science has expanded greatly over the past decade and a recent 2018 special issue in “Freshwater Biology” highlights many examples of the applications and case studies to illustrate the implementation of hydroecological approaches and provide a measure of efficacy from these approaches (Kennen and others, 2018).

As part of a recent update of state-wide surface-water metrics, McHugh and others (2024) computed high- and low-streamflow metrics for 97 continuous-record streamgages, and low-flow metrics for 719 partial-record gages. These gages were located throughout New Jersey. Continuous streamgages with a minimum of 20 years of record through 2017 were included in the study. In addition to high- and low-flow frequency and duration metrics computed using the entire period of record (POR) available for each continuous streamgage, trend tests on high and low n-day series (for example, the 7-day moving average of flow), baseflow and runoff, peak to 3-day mean-flow ratio, and annual skewness and kurtosis were also conducted to evaluate trends and variability in streamflow at each of the continuous streamgages.

Additionally, the U.S. Geological Survey (USGS) New Jersey Water Science Center has an ongoing study to compute a small subset of low-flow metrics for 31 continuous-record streamgages and 57 partial-record stations located throughout New Jersey for both a historical (1950–1979) and a more recent (1990–2019) 30-year period. The streamflow metrics being computed by this study will help identify changes in low-flow metrics over time and present a synopsis of streamflow conditions using this more recent data. The project will also present some land-use, precipitation, and water use changes in the associated watersheds to help users evaluate if changes to these characteristics can be connected to changes in streamflow metrics. Although the stations are located across New Jersey, the analysis does not include the four stations presented in this report.

Evaluation of Streamflow Record for the Selected Streamgages

The Barnegat Bay-Little Egg Harbor watershed has several active and discontinued streamgages within its boundaries (fig. 1). The four streamgages selected for this study are:

The numbers in parentheses following the name of the USGS streamgage are the USGS site numbers (table 1). For the remainder of the report, these streamgages will be referred to by their stream name. These four streamgages were selected for analysis because they have the longest period of streamflow record within the Barnegat Bay-Little Egg Harbor watershed, and they are all active streamgages with approved data available through the 2020 or the 2021 water year (table 1). Streamflow data for these stations were collected at 15-minute intervals using an electronic data logger and are available online from the USGS National Water Information System (U.S. Geological Survey, 2024). The daily streamflow data used in this study were collected, analyzed, and approved in accordance with standard operating procedures described in Rantz and others (1982), Mueller and others (2013), and Turnipseed and Sauer (2010) and can be accessed via the USGS National Water Information System (U.S. Geological Survey, 2024). The record lengths of the four streamgages are not all coincident as two of the four study streams (Cedar Creek and Westecunk Creek streamgages) have gaps within the span of the period of data availability. Given the varying periods of streamflow data available for each of the four streamgages and the gaps in record, these data were divided into three time periods, referred to in this report as periods of record (PORs). Historical period-of-record 1 (POR1) was defined as the period covering water years 1933–1958, historical period-of-record 2 (POR2) was defined as water years 1974–1988, and the most recent period-of-record 3 (POR3) was defined as water years 2004–2020. Using these time periods allows for a minimum of two streamgages’ records for comparison for each historical POR and all four streamgages for the most recent POR (table 3). Selection of these time periods also allows for streamflow metrics at the largest and most varied land-use sub-basin, Toms River, to be computed and compared for all three PORs. The streamflow record at Toms River is beneficial in offering a continuous look at streamflow conditions over time so higher-flow and lower-flow periods can be determined. The years selected for each POR represent the longest common time span for which data are available among the four study sites.

Selecting Streamflow Metrics to Help Describe Streamflow Conditions in the Barnegat Bay-Little Egg Harbor Watershed

This section provides an overview of the various methods used to calculate the streamflow metrics for the four study streams and highlights the subset of streamflow metrics that were selected to evaluate differences in streamflow between two different PORs.

Evaluating Differences in Streamflow Metrics for Selected Sub-basins

Streamflow metrics were computed and evaluated to help determine how streamflow characteristics may have changed over time. If a balanced aquatic ecosystem existed along the tributaries to the Barnegat Bay-Little Egg Harbor estuary in POR1, then it can be implied that the inherent flow and associated variability that existed during POR1 helped to shape and sustain that balance. Evaluating how streamflow metrics may have changed between the earlier periods of record and the current period of record may help show how the current system can be influenced by anthropogenic change and provide a way for the BBP to implement ecological-flow measures that provide adequate water supplies and flow in the Barnegat Bay-Little Egg Harbor watershed for ecological and human communities now and in the future (Barnegat Bay Partnership, 2021, p. 10).

In the following sections, differences in streamflow metrics between PORs are described for the four study sites. Several streamflow metrics (ML1–12, MA1, DL5, and seasonal 7Q10) are first discussed in detail, and then a comparison of historical and current streamflow conditions is presented by site. This comparison includes additional flow indices (DL1, DL3, DL12, DL14, DL15, and TL1). Combined, these streamflow metrics provide insight into how flow conditions in the four study streams may have changed over time. Boxplots which include the results of the Wilcoxon rank-sum test for several streamflow metrics are provided to help visualize differences between PORs, and the distribution of individual values are shown in corresponding scatterplots in appendix 1.

