National Water-Quality Assessment (NAWQA) Project
Frequently Asked Questions
Who conducted the study?
This study, the Effects of Urbanization on Stream Ecosystems (EUSE), was conducted by scientists from the U.S. Geological Survey (USGS) as part of the National Water-Quality Assessment (NAWQA) Program.
What was the purpose of the study?
The primary objectives of the EUSE study were to determine: (1) the magnitudes and patterns of physical, chemical, and biological responses of stream ecosystems to increasing urban development; (2) the manner in which these responses vary across the country; and (3) the response of physical, chemical, and biological processes as they relate to increasing urban development. Results from the EUSE study will contribute to our understanding of the stresses and adverse effects of urban development on stream ecosystems and how to minimize these stresses for healthier urban streams.
Why did the USGS initiate the study?
The USGS NAWQA Program is designed to answer: "What is the quality of our Nation's streams and ground water? How are conditions changing over time? How do natural features and human activities affect the quality of streams and ground water, and where are those effects most pronounced?" The EUSE study was initiated because there is still an incomplete understanding of the biological, physical, and chemical processes occurring in streams as watersheds undergo urban development, and this information is critical for proper protection and management of our Nation's water resources. The USGS study was the first to use a consistent National design (gradient) in different metropolitan areas, along with nationally-consistent methods for sampling and analysis, which allowed for a national comparison.
Where was the study done?
The study investigated streams in watersheds that represented varying levels of urban development in 9 urbanizing regions across the U.S., each of which was associated with a major metropolitan city: Portland, Oregon; Salt Lake City, Utah; Birmingham, Alabama; Atlanta, Georgia; Raleigh, North Carolina; Boston, Massachusetts; Denver, Colorado; Dallas, Texas; and Milwaukee, Wisconsin.
When did the study take place?
The study was completed during 1999-2004. Pilot studies in Salt Lake City, Birmingham, and Boston studies were completed in 1999-2000. For full studies, Atlanta, Raleigh, and Denver were conducted in 2002-2003; Portland, Dallas, and Milwaukee were completed in 2003-2004.
How did the individual study areas fit into a national design?
Data collected in each of the nine study areas followed a nationally-consistent approach and used uniform methods of sampling and analysis. In addition, data in each study area was collected at selected spatial scales—stream, watershed, and region—that were consistent across the Nation. Assessments also included modeling and identifying relations among the biological, physical, and chemical components of stream ecosystems that can be useful for estimating effects of urban development in unmonitored but similar watersheds within each region. This national distribution of study areas and the study design enabled the individual studies to be placed in a national context.
How were streams chosen for the study?
In each of the nine study areas, streams in about 30 similarly-sized watersheds were selected that represented a gradient from mostly urban land to mostly forested or agricultural land. The gradient approach was intended to provide an estimate of changes that might occur over many years by substituting space for time. This assumes that changes over time at a site, for example those due to the effects of urban development, will be similar to spatial trends found among sites with varying amounts of urban land. As much as was possible, each stream was selected to minimize variability of natural factors (such as soils, climate, elevation and slope) within each metropolitan study area. This was done by use of U.S. Environmental Protection Agency Level III ecoregions, which delineate regions representing a relatively homogeneous combination of climatic, topographic, geologic, and general land use/land cover characteristics. Watersheds studied in each metropolitan area were restricted to a single Level III ecoregion, and smaller Level IV regions were used to further constrain the selection of study watersheds in several metropolitan areas. Finally, preference was given to sites near active USGS gages.
How does the NAWQA design and sampling-site selection differ from a probabilistic-monitoring design?
NAWQA uses a "targeted" monitoring design in which sites are not selected randomly within a grid, but represent certain human activities, land uses, environmental settings, or hydrologic conditions during different seasons or times of year (see question above). Such monitoring is useful to answer questions related to water-quality conditions and the natural and human factors that cause those conditions. This approach is different than probabilistic monitoring, which involves random selection of sites across a certain geographic area and sampling each site once during seasonal (spring or summer) low-flow conditions. Such monitoring is useful for getting an unbiased, broad geographic snapshot of "whether or not there is a problem" and "how big the problem is." Probabilistic monitoring and more "targeted" monitoring (such as by NAWQA) answer different types of questions and provide different types of information that are both critical for understanding the ambient resource. Ideally, data-collection and laboratory analytical methods should be consistent and comparable between the two monitoring designs so that findings can be integrated and conclusions can go beyond what each can provide individually.
