Nutrient Criteria Technical Guidance Manual: Lakes and Reservoirs

United States Office of Water EPA-822-B00-001

Environmental Protection Office of Science and Technology April 2000

Agency Washington, DC 20460 www.epa.gov

Nutrient Criteria

Technical Guidance Manual

Lakes and Reservoirs

First Edition

Nutrient Criteria

Technical Guidance Manual

Lakes and Reservoirs

Written by

George Gibson, U.S. Environmental Protection Agency

Robert Carlson, Kent State University

Jonathan Simpson, Tetra Tech, Inc.

Eric Smeltzer, Vermont Department of Environmental Conservation

Jeroen Gerritson, Tetra Tech Inc.

Steven Chapra, University of Colorado

Steven Heiskary, Minnesota Pollution Control Agency

Jack Jones, University of Missouri

Robert Kennedy, U.S. Army Corps of Engineers

Acknowledgments

The authors wish to thank the following peer reviewers for their assistance in the preparation of this

manuscript: Roger Bachman, University of Florida; Nina Caraco, Institute of Ecosystem Studies;

Douglas Knaur, Wisconsin Department of Natural Resources; Matthew Liebman, USEPA Region I;

James Peterson, University of Wisconsin; and Val Smith, University of Kansas. Their detailed and

insightful comments and criticisms are appreciated.

Additional internal reviews, especially with respect to EPA methods, policy, and regulations were

provided by specialists from the Office of General Counsel, Office of Science and Technology -

Standards and Applied Science Division; Office of Wetlands, Oceans and Watersheds; Office of

Wastewater Management; and EPA Regional Nutrient Coordinators. State agencies and private interest

groups also offered comments. All timely opinions were considered and addressed where possible in this

edition. In this respect, Joanna Taylor, The CDM Group, Inc., patiently and graciously made repeated

format changes to this manuscript as it was being completed.

As with any prototype technical guidance, differences about methods and approaches are to be

expected. This and subsequent guidance manuals for other water body types are not intended to be

singular, one-time publications. As experience accumulates, future editions will be prepared.

Suggestions not presently incorporated may be revisited and appear in later versions.

Nutrient Criteria—Lakes and Reservoirs iii

Disclaimer

This manual provides technical guidance to States, Indian Tribes, and other authorized jurisdictions

to establish water quality criteria and standards under the Clean Water Act (CWA), to protect aquatic life

from acute and chronic effects of nutrient overenrichment. Under the CWA, States and Indian Tribes are

to establish water quality criteria to protect designated uses. State and Indian tribal decisionmakers

retain the discretion to adopt approaches on a case-by-case basis that differ from this guidance when

appropriate and scientifically defensible. While this manual constitutes EPA’s scientific

recommendations regarding ambient concentrations of nutrients that protect resource quality and aquatic

life, it does not substitute for the CWA or EPA’s regulations; nor is it a regulation itself. Thus, it cannot

impose legally binding requirements on EPA, States, Indian Tribes, or the regulated community, and

might not apply to a particular situation or circumstance. EPA may change this guidance in the future.

iv Nutrient Criteria—Lakes and Reservoirs

Contents

Executive Summary ........................................................................ viii

Chapter 1. Introduction .................................................................... 1-1

A. Purpose of This Document .................................................................1-1

B. Relationship Between Water Quality Standards and Criteria .......................................1-2

C. Uses of Nutrient Criteria ...................................................................1-3

1. Identification of Problems ..............................................................1-3

2. Management Planning .................................................................1-3

3. Regulatory Assessments ................................................................ 1-3

4. Project Evaluations ...................................................................1-4

5. Status and Trends of Water Resources ..................................................... 1-4

D. Overview of the Nutrient Criteria Development Process ..........................................1-4

1. Strategy for Reducing Cultural Eutrophication ..............................................1-6

2. Nutrient Criteria Development Process ....................................................1-7

Chapter 2. The Basis for Lake and Reservoir Nutrient Criteria ................................... 2-1

A. Historical Perspective .....................................................................2-1

B. The Nutrient Paradigm ....................................................................2-2

1. Phosphorus and Nitrogen as Limiting Factors for Algal Biomass ................................2-3

2. Role of the Watershed .................................................................2-4

3. Trophic Causal Chain .................................................................. 2-5

4. Lake Aging .......................................................................... 2-6

C. Connecting Watershed Loading to the Lake: A Mass Balance Model Approach .......................2-8

D. Trophic State Classification Systems ........................................................2-11

E. Uses of Trophic State Indices ..............................................................2-16

Chapter 3. Preliminary Steps for Criteria Development .......................................... 3-1

A. Defining the Resource of Concern ...........................................................3-1

B. Classification ...........................................................................3-2

1. Geographic Divisions .................................................................. 3-2

2. Nongeographic Classifications ..........................................................3-3

3. Covariates .........................................................................3-10

Chapter 4. Establishing an Appropriate Database ............................................... 4-1

A. Introduction ............................................................................4-1

B. Evaluating Existing Data ..................................................................4-1

1. Potential Data Sources .................................................................4-2

2. Quality of Historical Data ..............................................................4-4

3. Data Reduction ....................................................................... 4-5

C. New Data Collection......................................................................4-6

1. Types of Monitoring Associated With Nutrient Criteria ....................................... 4-6

2. Sampling Design ..................................................................... 4-8

D. Database Management ...................................................................4-13

1. Modernized STORET ................................................................4-13

2. EDAS .............................................................................4-14

Nutrient Criteria—Lakes and Reservoirs v

Chapter 5. Candidate Variables for Criteria Setting ............................................. 5-1

A. Introduction ............................................................................5-1

B. Nutrient Variables .......................................................................5-1

1. Phosphorus ..........................................................................5-1

2. Nitrogen ............................................................................5-3

C. Biological Variables ...................................................................... 5-4

1. Organic Carbon ...................................................................... 5-4

2. Chlorophyll a ........................................................................5-5

3. Secchi Disk Transparency ..............................................................5-7

4. Dissolved Oxygen .................................................................... 5-8

5. Macrophytes .........................................................................5-9

6. Biological Community Structure ........................................................ 5-10

D. Land Use .............................................................................5-12

Chapter 6. Identifying and Characterizing Reference Conditions .................................. 6-1

A. Significance of Reference Conditions ........................................................6-1

B. Approaches to Establishing Reference Conditions ............................................... 6-1

1. Direct Observation of Reference Lakes .................................................... 6-2

2. Paleolimnological Reconstruction ........................................................ 6-4

3. Model Prediction and Extrapolation ...................................................... 6-6

C. Initial Data Retrieval and Processing to Support Ecoregional

Nutrient Criteria Development .......................................................... 6-7

Chapter 7. Nutrient Criteria Development ..................................................... 7-1

A. Elements of Nutrient Criteria ...............................................................7-1

1. Investigation of the Historical Record .....................................................7-1

2. Establishment of the Reference Condition ..................................................7-2

3. Use of Models ....................................................................... 7-2

4. Expert Assessment of Information ........................................................7-2

5. Attention to Downstream Effects .........................................................7-2

B. Development of Ecoregional Nutrient Criteria ..................................................7-3

1. Regional and Lake/Reservoir Type Classification ............................................ 7-3

2. Conversion of Ecoregional Data to a Reference Condition and

Preliminary Nutrient Criteria ........................................................... 7-3

3. Refine Ecoregion Reference Condition Values ..............................................7-5

4. Evaluation by Regional Teams ...........................................................7-5

5. Examples of the Deliberation Associated with Development of a

Nutrient Criterion .................................................................... 7-7

C. Development of State Nutrient Criteria .......................................................7-8

1. Designated Use ......................................................................7-8

2. Hypothetical Illustration of the Relationship of State Criteria to Protect Designated Uses

Compared With an Ecoregional Criterion for TP .........................................7-12

3. Frequency Distribution Approach .......................................................7-14

D. Developing Nutrient Criteria Implementation Procedures ........................................7-14

1. Decisionmaking ..................................................................... 7-16

2. Multivariable Enrichment Index ......................................................... 7-16

E. Frequency and Duration ..................................................................7-17

vi Nutrient Criteria—Lakes and Reservoirs

Chapter 8. Using Nutrient Criteria To Protect Water Quality ..................................... 8-1

A. State Water Quality Standards ..............................................................8-1

1. Water Quality Standards and the Clean Water Act ........................................... 8-1

2. Protecting Designated Uses .............................................................8-2

3. Maintaining Existing Water Quality .......................................................8-2

4. Providing Flexibility in Implementation ...................................................8-3

B. Water Quality–Based Approach to Pollution Control ............................................8-4

Step 1: Identification of Impaired and Threatened Waters ........................................ 8-5

Step 2: Priority Ranking and Targeting ......................................................8-6

Step 3: Development of TMDLs ...........................................................8-6

Step 4: Implem entation of Controls ......................................................... 8-7

Step 5: Assessment of Controls ............................................................ 8-8

C. Nonpoint Source Pollution Management ...................................................... 8-9

1. Guidance for Controlling Nonpoint Sources of Nutrients ..................................... 8-10

2. Nonpoint Source Management Programs ..................................................8-11

3. Coastal Nonpoint Pollution Control Programs .............................................. 8-12

4. Farm Bill Conservation Provisions ......................................................8-13

D. A Comprehensive Procedure for Nutrient Management ..........................................8-15

1. Status Identification .................................................................. 8-15

2. Background Investigation ............................................................. 8-15

3. Data Gathering and Diagnostic Monitoring ................................................8-17

4. Source Identification .................................................................8-19

5. Management Practices for Nutrient Control ................................................ 8-19

6. Detailed Management Plan ............................................................8-21

7. Implementation and Communication .....................................................8-21

8. Evaluation Monitoring and Periodic Review ...............................................8-22

9. Completion and Evaluation ............................................................ 8-22

10. Continued Monitoring of the System ....................................................8-22

E. Resources .............................................................................8-23

Chapter 9. Modeling Tools ................................................................. 9-1

A. Introduction ............................................................................9-1

B. Review of Lake/Reservoir Eutrophication Modeling Frameworks ..................................9-1

1. Empirical Models ....................................................................9-1

2. Nutrient Budget/Mass Balance Models ....................................................9-3

3. Nutrient/Food Chain Models ............................................................9-5

C. Model Use for Aiding in the Establishment of Reference Conditions ................................9-7

1. Morphoedaphic Index .................................................................9-7

2. Mass Balance Models with Loading Estimation .............................................9-7

D. Watershed Management Models ...........................................................9-14

1. Mid-Range Watershed-Scale Loading Models ..............................................9-14

2. Detailed Watershed Loading Models .....................................................9-17

References ............................................................................... R-1

Appendix A. Nutrient Region Descriptions .................................................... A-1

Appendix B. Case Studies .................................................................. B-1

Nutrient Criteria—Lakes and Reservoirs vii

Executive Summary

Overenrichment of surface waters in the United States has been a long-standing problem to the extent

that approximately half of the waters reported by the States to be impaired are attributed to excess

nutrients and related biological growth. The EPA has established the National Nutrient Criteria Program

to address this water quality problem. The surface waters of concern are lakes and reservoirs, streams

and rivers, estuaries and coastal marine waters, and wetlands. Criteria representing enrichment

conditions of surface waters that are minimally impacted by human developmental activities will be

developed for each of the regions of the country. These will then become the basis for States and Tribes

of the United States to develop nutrient criteria to protect the designated uses of those waters. This

manual is designed to help accomplish this for lakes and reservoirs.

Nitrogen and phosphorus are the primary causes of overenrichment and are obvious nutrient criteria

variables, but biological response variables are also important in addressing the consequences of

overenrichment.

Limnologists and lake managers have developed a general consensus about freshwater lake responses

to nutrient additions, that essentially an ambient total phosphorus (TP) concentration of greater than

about 0.01 mg/L and or a total nitrogen (TN) of about 0.15 mg/L is likely to predict blue-green algal

bloom problems during the growing season. Similarly, chronic overenrichment leads to lake quality

degradation manifested in low dissolved oxygen, fish kills, algal blooms, expanded macrophytes, likely

increased sediment accumulation rates, and species shifts of both flora and fauna.

However, because some parts of the country have naturally higher soil and parent material

enrichment and different precipitation regimes, the application of that general consensus approach has to

be adjusted by region. Therefore, an ecoregional and reference condition approach is necessary to

develop nutrient criteria appropriate to each of the different geographical and climatological areas of the

country. Initially, the continental United States has been divided into 14 separate ecoregions of similar

geographical characteristics, and criteria will be developed for each.

While additional variables may be used as nutrient criteria, the initial effort will concentrate on TP,

TN, algal chlorophyll, and Secchi depth or similar measure of algal turbidity to reflect the primary

production response to overenrichment. Thus, the criteria involve four basic indicators of

overenrichment. Other indicators, such as dissolved oxygen (DO) and macrophyte growth or speciation,

and other flora and fauna changes are also deemed useful, but the first four are paramount, especially the

two limiting nutrients. Throughout the country, cultural eutrophication (or overenrichment) is largely

caused by either too much N or P or some combination of the two in their various forms. Nitrogen may

not be critical to many fresh water lakes, but it does become significant in estuaries and coastal waters

downstream. An essential part of the process for developing nutrient criteria is to pay attention to

downstream effects. Therefore, nitrogen as well as phosphorus reduction for lakes is needed to benefit

the lower reaches of the overall system.

