Reference Condition

This page describes the reference condition approach to develop numeric nutrient criteria (NNC). Remember that nutrient criteria development can be an iterative process and involve more than one method. If, after evaluating your data and analyses from an initial attempt at using the reference approach, you find that your data or analyses do not support the underlying assumptions (described in the sections) of the reference condition approach, consider stepping back in the process and reevaluate. The path forward might require using additional data sets and/or a different analytical approach.

Reference Condition Background

The reference condition approach involves computing criteria on the basis of water quality conditions present in least or minimally disturbed water bodies known to support designated uses. It is a scientifically defensible method for deriving NNC using measurements of the watershed setting, physical, chemical, and biological quality of water bodies to define reference, which is then applied to other similar water bodies. Reference waters can be defined from an existing spatial reference condition or a reference timeframe of water bodies that describe the desired conditions upon which to base your criteria.

The term reference condition refers to the condition that supports biological integrity, defined as “the capability of supporting and maintaining a balanced, integrated, adaptive community of organisms having a species composition, diversity, and functional organization comparable to that of the natural habitat of the region” (Frey 1977). EPA’s operational definition of integrity is similar and is defined as “the ability of an aquatic community to support and maintain a structural and functional performance comparable to the natural habitats of a region” (Frey 1977; Karr and Dudley 1981). Because the ideal reference condition for biological integrity rarely exists due to the history of human disturbance and existing pervasive impacts (e.g., atmospheric pollutants), biological integrity sensu stricto must often be estimated and the ideal reference condition replaced with a surrogate.

Differing levels of reference condition surrogates have been defined operationally—such as minimally disturbed, historical, least disturbed, and best attainable—to clarify consistency and communicate across applications (Stoddard et al. 2006). A set of regional least disturbed reference waters is typically identified using a combination of landscape, local habitat, and other factors as screens, which are selected to define the surrogate reference condition. Once that set of waters has been identified, statistical distributions of the parameter of interest from those waters are used to derive a numeric criterion. Using least disturbed reference sites to estimate the reference condition for NNC development was adapted from the process for developing biological criteria (biocriteria) and endorsed by EPA’s Science Advisory Board (USEPA 1992). Least disturbed reference conditions provide a baseline that should protect the inherent beneficial uses of the water bodies for which criteria are being developed. To date, EPA and several states have successfully applied reference condition approaches for NNC development either as a single approach or within a multiple lines of evidence structure (e.g., Florida, Hawai’i, Montana, South Carolina). Reference conditions often approximate what can be considered a natural existing condition for waters in a specific region and are, therefore, an important line of evidence. Resource managers can use reference condition water bodies and the levels of nutrients and other water quality constituents as a comparison against estimates for those constituents derived using mechanistic modeling or stressor-response analysis, and this approach is frequently applied in other water quality regulatory settings (e.g., TMDLs, USEPA 1999).

For some water body types (such as a reservoir or stream in a heavily modified landscape), defining a true natural state is impossible. Reference conditions that describe the desired state of the water body, however, can be developed using the established designated use of the water as the basis and the same biological integrity relationships described above used to define a reference condition. For example, truly natural conditions for a human-constructed water body (e.g., a reservoir) do not exist, so the best existing conditions, which support designated uses, can be used instead. Another case where this approach can apply is for waters that exist within large areas of the landscape that have been changed; such as the result of conversion of the landscape to agricultural fields. Again, basing reference conditions on the best existing conditions, which support the designated uses, can be used.

Application of the reference condition approach by states has varied in terms of (1) the variety of methods, variables, and thresholds applied to define reference; (2) the statistical characteristic of the reference water bodies used to derive a reference criterion; and (3) the relationship of (1) and (2) to supporting designated uses. States using a reference condition approach should provide a transparent description of how they apply the process.

Important Questions for Resources Managers

Are the waters supporting designated uses?

As required by Title 40 of the Code of Federal Regulations (CFR) Part 131.11, water quality criteria must be based on sound scientific rationale; must contain a sufficient number of parameters or constituents to protect the designated use; and, for multiple use designations, must support the most sensitive use.

In general, states and tribes have interpreted the meaning of protecting designated uses from nutrient pollution through their narrative criteria statements. EPA is approaching the Clean Water Act interpretation of integrity in waters by characterizing the existing condition of least disturbed reference sites and the best available information on populations of aquatic flora and fauna. Water bodies are a unique collection of physically and ecologically diverse environments. Some of those areas might have experienced changes from historical conditions as a result of physical modifications. Such modifications, however, do not necessarily preclude support of biological integrity or the use of a reference condition approach to derive NNC for those waters. The Water Quality Standards Handbook (EPA 1994) provides guidance for implementing these regulations.

Are there multiple uses? Do the criteria derived protect the most sensitive use of those waters?

For waters with multiple use designations, criteria must support the most sensitive use.