Monthly Minimums

Monthly minimum flows were summarized by site and POR to help evaluate patterns and changes in low-flow conditions for each of the four streams (table 4). Monthly minimum values offer basic information, such as the historical magnitude of low flow and how minimum flow values vary by month. The NB Metedeconk River and Toms River exhibit a similar pattern where minimum flows are lowest during the summer months, notably July–September, and are greatest during the winter and early spring months, notably February–April (fig. 3). For the NB Metedeconk River, monthly minimum flows ranged from medians of 19.0 to 47.0 cubic feet per second (ft³/s) in POR2 and 17.5 to 46.3 ft³/s in POR3. For the Toms River, monthly minimum flows ranged from medians of 81.5 to 199 ft³/s in POR1 and 88.7 to 188 ft³/s in POR3 (table 4).

When comparing median monthly minimum flows between PORs for NB Metedeconk and Toms Rivers, results of the Wilcoxon rank-sum test indicate that the month of May’s minimum flows are significantly lower in the Toms River for POR3 (median of 146 ft³/s) than for POR1 (median of 169 ft³/s), with a p-value of >0.05. There are other notable differences in monthly minimum flows between POR1 and POR3 for the Toms River. For example, June and April flows are noticeably lower in POR3 than in POR1, with minimum flows about 16 ft³/s (13.1 percent) and 22 ft³/s (11.1 percent) lower in POR3 than in POR1, respectively (table 4); however, none of these differences are statistically significant. NB Metedeconk River exhibited a mix of increases and decreases in monthly minimum flows between POR2 and POR3; the greatest percent changes (>10 percent) were decreases and these decreases occurred during the months of February, August, and October.

Patterns in monthly minimum flows for Cedar Creek differ between POR1 and POR3 (fig. 3) and more significant differences in monthly minimum flows occurred in Cedar Creek than any other study site (table 4). For POR1, the lowest median monthly flow occurred in October (27.0 ft³/s; table 4), with the lowest flows for this site (<40 ft³/s) occurring in almost all months of the year (fig. 3), including the winter months (December: 37.5 ft³/s). Patterns of monthly low flows follow a more traditional pattern for POR3, with the lowest flows occurring during the months of August and September. Monthly minimum flows at this site ranged from a median of 27.0 ft³/s to 75.4 ft³/s in POR1 and 49.5 ft³/s to 80.9 ft³/s in POR3.

The monthly minimum flows in POR3 are greater than in POR1 for Cedar Creek, except for the month of July (table 4). Statistically significant differences were identified in the months of April, May, June, October, November, and December. Prior to May 1, 2012, the Cedar Creek gage was located 2.2 miles further downstream than its current location and had a drainage area of 53.3 square miles. Daily mean streamflows prior to May 2012 were adjusted based on regression using concurrent streamflow measurements at the former and current sites. Potential reasons for lower flows in POR1 could be because the former gage location was subject to tidal backwater during lower flow conditions, which may have resulted in periods of reduced flow past the streamgage. Additionally, flow is periodically regulated by gate operations at lakes and cranberry bogs upstream.

For Westecunk Creek, monthly minimum flows ranged from a median of 17.0 ft³/s to 29.0 ft³/s in POR2 and 17.9 ft³/s to 28.0 ft³/s in POR3 (table 4). Minimum flows are lowest during the summer months and greatest during the months of January–April. No statistically significant differences were identified in the distribution of monthly minimum flows between POR2 and POR3 for Westecunk Creek (table 4; fig. 3).

Table 4.    

Median monthly minimum flow by period of record in cubic feet per second (ft³/s).

[POR1 for NB Metedeconk River, POR2 for Cedar Creek, and POR1 for Westecunk Creek have not been included in this table because no data exist for those periods of record. POR; period of record; NB, North Branch]

Table 4.    Median monthly minimum flow by period of record in cubic feet per second (ft³/s).

Month	Metric abbreviation	NB Metedeconk River, 01408120			Toms River, 01408500					Cedar Creek, 01408900			Westecunk Creek, 01409280
		POR2	POR3	Percent change, POR2–POR3	POR1	POR2	POR3	Percent change, POR1–POR3	Percent change, POR2–POR3	POR1	POR3	Percent change, POR1–POR-3	POR2	POR3	Percent change, POR2–POR3
January	ml1	40	41.7	4.3	166	171	182	9.6	6.4	71.7	73.5	2.5	29	27.6	–4.8
February	ml2	47	40.3	–14.3	181	193	177	–2.2	–8.3	71.7	79.9	11.4	23.5	27.5	17
March	ml3	46	46.2	0.4	197	179	188	–4.6	5	75.4	76.3	1.2	27.2	28	2.9
April	ml4	46	46.3	0.7	199	170	177	–11.1	4.1	*65.3	*80.9	23.9	27	27.8	3
May	ml5	34	33	–2.9	*169	152	*146	–13.6	–3.9	*47.6	*69	45	24.5	26.9	9.8
June	ml6	26	25	–3.8	122	118	106	–13.1	–10.2	*36.2	*57.5	58.8	24	22.5	–6.3
July	ml7	20	20.7	3.5	97.5	83	92.5	–5.1	11.4	59	51.7	–12.4	21	19.7	–6.2
August	ml8	20	17.5	–12.5	81.5	81.9	96.5	18.4	17.8	47.6	50.6	6.3	17	19.1	12.4
September	ml9	19	20.4	7.4	93	91	88.7	–4.6	–2.5	45.8	49.5	8.1	20	17.9	–10.5
October	ml10	25	21.7	–13.2	108	97	98	–9.3	1	*27	*53.5	98.1	22	20	–9.1
November	ml11	30	31	3.3	134	121	135	0.7	11.6	**51.2	**56	9.4	22	21.3	–3.2
December	ml12	37	34.8	–5.9	149	144	139	–6.7	–3.5	*37.5	*65.3	74.1	23	24.4	6.1

