A revised logistic regression equation and an automated procedure for mapping the probability of a stream flowing perennially in Massachusetts

A revised logistic regression equation and an automated procedure were developed for mapping the probability of a stream flowing perennially in Massachusetts. The equation provides city and town conservation commissions and the Massachusetts Department of Environmental Protection a method for assessing whether streams are intermittent or perennial at a specific site in Massachusetts by estimating the probability of a stream flowing perennially at that site. This information could assist the environmental agencies who administer the Commonwealth of Massachusetts Rivers Protection Act of 1996, which establishes a 200-foot-wide protected riverfront area extending from the mean annual high-water line along each side of a perennial stream, with exceptions for some urban areas. The equation was developed by relating the observed intermittent or perennial status of a stream site to selected basin characteristics of naturally flowing streams (defined as having no regulation by dams, surface-water withdrawals, ground-water withdrawals, diversion, wastewater discharge, and so forth) in Massachusetts. This revised equation differs from the equation developed in a previous U.S. Geological Survey study in that it is solely based on visual observations of the intermittent or perennial status of stream sites across Massachusetts and on the evaluation of several additional basin and land-use characteristics as potential explanatory variables in the logistic regression analysis. The revised equation estimated more accurately the intermittent or perennial status of the observed stream sites than the equation from the previous study. Stream sites used in the analysis were identified as intermittent or perennial based on visual observation during low-flow periods from late July through early September 2001. The database of intermittent and perennial streams included a total of 351 naturally flowing (no regulation) sites, of which 85 were observed to be intermittent and 266 perennial. Stream sites included in the database had drainage areas that ranged from 0.04 to 10.96 square miles. Of the 66 stream sites with drainage areas greater than 2.00 square miles, 2 sites were intermittent and 64 sites were perennial. Thus, stream sites with drainage areas greater than 2.00 square miles were assumed to flow perennially, and the database used to develop the logistic regression equation included only those stream sites with drainage areas less than 2.00 square miles. The database for the equation included 285 stream sites that had drainage areas less than 2.00 square miles, of which 83 sites were intermittent and 202 sites were perennial. Results of the logistic regression analysis indicate that the probability of a stream flowing perennially at a specific site in Massachusetts can be estimated as a function of four explanatory variables: (1) drainage area (natural logarithm), (2) areal percentage of sand and gravel deposits, (3) areal percentage of forest land, and (4) region of the state (eastern region or western region). Although the equation provides an objective means of determining the probability of a stream flowing perennially at a specific site, the reliability of the equation is constrained by the data used in its development. The equation is not recommended for (1) losing stream reaches or (2) streams whose ground-water contributing areas do not coincide with their surface-water drainage areas, such as many streams draining the Southeast Coastal Region—the southern part of the South Coastal Basin, the eastern part of the Buzzards Bay Basin, and the entire area of the Cape Cod and the Islands Basins. If the equation were used on a regulated stream site, the estimated intermittent or perennial status would reflect the natural flow conditions for that site. An automated mapping procedure was developed to determine the intermittent or perennial status of stream sites along reaches throughout a basin. The procedure delineates the drainage area boundaries, determines values for the four explanatory variables, and solves the equation for estimating the probability of a stream flowing perennially at two locations on a headwater (first-order) stream reach—one near its confluence or end point and one near its headwaters or start point. The automated procedure then determines the intermittent or perennial status of the reach on the basis of the calculated probability values and a probability cutpoint (a stream is considered to flow perennially at a cutpoint of 0.56 or greater for this study) for the two locations or continues to loop upstream or downstream between locations less than and greater than the cutpoint of 0.56 to determine the transition point from an A Revised Logistic Regression Equation and an Automated Procedure for Mapping the Probability of a Stream Flowing Perennially in Massachusetts By Gardner C. Bent and Peter A. Steeves intermittent to a