TN and TP are described as causal variables, and chlorophyll and algal turbidity are initial response

variables. Measuring just the response variables clearly shows the existence of a problem, but waters

with a short retention time could look clear and be aesthetically acceptable, and could still be sending an

unacceptable load of N and P downstream to be someone else’s problem. This is why EPA expects

downstream effects to be considered as part of the nutrient criteria development process.

viii Nutrient Criteria—Lakes and Reservoirs

Nutrient criteria development consists of five elements:

1. Historical data and other information to provide an overall perspective on the status of the

resource.

2. Present reference sites and their collective reference condition describing the current status.

3. Modeling to refine data implications and analysis above if necessary.

4. Objective assessment of all of the above information by the States and by the EPA Regional

Technical Assistance Groups (RTAGs), a board of State and Federal specialists established in

each EPA Region to help develop and administer the National Nutrient Criteria Program, to

establish the ecoregional criteria, and to review proposed State or Tribal nutrient criteria.

5. Attention to downstream consequences before the criterion is finally established.

Using this approach, EPA ecoregional benchmark criteria can be established that States and Tribes

can use to help set their own criteria to protect all their designated uses. A key responsibility of the

RTAGs, with their best knowledge of regional water quality and management potential, is the

development of these ecoregional criteria and review of subsequent State and Tribal criteria. A summary

of the procedural approach for ecoregional criteria development is as follows:

The RTAGs collect as much existing reference quality data for at least the four principal variables as

possible from STORET, States and Tribes, universities, local governments, and other Federal agencies.

Data collection is directed to the particular waterbody type of interest and to established physical classes

of those waters, e.g., small, medium, and large volume lakes. Because the States are all represented on

the RTAG, they are fully involved in the process.

• The data are reviewed for quality and utility and then the distribution of data points throughout

the ecoregion for each class is assessed and additional data gathered if needed.

• When satisfied with the adequacy of the data distribution for the classes, the reference sites

within each ecoregion are compared. If there are obvious shifts in reference values (e.g. through

cluster analysis) the ecoregion is subdivided accordingly or perhaps boundaries are shifted. The

same assessment should be made for temporal distribution to determine if seasonal criteria are

needed. Both of these divisions should help reduce variability in the reference condition as well,

albeit with the risk of reducing the population of applicable observations.

• In the process, the RTAGs are expected to coordinate with their adjacent counterparts to promote

consistent subregional boundaries and criteria. The EPA Headquarters nutrient criteria group

will play a mediating and coordinating role in this process, but the initial determinations will be

made by the RTAGs.

• The established reference conditions will then be incorporated with the other elements of criteria

development—historical perspective, possible modeling of data, and concern to protect

downstream waters—by the RTAG to set that particular ecoregional criterion for TP, TN,

chlorophyll a, and Secchi depth or similar measure of organic based turbidity.

Nutrient Criteria—Lakes and Reservoirs ix

These ecoregional criteria would typically serve as the basis for proposing and promulgating a water

quality standard when a State or Tribe fails to adopt an acceptable standard. EPA expects the States and

Tribes of the Continental United States to develop nutrient criteria for each class of surface water bodies

within three years of the establishment of the ecoregional criteria for those waters. It should be noted

that States and Tribes may elect to establish their criteria using methods other than those described in

these EPA guidance manuals. EPA promotes such flexibility so long as the proposed alternative is:

• Based on a scientifically defensible approach

• Contains sufficient parameters to address nutrient overenrichment causes and responses, i.e.,

consistent with the variables designated by EPA and with the five nutrient criteria elements listed

above

• Protects and maintains downstream water quality sufficient to preserve the beneficial uses of

those waters. In addition to criteria to protect the uses, States must adopt antidegradation

policies and procedures to protect and maintain existing water quality.

Hawaii, Alaska, and U.S. Trust Territories will develop separate ecoregions in conjunction with their

RTAGs and the National Nutrient Criteria Program.

This manual concludes with chapters describing data models, and management options available to

the States and Tribes to actively protect or restore their lake resources. Case histories illustrating nutrient

criteria development and management efforts are also appended with the names of individual specialists

to contact for more information.

Editorial Note

Throughout this text, reference is made to the roles and responsibilities of “States” or “States and

Tribes.” This term or phrase is intended to mean those jurisdictions with the appropriate responsibility

and authority and may also include the District of Columbia, Territories, or other governmental entities.

x Nutrient Criteria—Lakes and Reservoirs

CHAPTER 1

A. Purpose of This Document

B. Relationship Between Water Quality

Standards and Criteria

C. Uses of Nutrient Criteria

Introduction

D. Overview of the Nutrient Criteria

A. Purpose of This Document

Nutrient overenrichment is a major source of water pollution in the United States. The link between

eutrophication—the overenrichment of surface waters with plant nutrients—and public health risks has

long been presumed but not definitively documented (Bachman, personal communication, 1999).

However, corollaries with bacterial indicators such as (1) Escherichia coli and the spread of disease in

sewage-enriched waters, (2) trihalomethanes in chlorine-treated eutrophic reservoirs, and (3) recent

concerns over the incidence of Pfiesteria piscicida in eutrophic estuarine surface waters with suspected

attendant human illnesses, all suggest that overenrichment pollution is not only an aesthetic, aquatic

community problem, but a public health problem as well. A major element of the U.S. Environmental

Protection Agency’s (EPA

�

s) National Nutrient Strategy for the Development of Regional Nutrient

Criteria (U.S. EPA, 1998), also referred to as the National Nutrient Strategy, is the development of water-

body-type-specific technical guidance documents that can be used to assess potential nutrient-related

trophic state impairment and to develop region-specific nutrient criteria to help address this pollution

problem. This document provides this guidance for lakes and reservoirs. A similar document is being

prepared for rivers and streams, and future documents will be prepared for estuarine and coastal marine

areas and for wetlands.

Because of diverse geographic and climate conditions, single national nutrient criteria for lakes and

reservoirs are not appropriate. Instead, nutrient criteria must be developed at the State, regional, or

individual water body levels. This document, therefore, does not attempt to set national criteria, but

provides State and tribal water quality managers with guidance on how they can set criteria themselves.

The document provides background information on classifying water bodies and selecting variables that

can potentially be used as criteria, and it describes methods for developing appropriate values for these

criteria. The document also provides information on sampling, data processing, and appropriate

management techniques.

Because nutrient overenrichment consistently ranks as one of the top causes of water resource

impairment, this initiative is designed to address that particular water quality problem. It is important to

recognize what is meant by nutrient overenrichment. In the context of this guidance manual,

overenrichment means the addition of nutrients causing adverse effects or impairment to designated

use(s) of the water body or to the ecosystem. Symptoms of such impairment include but are not limited

to frequent nuisance algal blooms, fish kills, overabundance or decline of macrophytes, and loss of top

predators from the food chain.

It is also important to recognize that the best way to manage for nutrient control is to reduce the

human-caused fraction of the nitrogen, phosphorus, or related nutrients entering the waters. This often is

referred to as cultural eutrophication to distinguish this enrichment from the inherent nutrient load

entering the water body from soils and parent material indigenous to the area in the absence of disruptive

Nutrient Criteria—Lakes and Reservoirs 1-1

erosion. Cultural eutrophication results from such human endeavors as construction activities, sewage

discharges, agricultural practices, and residential development. This guidance manual is intended to help

the user develop criteria useful for abating cultural eutrophication.

B. Relationship Between Water Quality Standards and Criteria

States and authorized Tribes are responsible for developing water quality standards to protect the

physical, biological, and chemical integrity of their waters. A water quality standard defines the quality

goals for a water body by designating specific uses of a water body, setting criteria to protect those uses,

and establishing an antidegradation policy to protect existing water quality. The uses of a water body

include “existing uses” that were attained on or after November 28, 1975 (the date of the promulgation

by EPA of the first water quality standards regulations) and “designated uses,” which are desired uses

that may or may not already be attained. At a minimum, a water body

�

s uses must include recreation in

and on the water and propagation of fish and wildlife unless the State performs, and EPA approves, a use

attainability analysis justifying a different designated use. Other specific use categories such as boating,

trout propagation, or potable water supply also may be adopted.

After designating the uses of a water body, the State must adopt numeric and/or narrative criteria to

protect and support the specified uses (33 USC § 1313 (c) (2)). Such criteria must be based on a sound

scientific rationale and must contain sufficient parameters to protect the designated use(s). Narrative

criteria describe the desired water quality conditions in a qualitative context. They are the basis for water

quality assessments. An example is shown below:

All waters shall meet generally accepted aesthetic qualifications, shall be capable of supporting

desirable aquatic life, and shall be free from substances, conditions, or combinations thereof

attributable to human activities that produce objectionable color, odor, or taste or induce the

growth of undesirable aquatic life.

Numeric criteria, on the other hand, are quantitative values assigned to measurable components in the

water body. An example of a numeric criterion might be that a lake

�

s average total phosphorus

concentration should “not exceed 20 µg/L during the summer growing season.” Narrative and numeric

criteria should work in combination to:

• Form the basis for consistent measurement of environmental quality

• Provide distinct interpretations of acceptable and unacceptable conditions that can be debated by

concerned parties

• Reduce ambiguity for management and enforcement decisions

This document deals specifically with the establishment of nutrient criteria for lakes and reservoirs

(under the authority of the Clean Water Act Section 304) as a means of addressing nutrient

overenrichment problems. However, for these types of criteria to be effective, they should be

accompanied by responsive nutrient management approaches.

The EPA water quality standards regulations are at 40 CFR Part V31, and guidance on their implementation is in the EPA water

quality standards handbook (EPA-823-B-94-00Sa).

1-2 Nutrient Criteria—Lakes and Reservoirs

A responsible nutrient management plan should meet three practical conditions. First, the plan and

its component elements must be scientifically defensible, otherwise it might lead to well-intentioned

management actions that are unnecessary or harmful. This is like the admonition to physicians “above

all do no harm.” Second, effective nutrient management must strive to be economically feasible. The

public and local affected interests are likely to support approaches that are economically feasible and that

provide meaningful benefit compared with their cost. Finally, these approaches should be practical and

acceptable to the communities involved. The approaches should address appropriate social and political

issues, such as conflicts that might exist between public agencies and landowners or between watershed

residents and lake users. Any management plan may fail if these three general elements are not

sufficiently addressed, and it is almost certain to fail if they are all ignored.

C. Uses of Nutrient Criteria

1. Identification of Problems

EPA expects that the process of collecting current data and surveying more lakes and reservoirs than

have been investigated previously will produce new information revealing conditions not heretofore

recognized. By comparing the water quality criteria for nutrients to actual water quality, the resource

management decisionmakers may well recognize overenriched lakes or reservoirs or portions of these

water bodies for the first time. These new problems can be incorporated into the information system so

that remediation can be initiated.

2. Management Planning

The nutrient criteria development process not only establishes these benchmarks identifying

overenrichment, but it also makes it possible to rank the relative magnitude of the problems with respect

to each other. A scale of overenrichment with a frequency distribution can be created to readily identify

the scope of the enrichment problems to be addressed and the numbers of lakes or impoundments in each

state of degradation. Modeling plays a significant role here either to supplement existing data sets or to

assess the projected effect of various options and combinations of management approaches.

Thus, a form of triage can be practiced to assign scarce manpower and funds in an efficient way. For

example, a State may create a balance by (1) protecting many high-quality lakes, (2) restoring several

moderately degraded lakes by implementing cost-effective land use changes early on, and (3) designating

for restoration one or a few badly overenriched systems, realizing that only a long-term, protracted

project and budget will suffice.

3. Regulatory Assessments

Much of the management work done by EPA and the States is regulatory, and the nutrient criteria,

once established, should be incorporated into State standards to become the basis of enforceable tools.

These values are used to develop limits in National Pollutant Discharge Elimination System (NPDES)

permits for point source discharges. The permit limits for nitrogen, phosphorus, and other trace nutrients

emitted from waste water treatment plants, factories, food processors, and other dischargers can be

appropriately adjusted and enforced in accordance with the criteria.

Similarly, total maximum daily load (TMDL) estimates used to allocate remediation responsibilities,

especially regarding nonpoint sources on a watershed basis, can be established with respect to these

nutrient criteria. Knowing the optimum nutrient load for a lake (and its downstream recipient waters)

Nutrient Criteria—Lakes and Reservoirs 1-3

makes it possible to divide and allocate that load among the tributary subwatersheds of the system.

Resource managers then can begin land use improvements and other activities necessary to improve the

system in a methodical way and on a reasonable scale so that restoration can be achieved.

The criteria portion of water quality standards also may be used in antidegradation reviews and can

serve in the development of best management practices for State and local nonpoint source programs.

4. Project Evaluations

Nutrient criteria can be applied further to evaluate the relative success of management activities such

as those described immediately above. Although it may sometimes be expensive and frustrating, “before,

during, and after” measurements of nutrient enrichment variables in the receiving waters, when compared

with the criteria, provide an objective and direct assessment of the success of the management project.

5. Status and Trends of Water Resources

Throughout the continuing process of problem identification, response and remediation, and

evaluation to protect and enhance our water resources, the States and EPA are required by section 305(b)

of the Clean Water Act to periodically report to Congress on the status of the Nation

�

s waters. The

nutrient criteria would expand and refine that report by adding an additional set of both causal and

response parameters to the measurement process. The States and EPA will be able to compare the

measured enrichment conditions of their lakes and reservoirs and document the changes that have

resulted and the relative progress made.