^*

p-value is less than (<) 0.05 using the Wilcoxon rank-sum test

^**

p-value is greater than (>) 0.05 but <0.1 using the Wilcoxon rank-sum test

Flow Duration

The duration of flow represents the period of time associated with a specific flow condition (Poff and others, 1997). For this study, the annual minimum of the 90-day moving average flow (DL5) was evaluated as an indicator of potential changes in low-flow duration at the four study sites. This is a commonly used streamflow metric with ecological relevance (Kennen and Riskin, 2010; Archfield and others, 2014, Kennen and Cuffney, 2023). The DL5 was calculated for each year within each POR, and then the median of the minimum values was used to represent the minimum of the DL5 for the POR (fig. 4; tables 5, 6, 7, and 8).

For the NB Metedeconk River, Toms River, and Cedar Creek, the minimum 90-day moving average was greater during POR1 and POR2 than during POR3. For Westecunk Creek, the median DL5 value was greater during POR3. Results of the Wilcoxon rank-sum test indicate that DL5 values were significantly lower in Toms River for POR3 (median of 105 ft³/s) than for POR1 (119 ft³/s), with a p-value of >0.05 but <0.1 (table 6). This also occurred in Cedar Creek for POR3 (58.8 ft³/s) compared to POR1 (63.5 ft³/s), with a p-value of <0.05 (table 7; fig. 4).

At all four sites, the range in DL5 values decreased between the earlier PORs and POR3 (tables 9, 10; fig. 4). For example, NB Metedeconk River DL5 values ranged from 16.4–49.8 ft³/s in POR2 depending on the year and 23.5–41.4 ft³/s in POR3. At Toms River, the range in DL5 values for POR1 was 81.3–197 ft³/s, and for POR3, the range was 82.5–139 ft³/s. Cedar Creek DL5 values ranged from 40.8–95.1 ft³/s in POR1 and 51.5–66.3 ft³/s in POR3, and at Westecunk Creek, the range in DL5 values was 14.2–33.8 ft³/s in POR2 and 16.8–26.6 ft³/s in POR3 (tables 9, 10).

Table 5.    

Summary of selected flow metrics and periods of record at North Branch Metedeconk River.

[Flow values in this table were rounded as follows: values less than (<) 1.0 are reported to the hundredths decimal place, values between 1.0 and 100 are reported to the tenth decimal place, and values greater than (>) 100 are reported to three significant figures. POR1 for NB Metedeconk River has not been included in this table because no data exist for this period of record. POR, period of record; NB, North Branch; ft3/s, cubic feet per second; 7Q10, minimum consecutive 7-day average flow with a 10-year recurrence interval]

Table 5.    Summary of selected flow metrics and periods of record at North Branch Metedeconk River.

Flow metric description (unit of measurement)	Metric abbreviation	Metric type	Tested annual time series for statistical significance	NB Metedeconk River, 01408120
				POR2	POR3	Percent change, POR2–POR3
Mean of the annual minimum daily streamflow (ft3/s)	DL1-mean	Duration, low	no	18.8	17.2	–8.5
Median of the annual daily minimum streamflow (ft3/s)	DL1-median	Duration, low	yes	17.0	15.8	–7.1
Median of the annual minimum 7-day moving average streamflows (ft3/s)	DL3	Duration, low	yes	18.8	15.9	–15.4
Median of the annual minimum 90-day moving average streamflows (ft3/s)	DL5	Duration, low	yes	29.0	25.9	–10.7
Mean of the annual minimum 7-day moving average streamflows divided by median streamflow (Dimensionless)	DL12	Duration, low	no	0.41	0.34	–17.1
75-percent exceedance flow divided by median streamflow (Dimensionless)	DL14	Duration, low	no	0.67	0.67	0.0
90-percent exceedance flow divided by median streamflow (Dimensionless)	DL15	Duration, low	no	0.48	0.48	0.0
Mean of the daily streamflows (ft3/s)	MA1	Magnitude, average	yes	60.8	64.8	6.6
Median date of the annual minimum streamflow (Julian Day)	TL1	Timing, low	yes	239	268	12.1
Skewness (Dimensionless)	tau3	Distribution extremes	no	0.40	0.45	12.5
Kurtosis (Dimensionless)	tau4	Distribution extremes	no	0.27	0.32	18.5
7Q10 low-flow frequency, fall (ft3/s)	7Q10	Frequency, low	no	16.4	17.4	6.1
7Q10 low-flow frequency, winter (ft3/s)	7Q10	Frequency, low	no	26.9	30.7	14.1
7Q10 low-flow frequency, spring (ft3/s)	7Q10	Frequency, low	no	19.8	22.5	13.6
7Q10 low-flow frequency, summer (ft3/s)	7Q10	Frequency, low	no	12.9	13.7	6.2

Table 6.    

Summary of selected flow metrics and periods of record at Toms River.

[Flow values in this table were rounded as follows: values less than (<) 1.0 are reported to the hundredths decimal place, values between 1.0 and 100 are reported to the tenth decimal place, and values greater than (>) 100 are reported to three significant figures. POR, period of record; ft³/s, cubic feet per second; 7Q10, minimum consecutive 7-day average flow with a 10-year recurrence interval]

Table 6.    Summary of selected flow metrics and periods of record at Toms River.

