Elsevier

Journal of Hydrology

Volume 311, Issues 1–4, 15 September 2005, Pages 266-281
Journal of Hydrology

Effects of suburban development on runoff generation in the Croton River basin, New York, USA

https://doi.org/10.1016/j.jhydrol.2005.01.022Get rights and content

Abstract

The effects of impervious area, septic leach-field effluent, and a riparian wetland on runoff generation were studied in three small (0.38–0.56 km2) headwater catchments that represent a range of suburban development (high density residential, medium density residential, and undeveloped) within the Croton River basin, 70 km north of New York City. Precipitation, stream discharge, and groundwater levels were monitored at 10–30 min intervals for 1 year, and stream water and groundwater samples were collected biweekly for δ18O, NO3, and SO42− analysis for more than 2 years during an overlapping period in 2000–2002. Data from 27 storms confirmed that peak magnitudes increased and recession time decreased with increasing development, but lags in peak arrival and peak discharge/mean discharge were greatest in the medium density residential catchment, which contains a wetland in which storm runoff is retained before entering the stream. Baseflow during a dry period from Aug. 2001–Feb. 2002 was greatest in the high-density residential catchment, presumably from the discharge of septic effluent through the shallow groundwater system and into the stream. In contrast, moderate flows during a wet period from Mar.–Aug. 2002 were greatest in the undeveloped catchment, possibly as a result of greater subsurface storage or greater hydraulic conductivity at this site. The mean residence time of baseflow was about 30 weeks at all three catchments, indicating that human influence was insufficient to greatly affect the groundwater recharge and discharge properties that determine catchment residence time. These results suggest that while suburban development and its associated impervious surfaces and storm drains accelerate the transport of storm runoff into streams, the combined effects of remnant natural landscape features such as wetlands and human alterations such as deep groundwater supply and septic systems can change the expected effects of human development on storm runoff and groundwater recharge.

Introduction

Urbanization is a pervasive global trend. Nearly, half the world's population now resides in urban areas, and that percentage is expected to increase to 60% by the year 2030 (McGee, 2001). In the United States, 80% of the population now lives in urban metropolitan areas (US Census Bureau, 1999), and a dominant demographic trend is the growth of suburban areas into previously undisturbed forests, shrublands, and deserts (Katz and Bradley, 1999). Maintaining an adequate water supply and protecting water quality in suburban areas are growing problems whose solutions will require extensive effort and research.

The effects of suburban development has been characterized in several studies; increased flood frequencies in areas with impervious surfaces were reported in the late 1960s and early 1970s (Leopold, 1968, Seaburn, 1969, Anderson, 1970). More recent studies have focused on the effects of engineered aspects of catchments, (e.g. detention basins, riparian buffers and septic systems) on runoff volume and water quality (Robertson et al., 1991, Griffin, 1995, Chin and Gregory, 2001, Booth et al., 2002). The effects of suburban development on runoff characteristics are widely acknowledged to include (relative to the undisturbed pre-development condition): (1) decreased low flow and groundwater recharge, (2) increased surface runoff in annual streamflow, (3) increased magnitude of peak runoff, (4) decreased lag time between rainfall and runoff response, (5) increased rate of hydrograph rise and recession, and (6) decreased mean residence time of streamflow (Hirsch et al., 1990, McCuen, 1998, Rose and Peters, 2001). Most process-level studies have quantitatively documented these effects in suburban catchments in which impervious surfaces represent a large percentage of the total drainage area; but additional studies are needed that compare these effects in catchments with moderate suburban development to those in undeveloped catchments.

Impervious areas such as paved roads and roofs increase the rate of surface water runoff through storm sewers resulting in decreased groundwater recharge. Yet some suburban landscape features, such as lawns, parks, golf courses, and woodlands provide groundwater recharge rates similar to those that existed prior to development (Lerner, 2002). Another feature in many suburban areas is domestic septic systems that discharge to shallow groundwater, whereas other areas have sanitary sewers that transport treated domestic wastewater directly to surface waters (Hirsch et al., 1990).

The quantity and quality of surface runoff are of great concern in the Croton River Watershed of southeastern New York, a water supply area for New York City. This region has experienced extensive suburban development during the past 50 years resulting in large increases in impervious area. However, wooded and undeveloped land remains, and impervious area, generally, does not exceed 15% of the total watershed area (Center for Watershed Protection, 2001). Runoff processes within this region likely retain some characteristics from the era prior to European settlement when forest and wetland covered nearly the entire landscape (Schueler, 1987), however, more than 80,000 domestic septic systems in the watershed could potentially increase groundwater recharge and baseflow through discharge from leach fields (Heisig, 2000, Sherlock et al., 2002). Little is known about the net effect of these suburban features on baseflow, groundwater recharge, and stormflow generation in suburban settings that represent a broad range of development intensities.

This paper presents results of a study of the effects of suburban development on baseflow and runoff processes in three small catchments of similar size, geomorphology, and physiographic characteristics in the Croton River basin (Fig. 1). The three catchments represent a gradient of suburban conditions from forested (undeveloped) to medium and high density residential development. The study entailed (1) measurement of rainfall amount, stream discharge, and groundwater levels at wells within each catchment, and (2) calculation of mean residence time of stream water from 18O measurements in precipitation and baseflow from each of the three catchments. Our working hypothesis was that suburban development and its associated impermeable surfaces would increase runoff peaks and accelerate the hydrograph rise and recession during stormflow events, and also decrease the groundwater recharge rate and mean residence time.

The 971-km2 Croton River basin in southern New York State (Fig. 1) consists of 12 reservoirs that supply 492 million L of water per day to New York City and upstate communities equivalent to about 10% of the City's water supply (Galusha, 2002). The Croton basin encompasses parts of Dutchess, Putnam and Westchester Counties in New York, and part of the State of Connecticut. The basin had a total population of 189,912 in 2000 (Moffett et al., 2003).