perennial stream. If the first-order stream reach is determined to be intermittent, the procedure moves to the next downstream reach and repeats the same process. The automated procedure then moves to the next first-order stream and repeats the process until the entire basin is mapped. A map of the intermittent and perennial stream reaches in the Shawsheen River Basin is provided on a CD-ROM that accompanies this report. The CD-ROM also contains ArcReader 9.0, a freeware product, that allows a user to zoom in and out, set a scale, pan, turn on and off map layers (such as a USGS topographic map), and print a map of the stream site with a scale bar. Maps of the intermittent and perennial stream reaches in Massachusetts will provide city and town conservation commissions and the Massachusetts Department of Environmental Protection with an additional method for assessing the intermittent or perennial status of stream sites. Introduction The Commonwealth of Massachusetts Rivers Protection Act, Chapter 258 of the Acts of 1996 (The Commonwealth of Massachusetts, 1996), specifies that riverfront areas be protected on all rivers that flow perennially. The riverfront area is defined in 310 Code of Massachusetts Regulations (CMR) 10.58(2)(a) (hereafter referred to as the Regulations) (Massachusetts Department of Environmental Protection, 2002a, p. 393–402) as the 200-ft-wide area extending along the length of each side of perennial streams from the mean annual highwater line (determined from bankfull field indicators) on each side of perennial streams. Exceptions to the Regulations are provided for some urban areas. Streams that do not flow year round, intermittent streams, have no jurisdictional riverfront area along the stream. City or town conservation commissions are charged with administering the Regulations by determining the intermittent or perennial status of a stream site and by regulating work in the riverfront areas. The Massachusetts Department of Environmental Protection (MDEP) addresses appeals of decisions made by city or town conservation commissions concerning the intermittent or perennial status of stream sites. The logistic regression equation provides these agencies with an additional method for assessing the status of stream sites in Massachusetts. The Regulations define a river as any natural flowing body of water that discharges into an ocean, lake, pond, or another river, and which flows throughout the year (Massachusetts Department of Environmental Protection, 2002a, p. 394). By this definition, perennial streams are rivers, but intermittent streams are not. When an intermittent stream is not flowing, surface water may be present in isolated pools or be absent. Rivers start at the point where an intermittent stream becomes perennial or at the point where a stream flows perennially from a spring, pond, or lake. The revised Regulations of December 20, 2002 (Massachusetts Department of Environmental Protection, 2002b, p. 317–320; Massachusetts Department of Environmental Protection, 2002a, p. 394–395), specify that U.S. Geological Survey (USGS) topographic maps, or more recent maps provided by MDEP, will continue to be used for initial review of the intermittent or perennial status of a stream. Streams depicted as perennial on USGS topographic maps or more recent maps provided by MDEP will be classified as perennial. A stream site depicted as perennial, however, can be reclassified as intermittent with direct observations of no flow during any four days of any consecutive 12-month period. These observations cannot be made during a period of extended drought or on a stream measurably affected by withdrawals, impoundments, or other anthropogenic flow reductions or diversions. The definition of “extended drought” was amended to include the time periods during which the Massachusetts Drought Management Task Force declared an index level of “advisory, watch, warning, or emergency” (Massachusetts Executive Office of Environmental Affairs and Massachusetts Emergency Management Agency, 2001). The revised Regulations state that streams depicted as intermittent, or those not shown as a stream on USGS topographic maps or more recent maps provided by MDEP, will be mainly classified on the basis of the drainage area size upstream from the stream site. If an intermittent stream site’s upstream drainage area is greater than or equal to 1.00 mi, the stream site will be classified as perennial. If an intermittent stream site’s upstream drainage area is less than 0.50 mi, the stream site will be classified as intermittent. If an intermittent stream site’s upstream drainage area is greater than or equal to 0.50 mi and less than 1.00 mi, the stream site will be classified intermittent, with two exceptions. First, if the 99-percent flow duration estimated from low-flow statistics regression equations by the World Wide Web application STREAMSTATS (Ries and others, 2000) at the stream site is greater than or equal to 0.01 ft/s, then the stream will be classified as perennial. Second, if the st