Why are effects of urban development on stream ecosystems a concern?
Urbanization adversely affects stream ecosystems, and evidence is mounting that even low levels of urbanization can cause significant changes in stream biology, physical habitat, and chemistry. Healthy or natural streams provide many important benefits to society. For example, stream ecosystems capture and reduce nutrients, such as nitrogen and phosphorus, which might otherwise result in nuisance algal blooms. Diverse physical habitat, such as in the channel, stream banks, and stream bottom, is essential for healthy stream ecosystems. These physical characteristics affect shelter, food, and reproduction for all forms of organisms (biota), some of which provide shelter, food, and other materials needed by people. Undeveloped land in floodplains and near-stream or riparian areas can reduce erosion and help mitigate flooding. Well vegetated riparian areas around an urban stream provide recreational opportunities and beauty for people and add to their quality of life.
How has urban development changed over time?
Today, much more land is developed per person (for housing, schools, shopping, manufacturing, roads, and other land uses) than was the case 20 years ago, as suburbs spread out from dense city cores (Markham and Steinzor, 2006). The pattern of urban development in any watershed reflects the culmination of choices made by individuals, businesses, lending institutions, and governments about where, when, and how development occurs, such as housing, stores, businesses, and roads. The greater the suburban sprawl, the greater the need for support infrastructure to connect people to businesses, utilities and other services, and their work. Fortunately, planners in some metropolitan areas are trying to change the way land is developed so that land is used more efficiently, thereby minimizing sprawl.
What is impervious land cover?
With urban development, natural vegetation and topsoil are replaced with impervious cover (rooftops and pavement, such as roads, parking lots, and driveways) that block the infiltration of water from precipitation (rain and snow) into the ground. In addition, stream channels are sometimes straightened and lined with cement, an impervious cover which prevents erosion but eliminates once diverse physical habitat and connections to groundwater recharge for streams.
How can impervious cover on the land surface change the physical and chemical characteristics of a stream?
Large increases in impervious cover, together with increases in storm drains and cement-lined channels, can result in increases in the speed and amount of water flowing to streams. This change, combined with pollutants, such as sediment, nutrients, fertilizers, and other contaminants, have been linked to changes in streams including (1) hydrology, including the amount, movement, and distribution of water and (2) physical habitat, the actual structure of the stream that is home for organisms (biota), such as invertebrates and fish; and (3) chemistry.
How was the amount of impervious cover determined?
Impervious cover was estimated for the years 2000-2001, using a modeling method developed by National Oceanic and Atmospheric Administration (NOAA). The method involves using a 1-kilometer square grid and two indicators of impervious cover: population count and the brightness of satellite-observed nighttime lights; the model was calibrated using USGS data for impervious cover at a 30-meter resolution. More information on estimation of impervious cover can be found in the following publication available online: http://www.ngdc.noaa.gov/dmsp/pubs/ISAglobal_20070921-1.pdf
What land use/land cover data were used?
Land use refers to how people utilize the land to accommodate a demand for roads, houses, and businesses, as well as needs for other kinds of built infrastructure, such as waste water treatment plants and dams. Land cover refers to physical materials, such as vegetation and human-constructed roads and buildings that cover the earth's surface. The land use/land cover data were drawn from the 2001 National Land Cover Database, Version 1, developed by the Multi-Resolution Land Characteristics Consortium, a multi-agency effort (http://www.mrlc.gov/nlcd2001.php). Land cover in each study watershed was classified as belonging to one of several categories, including urban land, forested land, shrubland, agricultural land, and wetland. Land cover data and GIS-derived watershed boundaries were used to estimate the proportion of watershed area associated with these land cover types, as well as the amount of land cover in near-stream or riparian areas. Land cover data were also used to assess the spatial configuration of land cover types, including the degree of fragmentation. Detailed descriptions of these methods are available in Falcone and others (2007).