Flow metric description (unit of measurement)	Metric abbreviation	Metric type	Tested annual time series for statistical significance	Toms River, 01408500
				POR1	POR2	POR3	Percent change, POR1–POR2	Percent change, POR1–POR3	Percent change, POR2–POR3
Mean of the annual minimum daily streamflow (ft3/s)	DL1-mean	Duration, low	No	79.7	78.4	74.8	–1.6	–6.1	–4.6
Median of the annual daily minimum streamflow (ft3/s)	DL 1-median	Duration, low	Yes	74.5	68.2	78.4	–8.5	5.2	15.0
Median of the annual minimum 7-day moving average streamflows (ft3/s)	DL3	Duration, low	Yes	79.4	73.2	79.7	–7.8	0.4	8.9
Median of the annual minimum 90-day moving average streamflows (ft3/s)	DL5	Duration, low	Yes	**119	104	**105	–12.6	–11.8	1.0
Mean of the annual minimum 7-day moving average streamflows divided by median streamflow (Dimensionless)	DL12	Duration, low	No	0.42	0.42	0.42	0.0	0.0	0.0
75-percent exceedance flow divided by median streamflow (Dimensionless)	DL14	Duration, low	No	0.71	0.72	0.69	1.4	–2.8	–4.2
90-percent exceedance flow divided by median streamflow (Dimensionless)	DL15	Duration, low	No	0.53	0.52	0.52	–1.9	–1.9	0.0
Mean of the daily streamflows (ft3/s)	MA1	Magnitude, average	Yes	214	212	220	–0.9	2.8	3.8
Median date of the annual minimum streamflow (Julian Day)	TL1	Timing, low	Yes	250	256	264	2.4	5.6	3.1
Skewness (Dimensionless)	tau3	Distribution extremes	No	0.25	0.32	0.31	28.0	24.0	–3.1
Kurtosis (Dimensionless)	tau4	Distribution extremes	No	0.17	0.21	0.21	23.5	23.5	0.0
7Q10 low-flow frequency, fall (ft3/s)	7Q10	Frequency, low	No	81.5	70.5	77.1	–13.5	–5.4	9.4
7Q10 low-flow frequency, winter (ft3/s)	7Q10	Frequency, low	No	124	109	126	–12.1	1.6	15.6
7Q10 low-flow frequency, spring (ft3/s)	7Q10	Frequency, low	No	99.3	77.4	95.3	–22.1	–4.0	23.1
7Q10 low-flow frequency, summer (ft3/s)	7Q10	Frequency, low	No	65.9	58.1	60.0	–11.8	–9.0	3.3

^**

p-value is >0.05 but <0.1 using the Wilcoxon rank-sum test

Table 7.    

Summary of selected flow metrics and periods of record at Cedar Creek.

[Flow values in this table were rounded as follows: values less than (<) 1.0 are reported to the hundredths decimal place, values between 1.0 and 100 are reported to the tenth decimal place, and values greater than (>) 100 are reported to three significant figures. POR2 for Cedar Creek has not been included in this table because no data exists for this period of record. POR, period of record; ft³/s, cubic feet per second; 7Q10, minimum consecutive 7-day average flow with a 10-year recurrence interval]

Table 7.    Summary of selected flow metrics and periods of record at Cedar Creek.

Flow metric description (unit of measurement)	Metric abbreviation	Metric type	Tested annual time series for statistical significance	Cedar Creek, 01408900
				POR1	POR3	Percent change, POR1–POR3
Mean of the annual minimum daily streamflow (ft3/s)	DL1-mean	Duration, low	no	21.1	44.3	110.0
Median of the annual daily minimum streamflow (ft3/s)	DL1-median	Duration, low	yes	*19.2	*44.0	129.2
Median of the annual minimum 7-day moving average streamflows (ft3/s)	DL3	Duration, low	yes	42.9	46.6	8.6
Median of the annual minimum 90-day moving average streamflows (ft3/s)	DL5	Duration, low	yes	*63.5	*58.8	–7.4
Mean of the annual minimum 7-day moving average streamflows divided by median streamflow (Dimensionless)	DL12	Duration, low	no	0.49	0.58	18.4
75-percent exceedance flow divided by median streamflow (Dimensionless)	DL14	Duration, low	no	0.76	0.77	1.3
90-percent exceedance flow divided by median streamflow (Dimensionless)	DL15	Duration, low	no	0.62	0.65	4.8
Mean of the daily streamflows (ft3/s)	MA1	Magnitude, average	yes	99.8	91.4	–8.4
Median date of the annual minimum streamflow (Julian Day)	TL1	Timing, low	yes	249	255.0	2.4
Skewness (Dimensionless)	tau3	Distribution extremes	no	0.25	0.29	16.0
Kurtosis (Dimensionless)	tau4	Distribution extremes	no	0.20	0.18	–10.0
7Q10 low-flow frequency, fall (ft3/s)	7Q10	Frequency, low	no	36.2	45.7	26.2
7Q10 low-flow frequency, winter (ft3/s)	7Q10	Frequency, low	no	52.8	57.4	8.7
7Q10 low-flow frequency, spring (ft3/s)	7Q10	Frequency, low	no	51.9	50.2	–3.3
7Q10 low-flow frequency, summer (ft3/s)	7Q10	Frequency, low	no	35.7	41.5	16.2

^*

p-value is <0.05 using the Wilcoxon rank-sum test

Table 8.    

