Parsing Ecological Impacts in Watersheds

In the United States and elsewhere, watershed-scale risk assessment is emerging as a new discipline for investigation, evaluation, and restoration of rivers, bays, and coastal estuaries. Using the three-tiered paradigm of ecological risk assessment (problem formulation, analysis, and risk characterization) developed by the U.S. Environmental Protection Agency (USEPA) nearly $30\text{years}$ ago, many state and federal agencies in the U.S. are adopting a more holistic approach to environmental management, one that seeks to identify natural and anthropogenic chemical, physical, and biological stressors and provides a framework for preservation or restoration of the integrity of waterways and their natural resources.

In the U.S., the stressor identification process developed by the USEPA provides a framework to assist water quality experts in identifying unknown causes (stressors) of biological impairment through the use of a logical, scientific process by which available information can be evaluated (e.g., http://cfpub.epa.gov/caddis/guide.cfm).

California and Oregon (USA) have developed watershed assessment manuals to provide guidance for assessment, monitoring, and planning for watershed protection, conservation, and restoration and to coordinate state and federal planning needs, such as recovery plans and total maximum daily loads (TMDLs). In both cases, the emphasis is on baseline surveys and on developing an understanding of existing physical, biological, and ecological conditions of the ecosystem; discerning the historical evolution of the watershed; and developing a strategy to improve, monitor, and protect the watershed (e.g., http://cwam.ucdavis.edu/index.html).

In Europe and elsewhere there is increasing use of the Drivers, Pressures, States, Impacts, and Responses (DPSIR) approach (Elliott 2002; Joziasse et al. 2005; Mairate et al. 2003; Marsili-Libelli et al. 2004; Smeets and Weterings 1999), which builds on an existing model by the Organization for Economic Co-operation and Development (OECD) to organize and make sense of the broad array of information, including spatial and temporal features needed to understand the ecological health of large,complex waterways (e.g., http://org.eea.eu.int/documents/awp99/awp99̱annex1.html).

This editorial highlights the reemergence of watershed-level thinking by the regulatory community; identifies new and emerging watershed-level assessment methods; and underscores the benefits to stakeholders, including industrial, regulatory, environmental, and citizen groups, from the integration of a watershed-level assessment with economic, social, legal, and political factors that affect management and engineering decisions in watersheds.

Watershed-level thinking implies recognition of chemical and nonchemical stressors, as well as other natural or anthropogenic stressors, in an entire drainage area; these may affect the integrity of a lake, river, or coastal area. As the regulatory community and the environmental industry as a whole begin to take a more holistic approach to environmental ecology and protection, the consideration of these stressors will become increasingly important when identifying and selecting mitigation options that can meaningfully reduce or eliminate long-term potential impacts on waterways. For example, estuarine studies often emphasize the potential impacts of chemical releases associated with industrialization and urbanization (Elliott 2002).

Chemicals that receive particular attention typically include bioaccumulative compounds such as polychlorinated biphenyls (PCBs), chlorinated pesticides, and methyl mercury; petroleum and coal-derived hydrocarbons such as polycyclic aromatic hydrocarbons (PAHs); and metals such as chromium, lead, copper, and zinc. Parsing chemical stressors and their sources from one another and from other types of stressors, such as nonindustrial stressors, remains a significant challenge. Nonindustrial physical, chemical, and biological stressors (e.g., introduced species, overfishing, river channelization, and housing development) have been shown to confound the interpretation of historical and current biological and ecological conditions in many estuarine environments (Apitz et al. 2005a; Elliott and Cutts 2004; Landis 2005). For example, Landis (2005) has developed and applied a relative risk model at the watershed scale to carry out regional risk assessments that address multiple habitats with multiple stressors affecting multiple endpoints. Although a single stressor may drive site characterization and remedial requirements, Landis (2005) points out that at a regional scale other stressors acting upon the assessment endpoints also must be considered.

Effective and sustainable management strategies must focus on the entire ecosystem, rather than on one site or environmental medium at a time. Without a watershed-level perspective, environmental managers and regulators risk misrepresenting known stressors because they lack an understanding of the broad range of potential ecological stressors and the natural and anthropogenic sources that adversely affect the ecosystem.