What kinds of data were collected and how often at each site?
At each site, biological, physical (hydrology and habitat), and chemical (water chemistry) components were measured or estimated using a standard set of methods along about a 150 meter length of stream referred to as a "reach."
Physical components measured included stream hydrology and physical habitat (in stream and near-stream or riparian areas). For hydrology, the water level or "stage" and water temperature were measured hourly for one year using sensors that made automatic measurements at or near where other physical, chemical and biological data were collected. Physical habitat characteristics, such as channel shape and size, substrate size and distribution, water velocity, shaded cover, aquatic and riparian vegetation, and bank and riparian stability, were measured once during seasonal low-flow periods (typically summer).
Chemical samples were collected two to six times at each site (during spring and summer low-flow conditions) and analyzed for nutrients, pesticides and pesticide breakdown products, suspended sediment, sulfate, and chloride in water; water temperature, dissolved oxygen, pH, specific conductance, and discharge were measured in the field at the time of biological sampling. During the year in which biological sampling occurred, water samples were also collected bi-monthly at 10 stream sites in each of full studies (Portland, Atlanta, Raleigh, Denver, Dallas, and Milwaukee). Passive sampling devices were deployed for about six weeks at selected sites to estimate bioavailable amounts of selected waterborne synthetic chemicals, such as polychlorinated biphenyls (PCBs), polycyclic aromatic hydrocarbons (PAHs), and certain insecticides, which are known to be toxic to biota.
To assess the biological component of stream ecosystems, fish and benthic (attached) invertebrates and algae were collected one time during summer low-flow conditions. Fish were collected throughout the entire stream reach, identified in the field, counted, weighed, measured and, except for fish needing further identification, released. Invertebrate and algal samples were collected from five discrete locations within each reach and combined into a single algal sample and single invertebrate sample to represent the site. Most algal and invertebrate samples were collected from cobble rocks in riffle areas; however, where rocks were not readily available, woody snags were sampled.
What types of measures were computed to assess the conditions of stream ecosystems?
From streamflow stage data, measures or metrics representing hydrologic condition were calculated which summarize several characteristics of stream flashiness, including the magnitude, duration, and frequency of high flow conditions, and the rate of change in extreme high stage conditions. Other streamflow-summary statistics, such as mean annual discharge and maximum daily discharge, were computed based upon daily discharge values for sites with active USGS gages. Additional modeling was done at selected sites using the USACE Hydrologic Engineering Centers-River Analysis System (HEC-RAS) (v. 3.0) 1-dimensional steady-flow hydraulic model (Brunner, 2001).
A pesticide toxicity index (PTI), which represents the potential acute toxicity of multiple pesticides in a water sample, was calculated by combining the relative toxicity values of individual pesticide concentrations. Chemicals collected by the passive sampling devices were analyzed for amounts of hydrophobic contaminants, such as polychlorinated biphenyls (PCBs) and polycyclic aromatic hydrocarbons (PAHs), and for potential toxicity to aquatic organisms.
Differences among sites in the distribution and abundance of stream organisms were represented by metrics such as number of individual organisms, numbers and types of species, relative abundance, diversity, community composition scores (representing the degree of similarity among selected biological communities, such as invertebrate communities, across all sites in a study area), the percentage of species that can or cannot tolerate pollution or disturbance, and other characteristics.
What were the main findings of the study?
Comparisons among the nine metropolitan areas show that not all urban streams, and not all groups of stream biota, respond in a similar way to urban development. Land use/land cover prior to urbanization can affect how stream ecosystems respond with greater amounts of urban development. The response of stream biota was generally stronger in areas where urban development occurred at the expense of forested lands rather than agricultural lands; stream ecosystems in agricultural lands were previously disturbed from human activities and tended to be in poorer condition than stream ecosystems in non-agricultural lands.