The rest of this guidance manual presents detailed information that elaborates on this important

material. The intent is to present essentially a two-part guidance document, the first half of which is a

presentation of the science and technology associated with the measurements required and processes

associated with the development of the benchmark nutrient criteria needed to make enrichment

identifications. The second part addresses the equally important process of making management

decisions to protect and enhance the trophic state of our Nation

�

s waters and to evaluate the relative

success of that management so we can know what works and what does not, so that the next round of

criteria development and management will be conducted from a truly expanded base of knowledge.

D. Overview of the Nutrient Criteria Development Process

A distribution of lakes may exhibit a range of nutrient conditions. Using total phosphorus as an

example, some lakes may have little or no enrichment and consequently a limited number of species and

individuals or biomass. Lake Superior is a classic example of such conditions of oligotrophy. At the

other extreme of phosphorus enrichment is massive overenrichment with so much phosphorus in the

water column that algal blooms or choking macrophyte growth and frequent fish kills are common.

Species diversity in these hypereutrophic lakes is also low even though biomass is usually very high.

These are the “pea soup” lakes most communities associate with badly degraded conditions.

Phosphorus concentrations typical for both extremes can be measured, but an in-lake total

phosphorus concentration of less than 10 µg/L generally is considered to be oligotrophic. Conversely,

100 µg/L often is used as the threshold for hypereutrophication (Vollenweider, 1968; Wetzel, 1975;

Carlson, R., personal communication, 1999). Although such levels are known to exist naturally, more

often concentrations of this magnitude are associated with extensive or intensive cultural development.

1-4 Nutrient Criteria—Lakes and Reservoirs

Natural enrichment ranges throughout these magnitudes of concentration according to geographic

and geological regions of the country. Consequently, it would be necessary to determine the natural

ambient background for each lake so that the eutrophication caused by human development and abuse

can be addressed. Addressing this cultural eutrophication is the objective of this manual, but the

development of nutrient criteria on a lake-by-lake basis may be prohibitively time consuming and

expensive for States and Tribes.

Alternatively, these lakes or reservoirs can be divided into regionally similar groups based on their

physical characteristics within a proximal geographic area. Those lakes of each established group having

the least land development or other human impact can be identified as the reference lakes for measuring

relatively undisturbed nutrient conditions appropriate for that class and region. This reference condition

information, within an appropriate historical context and objectively interpreted, then can become a

candidate criterion for use as a benchmark against which other similar lakes may be compared. Before

the criterion is finally established, however, the scientists and resource managers involved should assure

themselves that it also will have a beneficial or at least neutral downstream effect on the lakes, reservoirs,

streams, or estuaries within or just below the area of application. This concept, as illustrated in Figure

1.1, is essentially the basis for the National Nutrient Criteria Program and is described variously

throughout this text.

Figure 1.1. Conceptual basis for the National Nutrient Criteria Program using TP as an

example variable.

Nutrient Criteria—Lakes and Reservoirs 1-5

The two extreme values of hypereutrophy and pristine or presettlement conditions can be estimated

from monitoring, historical records, and paleolimnological determinations. The reference condition,

located within this measured range, and the derived criteria are scientifically based. But they also

include a conscious decision to use areas of least human impact as indicators of low cultural

eutrophication. A measure of best practical judgment is also necessary where scientific methods and data

are not adequate for the decisions necessary for water resource protection.

The use of minimally impacted reference sites has been adapted from biological criteria development

and is endorsed by EPA’s Science Advisory Board (1992). Conditions that represent minimal impacts

provide a baseline that should protect the inherent beneficial uses of the Nation’s waters. The use of

scientific rationale together with practical resource management is called for in the selection of a

percentile distribution of values as a reference condition. The term “minimally impacted” implies a high-

end percentile of conditions in reference lakes and a low-end percentile of the conditions in all lakes (i.e.,

some enrichment is allowed, but not enough to cause adverse in-lake effects or adverse downstream

effects). The upper end of the range of data from purposely identified reference sites represents the bare

minimum threshold of a reference condition, whereas lower percentiles of the reference site data

represent high-quality conditions that may not need to be achieved or cannot be achieved in the entire

population of lakes in an ecoregion. The upper 25th percentile represents an appropriate margin of safety

to add to the minimum threshold, excludes the effect of spurious outliers, and serves as a first-order

recommendation of a sufficiently protective value. Data analyses performed to date indicate that the

lower 25th percentile of data from a sample representative of the entire population of lakes in an

ecoregion roughly approximates the upper 25th percentile of the reference data (see Chapter 6, section C,

Minnesota case study). Where sufficient data are available, comparison and statistical analysis of causal

and response variables can help determine effect thresholds and further refine reference conditions.

Establishing the reference condition is but one element of the criteria development process. Reference

condition values are appropriately modified based on examination of the historical record, modeling,

expert judgment, and consideration of downstream effects.

1. Strategy for Reducing Cultural Eutrophication

Six key elements are associated with the strategy for reducing cultural eutrophication (U.S. EPA,

1998):

• EPA believes that nutrient criteria need to be established on a regional basis and need to be

appropriate to each water body type. They should not be established as a single set of national

numbers or values because there is simply too much natural variation from one part of the

country to another. Similarly, the expression of nutrient enrichment and its measurement by

necessity varies from one water body type to another. Streams do not respond to phosphorus and

nitrogen the same way as lakes or coastal waters.

• Consequently, EPA has prepared guidance for these criteria on a water -body-type- and

region-specific basis. With detailed manuals available for data gathering, criteria development,

and management response, the goal is for States and Tribes to be able to conduct surveys and

develop criteria to help them deal with the problem of nutrient overenrichment of their waters.

• To help achieve this goal, the Agency has initiated a system of EPA Regional technical and

financial support operations each led by a Regional Nutrient Coordinator—a specialist

responsible for providing the help and guidance necessary for States or Tribes in his or her

region to accomplish the necessary environmental investigations and remediations. These

1-6 Nutrient Criteria—Lakes and Reservoirs

regional coordinators are guided and assisted in their duties by a team of inter-Agency and

intra-Agency specialists from EPA Headquarters. This team is responsible for providing both

technical and financial support to the Regional Technical Assistance Groups (RTAGs) created by

these coordinators so the job can be completed and communication established and maintained

between the policymaking function in Headquarters and the actual environmental management in

the Regions.

• EPA will develop basic ecoregional nutrient criteria values for water body types. The regional

teams and States/Tribes can use these values as guidance for developing criteria protective of

designated uses; the Agency also may use these values if it elects to promulgate criteria for a

State or Tribe. These criteria will have value in two contexts: (1) as the basis of water quality

standards, NPDES permit limits, and as TMDL target values and (2) as decisionmaking

benchmarks for management planning and assessment.

• EPA plans to provide sufficient information for States and Tribes to begin adopting nutrient

standards by 2003.

• States/Tribes are expected to monitor and evaluate the effectiveness of nutrient management

programs implemented on the basis of the nutrient criteria. EPA intends the criteria guidance to

reflect the “natural,” minimally impaired trophic condition of a given regional class of water

body. Once water quality standards are established for nutrients based on these criteria, the

relative success or failure of any management effort, either protection or remediation, can be

evaluated.

Thus, the six elements of the National Nutrient Criteria Program describe a process that encompasses

taking measurements of the collective water resources of an area; establishing nutrient criteria that can

be used for evaluating the discrete waters within that region or area; assessing individual water bodies

against these criteria and associated standards; designing and conducting the appropriate management;

and, finally, evaluating its relative success.

2. Nutrient Criteria Development Process

Provided below is a discussion of the activities that generally comprise the nutrient criteria

development process. They are listed in the order generally followed and the subsequent chapters of this

document follow this sequence. Figure 1.2 presents a schematic illustration of the criteria development

process with parallel, appropriate chapter headings.

�

Preliminary Steps for Criteria Development (Chapter 3)

Establishment of Regional Technical Assistance Groups

The Regional Nutrient Coordinator in each EPA Region will contact and obtain the involvement of

key specialist (e.g., limnologists, water resource managers, oceanographers, stream and wetland

ecologists, water chemists, and land use specialist) in that Region with respect to the water bodies of

concern, and these experts should be recruited from other Federal agencies, State agencies, universities

Nutrient Criteria—Lakes and Reservoirs 1-7

Figure 1.2. Flowchart of the nutrient criteria development process.

1-8 Nutrient Criteria—Lakes and Reservoirs

and colleges. Particular Federal agencies of interest are the U.S. Geological Survey (USGS), Natural

Resources Conservation Service (NRCS), National Oceanic and Atmospheric Administration (NOAA),

U.S. Forest Service (USFS), and the U.S. Fish and Wildlife Service (USFWS). In certain areas of the

country, the U.S. Army Corps of Engineers (USACOE) or the Bureau of Land Management (BLM) or

special government agencies such as the Tennessee Valley Authority (TVA) may be pertinent. Similarly,

for information and education activities, especially with respect to agriculture, the USDA Cooperative

Extension Service is a valuable resource. State agencies with responsibilities relevant to this effort are

variously named, but are commonly referred to as: Department of Natural Resources, Department of

Water Resources, Department of the Environment, Department of Environmental Management, Fisheries

and Wildlife Management, State Department of Agriculture, State Department of Forestry, and other land

use management agencies. Most State land grant universities have faculty talent important to nutrient

management, and almost all colleges and universities have applied science faculty with research interests

and talents appropriate to this initiative. In selecting participants for the group, diverse expertise is an

obvious prerequisite, but willingness to cooperate in the group effort, integrity, and a lack of a strong

alternative interest are also important factors to consider for selecting these essential people who must

make collective and sometimes difficult determinations.

The experts chosen will constitute the RTAG, which will be responsible for major decisions in

regional implementation of the program. And the group should be sufficiently large to have the

necessary breadth of experience, but small enough to effectively debate and resolve serious scientific and

management issues. A membership of about 30 approaches an unwieldy size, although that number may

initially be necessary to maintain an effective working group of half that size.

The RTAG is intended to be a regional, Federal agency advisory body consisting of a viable subset of

scientists and resource managers from each pertinent agency as described above together with their State

counterparts. The RTAG has a Federal responsibility and as such should not delegate or share this

obligation with the private sector. The perspectives of private citizens, academicians, and special interest

groups are important, and these and other members of the public may attend RTAG meetings and offer

opinions when invited, but the final deliberations and decisions are the responsibility of the Federal and

State members of the RTAG. They must also be able to meet and debate the issues without undue

outside influence.

However, as a matter of policy, EPA encourages the RTAGs to regularly provide access and reports

to the public. The meetings should generally be open to the public and the schedule of those meetings

published in the local newspapers. At a minimum, RTAGs are encouraged to hold regular “stakeholders”

meetings so that environmental, industrial, and other interests may participate via a separate public forum

associated with responding to the group’s efforts. It is important that citizens and public groups be

involved, and any significant determinations of the Group should include a public session at which a

current account of activities and determinations is presented and comments acknowledged and

considered. In addition, where specific land uses or practices are addressed, those property owners,

farmers, fishermen, or other involved parties should be consulted in the deliberation and decisionmaking

process.

It is reasonable to expect monthly or at least quarterly meetings of the RTAG, with working

assignments and assessments conducted between these meetings. To coordinate activities among the 10

RTAGS, and with the National Nutrients Team, regular conference calls have been established. At these

sessions, new developments in the Program, technical innovations and experiences, budgets, and policy

evolutions will be conveyed and discussed. In the same context, an annual meeting of all Regional

Nutrient Coordinators, State representatives, and involved Federal agencies is also held each spring in or

Nutrient Criteria—Lakes and Reservoirs 1-9

near Washington, DC. At this meeting, major technical reports are presented by specialists and issues

significant to the Program are discussed.

The composition and coordination discussed above are intended to establish the shortest possible line

of communication between the State, Region, and National Program staff members so that a rapid but

reasoned response is promoted to changing issues and techniques affecting nutrient management of our

waters. It is also designed to be responsive to the water resource user community without becoming a

part of user conflicts.

Delineation of Nutrient Ecoregions Appropriate to the Development of Criteria

The initial step in this process has been taken through the creation of a national nutrient ecoregion

map consisting of 14 North American subdivisions of the coterminous United States (Figure 1.3). These

are aggregations of Level III ecoregions revised by Omernik (1998). Alaska, Hawaii, and the U.S.

Territories will be subdivided into nutrient ecoregions later, with the advice and assistance of those

States and governments.

The initial responsibility of each RTAG will be to evaluate the present ecoregional map with respect

to variability based on detailed observations and data available from the States and Tribes in that EPA

Region. This preliminary assessment of the nutrient ecoregional boundaries will further depend on the

additional nutrient water quality data obtained by those States. The databases, especially with respect to

selected reference sites, will be used to refine the initial boundaries of the map in each EPA Region.

It is expected that the collective effect of these evaluations by all 10 EPA RTAGs will result in the

further refinement and subdivision of many of the 14 ecoregions, especially the large, multi-State ones.

The boundaries will shift or be subdivided in accordance with the inherent trophic conditions and

nutrient indicators of similar water bodies in each locality.

Physical Classification

The next step in evaluating the data is to devise a classification scheme for rationally subdividing the

population of lakes in the State. Because identification of overenrichment is the objective of nutrient

criteria development, trophic classification per se should be avoided, as should any classification based

on levels of human development.

Physical characteristics independent of most cultural enrichment sources are far more appropriate.