This focus requires a conceptual basin model (CBM), which describes and inventories the mass flow of water, contaminants, and particles within a river basin. The CBM should incorporate a decision-making hierarchy, setting priorities at a basin scale followed by site-specific risk assessment and control of point and diffuse contaminant sources at a local level (Apitz et al. 2005b). Thoughtfully developed CBMs are increasingly recognized as more effective than site-limited approaches in terms of thoroughly understanding a watershed ecosystem and for river-basin management (see, e.g., USEPA 2001).

Initial development of the CBM is informed by data gathering and data management efforts; as it coheres, the model itself guides further field data collection. Although the model focuses on evaluating and parsing specific environmental stressors, the identification and implementation of long-term mitigation measures for improvement of watershed health must continually guide model development.

A comprehensive CBM incorporates all known and potential stressors, including their relative magnitudes, spatial scales, and impacts on the local ecology. The CBM also incorporates the potential interactions of different stressors on biological receptors. The initial development of the CBM is informed by an understanding of the ecology, historically recognized stressors, other potential estuarine stressors, and their potential ecological impacts. The CBM links potential causes and effects of stressors, and enables prioritization of stressors with respect to the ecological health of the estuary. In the process, the CBM also communicates hypotheses and assumptions and identifies where additional data collection could provide useful information.

The EPA favors the stressor identification (SI) process (USEPA 2000) as a means of identifying and differentiating causes of biological impairments in watersheds where there are multiple stressors (e.g., sedimentation, low dissolved oxygen, and toxics). The SI framework highlights the conceptual links between diverse natural and anthropogenic stressors and drivers on the ecological functions of a watershed.

The SI framework is illustrated in Fig. 1. The approach begins by characterizing the nature of the impairment and proceeds by identifying candidate causes, eliminating improbable causes, and identifying the most probable causes based on the strength of evidence. A logical framework is used to establish strength of evidence based on spatial and temporal patterns, available knowledge of the stressors, and plausibility arguments. This sharpened focus enables industry and regulators to consider appropriate mitigation measures and to assess the efficacy of such measures in reducing potential short- and long-term impacts (USEPA 2004; 2005). Furthermore, the SI approach can be used to support highly effective communication with environmental managers and other stakeholders as part of a public outreach process.

The DPSIR Framework endorsed by the European Environment Agency (EEA) and OECD provides a similar mechanism for analyzing environmental problems. Particularly useful for policymakers, DPSIR offers a basis for analyzing the interrelated factors that impact the environment. The DPSIR approach defines the interactions between various parameters, includingDrivers may be defined as human activities, and the pressures, states, and impacts are reflected in the damage done to the environment by those activities; responses focus on reducing those impacts by controlling or eliminating drivers. Sustainable management decisions require that responses seek to balance environmental, regulatory, and socioeconomic goals (Apitz et al. 2005a; Joziasse et al. 2005).

Ecological health (or ecological integrity) is typically determined by evaluating key indicators of the biological, chemical, and physical integrity of a watershed. While watershed assessments have focused largely on industrial impacts on ecological health, this editorial has identified the key considerations and emerging approaches to facilitate the parsing of diverse ecological stressors on watersheds, including nonchemical stressors.

Civil and environmental engineers have a critical role to play in the further evolution of watershed-level thinking in North America, Europe, and elsewhere around the globe where water resources are increasingly threatened by human activities. In well-designed watershed assessment studies, tools are needed to identify the likely contributors to key assessment benchmarks representing indications, or symptoms, of an ecosystem’s impairment (e.g., loss of carrying capacity, reduced biodiversity, reduced habitat (e.g., wetland loss), degraded water or sediment quality, limitations on the ability to provide recreational services, and biological or physical injury).

The selection of appropriate remedial responses may involve a complex comparative risk assessment that considers the financial, regulatory, scientific, and technical aspects of the remedy (Magar and Wenning 2005). Appropriate approaches may be affected by the scale of the problem, the possibilities for source control and/or natural recovery, costs, and the presence and impacts of other stressors. The success or failure of a response must also be evaluated over time, and with respect to the range of stressors that may exist at a site, to determine changes in drivers and pressures and to evaluate the efficacy of the response.

This watershed-level approach provides a model for environmental stewardship at facilities located in estuarine environments. If the engineering community is successful, we will have enabled property owners, large industry, regulators, and other stakeholders to identify and prioritize mitigation options that can meaningfully and cost-effectively reduce or eliminate stresses and their long-term potential impacts.