All study areas showed hydrologic responses to urban development for one or more of four measures of streamflow flashiness, the rapid response of streamflow to storm events, characterized by frequent, large increases in flow, rapid changes in flow rates, and short duration of high flow. An increase in the frequency of high flow was the most common and strongest hydrologic response. Although physical habitat changes were not always found to be related to urban development, changes in channel size, width-to-depth ratio, and (or) the size of streambed sediment were the most common changes associated with urban development in the nine study areas.
Amounts of chloride, pesticides (particularly insecticides), and hydrophobic organic contaminants in stream water generally was higher (either in concentration, or number of compounds, or both) at higher amounts of urban development in most study areas. For nutrients such as nitrogen and phosphorus in stream water, levels of nitrogen increased in study areas where urban development occurred on forested land but not on agricultural land; phosphorus levels generally showed little relation to urban development. Results from the passive sampling devices also showed that the relative toxicity of waterborne chemical compounds, such as PAHs that are often associated with coatings used on roads and parking lots, increased with urban development.
Among the different groups of stream biota assessed, invertebrate communities appeared to be most sensitive to the effects of urban development by showing moderate or strong changes in community composition in eight of the nine study areas (Denver was the exception). Algal and fish communities showed a moderate to strong change in composition in four of the nine study areas.
Findings show that aquatic invertebrates demonstrate adverse effects to low levels of urban development that were previously thought to be protective of aquatic life. By the time a watershed reaches about 10 percent impervious cover in urban areas, aquatic invertebrate communities are degraded by as much as 33 percent in comparison to aquatic invertebrate communities in primarily forested watersheds. At higher levels of urban development, the continuing change in community composition suggests that ecological function of a community is reduced by the overall loss in number of aquatic species, which results in a decline in food-web complexity.
What linkages between measured physical, chemical, and biological components in stream ecosystems did the study find?
Previous studies have found that urban development can alter streamflow or other hydrologic patterns and lead to an increase in the frequency of extreme floods or prolonged low-flow conditions. The altered streamflow, in turn, can influence physical habitat and water chemistry. These cumulative physical and chemical changes can "stress" the biological community and can lead to changes in the species compositions of a stream's biological communities and its foodweb.
Understanding the response of a stream ecosystem to urban development is often complicated by interactions among multiple physical and chemical factors. These interactions can make it challenging to determine which stressors are principally responsible for changes in the biological community in an urbanizing watershed. The USGS study found that that no single stressor was universally important in explaining responses to urban development across all the study areas or biological communities across the nation.
The Biological Condition Gradient provides a common frame of reference for assessing the biological condition of streams in a way that helps communicate the concept of "stream health". The Biological Condition Gradient is in essence a standardized, descriptive ranking system designed to communicate stream health as assessed from biological community data sampled in a stream. The Biological Condition Gradient level for any stream is reported using six tiers that represent biological condition endpoints, ranging from natural to highly disturbed.
In a pilot investigation using data from the Boston study area, an innovative regional model was developed for New England to predict how different combinations of urban-related stressors associated with stream hydrology, habitat, and chemistry affect stream health, as measured by changes in the Biological Condition Gradient. An important feature of the model is its capacity to predict how multiple urban stressors can affect the level of stream health, which is defined by one of the six tiers (ranging from excellent to poor) along the Biological Condition Gradient. Thus, the model can be used as a modeling tool to evaluate different management scenarios for protecting stream health in urbanizing areas. For example, the model predicted that when the levels of urban development in a watershed exceeded 31 percent, the likelihood of attaining a healthy stream would be only about 25 percent. However, if management actions were implemented to improve water quality and reduce stream flashiness, the likelihood of attaining a healthy stream increased to about 70 percent.
What are the watershed management implications of the study?
Streams in different regions of the country respond differently to urban development. This is because of complexity in these ecosystems and because of regional differences in the overall suite of factors that affect stream ecosystems, such as climate and the types of land being developed for urban uses. These regional differences also affect the response of hydrology, physical habitat, water chemistry, and aquatic biota. Because of pronounced regional differences in background, or natural, stream ecosystem conditions, "one size fits all" protection and rehabilitation strategies will not protect all stream ecosystems. Management goals for urban stream ecosystems should reflect what is possible to achieve in terms of stream water quality and biological condition in a given region of the country.