Such classification may be done initially on a size basis (e.g., acres of surface area or square miles of

watersheds). A volumetric variable that may be used for further subclassification based on median or

maximum depth. Similarly, inherent water quality characteristics such as marl or bog lakes may also

apply. In fact, such lakes, especially if few in number, are usually separated out of the general population

and identified as a separate and unique class. Hydroelectric reservoirs and effluent dominated systems

are also examples.

Once lakes have been classified, it is important to determine how much information is available

describing the enrichment status of these lakes. State agency records are the basis for an initial data

1-10 Nutrient Criteria—Lakes and Reservoirs

Figure 1.3. Draft aggregations of level III ecoregions for the National Nutrient Strategy.

Nutrient Criteria—Lakes and Reservoirs 1-11

search. In many States, water quality information resides in more than one agency. For example,

Maryland has a Department of Natural Resources and a Department of the Environment, both of which

retain water quality records. To compound the search further, States may also have pertinent data sets in

their Departments of Fisheries and Public Health. It is wise to initiate the search for information with

calls and questionnaires to colleagues in the State or Tribal agencies likely to be involved so an

appropriate list of contacts and data sets can be compiled. In doing so, regional Federal agencies should

not be overlooked either. These include the USFWS, Park Service, and U.S. and State Geological

Surveys.

�

Establishing an Appropriate Database (Chapter 4)

Review of Historical Information

Historical information is important to establish a perspective on the condition of a given waterbody.

Has its condition changed radically in recent years? Is the system stable over time? Has there been a

trend up or down in trophic condition? Only an assessment of the historical record can provide these

answers. Without this information, the manager risks setting reference conditions and subsequent criteria

on the basis of present data alone, which may in fact be a degraded state. Valid historical information

places the current information in its proper perspective.

Data Screening

The first step in the process of either assessing historical observations and data sets or more current

data is to review this material to determine the suitability of that information to support nutrient criteria

development. Anecdotal information and observations are valuable, but the sources must be carefully

considered. Fishermen’s accounts, local sport fishing news stories, and the observational logs of

scientific field crews are all legitimate sources of information, but they are subject to different levels of

scrutiny before a trend is determined. The same applies to different databases. Nutrient information

gathered for the purpose of identifying failing waste water treatment plants cannot be assessed in the

same light as similar data collected to determine overall lake quality or trophic state. The analytical

procedures used, type of sampling design and equipment, and sample preservation are other variables that

must also be considered in any data review and compilation. Once this screening is done, the compiled

data may be sorted according to named lakes or reservoirs.

Nutrient Data Collection and Assessment

EPA has initiated the data collection and assessment process by screening the existing STORET

database for information on lakes, reservoirs, streams, and coastal waters with respect to four initial

parameters of concern: total nitrogen (TN), total phosphorus (TP), chlorophyll a, and Secchi depth.

These four parameters were originally selected for robustness and conservativeness of estimation;

however the preliminary screening of the STORET data revealed that these measurements are also

relatively abundant in the database. Although this is an entirely appropriate starting point for nutrient

criteria development, States and Tribes are not required to confine their investigations and data selection

for enrichment assessment and criteria development to only these variables.

States and Tribes are encouraged to select measures above and beyond these initial characteristics

that contribute to the most appropriate and reliable assessment of the enrichment of the waters of their

region. In particular, it is advisable to use both causal indicators (the nutrients introduced to the system

especially species of nitrogen and phosphorus, and perhaps silica and carbon as indicated) and response

1-12 Nutrient Criteria—Lakes and Reservoirs

indicators (those measures of biotic productivity and activity reflecting the enrichment of the system)

including chlorophyll a; Secchi depth; turbidity; algal taxa; plankton taxa; dissolved oxygen; macrophyte

taxa, extent, and biomass; and fish taxa and numbers.

The combination of nutrient and biological system response information will yield the most

definitive and comprehensive criteria. To use only causal or only response variables in the criteria leaves

the State or Tribe in jeopardy of not protecting the waters from overenrichment. For example, an

offensive water body covered with an algal scum may be low in the causal variables of reactive nitrogen

and phosphorus because they are tied up in biomass (in fact, TN and TP were selected by EPA to avoid

this problem). Therefore, the lake in question may meet those criteria, but not its designated or existing

use. The converse may also occur, in which a highly enriched system with a rapid flushing rate appears

to be acceptable when only the biota and dissolved oxygen are measured, but the load of nutrients being

delivered downstream is degrading the receiving waters. Using a balanced combination of both causal

and response variables in the criteria together with careful attention to seasonal variability should

mitigate against these false positive and false negative results.

�

Candidate Variables for Criteria Setting (Chapter 5)

EPA is beginning the National Nutrient Criteria Program with a survey of national computerized data

sets such as STORET and NAWQA for TP, TN, chlorophyll a, and Secchi depth. These are believed to

be the most common variables recorded with respect to enrichment investigations. The information will

be screened for suitability and then plotted o n regional maps of the United States for use by the Regional

Nutrient Coordinators and RTAGs described above, and by the States. This being the case, it is

reasonable for individual States and Tribes to begin with the same four indicators, although other causal

and response variables are also discussed later in this manual (see Chapter 5).

�

Establishing Reference Conditions (Chapter 6)

Candidate reference lakes can be determined from compiled data and with the help of Regional

experts familiar with the lake resources of the area. There are two recommended ways to go about this.

One is to select those lakes believed to be minimally impacted by human activity (e.g., with little or no

riparian or watershed development). These lakes should be reviewed and visited to confirm their

“natural” status. When satisfied with this list, a median value (adjusted for seasonal and spatial

variation) for TP, TN, chlorophyll a, Secchi depth, and other appropriate enrichment indicators can be

prepared for each lake based on existing and/or new data collections. The upper 25th percentile of the

frequency distribution of these reference lakes can then be selected as the reference condition for each

value (because these lakes represent the best obtainable and most “natural” condition, some allowance

for variation should be made) (Figure 1.4(a)).

Another option is to plot the frequency distribution of all of the lake data presently available by each

variable and selecting percentiles for TP, TN, chlorophyll a, Secchi depth, and other similarly

appropriate variables. The lower 25th percentile, reflecting high nutrient quality can be selected as the

reference condition for each value (because in this instance the pool of information likely includes lakes

of considerably less than “natural” trophic condition) (see Figure 1.4(b)).

Nutrient Criteria—Lakes and Reservoirs 1-13

Figure 1.4. Two approaches for establishing a reference condition value using total phosphorus as

the example variable.

The choice of the upper 25th and the lower 25th percentiles for the selected reference lakes and the

random sample reference or census of all lakes in a class, respectively, is a rational but qualitative

decision. It represents the effort to avoid imposing an undue penalty on high-quality mesotrophic lakes

in regions where the lakes are predominantly oligotrophic. By selecting an upper percentile of the

reference lakes, there is a greater likelihood that more of the broader population of lakes will comply.

Conversely, in regions of intense cultural enrichment, a lower percentile of the distribution of the

remaining lakes used as reference must be selected to avoid establishing criteria based on degraded

conditions. The quarterly increments were chosen as a reasonable division of the data sets recognizable

by the public, and the upper 25th percentile and lower 25th percentile as reasonable and traditional

fractions of the range and frequency of distribution. This approach promotes water quality enhancement

and has broad application over the country.

Although these quantitative values are believed appropriate to the objective of the program, we

recognize that some variation about such percentiles may be necessary. Certainly, in severely degraded

areas even a 25th percentile may be insufficient, and some lower fraction of the remaining reference

values may be required. On the other hand, where all lakes or reservoirs are in remarkably good

condition relative to cultural enrichment, the acceptable fraction of the reference condition may be

justifiably increased. The key point here is the presentation of a defensible scientific rationale for the

determination. Otherwise, EPA presumes the above guidance will be appropriate.

It is intended that these two frequency distributions, with different quartiles, will produce a similarly

appropriate reference condition—all other factors being equal. In either case, a number is generated that

can be used as an initial reference preliminary to criteria development, and as a source of comparison for

individual lakes in the class.

1-14 Nutrient Criteria—Lakes and Reservoirs

This is the beginning of the process that eventually leads to adoption of nutrient criteria as part of the

State or Tribal water quality standards. Other factors that must be addressed along the way are gaps in

the database that must be filled by additional data collections, possible biases in the data or data

interpretation (especially if the information was originally collected for another purpose such as fishery

management or waste water investigations), sampling errors by field teams, and equipment changes or

measurement errors introduced by changes in analytical techniques.

�

Nutrient Criteria Development (Chapter 7)

Nutrient Criteria Components

The move from data review and data gathering to criteria development involves a sequence of five

interrelated elements:

• Examination of the historical record or paleolimnological evidence

• Compilation of reference condition data

In situations where a class of lakes in question are all significantly impaired and none can be

perceived as approximately “natural,” then the best quality remaining constitutes the present day

example of a reference condition. In this instance the reverse of the earlier example of

preselected minimally impaired reference lakes is true. Because most of the chosen lakes are

assumed to be at least somewhat degraded, a lower percentile should be selected as a basis for

the reference condition (e.g., the 25th percentile). To do otherwise is to ultimately lower the

criteria to the level of present degradation and no restoration of the overenrichment condition

will be achieved.

Remember that the present day reference condition is only part of the criteria development

process; historical conditions, data extrapolations, and the best objective judgment of the RTAG,

including concern for downstream impacts are the other components that will collectively

establish the criteria.

• In some instances empirical modeling or surrogate data sets may be used where insufficient

information exists. This may be the case especially with reservoirs or significantly developed

watersheds.

• The objective and comprehensive interpretation of all of this information by a panel of specialists

selected for this purpose (i.e., the RTAG). These experts should have established regional

reputations and expertise in a variety of complimentary fields such as limnology, ecology,

nutrient chemistry, and lake management.

• Finally, the criterion selected should first meet the optimal nutrient condition for that class of

waterbody in the absence of cultural impacts and protect the designated use of that waterbody.

Second, it must be reviewed to ensure that the level proposed does not entail adverse nutrient

loadings to downstream waterbodies. In designating uses for a waterbody and developing criteria

to protect those uses, the State or Tribe must take into consideration the water quality standards

of downstream waters. This concern for downstream effect can be extended all the way to

coastal waters, but in practice the immediate downstream receiving waters will be the area of

greatest attention for the resource manager. This impact supersedes the level of optimal

Nutrient Criteria—Lakes and Reservoirs 1-15

enrichment for the target lake waters. If a downstream impact is expected, the criteria for that

lake or class of lakes should be revised downward accordingly.

Once the initial criteria (either Regional or State/Tribal) have been selected, they can be verified and

calibrated by testing the sampling and analytical methods and criteria values against waterbodies of

known conditions. This ensures that the system operates as expected. This can be accomplished either

by field trials or by use of an existing database the quality of which has been assured. This process may

lead to refinements of either the techniques or the criteria.

It should be noted that criteria may be developed for more than one parameter. For example, all

reference lakes of a given class may be determined to manifest characteristics of a particular level for TP

concentration, TN concentration, chlorophyll a, and Secchi depth. These four measures will comprise

four criteria levels appropriate to optimal nutrient quality. EPA expects a given test lake to meet or

surpass these levels for at least TN and TP and one of the two response variables, and that a scientifically

valid explanation will be derived for the remaining exception before it can be determined to meet the

criteria. The policy for such application will be developed by the State or Tribe in consultation with

EPA. The point here is that these four (or more) parameters used in this illustration are expected to be

interrelated, and a consistent response for most if not all of them gives a level of confidence to the

resource manager that he has evaluated the lake properly.

When the lake in question reveals high TN and TP concentrations, but not the expected high

chlorophyll a or low Secchi depth measurements, further investigation is indicated before deciding on

whether criteria have been met. Flushing rates or inorganic turbidity or water color may be additional

factors influencing the condition of the lake.

Assessing Attainment with Criteria

A rule of compliance is then established for the criteria that have been selected for each indicator

variable. The four initial variables include two causal variables (TN and TP) and two response variables

(chlorophyll a and Secchi depth or a similar indicator of turbidity). Failure to meet either of the causal

criteria should be sufficient to prompt action. However, if the causal criteria are met, but some

combination of response criteria are not met, then there should be some form of decisionmaking protocol

to resolve the question of whether the lake in question meets the nutrient criteria or not. There are two

approaches to this:

• Establish a decisionmaking rule equating all of the criteria

• Establish an index that accomplishes the same result by inserting the data into an equation that

relates the multiple variables in a nondimensional comprehensive score

With regard to more stringent State or Tribal criteria or standards, Agency policy on antidegradation

generally requires that no lake or reservoir be allowed to degrade below its existing condition regardless

of designated use or State or Regional criterion. (See Chapter 7, Section F, “Maintaining Existing Water

Quality.”) This protects against the degradation of unique lakes of higher relative nutrient quality than

might be stipulated in State or regional nutrient criteria.

1-16 Nutrient Criteria—Lakes and Reservoirs

�

Management Response (Chapter 8)

There are a variety of management responses possible to the overenrichment problem identified by

the use of nutrient criteria. Chapter 8 describes a 10-step process that permits the resource manager to

evaluate and select the best of these approaches to accomplish improvements in water resource condition.

The emphasis is on developing a scientifically responsible, practical, and cost-effective management

plan.

The chapter also describes three basic categories that encompass all management activities:

education, funding, and regulation. It closes with the admonition to always carefully evaluate the relative

success of the management project, report results, and continue monitoring the status of the water

resource.