Summary of selected flow metrics and periods of record at Westecunk Creek.

[Flow values in this table were rounded as follows: values less than (<) 1.0 are reported to the hundredths decimal place, values between 1.0 and 100 are reported to the tenth decimal place, and values greater than (>) 100 are reported to three significant figures. POR1 for Westecunk Creek has not been included in this table because no data exists for this period of record. POR, period of record; ft³/s, cubic feet per second; 7Q10, minimum consecutive 7-day average flow with a 10-year recurrence interval]

Table 8.    Summary of selected flow metrics and periods of record at Westecunk Creek.

Flow metric description (unit of measurement)	Metric abbreviation	Metric type	Tested annual time series for statistical significance	Westecunk Creek, 01409280
				POR2	POR3	Percent change, POR2–POR3
Mean of the annual minimum daily streamflow (ft3/s)	DL1-mean	Duration, low	no	13.1	15.8	20.6
Median of the annual daily minimum streamflow (ft3/s)	DL1-median	Duration, low	yes	12.0	15.8	31.7
Median of the annual minimum 7-day moving average streamflows (ft3/s)	DL3	Duration, low	yes	16.0	16.9	5.6
Median of the annual minimum 90-day moving average streamflows (ft3/s)	DL5	Duration, low	yes	19.3	22.0	14.0
Mean of the annual minimum 7-day moving average streamflows divided by median streamflow (Dimensionless)	DL12	Duration, low	no	0.55	0.58	5.5
75-percent exceedance flow divided by median streamflow (Dimensionless)	DL14	Duration, low	no	0.79	0.79	0.0
90-percent exceedance flow divided by median streamflow (Dimensionless)	DL15	Duration, low	no	0.64	0.67	4.7
Mean of the daily streamflows (ft3/s)	MA1	Magnitude, average	yes	31.9	32.0	0.3
Median date of the annual minimum streamflow (Julian Day)	TL1	Timing, low	yes	253	253	0.0
Skewness (Dimensionless)	tau3	Distribution extremes	no	0.23	0.29	26.1
Kurtosis (Dimensionless)	tau4	Distribution extremes	no	0.19	0.22	15.8
7Q10 low-flow frequency, fall (ft3/s)	7Q10	Frequency, low	no	12.7	15.6	22.8
7Q10 low-flow frequency, winter (ft3/s)	7Q10	Frequency, low	no	17.3	18.9	9.2
7Q10 low-flow frequency, spring (ft3/s)	7Q10	Frequency, low	no	13.8	20.4	47.8
7Q10 low-flow frequency, summer (ft3/s)	7Q10	Frequency, low	no	11.6	12.4	6.9

Table 9.    

Range of the 90-day moving average flow (DL5) values by period of record and site for NB Metedeconk River and Toms River.

[POR1 for NB Metedeconk River has not been included in this table because no data exist for these periods of record. Min, minimum; max, maximum; POR, period of record; ft³/yr, cubic feet per year; NB, North Branch]

Table 9.    Range of the 90-day moving average flow (DL5) values by period of record and site for NB Metedeconk River and Toms River.

Metric abbreviation	Metric type	NB Metedeconk River, 01408120				Toms River, 01408500
		POR2	Year	POR3	Year	POR1	Year	POR2	Year	POR3	Year
DL5	min	16.4 ft3/yr	1981	23.5 ft3/yr	2016	81.3 ft3/yr	1957	75.5 ft3/yr	1981	82.5 ft3/yr	2008
DL5	max	49.8 ft3/yr	1978	41.4 ft3/yr	2006	197 ft3/yr	1938	204 ft3/yr	1978	139 ft3/yr	2018

Table 10.    

Range of the 90-day moving average flow (DL5) values by period of record and site for Cedar Creek and Westecunk Creek.

[POR2 for Cedar Creek and POR1 for Westecunk Creek have not been included in this table because no data exist for these periods of record. Min, minimum; max, maximum; POR, period of record; ft³/yr, cubic feet per year]

Table 10.    Range of the 90-day moving average flow (DL5) values by period of record and site for Cedar Creek and Westecunk Creek.

Metric abbreviation	Metric type	Cedar Creek, 01408900				Westecunk Creek, 01409280
		POR1	Year	POR3	Year	POR2	Year	POR3	Year
DL5	min	40.8 ft3/yr	1957	51.5 ft3/yr	2008	14.2 ft3/yr	1985	16.8 ft3/yr	2016
DL5	max	95.1 ft3/yr	1939	66.3 ft3/yr	2007	33.8 ft3/yr	1978	26.6 ft3/yr	2006

Mean Streamflow

The mean of the daily mean flow values (MA1) for each water year serves as a metric of average streamflow conditions at each of the sites. MA1 values for each site and PORs are presented in tables 5, 6, 7, and 8. Hydrographs also are presented for each of the sites and PORs to show fluctuations in daily mean streamflow over time (fig. 5). For the most recent period, POR3, overall mean streamflow was about 65 ft³/s at NB Metedeconk River (table 5), 220 ft³/s at Toms River (table 6), 91 ft³/s at Cedar Creek (table 7), and 32 ft³/s at Westecunk Creek (table 8). The mean of the daily streamflows indicated an overall increase of about 2.8 percent at the Toms River streamgage from POR1 to POR3 (table 6). The daily mean streamflow from POR2 to POR3 also indicated an increase of 3.8 percent and 6.6 percent at Toms River (table 6) and NB Metedeconk River (table 5), respectively. The Cedar Creek and Westecunk Creek streamgages showed slightly different patterns with a decrease in the daily mean streamflow of 8.4 percent from POR1 to POR3 for Cedar Creek and essentially no change in daily mean streamflow between POR2 and POR3 at Westecunk Creek. When comparing the distribution of annual MA1 values between PORs (fig. 6), no statistically significant differences in the distribution were indicated by results of the Wilcoxon rank-sum test for any of the sites.