A major challenge in assessing relative biological conditions for streams in different parts of the country is setting expectations for the assessed sites that account for regional differences in best available, or "reference," conditions of aquatic biota. The Biological Condition Gradient, which provides a frame of reference for assessing aquatic biological conditions, can be applied in all parts of the United States to address this inconsistency. Best available stream conditions in a part of the country that include high quality biological communities may be associated with top level Biological Condition Gradient tiers; other parts of the country, with degraded biological communities, may have reference conditions in mid-level Biological Condition Gradient tiers. Informative cross regional comparisons can only be made when they are based on a common frame of reference.
Results of our study demonstrate that physical and chemical changes associated with urban development at low levels, previously thought to be protective, can have adverse effects on of aquatic life. Watershed planning and management efforts that minimize physical and chemical changes may in turn minimize biological changes in streams. Successful efforts to limit adverse effects of urban development on stream ecosystems have included restricting development in undisturbed watersheds, conserving forest land and increasing tree canopy on urban land, limiting pollutants leaving development sites and reducing pollutants from urban land, encouraging re-use of existing urban areas to accommodate increased growth, re-using existing urban lands through redevelopment, and limiting, disconnecting and (or) treating impervious cover.
What are the key challenges for efforts toward improving the health of urban streams?
There are three recurring challenges related to (1) scale and endpoint, (2) uncertainty and knowledge gap, and (3) implementation.
For scale and endpoint, land cover that results in adverse effects on urban streams is controlled at the local level, while most environmental regulations to protect water quality are enforced at the state or federal level. In addition, the scale at which urban stream impacts are most effectively measured and managed--the watershed scale--does not align with the jurisdictional boundaries within which most regulations and policies are applied. Finally, the endpoints that are the focus of management strategies are not necessarily related to conditions that people care about. Some uncertainty is unavoidable in assessing the influence of management practices on stream systems because interactions between regional, urban, and stream systems are complex and dynamic. Knowledge gaps exist because data are often needed to select and design and most effective management strategies, but these are often not readily available. Results of the USGS study contribute directly to reducing uncertainty and knowledge gap.
The implementation challenges are as complex as the ecosystems we seek to protect and include, among others, the need for updating codes and ordinances related to urban development and streams, educating the public about effective, science-based regulations to protect stream ecosystems, and increasing public awareness of the link between the actions they take on their land and the eventual impacts on a stream.
Fortunately, a wide array of policies, technologies, incentives and educational strategies are available to watershed managers that, collectively, can prevent or mitigate the impacts of urban development on stream ecosystems, so that these important resources can be enjoyed by the many who live, work and play near them.
How can I obtain data, maps, and reports from this study?
Links to the interpretive and data reports, and to related journal papers and method reports, are at: http://water.usgs.gov/nawqa/urban/html/publications.html. All data and additional information about the study design are available at Giddings and others, 2009, http://pubs.usgs.gov/ds/423/. Maps and descriptions of individual study areas are available at: http://water.usgs.gov/nawqa/urban/html/wherewestudied.html.
What other agencies or groups contributed to the study?
The USGS was the principal agency responsible for the overall design of the study, but the USGS also cooperated with other agencies and groups that contributed to planning, site selection, sampling, data analysis, and preparation of publications. The Center for Watershed Protection, Ellicott City, MD contributed to chapter 7 on management techniques for minimizing effects of urban development on streams. The Nicholas School of the Environment, Duke University, Durham, North Carolina completed the New England regional model to predict stream health. The State of Maine Department of Environmental Protection and the U.S. Environmental Protection Agency provided guidance on the use of the Biological Condition Gradient. The Chesapeake Stormwater Network contributed information on aspects of the Impervious Cover Model used in the USGS study.
Whom can I contact for more information?
James F. Coles (Email: firstname.lastname@example.org; phone: 603-226-7845)
Where can I learn more about the effects of urban development on aquatic environments?
Where can I learn more about other NAWQA studies?