Finally, Chapter 9 and the appendices offer illustrations of the uses of models in the nutrient

management process, a narrative description of the nutrient ecoregion map as presently developed, and

examples of lake and reservoir nutrient management experiences. Future editions of this manual will

incorporate actual Regional criteria values and State/Tribal accounts of their use of the manual for

nutrient criteria development.

Nutrient Criteria—Lakes and Reservoirs 1-17

CHAPTER 2

The Basis for Lake and

Reservoir Nutrient

Criteria

A. Historical Perspective

B. The Nutrient Paradigm

C. Connecting Watershed Loading to the Lake:

A Mass Balance Model Approach

D. Trophic State Classification S

ystems

E. Uses of Trophic State Indices

Large-scale comparative studies of lakes have enabled scientists to identify key variables that

influence lake structure and processes (Peters, 1986). This empirical approach has its roots in regional

studies (e.g., Naumann (1929) and Thieneman (1921)) and in early among-lake comparisons of lake

function such as the role of morphometry on lake productivity (Rawson, 1955) and nutrient input

regulation of lake fertility (Edmondson, 1961).

The eutrophication process was quantified by Vollenweider (1968, 1975, 1976), which brought the

large-scale comparative approach to the forefront of limnology. Vollenweider developed a mass

balance model using literature data from a diverse population of temperate lakes to demonstrate a

surprisingly strong relationship between nutrient inputs to lakes and concentration of nutrients within the

lake. This relationship was sufficiently powerful to stand out against other sources of among-lake

variation and signaled that nutrient loading, as modified by hydrology, morphology, and in-lake

sedimentation, was the dominant factor explaining lake eutrophication.

The element phosphorus was the focus of study because overwhelming evidence suggested that

phosphorus limited algal growth in many aquatic systems (Schindler, 1977). Phosphorus values were

highly correlated with algal biomass in lakes (Sakamoto, 1966); in turn, water clarity was shown to vary

with algal levels (Edmondson, 1972). With these linkages quantified, the science of lake management

arose around the premise that reductions in nutrient loads would reverse eutrophication, as measured by

reduced nutrient concentrations, algal levels, and greater water clarity.

Empirical models provided limnologists with a quantitative basis for estimating the level of response

to be expected from a given change in nutrient load from point and nonpoint sources. Models were the

tools for forecasting the capacity of a lake to withstand change in its trophic state with various degrees of

human development in its catchment (Dillon and Rigler, 1975). Recently, land use, as a surrogate for

external nutrient loading, was used to effectively predict algal chlorophyll in lakes (Meeuwig and Peters,

1996); the strength of this approach stems from the strong correlation between nutrient losses and land

use practices in catchments (Smart et al., 1985). This linkage of land use to chlorophyll, a widely

accepted measure of lake trophic state, is additional evidence for the importance of external control on

lake processes.

The large-scale comparative approach placed individual lakes within a continuum, from least to most

fertile. With this understanding, lakes lost some of their individuality because scientists now viewed

them within the context of this continuum. The functional relations between external nutrient loading,

algal biomass, and water clarity were summarized in a small number of general models. These models

were typically based on regression analyses of data from individual lakes, averaged over a sampling

Nutrient Criteria—Lakes and Reservoirs 2-1

season. The models quantified large-scale lake functions and provided the conceptual basis for lake

management and restoration.

Because these early data were drawn from a diverse group of lakes, both in terms of lake type and

geographic location, these models often are referred to as “global models.” An underlying assumption

was that processes responsible for the large cross-sectional patterns in these global relations also operate

within single systems over time (Prairie and Marshall, 1995). About the time of Vollenweider

�

s work,

Edmondson (1972) applied virtually the same concepts to data covering the enrichment and recovery of

Lake Washington. Edmondson

�

s work was tangible confirmation that a single lake responded to nutrient

loading, as the pattern drawn from the data of many lakes would suggest. The remarkable feature during

the 1970's was that a quantitative paradigm for lake function had been proposed based largely on data

drawn from the literature. It was a synthesis of ideas from earlier descriptive and empirical studies. A

feature of predictions from empirical models was that there was a great uncertainty in them; many

models exhibited an order of magnitude variation. This variation was a point of concern and the focus of

subsequent study.

During the two decades since the empirical period of the mid-1970's, lake management has been

influenced by several major thrusts that have modified, but not invalidated, the work of that period. With

expanded data sets over the past 20 years, the original global generalizations have been modified

showing that in highly enriched lakes algal biomass does not increase in a uniformly linear relationship

to phosphorus in all lakes (McCauley et al., 1989; Prairie et al., 1989; Watson et al., 1992) because other

environmental factors also play a role. The Organization for Economic Cooperation and Development

(OECD, 1982) project was an early effort to systematically gather data and quantify the relationship

between nutrient load in waters and their trophic reaction. This project, composed of four regional

studies (Alpine, Nordic, Reservoir and Shallow Lakes, and North American), attempted to corroborate

Vollenweider

�

s generalizations. Its approach shifted the focus of among-lake comparisons from a global

scale to studies within regions and studies of specific lake types.

Since then, several regional studies have used the comparative approach to generate empirical models

specifically for local conditions. These regional studies have demonstrated the importance of other

factors regulating algal biomass in lakes. Four other factors include nitrogen (Canfield, 1983; Pridmore,

1985), light limitation due to suspended solids (Hoyer and Jones, 1983; Jones and Knowlton, 1993), lake

morphometry (Riley and Prepas, 1985), and grazing by herbivores (Quiros, 1990).

B. The Nutrient Paradigm

The concept of nutrient criteria is based on the idea that nutrients produce changes in lakes and

reservoirs that are considered to be detrimental to the function or use of the water body. This idea of

nutrient control of water body function is not new; it can be traced back to when Einar Naumann, the

Swedish limnologist, elucidated the major part of the nutrient paradigm in 1929. His ideas of the

relationship between nutrients and lakes can be summarized in the following four statements:

• The primary factors that determine algal biomass (the amount of plant organic material) are the

plant nutrients phosphorus and nitrogen.

• The geology (and land use) within the lake

�

s watershed determines the amount of nutrients that

enter the lake and, therefore, plant biomass.

2-2 Nutrient Criteria—Lakes and Reservoirs

• Changes in the plant biomass affect the entire lake

�

s biology.

• There is a natural ontogeny to lakes; the amount of plant biomass and, therefore, the entire

biology of the lake increases as the lake ages.

Although there have been many significant additions and improvements in our understanding of lakes

since Naumann, his original concept of nutrients remains the basis of the nutrient paradigm. Below, each

statement is examined as it refers to the need for and the development of nutrient criteria for lakes.

1. Phosphorus and Nitrogen as Limiting Factors for Algal Biomass

The primary factors that determine algal biomass (production) are the plant nutrients phosphorus and

nitrogen. When Naumann (1929) suggested this concept, he was probably drawing on a much older

concept, Justus von Leibig

�

s Law of the Minimum. The law, as it is formulated today, states that the

factor that is in shortest supply relative to the needs of the plants limits the growth of those plants. The

concept is central to the nutrient paradigm in lakes because it insists that very few factors (usually only

one factor, often a plant nutrient such as nitrogen or phosphorus) will actually limit plant growth.

If only one factor, such as phosphorus, was always limiting, the task of developing nutrient criteria

would be a simple matter of determining limits on that single factor. Unfortunately, the factor that limits

plant biomass may (1) change seasonally or over longer periods of time, (2) vary depending on the land

use, or (3) vary regionally. It would make little sense to construct a single nutrient criterion when that

nutrient may not necessarily limit a target lake or lakes. It is for that reason that the emphasis of this

document is the development of nutrient criteria based on both the nutrient inputs and the biological

response.

The causal variables such as phosphorus and nitrogen are essential criteria because they will be the

limits necessary to establish management objectives and are usually directly related to discharge runoff

abatement efforts. Although phosphorus is the limiting factor for most lakes and reservoirs, in some

regions the nutrient paradigm centers on nitrogen rather than phosphorus, especially where sewage

treatment plant effluent is involved (Canfield, 1983; Pridmore, 1985; Jones et al., 1989). These regions

are often in the subtropics or at high latitude or altitude (Wurtsbaugh et al., 1985; Morris and Lewis,

1988) but are also found in parts of Britain. In these lakes, nitrogen rather than phosphorus explains the

among-lake variance in algal chlorophyll, and chlorophyll-total nitrogen regressions match the “fit and

form” of chlorophyll-total phosphorus regressions developed for phosphorus-limited temperate lakes.

The reason for nitrogen limitation is not yet understood because of a long-held tenet in limnology that

states that nitrogen fixation will compensate for shortfalls (Schindler, 1977) and that nitrogen limitation

is not a persistent condition. This belief does not seem as universal as once thought (Knowlton and

Jones, 1996). In some regions, nitrogen limitation may be a function of abundant phosphorus in the

geological formation of the region (Canfield, 1983).

Nitrogen limitation also may be tied to efficient nitrogen cycling in subtropical forests or may be a

function of nitrogen uptake by rice and other crops in the subtropics. In high-elevation lakes, phosphorus

may be contributed by soil weathering, whereas nitrogen is rare in these low organic soils. A recent

literature review showed that nitrogen limitation was about as common as phosphorus limitation (Elser et

al., 1990). Detailed water chemistry data from the midwestern lakes suggest that nitrogen values in the

epilimnion fall during summer but that phosphorus values remain more constant. These data suggest

phosphorus may be cycled more efficiently than nitrogen and that without external inputs, late summer

nitrogen limitation can be expected. These results do not imply that continued focus on phosphorus for

Nutrient Criteria—Lakes and Reservoirs 2-3

eutrophication control is unwarranted; however, a better understanding of the frequency and extent of

nitrogen limitation is needed to discern lake function. Nitrogen criteria as well as phosphorus criteria are

appropriate.

Response variables such as chlorophyll a and algal or macrophyte species or biomass indicate the

relative success of the nutrient management effort. By carefully incorporating both the causal and

response elements, a State or Tribe should be able to fine-tune its criteria to meet the necessary

enrichment levels for a given class of lakes. These variables are described in more detail in Chapter 5.

2. Role of the Watershed

The geology (and land use) within a lake

�

s watershed determines the amount of nutrients that enter

the lake and, therefore, plant biomass. This statement is probably the primary reason for the

development of nutrient criteria: human activity in the watershed affects a lake

�

s function. It is the

reason behind the National Nutrient Assessment Workshop

�

s conclusion that changes in land use can

serve as an early warning system for changes in lakes (U.S. EPA, 1996).

In simplest terms, a lake

�

s nutrient concentration is affected primarily by the rate of weathering and

erosion from the soils in the watershed. If the underlying geological structure is granitic, then the rates of

weathering will be low and both the productivity of the terrestrial vegetation and the concentration of

nutrients in the runoff from the watershed will be low. On the other hand, if the underlying bedrock is

sedimentary, the weathering rates will be higher and the fertility of the soil and the nutrient content of the

runoff water will be higher as well. Consequently, Naumann (1929) observed that lakes in regions of

sedimentary rock had higher algal densities (were greener) than lakes in granite-based watersheds. (For

the purpose of this manual, atmospheric deposition of nitrogen and phosphorus, while possibly

important, is accepted as a regional constant subject to further attention as our management technology

improves.)

Human activity has at least two effects on the natural load of nutrient input to lakes: (1) it disturbs

the overlying vegetation, exposing the soil to increased weathering and erosion, and (2) it adds easily

erodible nutrient-containing material, such as fertilizers and animal waste, into the watershed. As the

biological surface of an undisturbed watershed is disrupted, and as people move into the watershed, it can

be expected that there will be increased soil and nutrient runoff.

Of course the degree of disturbance relative to the size of the lake will affect the impact of the

disturbance; building a summer cottage would not have the same impact on a lake as would clear-cutting

a forest or developing a condominium complex. Sometimes the term “assimilative capacity” is used to

imply that the lake has a certain capacity to absorb the impact of disturbance. This concept, although

comforting, probably has little basis in fact. Impact, until demonstrated otherwise, is probably better

thought of as a continuous response to nutrient increases. The degree of change will depend on other

factors, such as the size of the lake, and the change may not be immediately or even ever detectable to

humans or their monitoring instruments. However, whether detected or not, changes do occur. It is for

this reason that watershed disturbance is a sensitive early warning of lake change. Clearly, biological

impact within the lake will be directly related to the increased amount of nutrient loading, and that

impact will occur, whether or not it is detected.

Naumann (1929), nonetheless, used the relationship between nutrients and plants to establish a

trophic state classification. He probably began his classification scheme with the perfectly reasonable

goal of classifying lakes into those with low (oligotrophic) and high (eutrophic) plant biomass.

2-4 Nutrient Criteria—Lakes and Reservoirs

Oligotrophic lakes were clear with little algae, whereas eutrophic lakes were green. He then added to his

classification system the causal factors that produced this degree of greenness, for example, the amounts

of nitrogen or phosphorus. He called these the “factors of production.” Oligotrophic lakes were those

that had low production (biomass) because they were low in nutrient concentrations. Eutrophic lakes

were green because there were abundant nutrients to support the growth of algae.

The combination of the factors that affected production (causal factors) with plant production itself

(response variable) allowed for a suite of trophic classes that dissected lakes into groups of varying

production based on the factor or factors that were thought to limit that productivity. The classic

oligotrophy-eutrophy axis was based on limitation by nutrients. The mesotrophic category was added to

describe situations intermediate between oligotrophy and eutrophy. The term “hypereutrophic,” or

hypertrophic, was added by Wetzel (1966) to describe situations of extreme eutrophy where light, not

nutrients, is the dominant environmental factor controlling growth. This continuum of trophic states is

illustrated in Figure 2.1.