Types of Reference Condition Approaches

Spatial and Temporal Reference Condition Approaches

Ideally, a reference site should be unimpacted by human activity or disturbance. Typically, the reference condition approach is applied spatially, using a set of existing least disturbed sites known to be minimally impacted by nitrogen and phosphorus pollution (comparative/spatial reference condition). The reference condition also can be historic, however, and based on data collected in the past, when the water body or water bodies were determined to be least disturbed by nitrogen or phosphorus pollution (historical/temporal reference condition).

The general approach for establishing a reference condition involves direct observation (data collection) of water quality and the physical and biological attributes of the selected reference sites and estimation of reference condition values for the parameters of interest (in this case, nutrients and parameters associated with nutrient pollution). If historical data are available at sites of interest from a period of time predating disturbance, then pursuing a temporal reference approach is appropriate. If historical data are unavailable, then a spatial reference approach could be pursued by identifying sites attaining designated uses and desired reference conditions at all times. Resource managers often select a percentile threshold of the data distribution representing the desired reference condition to account for uncertainty in the measurements. Some other characteristics are described in Table 1.

Table 1. Comparison of temporal and spatial reference condition approaches

  Temporal Reference Approach Spatial Reference Approach
Characteristics of Approach · Identifies temporally defined conditions fully representative or supportive of the designated uses at one or more sites

· Accounts for temporal variability

· Identifies geographically defined reference sites that exhibit conditions supporting designated uses currently as well as in the recent past

· Accounts for spatial variability; can account for temporal as well, if samples are collected over time

Available Data · Historical data available at sites from a period of time predating disturbance

· Data available include times that meet reference site requirements

· Limited historical data available that correspond to known conditions of water bodies

· Data available include sites that meet reference site requirements

Applicable Area · Site-specific or regional · Regional

A combination of the two approaches involves using conditions that existed during identified times and space to set the expectation for all other times. If the actual number of reference sites is limited, ecoregions or surrogate data from areas with comparable ecological characteristics have been successfully used to develop nutrient criteria (USEPA 2000a, 2000b, 2000d, 2000e, and 2000f).

How to Apply the Reference Condition Approach

This approach should be iterative and revisit initial assessment endpoints and conceptual models as needed.

Step 1. Identify Reference Water Bodies

Reference waters are waters in a region (or time period) that are the least disturbed by human activities. They should also exhibit the range of characteristics (e.g., discharge, latitude, longitude, size, and volume) that can be expected of most of the other water bodies in the region. Screens are often defined to objectively identify a region’s reference water bodies. These screens can include factors such as the degree of anthropogenic disturbance in the watershed, whether water bodies have been placed on the 303d list, and chemical characteristics that indicate the presence of pollution.

Step 2. Classify Water Bodies (Classification)

Classification is the sorting of established groups of water bodies within which natural expectations for nutrient concentrations are similar. For spatial reference approaches, classification is a key step in the process of developing criteria. Classification of reference water bodies should reduce the variability in ambient measurements within each of the groups and help ensure that derived criteria are appropriate. For example, nutrient concentrations in shallow, productive lakes naturally differ from deep, unproductive lakes, and reference criteria derived for the first group should not be applied to the second.

Step 3. Choose a Percentile (Computation)

Once the reference site(s) are vetted and pass the screening process, compute a statistical distribution of values for the variables of interest (e.g., total nitrogen, total phosphorus, chlorophyll a). Then, choose a percentile of the distribution that sets the NNC. Water bodies with values below the criterion value are in compliance.

Selecting the percentile is a policy decision informed by the data and supported with documentation of a sound scientific rationale for the selected percentile. Elements involved in justification include the confidence that the reference set of water bodies reflect integrity and use support. The size of the dataset may also influence the selected percentile.

As an example, in past efforts, some states have chosen the 75th percentile of the variable distributions, while others have chosen the 90th percentile. In general, using a higher percentile of conditions (e.g., the 90th percentile) reflects high confidence that the sites represent the a priori conditions defined in step 1, the data set used to derive the criteria is supportive of designated uses (e.g., support healthy biological communities) and that data from a sufficient number of sites are available to estimate the percentile with confidence (i.e., there is little uncertainty associated with the estimated data distribution). Selecting higher percentiles excludes the effect of spurious outliers and reflects confidence that the data have captured the expected range of variable values for each water body segment. Policy decisions regarding percentile selection may also be adjusted via an iterative process that includes feedback from stakeholders who are familiar with the water bodies. It might be useful to combine or compare reference condition-based candidate values with those from other lines of evidence, including stressor-response relationships, mechanistic modeling, scientific literature, and existing criteria from surrounding states of EPA 304(a) recommendations. In any case, a sound scientific justification should accompany the documentation of percentile selection.


Additional Challenges

There are a few challenging decision points when it comes to establishing a defensible reference condition approach for deriving numeric criteria values. Establishing the reference condition is but one component in the criteria development process. Reference condition values are appropriately modified based on examination of the historical record, antidegradation considerations, downstream protection, and percentile selection. You also must consider the duration and frequency components of the criteria.