When attempting to establish minimum-flow guidance for the protection of the local ecosystem, it may be necessary to investigate more complex statistics, but using daily streamflows is a good first step in understanding the variability of streamflow. Although statistically significant changes in the annual streamflow were not identified, utilizing a statistic like the 90-percent exceedance streamflow or the median of the August minimum flow (ML8) as an experimental threshold can highlight what percentage of the POR was below the 30-percent target for each streamgage indicated by the BBP. For example, if the median of the August minimum flow (ML8) was selected as the ecological-flow limit for the Toms River streamgage, then a proposed buffer of 30 percent above this limit would be the targeted zone where flow alterations would be reduced or paused to protect the aquatic ecosystem. The ML8 value for POR1 at Toms River was 81.5 ft³/s (table 4); adding a 30-percent buffer increases the threshold to 106 ft³/s (table 11). Applying this buffer to the approved daily streamflow record for POR1 at the Toms River streamgage results in daily streamflows below the 30-percent buffer 11.4 percent of the time (table 11). Figure 5 shows daily streamflow data, the ML8 flow as an example flow limit, as well as a 30-percent buffer line for each station and POR to illustrate the variability and extremes in the historical-streamflow record.

Table 11.    

Percentage of days that mean daily flow values were below an example threshold for each site and period of record.

[USGS, U.S. Geological Survey; ID, identifier; POR, period of record; ML8, August minimum flow; %, percent; ft³/s, cubic feet per second, N.J., New Jersey]

Table 11.    Percentage of days that mean daily flow values were below an example threshold for each site and period of record.

USGS Station Number	USGS Station Name	POR	Median August minimum flow (ML8), in ft³/s	Percentage of days the mean daily flow was below the median August minimum flow	30% buffer above the median August minimum flow, in ft³/s	Percentage of days the mean daily flow was below the 30% buffer above the median August minimum flow
01408120	North Branch Metedeconk River near Lakewood, N.J.	2, 3	20	5.7	26	15
			17.5	3.3	22.8	9.9
01408500	Toms River near Toms River, N.J.	1, 2, 3	81.5	3.5	106	11.4
			81.9	6.7	106	16.4
			96.5	9.4	125	23.5
01408900	Cedar Creek at Western Blvd near Lanoka Harbor, N.J.	1, 3	47.6	5.8	61.9	18.2
			50.6	6.3	65.8	29.5
01409280	Westecunk Creek at Stafford Forge, N.J.	2, 3	17	5.4	22.1	23.4
			19.1	9.1	24.8	33.5

Frequency

The 7Q10 was computed for each site by season and POR to investigate potential sub-annual patterns in low-flow frequency. When comparing the earliest period, POR1, to the most recent period, POR3, the results for Toms River showed a net decrease in the magnitude of 7Q10 frequency for the spring, summer, and fall months (table 6; fig. 7). Of these, the greatest percent decrease (9 percent) occurred during the summer months, which had a 7Q10 of about 66 ft³/s in POR1 and 60 ft³/s in POR3. The winter season showed a smaller net increase (<2 percent) in the 7Q10 flow at Toms River. When comparing POR1 to POR2, the 7Q10 showed a marked decrease in magnitude for all four seasons, ranging from about 12 percent to 22 percent. This was followed by increases of varying degrees across all seasons between POR2 and POR3.

When comparing POR2 to POR3, the 7Q10 indicated a net increase in magnitude for the three streams with data available for both POR2 and POR3: NB Metedeconk River, Toms River, and Westecunk Creek (fig. 7). The increase in 7Q10 magnitude for all seasons between POR2 and POR3 ranged from 3.3 percent in the summer months at Toms River to greater than 47 percent in the spring at Westecunk Creek (tables 5, 6, and 8). Lastly, Cedar Creek showed large increases in the 7Q10 between POR1 and POR3, except for the spring months, for which there was a small decrease (table 7). These metrics help to illustrate how the shape of flow distribution may be changing throughout the period of record and help to explain the complex interactions that exist between changes in the magnitude of the 7Q10 and other flow metrics such as the mean daily streamflow, the daily minimum, or the normalized 75- and 90-percent exceedance streamflow.

Analysis of Statistical Results to Inform Changes in Streamflow Conditions

Results of the selected historical analysis periods were examined separately for each site (tables 5, 6, 7, and 8) to help describe changes in streamflow patterns. Streamflow metrics⁵ were evaluated for each site. The streamgage records were not coincident for all periods; therefore, the primary focus was to understand how current streamflow conditions differ from past conditions.

The Toms River streamgage is the only study site with a complete record for all three time periods, giving insight on how the general streamflow patterns varied across the three PORs. Using the mean of the daily mean streamflow (MA1) values for each POR as a descriptor, Toms River showed a slight decrease in the average streamflow from POR1 to POR2 and a larger increase in streamflow from POR2 to POR3, with the average streamflow being greater for POR3 than for POR1. These results indicate that POR2 had the lowest flow of the three PORs. In addition to general streamflow increases, the Toms River sub-basin was identified earlier in the report as having the largest increase in development during the largest increase in urban land use between 1986 and 2015 (table 2).