3. Trophic Causal Chain

Naumann (1929) was very insightful in recognizing that the components of the lake are an

interconnected system; as one component—the plants—responds to nutrient inputs, other biological,

chemical, and even physical components would be affected as well. Increases in nutrient loading do not

necessarily directly affect any component other than the plants, but by various pathways, other

components of the lake ecosystem, such as zooplankton, fish, and hypolimnetic oxygen concentration,

are affected as well. This trophic state cascade is depicted in Figure 2.2.

Because of these linkages between components, numerous variables may respond to varying degrees

to increases in nutrients. Not only will algae or macrophytes increase, but zooplankton and fish biomass

may increase as well, plant and animal species may change (with some going extinct), and hypolimnetic

oxygen may be depleted. People react to the various changes or symptoms of lake condition reflected in

the chemistry and biology that cascade from the change in loading, not directly to the change in loading

or nutrient concentration itself.

Figure 2.1. A trophic continuum.

Nutrient Criteria—Lakes and Reservoirs 2-5

Figure 2.2. The trophic causal chain.

This cascade of biological and chemical changes produces a variety of choices for the response

variables needed to supplement the causal variables in the formulation of nutrient criteria. Choice of the

response variable can be made based on sensitivity to change, cost of measurement and analysis, or

importance to designated use.

4. Lake Aging

The idea that lakes undergo directional change in plant production as they age was probably related

to the observation that shallower lakes appeared to have more plant biomass in them than deeper lakes.

This observation later translated into the idea that increases in plant biomass were inevitable as a lake

ages and fills in. This concept has lead to terms such as “natural eutrophication” to describe inevitable

increases in plant biomass as a lake becomes older and shallower. If natural eutrophication is thought to

proceed at a rate related to inputs from the watershed, then we might expect to see accelerated rates of

eutrophication if cultural influences occur in the watershed (cultural eutrophication).

If trophic state is a description of the biological condition of the lake, eutrophication describes a lake

that is becoming more eutrophic (Figure 2.3). Specifically, it describes a change in the direction of

eutrophy. A lake does not have to become eutrophic to have undergone eutrophication; it only has to

move in the direction of eutrophy. Oligotrophication describes the process of a lake moving in the

alternative direction, towards oligotrophy.

2-6 Nutrient Criteria—Lakes and Reservoirs

Figure 2.3. Eutrophication and oligotrophication in relation to the trophic continuum.

The term “nutrient enrichment” is in many, if not most, instances an alternative term for

eutrophication. However, the emphasis in that term is on the increase in nutrients rather than on the

lake

�

s response to that enrichment. If causal factors such as nutrient enrichment are closely linked via

our terminology with the lake response, problems can arise if, for example, a lake

�

s biology changes

without any change in nutrient loading or, conversely, if nutrient loading occurred without a coincident

change in biology. These are not just hypothetical situations. The addition or removal of a benthivorous

fish, such as bullhead or carp, can change the internal regeneration of nutrients and change the biological

condition of the lake without any enrichment, or at least external enrichment, of the lake. The increase in

grazing on algae because of the addition of a piscivore or the removal of the zooplanktivores also can

alter the amount of plant biomass without the need of an alteration of nutrient loading. These

manipulations, often called biomanipulations (Shapiro et al., 1975), are a type of lake manipulation that

can alter the state of the lake without a change in nutrient load.

Natural eutrophication should not be confused with naturally eutrophic. The latter term describes

lakes in watersheds where the natural load of nutrients is high despite the absence of human activity.

Natural eutrophication describes a belief that lakes, presumably all lakes, have more plant material within

them (become more eutrophic) as a natural part of becoming older.

The concept of natural eutrophication is probably correct to the extent that processes within the lake

such as nutrient regeneration may enhance the effect of inputs from the watershed. The importance of

these internal processes are still not well understood, especially along a gradient of lake aging. If

becoming shallower were the only consideration, then the internal concentration of nutrient would

increase only to a level as high as that in the incoming water. A lake in a watershed in which the

concentrations of nutrients were very low in the incoming water would not become that much more

eutrophic even if it did become shallower. However, a shallower lake also may have increased

macrophyte growth and increased regeneration of nutrients from the sediments. In this case, the

biological response would become increasingly independent of the external supply of nutrients.

A difficulty with dwelling on the possibility of natural eutrophication is that it emphasizes an

inevitability of the eutrophication process; it also takes the focus off the immediate effects of the

Nutrient Criteria—Lakes and Reservoirs 2-7

watershed on the lake. Natural eutrophication is a process that is measured in terms of thousands of

years, whereas the problems we encounter most often with lakes are the effects of processes that take

only a few years to develop. Watershed disturbance can rapidly move a lake to a new level of nutrient

concentration and biological response. More important, in most or many lakes, that response is, to some

extent, reversible; we have not just moved rapidly down an irreversible path. Cultural eutrophication is,

in fact, a reversible process, and nutrient criteria are an important element in this reversal.

C. Connecting Watershed Loading to the Lake: A Mass Balance Model Approach

Like many earlier nutrient loading models, Vollenweider used a mass balance model for the basis of

the prediction (see Chapter 9 Section B). Below is a basic review of Vollenweider

�

s model and how it is

used to link loading to concentration in a lake or reservoir.

The term “mass balance model” comes from the assumption that a substance such as phosphorus

cannot just appear or disappear from a reservoir; it must come from somewhere and it must go

somewhere. The phosphorus going into the reservoir must either go out again through some outflow, be

sedimented to the bottom, incorporated into macrophyte biomass, or remain in the water in either

dissolved or particulate forms. It is this phosphorus that remains in the water that is of interest because it

is the amount that is available for algal growth.

Mass balance modeling is done in a manner similar to keeping a checking account. The total amount

of phosphorus entering the reservoir (loading) each year is measured. Loading describes the total amount

of material being moved in a stream in a given amount of time. The loading from any source (i) is

calculated as:

Loading (J

) = Water Discharge (Q

) × Concentration (C

)

= Q

The external loading, symbolized by Vollenweider by the letter J, can be calculated as the sum of the

loading from all the external sources (i) to the lake:

∑

Loading often is used to measure export of a nutrient or sediment from a watershed. For example, it

is important to gauge the effects of farming practices on erosion, so we might calculate the tons of

sediment removed from a watershed over a year. On the other hand, if the stream enters a lake or

reservoir, we might want to know how much material is entering that body of water. In this case, nutrient

loading might affect water quality, and sediment loading might affect the fill-in rate.

Predicting the internal concentration of a substance in a lake is also done using a mass balance

equation. The appropriate mass balance for this prediction is based on the idea that the rate of change of

the total amount of a material (M) in a lake (dM/dt) is dependent on the total amount of material that

enters a lake (J) and the total amount that leaves in the same time period:

2-8 Nutrient Criteria—Lakes and Reservoirs

dM/dt = (



)



dM/dt = J



If the reservoir is not rapidly changing from one year to the next, the amount coming in one year

should be equal to the amount going out:

Inflow = Outflow (i.e., dM/dt = 0)

J = Q

If we assume, as did Vollenweider, that the lake is completely mixed, then the lake concentration is

equal to the outflow concentration:

lake

= C

Therefore, the amount entering the lake will be equal to the amount leaving:

J = Q

lake

Rearranging this equation, we obtain an equation predicting the concentration in the lake based on

the external loading of the substance and the outflowing discharge of water:

lake

= J/Q

Notice that the term (J/Q

) has the units of concentration. Consider that it is the incoming loading

divided by the outgoing water discharge.

This model is designed to predict the concentration of any conservative material. A conservative

material, such as chloride or sodium, does not sediment within the lake basin, and the amount leaving the

basin over the outflow should be equal to the amount entering. Conservative materials are not very

interesting in themselves, but they are used as indicators of the accuracy of budgets of materials that do

sediment within the lake. If the input of a conservative element is not equal to the output, then some

other source of water and/or material has been neglected.

A nonconservative material, such as phosphorus, is one that is lost from the water column (e.g.,

sedimentation) within a lake basin. Because some material is lost from the water column, the input

loading is not equal to the output loading. To model a nonconservative material, a sedimentation term

must be added to the equation:

Input = Output + Sedimentation

Sedimentation was considered by Vollenweider to be proportional to the mass of the substance in the

lake (M). The total amount of material in the lake (M) is calculated as:

Nutrient Criteria—Lakes and Reservoirs 2-9

M = C

lake

where V = volume of the lake (m

)

Vollenweider considered the amount of material lost to the sediments. This is designated by sM,

where s is a first order fractional loss of the mass settled per unit time (1/t) and M is the mass of

substance in lake (C

lake

V). The sedimentation coefficient, s, is really a net sedimentation term, because

the material may not only settle out of the water column but also may be resuspended into the column

from the sediments.

The mass balance equation, with the added sedimentation term, becomes

dM/dt = J



Vollenweider then assumed that over a calendar year the system would be near or at steady state, and

the mass balance equation becomes:

J = Q

+ sM

Note that all the terms still have the dimensions of amount per time.

An equation for predicting the lake concentration from loading can be produced by substituting C

lake

for C

(again assuming that the lake concentration is equal to the outflow concentration) and substituting

lake

V for M:

J = Q

lake

+ sC

lake

Rearranging, we obtain the predictive equation:

lak e

The equation can be further rearranged into the form:





lak e









With this equation, several things became clearer about loading:

• Vollenweider considers only the “total” form of the substance. He does not discriminate

between dissolved and particulate forms.

2-10 Nutrient Criteria—Lakes and Reservoirs

• The term (J/Q

) has the dimensions of concentration (mg/m

-3

) and represents the average

incoming concentration of the substance (Vollenweider, 1976) assuming evaporation is minimal.

This term is sometimes replaced by a symbol for incoming concentration, C

• The term (1/(1 + s(V/Q

)) is really a description of the fraction of the incoming concentration

that is not retained within the basin. In some models, retention is represented by the symbol R

)), by (1-R).

and the term (1/(1 + s(V/Q

• The term (V/Q

) has the units of time and is the hydrologic residence time, T or t, which

represents the average time that water remains within the lake.

Using these simpler symbols, the equation can be reduced to a simple statement of the relationship

between loading and lake concentration:





lak e

C i









where C

= average inflow concentration.

Although relatively simple, the equation illustrates the major aspects of prediction with mass balance

models and trophic state:

• The concentration of a substance such as phosphorus in the reservoir (C

lake

) is directly

determined by the concentration of that substance in the incoming streams (C

). The higher the

concentration in the streams entering a reservoir, the higher the nutrient concentration will be in

the reservoir itself.

• Internal factors such as water residence time (T) and the net sedimentation coefficient (s)

determine the amount of material that is sedimented, and therefore lost, from the water column.

The longer the water residence time, the greater the amount of material that will be sedimented

within the reservoir, and the lower the reservoir concentration will be.

Additional terms have been added to the equation to account for release of a nutrient from the

sediments into the open water or for the biological availability of the incoming phosphorus. These

additional terms can make the predictions more specific to the particular reservoir being modeled. See

Chapter 9 for a discussion of these models.

D. Trophic State Classification Systems

The concept of trophic state, with its relationship of the watershed to the chemistry and biology of

the water body, has become one of the primary methods of classifying lakes. Despite controversies of

definition, it has endured and probably will endure because of several important reasons:

• History and tradition. The language and implications of trophic state are deeply ingrained in

limnology. In a sense the concept of trophic state is the nutrient paradigm.

Nutrient Criteria—Lakes and Reservoirs 2-11

• Communication. When a trophic state term, such as eutrophic or eutrophication, is used, there is

a general agreement as to what a lake is like in terms of nutrients and biology. This implication

of interrelationships tends to communicate far more information than can be implied with the

statement of the value of a single variable.

• Education. The trophic state concept, even in qualitative terms, is a convenient vehicle to

educate the public on the simplicity, and indeed the complexity, of the relationship between land

use and the biological consequences.

Trophic state classification may have begun as a continuum concept, but it rapidly evolved into a

classification of “types.” Most, but not all, existing trophic classification systems, or indices, reflect this

typological emphasis. The representation of this type of classification scheme is simply of list of

characteristics for a specific trophic type (Table 2.1). Lakes are assigned to a trophic class based on their

agreement with the characteristics on the list. This type of classification runs into difficulty when

specific variables may classify the lake in different categories. This happens when the correlation

between variables is not strong.

The essence of a typological trophic classification is the belief that there is a real type of lake called

eutrophic in the sense that there is a real type of human called young, middle aged, or elderly. Lakes can,

therefore, be classified and placed into one of these types. Eutrophication is the progressive directional

change of a lake out of one type and into another. Once in a type, the lake takes on certain characteristics

by which it can be recognized and, therefore, classified. Such classifications are easily recognized from

lists of characteristics typical for each trophic state heading (Table 2.1).

The OECD index (Vollenweider and Kerekes, 1980) used a statistical approach to quantify the

ranges of several variables within each trophic designation (Table 2.2). This index was derived by

asking a group of scientists their opinion as to what was the average value for each trophic class for each

variable. The summarized data were used to produce bell-shaped curves for each variable for each class

(Figure 2.4). The overlap that resulted emphasized that lakes of the same concentrations may be in more

than one trophic class.