Referencing Historical Data

The reference condition approach should be interpreted in light of the historical condition of the resource and projections of its future potential. A good historical database is important for setting proper nutrient criteria and standards because it establishes a perspective on the condition of a given water body, provides knowledge of what has happened before placing more current information in proper perspective, and establishes what the normal natural trends are and what disruption human activities might have caused. Important questions include:

    • Has the condition changed radically in recent years?
    • Is the system stable over time?
    • Has there been an upward or downward trend in trophic condition?

Only an assessment of the historical record can provide answers to these questions. Without that information, you risk establishing criteria that are not truly meeting long-term conditions that define the desired reference conditions.

Antidegradation Policy and Attention to Downstream Effects

A critical requirement for the use of reference conditions associated with nutrient criteria is the EPA antidegradation policy, which protects against incremental deterioration of water bodies and reference conditions. An observed downward trend in the conditions of reference sites cannot be used to justify relaxing reference expectations, reference conditions, and associated nutrient criteria. Once established, nutrient criteria should be refined in a more stringent direction only in response to improved conditions. Without antidegradation safeguards, even establishing reference conditions and nutrient criteria could still allow for continual deterioration of water quality. To combat this, the states should implement an effective antidegradation policy that promotes continually improving conditions. As an example, Maine has an antidegradation policy that requires that water bodies remain stable or improve in trophic state (Courtemanch et al. 1989; NALMS 1992).

Selection of a Duration and Frequency

Water quality criteria include three components:

  • Magnitude: The concentration of a pollutant that can be maintained over time in the ambient receiving water without adversely affecting the designated use that the criteria are intended to support.
  • Duration: The time period over which exposure is averaged (i.e., the averaging period) to limit the time of exposure to elevated concentrations. This accounts for the variability in the quality of the ambient water caused by variations in constituent inputs, flow, and other factors.
  • Frequency: How often the magnitude/duration condition can be exceeded and still protect the designated use.

Combining the criterion-magnitude with the duration and frequency prevents harmful effects from infrequent exceedances of the criterion-magnitude by ensuring compensating periods of time during which the concentration is below the criterion-magnitude (USEPA 2012c).

EPA has recommended in the past that, for a given water body, the criteria parameter must not exceed the applicable criterion concentration more than once in a 3-year period, or twice once in a 5-year period. A frequency of exceedances chosen should allow water bodies enough time to recover from occasionally elevated levels of nitrogen and phosphorus concentrations.

Case Studies

Minnesota Lakes

  • Values within the interquartile range of reference conditions were considered typical
  • Most sensitive subuses of lakes were considered
  • Emphasis was placed on 75th percentile for reference conditions

Red River of the North

  • The reference condition approach was considered
  • Found no reference sites along the mainstem or tributaries of the Red River
  • Reference condition approach deemed not appropriate for setting nutrient criteria for the Red River

Florida’s Streams

  • Reference condition approach used to derive TN and TP criteria
  • Least disturbed conditions determined by human disturbance and biological health
  • Regional stream classifications used to inform criteria derivations

Florida’s Lakes

  • Weighted-line-of-evidence approach used to derive chlorophyll a criteria
  • Regression relationships with chlorophyll a used to derive TN and TP NNC
  • Lake color and alkalinity used to create lake classifications

Wisconsin Streams

  • A multiple linear regression model was used to estimate reference conditions
  • The 75th percentile approach was used to determine reference values for biotic indices

Maine Fresh Surface Water

  • Maine DEP used 90th-percentile reference conditions for the most protected water classes
  • 75–90th percentile was used on a case-by-case basis for other water classes
  • Reference conditions were part of a line-of-evidence approach

Remote Sensing in FL

  • A reference condition approach using satellite-derived chlorophyll a was used to derive nutrient criteria
  • Used ocean-color satellite data (1998–2009) from coastal waters meeting designated uses

Tidal James River Chl-a Criteria

  • Reference phytoplankton community conditions were used to derive chlorophyll a criteria
  • Reference conditions were linked to light penetration and nutrient concentrations

TN Ecoregional Nutrient Criteria

  • Initiated new method to establish reference conditions based on ecoregions
  • Delineated subecoregion boundaries and selected reference streams from unimpacted watersheds
  • A database containing chemical, physical, and biological data can be used in the future for NNC

Florida Everglades Phosphorus

  • Determined that minimally impacted areas best represented historical P concentrations
  • Looked at historical evidence
  • Considered water column and soil P levels in all hydrologic units to account for spatial variability

Estuarine Criteria in Florida

  • Focused on current conditions that supported balanced natural populations of flora and fauna
  • Examined peer-reviewed literature, state and local reports, and the Florida Impaired Waters Rule (FL IWR) database
  • Considered upper percentile and median concentrations as reference points
This website is in beta. Information on this website is not final and is subject to change