For the Toms River, changes from POR1 to POR3 indicate increases in the mean of daily streamflows (MA1; fig. 6; table 6) and increases in the median of the annual daily minimum flow (DL1-median; fig. 8). However, little to no changes were identified in the minimum 7-day moving average flows (DL3; fig. 9) and the minimum 7-day moving average flows normalized to median streamflow (DL12; table 6). In contrast, the mean of the annual daily minimum flow (DL1-mean) decreased from POR1 to POR3, which indicates that the flow distribution may have changed. The pattern of an increasing MA1 and an increasing DL1-median flow contrasted with a decreasing DL1-mean may indicate more extreme daily minimum flows are occurring. Statistically significant decreases in the minimum 90-day moving average of flows (DL5) were also evident at Toms River between POR1 and POR3 (fig. 4) along with lower minimum monthly flows of about 20 ft³/s each for the months of April, May, and June (table 4). For Toms River, the dates of the annual minimum flows by POR were September 7, 13, and 21 (Julian days 250, 256, and 264, respectively), indicating a small but important shift in timing of low flows. These patterns can be further examined by reviewing the results of the seasonal low-flow frequency values for Toms River.

When comparing 7Q10 values between POR1 and POR3 at Toms River, the 7Q10 decreased between 4 percent and 9 percent for the spring, summer, and fall months, and winter low-flow frequency values had a small increase of 1.6 percent (table 6). Although these statistics continue to describe changes in low-flow conditions, in some cases simply investigating the change in magnitude for a single frequency (such as 7Q10) may not provide enough insight to inform how the streamflow is changing. If the fitted frequency curve changes in shape from slightly concave to straighter, stepped, or slightly convex, that implies a potential change to the distribution of the 7-consecutive day minimum average streamflow over time (Vogel and Fennessey, 1995). An in-depth analysis of changes in low-flow frequency curve shape and statistics are beyond the scope of this study, but an example of this type of change was identified at the Toms River streamgage. A brief analysis of the frequency curve for the summer 7Q10 low-flow statistics for POR1 and POR3 at Toms River shows a change in the curve shape (fig. 10). In POR1, the frequency curve fitted to the 7Q10 was smooth and slightly concave in shape, consistent with a slightly positive-skewed distribution. In POR3, the fitted frequency curve shows a distinct step change between two flatter ranges of ranked data values. The step change occurs between the 30- and 40-percent non-exceedance probabilities so it does not impact the 10-percent probability and associated 7Q10, but the step change provides insight about other potential changes in low-flow conditions (fig. 10). The magnitude for the 10-year recurrence interval was slightly reduced, and the change in curve shape may be a signal that other changes are occurring at lower flows within the sub-basin. Coincident with this finding, decreases of about 2 percent and 3 percent also were observed in the normalized DL14 and DL15, respectively, at Toms River from POR1 to POR3 (table 6). The low-flow frequency and exceedance results help describe the magnitude and distribution of rarer streamflows, or streamflows along the left tail of the flow-distribution curve.

Examining the change from POR2 to POR3 for Toms River showed a pattern that was generally consistent—increasing flows from the low period of POR2 to POR3. Some of the changes were seen in low-flow frequency, including increases in the 7Q10 magnitude for all seasons; these changes are being influenced by the lower flows in POR2. Reviewing the collection of streamflow metrics for Toms River indicates that, although overall annual flows are increasing, the magnitudes of statistically rarer low flows and spring and summer low flows are getting lower. These types of increased variability at the lower end of the flow distribution tend to be consistent with increased water-supply demand or other types of alteration that is potentially influenced by increased development in the sub-basin (Kennen and others, 2014; Singh and Basu, 2022).

Results that compared POR2 to POR3 in the NB Metedeconk River followed a similar pattern to that seen in the Toms River for MA1 (tables 5 and 6). NB Metedeconk River also showed decreases in DL1, DL3, and DL5, and no change in the normalized 75- and 90-percent exceedance streamflows. The 7Q10 low-flow frequency metrics all showed increases for each season, indicating that streamflow throughout POR2 was lower. Collectively, results show a decrease in DL1, DL3 and DL5 when the 7Q10 magnitude was increasing from POR2 to POR3. These results may indicate multiple alterations occurred during lower-flow conditions that may require a more advanced analysis to better explain the changes. It is possible that increased downward spikes in low flows are meaningful enough to alter minimum flows, but other mechanisms (such as anthropogenic effects or shifts in precipitation) are altering slightly higher flows (Freeman and others, 2007; Falcone and others 2010; Rice and others 2015; Singh and Basu, 2022), increasing the overall skew of the minimum 7-consecutive day minimum streamflow, therefore changing the frequency curve that resulted in a slightly higher 7Q10 magnitude. Monthly minimums showed a mix of increases and decreases between POR2 and POR3; the greatest percent changes (>10 percent) were decreases and occurred during the months of February, August, and October. Lastly, for NB Metedeconk River, the median dates of the annual minimum flows for POR2 and POR3 were August 27 and September 25 (Julian days 239 and 268, respectively). The nearly 4-week difference was not statistically significant (p-value >0.1); however, it represented the greatest change in timing of minimum flows among the four sub-basins with a shift toward the fall season (fig. 11; tables 5, 6, 7, and 8).