The second approach to trophic classification assumes that trophic types are not real but abstractions

and, to some extent, arbitrary divisions of a continuum. This approach is similar to Naumann

�

s original

classification. In this case, the discussions have been generally along the lines of what is the appropriate

trophic state variable that should be divided into trophic state classes. The discussion of appropriate

variables for classification is continued in Chapter 5.

Some continuum-based classification indices emphasize that trophic state reflects a number of

variables, recalling the multiple variable approach of typological schemes. For example, Huber et al.

(1982) stated that “trophic state is the integrated expression of the nutritional status of a water body. As

such, it is widely accepted that no single trophic indicator or parameter is adequate to completely

describe and/or quantify the concept.” Multiple variable indices differ from the typological indices

largely in that they quantify the multiple variables found in the trophic state list of characteristics. These

approaches emphasize the collection of quantitative data and are a major advance over qualitative

listings.

2-12 Nutrient Criteria—Lakes and Reservoirs

Table 2.1. Illustration of a Typological Trophic Classification System Based on Lake

Characteristics (adapted from Rast and Lee, 1987)

Variable

General Characteristics

Oligotrophic Eutrophic

Total a

quatic plant production

Low High

Number of algal species

Many Few

Characteristic algal groups

Greens, diatoms Blue-greens

Rooted aquatic plants

Sparse Abundant

Oxygen in hypolimnion

Present Absent

Characteristic fish Deep-dwelling, cold water fish

such as trout, salmon, and

cisco

Surface-dwelling, warm water fish

such as pike, perch, and bass; also

bottom-dwellers such as catfish and

carp

Water quality for domestic

and industrial use

Good Poor

Table 2.2. OECD Ranges Based on Scientists’ Opinions (after Vollenweider and Carekes, 1980)

Variable Oligotrophic Mesotrophic Eutrophic

Total phosphorus

mean

range (n)

3-18 (21)

11-96 (19)

16-390 (71)

Total nitrogen

mean

ran

ge (n)

660

310-1600 (11)

750

360-1400 (8)

1,900

390-6100 (37)

Chlorophyll

mean

range (n)

1.7

0.3-4.5 (22)

4.7

3-11 (16)

2.7-78 (70)

Peak chlorophyll

mean

range (n)

4.2

1,3-11 (6)

5-50 (12)

10-280 (46)

Secchi depth (m)

mean

range (n)

9.9

5.4-28 (13)

4.2

1.5-8.1 (20)

2.4

0.8-7.0 (70)

Note: Units are

g/l (or mg/m

), except Secchi depth; means are geometric annual means (log 10), except peak chlorophyll

Nutrient Criteria—Lakes and Reservoirs 2-13

Figure 2.4. Probability distribution curves for total phosphorus by trophic status class.

Probably the most sophisticated of the multivariate indices is that of Brezonik and Shannon (1971),

which uses principal components analysis to derive a trophic state index (TSI) based on seven variables:

(1) TP, (2) primary production, (3) inverse of Secchi depth, (4) total organic nitrogen, (5) chlorophyll a,

(6) specific conductance, and (7) the inverse Pearsall cation ratio ([Ca]+[Mg]/[Na]+[K]). Other less

sophisticated indices generally combine unweighted variables by one means or another. The EPA Index

(U.S. EPA, 1974) ranked lakes based on “the percentage of the 200+ lakes exceeding Lake X in that

parameter”; the index was “simply the sum of the percentile ranks for each of the parameters used.” The

variables used were TP, dissolved phosphorus, inorganic nitrogen, Secchi depth (500-Value [inches]),

chlorophyll a, and minimum dissolved oxygen (15-DO

min

These multivariate quantitative indices move trophic classification from a typological concept to one

assuming a continuum of values, but they suffer from several drawbacks. The indices require that all the

variables be measured before an index value is derived, thus greatly increasing the cost and analytical

time required. A missing value eliminates a TSI determination. Changes in a single variable often will

be overlooked in the index if other indices do not change. Conversely, if index variables are correlated,

then a change in one may trigger a change in a number of variables causing an exaggeration of the

amount of change. Finally, a change in the index does not tell the reader what has changed; information

2-14 Nutrient Criteria—Lakes and Reservoirs

is lost.

Carlson (1977) suggested returning trophic state to its first principles: a quantifiable plant

biomass-based concept that could fit easily into existing and future nutrient and lake models. He did not

redefine trophic state but used Naumann

�

s (1929) original idea of a classification according to plant

biomass. Instead of the distinct typological classes, Carlson assumed algal biomass to be from a

continuous range of values. He suggested that the commonly used trophic classes were arbitrary

divisions of the biomass continuum. To emphasize the continuum nature of a biomass-based trophic

state, he used a numeric rather than a nomenclatural scale, dividing the range of algal biomass based on a

doubling of Secchi depth, a variable that is affected by algal density.

The original Secchi depth equation in Carlson (1977), reproduced below, illustrates how the index

was constructed:





TSI

(

S D

)



−

ln 2







The basic Secchi disk index was constructed from a doubling and halving of Secchi disk

transparency. The base index value is a Secchi depth of 1 m, the logarithm of which is 0.

ln 1 = 0

�

0 = 6

10 × 6 = 60

Therefore, the TSI of a 1 m Secchi depth is 60. If the Secchi depth were 2 m,

ln 2 / ln 2 = 1

�

1 = 5

10 × 5 = 50

The index utilizes relationships between trophic variables to produce equations that allow the index

to be calculated from variables other than Secchi depth. The indices for the chlorophyll and TP are

derived in a similar manner, but instead of a Secchi depth value in the numerator, the empirical

relationship between chlorophyll or TP and Secchi depth is given instead. For example, the TSI equation

for chlorophyll is:



204

−

0.6 8 ln

CHL



TSI

(

C H L

)



−







Nutrient Criteria—Lakes and Reservoirs 2-15

The above forms of the equations illustrate how the indices were derived, but they can be simplified

for everyday use:

TSI (SD) = 60



14.41 ln SD

TSI (CHL) = 9.81 ln CHL + 30.6

TSI (TP) = 14.42 ln TP + 4.15

The value of multiple equations is that the same TSI value should be obtained no matter what

variable is used to calculate it (i.e., a common scale). This means that if data are missing for chlorophyll,

for example, a similar value could be obtained from transparency.

Although these three variables should covary, they should not be averaged because neither

transparency nor TP are independent estimators of trophic state. Using transparency or phosphorus as an

estimator of chlorophyll is very different from assuming equal and independent status of the variables.

Secchi depth and TP should be used as a surrogate, not a covariate, of chlorophyll.

In essence, this TSI scale is an indexed scale of algal biomass. Because it is directly related to lake

phosphorus concentration, it fits easily into phosphorus loading models such as that of Vollenweider

(1976). If a loading model can predict phosphorus concentration in the water, then the trophic state can

be predicted easily as well. Work by Kratzer and Brezonik (1981) allows the index to be predicted from

nitrogen concentrations as well.

TSI(TN) = 54.45 + 14.43 ln(TN)

[Nitrogen values must be in units of mg/L]

Their index could be used especially if there is any indication that nitrogen, rather than phosphorus,

is limiting.

E. Uses of Trophic State Indices

Indices have several purposes. In some instances, indices take uncorrelated variables and aggregate

them into a single word or value so that a general condition may be easily communicated. For example, a

pollution index might include concentrations of heavy metals, pesticides, and phosphorus, which may or

may not be correlated but could contribute to what the public considers to be pollution. The multivariate

trophic state indices are of this type. These indices assume that trophic state consists of a number of

possible attributes of lakes, ranging from nutrient concentration to hypolimnetic oxygen depletion. An

index is necessary to somehow combine these various ingredients into the trophic stew and relate lakes to

one another in a national continuum.

Alternatively, indices such as that of Carlson (1977) use the term “index” to mean that the variable

measured is not trophic state, but an indicator of trophic state. For Carlson, trophic state is plant

biomass. Chlorophyll, transparency, or even TP are variables that can estimate biomass but are really not

living plant (autotroph) carbon. Even the measurement of organic carbon is not free from interferences

from detritus or nonplant carbon. Trophic state is used as a surrogate for a real entity, plant biomass, that

cannot be measured directly.

2-16 Nutrient Criteria—Lakes and Reservoirs

A third use of the term “index” that combines aspects of the first and second definition is that of

simplification of a concept of measurement. For example, few readers know that the Richter scale, used

to describe the magnitude of an earthquake, is the maximum deviation of a needle on a seismograph 62

miles from the epicenter. Actually, most people do not need to know the mechanics of calculating the

Richter scale to have a sense of the severity of an earthquake. In the same sense, trophic state indices are

shorthand methods to convey information. Total phosphorus or chlorophyll has little value in

communication unless there is some standard to which the reader or listener can compare the value. In

this case, saying eutrophic or TSI of 60 rather than “chlorophyll concentration is 20 µg/L” may convey

information more easily because there are fewer terms to explain to an audience and fewer terms for the

audience to put into the context of their own experiences.

In each of these instances, the index helps the reader equate several parameters in one indicator of

enrichment condition so that an interpretation of condition is conveyed and conclusions can be made.

Whether the measures of water quality compiled in a data set are used individually or collectively in an

index, they are the essential, objective information the resource manager needs to determine the status of

a lake or reservoir. With the representative information and a guide as benchmark criterion defining the

relatively unimpaired and attainable water quality condition for comparison, the manager can classify,

select, and plan for the restoration and protection of the lakes and reservoirs in his or her region.

Nutrient Criteria—Lakes and Reservoirs 2-17

CHAPTER 3

A. Defining the Resource of Concern

B. Classification

Preliminary Steps for

Criteria Development

A. Defining the Resource of Concern

Defining the resource of concern begins the overall process of establishing nutrient criteria.

Resources of concern here are lakes and reservoirs, and managers must decide which water bodies are to

be included in the population to which criteria will be relevant and applicable. Many States define

jurisdictional lakes (“waters of the State”) as those above a size threshold. For example, the inclusion of

farm ponds and other similar small ponds can potentially result in an inordinately large population of

lakes that would be required to be considered during the criteria establishing process. These practical

considerations often make it desirable to eliminate small water bodies from the resource population.

States may have already established a regulatory size threshold that specifies what should be

considered a lake from the State management perspective. For example, the Florida Department of

Environmental Protection routinely samples only lakes larger than 10 acres, because there are more than

7,000 lakes of 10 acres or more in Florida (Huber et al., 1982) and lakes under 10 acres are thought to

number 10,000 or more. Florida surface water quality criteria apply to all lakes not wholly owned by a

single person other than the State (Florida Amended Code, 62-340). States are encouraged to determine

if the established threshold is appropriate for the nutrient criteria setting procedure described in this

document and to adjust it as necessary. If States have not set size limitations that define a lake, State

water resource agencies should evaluate the lake resources in the State to determine appropriate size

limitations. The goal of such an exercise is to eliminate small water bodies that, because of their size

(and resulting hydrology) or uses (e.g., small agricultural impoundments), do not accurately represent

typical lake conditions or do not exhibit expected responses to stressors.

For the purpose of this document, lakes are defined as natural and artificial impoundments with a

surface area greater than 10 acres and a mean water residence time of 14 or more days. Man-made lakes

(i.e., artificial lakes) with the same characteristics may be viewed as part of the same system. Reservoirs

are man-made lakes for which the primary purpose of the impoundment is other than recreation (e.g.,

boating, swimming) or fishing, and the water retention time and water body depth and volume vary

widely. Hydroelectric power generation, drinking water supply, and flood control are examples of typical

uses of reservoirs.

Impoundments on rivers, especially ones on larger rivers, also require specific definition.

Impoundments behind low-head dams for navigation, as on the Ohio and Mississippi Rivers, are hardly

lakelike in their characteristics; in fact, they are called navigational pools. At what point does a pool on a

river become a lake? Limnologists generally consider lakelike characteristics to increase with water

mean residence time. Many studies suggest that phytoplankton do not accumulate at retention times less

than 7 days (e.g., Kimmel et al., 1990).

Nutrient Criteria—Lakes and Reservoirs

3-1

These definitions are provided for the purpose of illustration and consistency. States with legal

definitions of their lakes or reservoirs should obviously adhere to their own terms and interpret this

guidance accordingly.

B. Classification

1. Geographic Divisions

The establishment of a single, national nutrient criteria for lakes is not a realistic goal because of the

significant variability of water bodies that exist across the country in a variety of climates, geographic

locations, and ecosystems. On a national basis, individual lakes and reservoirs are affected by varying

degrees of development, and user perceptions of water quality can differ even over small distances. As a

result, the nutrient criteria development process discussed in this document is based on an approach that

acknowledges geographic differences in lakes across the country and within States and that uses a

classification system to clarify those differences. The initial classification scheme used in this manual is

the ecoregion approach (Omernick, 1987, 1988, 1995). However, many viable regionalization techniques

exist for delineating geographic regions.

The process of identifying geographic divisions (i.e., regionalization) is part of a hierarchical

classification procedure with the purpose of grouping similar lakes together (i.e., to prevent comparison

of unlike lakes). Classifying lakes reduces the variability of lake-related measures (e.g., physical,

biological, or water quality variables) within classes and maximizes the variability among classes.

Classification invariably involves professional judgment to arrive at a workable system that separates

clearly different ecosystems, yet does not consider each lake a special case. The intent of classification is

to identify groups of lakes that under ideal conditions would have comparable characteristics (e.g.,

biological, ecological, physical). To the extent possible, classification should be restricted to those

characteristics of lakes that are intrinsic, or natural, and are not the result of human activities. These

characteristics include size, maximum or mean depth, detention time, and shape.