The Croton basin is largely underlain by Precambrian sedimentary and igneous rock of the New England Upland province; elevations range from 200 to 500 m above sea level. Soils are developed on glacial till and are medium to moderately textured and generally well drained. The Croton basin is 56.7% forested, 25.0% residential land, 7.4% agricultural land, 4.1% commercial land, 5.7% lakes and reservoirs, and 0.8% undeveloped land (Linsey et al., 1999). Mean annual precipitation is 1299 mm, and mean annual temperature is 9.9 °C at Yorktown Heights, New York in the southern part of the Croton basin at an elevation of 204 m (1971–2000 mean; Northeast Regional Climate Center; climod.nrcc.cornell.edu). During the principal winter of the study, 2001–2002, total snowfall was only 368 mm, compared to a 30-year mean of 960 mm at Yorktown Heights. The dry winter with a mean temperature that was 3.4 °C above normal at Yorktown Heights combined to provide little snowmelt to streams in the late winter/early spring of 2002.

Each of the three catchments selected for study represents a different degree of development (Linsey et al., 1999). One is undeveloped and has second growth forest cover (UND), and the other two are dominated by suburban residential development (Fig. 2). A US National or global standard does not exist for classifying urban or residential land use based on population or housing density (Hitt, 1994). Both of the developed catchments in this study would be classified as ‘high density residential’ according to criteria developed by the US Geological Survey's National Water-Quality Assessment (Hitt, 1994), however, we have classified these catchments as medium density residential (MED, 1.6 houses/ha) and high-density residential (HIGH, 2.8 houses/ha) to distinguish them. Pertinent characteristics of the three catchments are given in Table 1.

The developed catchments consist primarily of single family detached homes that are supplied by local groundwater. All houses in the HIGH catchment and about one-third of those in the MED catchment have individual wells, but the other two-thirds of the homes in MED are supplied by four nearby wells that pump and store water in an above-ground tank for later distribution. All wells in this area are cased through unconsolidated till or alluvium, are finished in fractured bedrock, and have an average depth of 120 m (Linsey et al., 1999). All houses in the two developed catchments have septic systems with leach fields. Runoff from roads and other impervious surfaces flows through storm drains to culverts that empty into the study streams.

The locations of residential areas relative to the storm-drain network and stream in HIGH are different from those in MED. The residences form a grid-like standard housing layout (Arnold and Gibbons, 1996) throughout the HIGH catchment (Fig. 2A), whereas residences in the MED catchment form a cluster-housing layout in the upper northern part that borders a headwater wetland to the south through which stormwater flows before entering the stream (Fig. 2B). The regular housing layout at HIGH, with a row of properties in direct contact with the principal stream, implies more direct delivery of septic wastewater from leach fields to the shallow groundwater system and to the stream, whereas wastewater from leach fields at MED discharges water through the headwater wetland, which flows into a stream at the lower end of the catchment.

Section snippets

Field monitoring and data collection

All three streams were sampled weekly or biweekly during baseflow conditions (at least three rain-free days prior to sampling) for chemistry and isotope analyses from Mar. 2000–Aug. 2002; all other hydrological and meteorological measurements occurred from Aug. 2001–Aug. 2002. Air temperature was measured by an automated system in each catchment that provided mean values every 10 min, and precipitation amount was summed over the same 10-min interval. Data are analyzed for a dry (Aug. 2001–Feb.

Results and discussion

The following section addresses the effects of: (1) impervious area on hydrograph peaks and recessions, (2) a wetland in the MED catchment on hydrograph peak lags and shape, and (3) septic discharge in the HIGH catchment on low flow. We also discuss runoff differences among the catchments during a wet period and the residence time of baseflow in each catchment.

Summary and conclusions

The effects of impervious area, a headwater wetland, and septic leach field discharge on stormflow runoff, baseflow generation, and groundwater recharge/discharge were compared on an episodic and seasonal basis in three catchments representing a gradient from undeveloped to high-density residential suburban development. The results indicate that some aspects of stream runoff appear to be affected by human development whereas other aspects appear unaffected. During storms, peak flows generally

Acknowledgements

We thank the following colleagues for field and technical support: Steven Wolosoff, Debra Curry, Andrew Holloway, Albert Zumbuhl, Christopher Somerlot, Myriam Adam, David Lyons, Jun Wang, Nina Lee, and Donald Cuomo. Thanks to Theodore Endreny and James Shanley for reviewing an early version of this manuscript. The project was funded by the New York City Department of Environmental Protection and cooperative funds from the US Geological Survey provided to the first author.

References (36)

  • A. Chin et al.

    Urbanization and adjustment of ephemeral stream channels

    Ann. Assoc. Am. Geogr.

    (2001)
  • D. Galusha

    Liquid Assets: a History of New York City's Water System

    (2002)
  • C.B. Griffin

    Uncertainty analysis of BMP effectiveness for controlling nitrogen from urban nonpoint sources

    Water Res. Bull.

    (1995)
  • Heisig, P., 2000. Effects of residential and agricultural land uses on the chemical quality of small streams in the...
  • R.M. Hirsch et al.

    The influence of man on hydrologic systems

  • Hitt, K.J., 1994. Refining 1970's land-use data with 1990 population data to indicate new residential development. US...
  • B. Katz et al.

    Divided we sprawl

    Atlantic Monthly

    (1999)
  • R. Lamb et al.

    Using interactive recession curve analysis to specify a general catchment storage model

    Hydrol. Earth Sys. Sci.

    (1997)
  • Cited by (227)

    View all citing articles on Scopus
    View full text