Cedar Creek and Westecunk Creek streamgages are in the southern and less developed part of the Barnegat Bay-Little Egg Harbor watershed (fig. 2). Data were not available for all 3 PORs at these two streamgages, so comparisons were made for POR1 to POR3 and POR2 to POR3, respectively. The mean of the daily streamflow (MA1) was used again as a simple measure of the general conditions during each analysis period. In general, as the amount of human water-use demand goes up, streamflows tend to decrease concomitantly (Wada and others, 2013). As mentioned previously, results in NB Metedeconk River, Toms River, and Westecunk Creek sub-basins indicated that streamflow has increased or remained relatively flat over time (tables 5, 6, and 8). The Westecunk Creek streamgage showed a <1-percent increase in mean daily streamflow from POR2 to POR3 and there is a 3.8- to 6.6-percent increase in mean daily streamflow from POR2 to POR3 at Toms River and NB Metedeconk River. Conversely, there was an 8.4-percent decrease in MA1 in Cedar Creek between POR1 and POR3 (table 7).

For Cedar Creek, the mean and median of the annual daily minimums (DL1-mean and DL1-median), and the annual minimum 7-day moving average of flows (DL3) all increased from POR1 to POR3 (table 7). Both DL1 metrics more than doubled between POR1 and POR3. The monthly minimums increased for all months between POR1 and POR3, except for July, with notable increases occurring in the months of April, May, June, October, and December (23.9, 45.0, 58.8, 98.1, and 74.1 percent, respectively); the median of the minimum monthly flows increased by 15.6–28.0 ft³/s for these months (table 4). The large percent increase in DL1 and monthly minimums is likely related to gage location and the occurrence of low values year-round during POR1, as previously described in the “monthly minima” section of this report. The timing of the annual minimum flow (TL1) for Cedar Creek shifted toward later in the year, similar to NB Metedeconk and Toms River, albeit slight. The dates of the annual minimum flows for POR1 and POR3 were September 6 and September 12 (Julian days 249 and 255, respectively, table 7). In addition, DL12, DL14, and DL15 each increased between POR1 and POR3; the only duration metric that decreased at Cedar Creek was the 90-day moving average (DL5; 7.4 percent decrease). Low-flow frequency results at Cedar Creek also generally indicate increasing magnitudes for all the seasonal 7Q10s, with the exception of the spring 7Q10, which had a 3.3-percent decrease. In the case of Cedar Creek, results at the low end of the flow range may be influenced by improved flow monitoring conditions resulting from the relocation of the streamgage and may not be solely a result of differences between PORs. The streamflow past the gage at the previous location was occasionally affected by tide, resulting in short temporary periods of reduced or zero flow over the weir. The resulting daily flow should include the combined effect of water going into temporary storage and later being released downstream; however, it is possible that during low-flow periods, the methods for adjusting the computed streamflow introduced uncertainty, so the gage was moved upstream in 2012 and away from the influence of daily tide.

Similar to NB Metedeconk and Toms River, the seasonal 7Q10 for Westecunk Creek was greater in POR3 than in POR2 for all four seasons (table 8). Westecunk Creek exhibited some of the largest percent increases in seasonal 7Q10 values during these two time periods (up to 47.8 percent during spring). Summer 7Q10 values showed among the smallest percent increases for each of the study sites. Unlike the NB Metedeconk and the Toms River, which had decreases in one or more of the n-day duration metrics (DL1, DL3, and DL5) as well as DL12 between POR2 and POR3, Westecunk Creek exhibited an increase for each of the metrics, ranging from a 5.5 percent increase for DL12 to 31.7 percent increase for DL1-median. Monthly minimum values showed a mix of increases and decreases between POR2 and POR3, with the greatest percent increases occurring during the months of February, May, and August (12.4, 9.8, and 17.0 percent, respectively), and the greatest percent decreases occurring in September and October (10.5 and 9.1 percent, respectively, table 4). The timing of the annual minimum flow (TL1) for Westecunk Creek did not change between POR2 and POR3, with September 10th (Julian day 253) as the date of the annual minimum flow for both PORs (table 8).

Given that there has been no increase in development in the Westecunk Creek sub-basin, and LU/LC is predominantly forest and wetlands (table 2), this sub-basin may be viewed as a sub-basin with reference-like (relatively undisturbed) flow conditions in the study area. This sub-basin generally showed increases in flow metrics between POR2 and POR3. In contrast, the NB Metedeconk River showed a decrease in every low-flow duration metric that was evaluated (table 5) despite an increase in mean annual streamflow, and results of analysis of streamflow metrics in the Toms River indicated that some low-flow metrics increased while others decreased (table 6).

Skewness and kurtosis were computed from the daily streamflows for the four streamgages used in this study and evaluated for deviation from what would be expected for coastal streams (Armstrong and others, 2008). The skewness remained relatively close to zero (that is, a slightly negative skew) for all periods, and while no statistically significant changes were found, some patterns were identified. For example, the change in skewness from POR1 to POR3 for Toms River and Cedar Creek showed increases of 24 percent and 16 percent while changes for POR2 to POR3 at NB Metedeconk River and Westecunk Creek also showed increases of about 12 percent and 26 percent. The most recent period of POR2 to POR3 at the Toms River streamgage was the only skewness value to show a small decrease. Kurtosis results were a little more mixed, with generally increasing changes for most periods except POR1 to POR3 at Cedar Creek (a 10-percent decrease).