[1]  D. Helsel,et al.  Statistical methods in water resources , 2020, Techniques and Methods.

[2]  Dennis R. Helsel,et al.  Risk of Nitrate in Groundwaters of the United StatesA National Perspective , 1997 .

[3]  B. Nolan,et al.  Relating Nitrogen Sources and Aquifer Susceptibility to Nitrate in Shallow Ground Waters of the United States , 2001, Ground water.

[4]  P. Stackelberg,et al.  Relation of Ground-Water Quality to Land Use on Long Island, New York , 1995 .

[5]  Kernell G. Ries Estimation of low-flow duration discharges in Massachusetts , 1993 .

[6]  M. Adams,et al.  Long-term impacts of forest treatments on water yield: a summary for northeastern USA , 1993 .

[7]  Gardner C. Bent,et al.  Effects of forest-management activities on runoff components and ground-water recharge to Quabbin Reservoir, central Massachusetts , 2001 .

[8]  M. Acreman,et al.  The role of wetlands in the hydrological cycle , 2003 .

[9]  Michael G. Rupert,et al.  Probability of detecting atrazine/desethyl-atrazine and elevated concentrations of nitrate (NO 2 +NO 3 -N) in ground water in the Idaho part of the upper Snake River basin , 1998 .

[10]  B. Ruddy,et al.  Probability of nitrate contamination of recently recharged groundwaters in the conterminous United States. , 2002, Environmental science & technology.

[11]  J. Shanley,et al.  Effects of selective forest clearing fertilization, and liming on the hydrology and water quality of a small tributary to the Quabbin Reservoir, central Massachusetts , 1995 .

[12]  FACTORS RELATED TO THE JOINT PROBABILITY OF FLOODING ON PAIRED STREAMS , 1998 .

[13]  F. Voss,et al.  Predicting the Probability of Elevated Nitrate Concentrations in the Puget Sound Basin: Implications for Aquifer Susceptibility and Vulnerability , 1997 .

[14]  Seth Rose,et al.  Effects of urbanization on streamflow in the Atlanta area (Georgia, USA): a comparative hydrological approach , 2001 .

[15]  R. Flynn Development of regression equations to estimate flow durations and low-flow-frequency statistics in New Hampshire streams , 2003 .

[16]  Flow origin, drainage area, and hydrologic characteristics for headwater streams in the mountaintop coal-mining region of Southern West Virginia, 2000-01 , 2003 .

[17]  T. C. Winter,et al.  Ground Water and Surface Water: A Single Resource , 1999 .

[18]  W. B. Langbein,et al.  General introduction and hydrologic definitions , 1960 .

[19]  Kernell G. Ries August median streamflows in Massachusetts , 1997 .

[20]  R. Kolka,et al.  Defining perennial, intermittent and ephemeral channels in eastern Kentucky: application to forestry best management practices , 2005 .

[21]  J. D. Kliever Low-Flow Characteristics of Selected Streams in Northern Rhode Island , 1995 .

[22]  W. Battaglin,et al.  Logistic model of nitrate in streams of the upper-Midwestern United States , 1997 .

[23]  Luna Bergere Leopold,et al.  The View from the River: , 1994 .

[24]  C. Eagar,et al.  Summary of water yield eperiments at Hubbard Brook Eperimental Forest, New Hampshire , 1997 .

[25]  Gardner C. Bent,et al.  A logistic regression equation for estimating the probability of a stream flowing perennially in Massachusetts , 2002 .

[26]  William A. Battaglin,et al.  Estimating the Susceptibility of Surface Water in Texas to Nonpoint-Source Contamination by Use of Logistic Regression Modeling , 2003 .

[27]  R. Clawges,et al.  Volatile Organic Compounds in Untreated Ambient Groundwater of the United States, 1985-1995 , 1999 .

[28]  D. L. Simmons,et al.  Base flow of 10 south-shore streams, Long Island, New York, 1976-85, and the effects of urbanization on base flow and flow duration , 1992 .

[29]  Kernell G. Ries Development and application of generalized-least-squares regression models to estimate low-flow duration discharges in Massachusetts , 1994 .

[30]  Eric R. Ziegel,et al.  Logistic Regression Examples Using the SAS System , 1996 .

[31]  C. Jackson,et al.  AVERAGE DISCHARGE, PERENNIAL FLOW INITIATION, AND CHANNEL INITIATION ‐ SMALL SOUTHERN APPALACHIAN BASINS 1 , 2004 .

[32]  E. T. Oaksford,et al.  Relation between land use and ground-water quality in the upper glacial aquifer in Nassau and Suffolk Counties, Long Island, New York , 1989 .

[33]  Michael G. Rupert,et al.  Probability of detecting atrazine/desethyl-atrazine and elevated concentrations of nitrate in ground water in Colorado , 2003 .