The general approach to the regionalization process is to establish divisions at the broadest level and

then to continue to stratify to a reasonable point. In this section, a regionalization system for the national

scale is presented to provide a framework for developing nutrient criteria. EPA encourages States to

identify State-specific subregions, if appropriate, and to use the national regionalization scheme

discussed below as the basis for further subdivisions.

�

National Nutrient Ecoregions

Ecoregions are a mapped classification system of ecological regions, that is, regions with assumed

relative homogeneity of ecological characteristics (Omernik, 1987). EPA has developed maps of

ecoregions of the United States at various levels of resolution and aggregation (Omernik, 1987). The

most commonly used is the Level III ecoregions, consisting of 79 ecoregions in the conterminous United

States. Ecoregions were based on interpretations of the spatial coincidence in all geographic phenomena

that cause or reflect differences in ecosystem patterns. These phenomena include geology, physiography,

vegetation, climate, soils, land use, wildlife, and hydrology. The relative importance of each

characteristic varies from one ecoregion to another regardless of the hierarchical level.

For the National Nutrient Criteria Program, a map of aggregations of the Level III ecoregions was

developed to define broad areas, within which there are general similarities in the quality and types of

ecosystems as well as in natural and anthropogenic characteristics that affect nutrients (see Figure 1.1).

3-2

Nutrient Criteria—Lakes and Reservoirs

The regions are intended to provide a spatial framework for general guidance and reporting for the

National Nutrient Criteria Program.

These nutrient regions and their component Level III ecoregions are described more fully in

Appendix A. The nutrient regions delineated in Figure 1.3 are not intended to be homogeneous. They

are aggregations of ecoregions where expectations within a nutrient region are more similar than

expectations among nutrient regions. Some regions may be characterized by relative homogeneity; other

regions may be characterized by extreme heterogeneity. An example of a heterogeneous region is Region

XII, the Southern Coastal Plain, which has lakes ranging from ultra-oligotrophic lakes in sandy ridges

and hills to highly eutrophic solution lakes in areas with phosphatic soils (Griffith et al., 1997). By

comparison, Region VI, the Corn Belt and Northern Great Plains, is more homogeneous and is expected

to be dominated by mesotrophic to eutrophic lakes, owing to the fertile plains soils and extensive

agriculture. Region VIII, the Nutrient Poor Largely Glaciated Upper Midwest and Northeast, is

dominated by oligotrophic lakes, but it also has small subregions with higher nutrient concentrations and

mesotrophic lakes (Omernik et al., 1988; Rohm et al., 1995).

The nutrient regions can form the basis for initial development of nutrient criteria. Expectations can

be developed for nutrient concentrations and loadings in each of the regions and criteria derived from

those expectations.

�

Further Subregionalization

The heterogeneity within many of the nutrient regions will require further subregionalization or

subclassification to implement nutrient criteria. Using the ecoregion concept as a basis, EPA has

developed lake regions based on phosphorus and other considerations for three areas: the Upper

Midwest, comprising parts of nutrient regions VI, VII, and VIII in Minnesota, Wisconsin, and Michigan

(Omernik et al., 1988); the Northeast, comprising nutrient regions VII, VIII, and XIV ranging from

northern Pennsylvania and New Jersey through New York and the New England States (Rohm et al.,

1995); and Florida, comprising a small part of nutrient region IX, most of Region XII, and all of Region

XIII (Griffith et al., 1997). The regionalizations for the Upper Midwest and Northeast are based on total

phosphorus concentration because of the dominance of phosphorus as the principal limiting nutrient in

cool temperate lakes of the world (e.g., Schindler, 1978). The regionalization for Florida also takes into

account total nitrogen concentration, algal chlorophyll, pH, color, Secchi depth, lake origin, and lake

hydrology. In warm temperate and subtropical lakes, nitrogen concentration is often the principal limiting

nutrient (e.g., Shannon and Brezonik, 1972; Carlson, 1992).

These subregionalizations were developed from data on nutrient concentrations of sampled lakes in

the regions, soils, and land use (Omernik et al., 1988; Rohm et al., 1995; Griffith et al., 1997). The

distributions of nutrient concentrations of each subregion were characterized (usually with a histogram)

if data were available. In subregions where no data were available, the nutrient distributions were

estimated based on similarity of soils and land use to regions where data were sufficient to characterize.

It is expected that as more data are developed through the National Nutrient Criteria Program, more

nutrient ecoregions will be similarly subdivided.

2. Nongeographic Classifications

Many lake classifications have been proposed in addition to trophic state and geography

(Hutchinson, 1957). Lake classification can be further complicated by natural or human- induced

conditions that can intrinsically affect the state of a lake and, therefore, how it can be classified. For

Nutrient Criteria—Lakes and Reservoirs

3-3

example, acidic lakes (whether naturally acidic or from acid deposition) are commonly found in the

Adirondacks of New York, Pennsylvania, and West Virginia and in sand ridges of Florida. High mineral

turbidity is found in reservoirs where streams have a high load of suspended fine sediment, typically in

arid and semiarid regions.

Although lake types can be explained to greater or lesser extent on geographic considerations, it may

be more convenient to classify lakes by nongeographic variables, which may yield more explanatory

power than geographic locations. Discussed below are certain factors that potentially can affect the

classification process but that generally fit the geographic-oriented focus of the above geographic

approaches (e.g., ecoregional, water quality characteristic).

�

Lake Origin

Hutchinson (1957) lists 76 different types of lakes based solely on their origins. Although we often

think of a lake simply as a hole in the ground with water in it, the number of different lake types should

make us pause to consider how many ways the origin shapes the area, the volume, and the shape of the

lake basin. Lakes of volcanic origin are probably very deep, with virtually no littoral zone and small

watersheds. Crater Lake (lake type 10), for example, is extremely deep and very clear and has only the

crater walls for a watershed. However, it is susceptible to nutrients introduced by septic leakage because

of the very small water load.

Lakes of tectonic origin such as those found in faults (lake type 9; e.g., Lake Baikal) might behave

similarly. Other lakes may be extremely shallow such as oxbow lakes (lake type 55) or maritime coastal

lakes (lake type 64). In these instances, there may be extensive shallow areas and considerable

interaction of the sediments with the overlying water. The shape of the basin and watershed help

determine the controlling variables of surface area, depth, volume, and retention time. Rather than use

discrete classes (e.g., large lakes, small lakes), it may be more effective to treat the shape-related

variables as a continuous characteristic. This is discussed in more detail in Section 3.

�

Reservoirs

Reservoirs and impoundments, created by the damming of a stream, have characteristics of both

rivers and lakes (Thornton, 1990). Reservoirs are divided into three zones—riverine, transitional, and

lacustrine—which correspond to (1) flowing, riverlike conditions, (2) transition to lake conditions, and

(3) nonflowing lakelike conditions near the dam, respectively. With expected life spans ranging from

several decades to a century or more, reservoirs are more ephemeral than most natural lakes and have

several physical characteristics unique to reservoirs and natural reservoirs formed by natural dams (e.g.,

beaver dams, terminal moraines, landslides).

Reservoirs vary widely in physical characteristics of shape, size, and hydrology. They can range

from small, shallow impoundments (farm ponds) to deep storage reservoirs to “run of the river”

flowthrough navigational pools and hydroelectric reservoirs on large rivers. They are built and managed

for widely different purposes, including flood control, navigation, municipal or agricultural water

storage, hydroelectric generation, and gamefish production. Many dams are constructed to allow

discharge from the epilimnion, metalimnion, and/or hypolimnion, depending on management goals of the

water bodies. This must be known before understanding the limnology of the reservoir. The management

practices in turn affect physical, chemical, and biological characteristics of the reservoir.

3-4

Nutrient Criteria—Lakes and Reservoirs

Although no natural reservoir reference conditions exist, the operational determination of nutrient

reference conditions for reservoirs is the same as for natural lakes. Reservoirs can be classified

according to hydrology, morphometry, management objectives, and other factors. Age of the reservoir

may be important in determining expectations. Several considerations affect the classification of

reservoirs as opposed to natural lakes:

• Distribution. Reservoirs and impoundments are most numerous in regions with few or no natural

lakes. The nonglaciated parts of North America have the largest number of reservoirs (except

Florida, which is a Karst landscape).

• Form. The form or shape of a basin and watershed may be the most important distinction

between natural and artificial lakes. Shape substantially influences hydrology and water quality

of impoundments. Large reservoirs are drowned river valleys and tend to be long and deep with

numerous embayments of tributary streams. The watersheds of reservoirs are relatively much

larger than those of natural lakes and contribute relatively greater sediment loads.

• Longitudinal gradient. Reservoirs have characteristics typical of both lakes and streams. They

are streamlike at the head where major tributaries enter and are more lakelike near the dam

(Thornton, 1990).

• Turbidity and loading. Many reservoirs are more turbid than natural lakes, and they receive

more nutrients and organic matter from their tributary streams than do natural lakes. This is

partly related to the greater relative size of reservoir watersheds.

• Management. Reservoirs were built and are managed for specific purposes: hydropower, water

supply, and flood control. Fisheries and other uses are usually secondary. Management might

include extreme water level fluctuations and discharge depth controls, effects not present in most

natural lakes.

Most of the differences between reservoirs and natural lakes are resolved in the classification of the

lake resource. The needs for which they were designed dictate many of the attributes of artificial water

bodies. Operational strategies can influence reservoir characteristics and resultant water quality

(Kennedy and Walker, 1990; Kennedy et al., 1985). The release of water from deep in the water column

(hypolimnetic release) increases heat gain and the dissipation of materials accumulated in bottom waters

(Martin and Arneson, 1978; Wright, 1967). Surface releases dissipate heat and retain materials. These

and other operational differences can provide a basis for grouping reservoirs within and among regions.

Relationships between why a dam is built, how and where it is built, how it is subsequently operated,

and the characteristics of the resulting reservoir are reasonably well defined (Kennedy, 1999a). With

regard to the establishment of nutrient criteria, can we utilize these relationships to define appropriate

groups within which to identify reference conditions? Three categories of reservoir characteristics seem

germane for this purpose: location within a drainage basin; structural and operational characteristics of

the dam; and hydraulic residence time.

Location in Drainage Basin

Decisions about where dams are located broadly define physical attributes of the resulting reservoir,

which, in turn, strongly influence its limnological character (Kennedy, 1999a). For example, tributary

storage reservoirs are located on lower order rivers in the upland areas of drainage basins and, thus, often

Nutrient Criteria—Lakes and Reservoirs

3-5

reside in steeply sloping and dendritic basins with long, complex shorelines. Such reservoirs are

frequently relatively deep and strongly stratified. Inflows are often lower for suspended sediment

concentrations and may exhibit great seasonal or short-term variability. Changes in storage can result in

marked changes in pool elevation, and water residence times are often long.

By contrast, run-of-the-river and mainstem storage reservoirs, frequently operated to meet navigation

and hydropower objectives, are located on higher order river reaches. Run-of-the-river reservoirs often

are limited in lateral extent to the areas immediately adjacent to the original river channel and seldom

experience frequent or extensive changes in pool depth. Because they are commonly located at the

downstream extent of large drainage basins, they receive high suspended sediment loads, are turbid, and

flush rapidly. Dams for mainstem storage reservoirs, on the other hand, generally inundate broad river

flood plains, offering extensive storage volumes. Despite relatively high inflow rates, water residence

times can be long because of the large potential storage volume. Moderate changes in pool depth occur,

and in-reservoir inorganic turbidity levels, while initially relatively high due to riverine influences, often

are reduced because of long water residence times.

Dam Structure and Operation

The purpose or purposes for which dams are built determine, in general, their structural design and

their mode of operation. The location of outlet structures relative to the depth of the water column, as

well as the thermal structure of the water column, determine the depth strata from which water is

released. As discussed previously, withdrawal depth can have significant implications for reservoir

thermal cycles and the expression of trophic responses to changing nutrient levels. Thus, interactions

between reservoir depth, depth from which water is released, the volume of water released, and thermal

structure of the water column must be considered when assessing relative similarities between reservoirs

and lakes or among reservoirs.

As engineered systems designed to accomplish specific and often narrowly defined water control

objectives, dams and the reservoirs they impound exhibit prescribed characteristics dictated by functional

requirements. The scheduling of changes in reservoir volume (and, therefore, depth), for example, is

determined by basin capacity, hydrology, and water uses. From an operational standpoint, this often

involves the development and application of “rule curves,” or predetermined changes in reservoir surface

elevation. For tributary storage reservoirs, particularly those operated for flood control, rule curves

frequently require the lowering of surface elevations as a means to allow storage of subsequent flood

waters, which may be retained for extended periods of time before their release downstream. The result

is marked seasonal fluctuations in water column depth, reservoir volume, and water retention time.

Operational requirements for run-of-the-river reservoirs offer a contrasting example. Because the

primary purpose of such reservoirs is navigation, reservoir surface elevations must be controlled within

narrow limits. Thus, despite larger inflow volumes, rule curves for run-of-the-river reservoirs dictate

minimal fluctuations in water column depth. In the absence of changes in water storage volume, water

residence times are determined by hydrologic conditions and are uncoupled from dam operation.

From the above discussion and examples, it is obvious that dam structure and operation need to be

considered when evaluating factors that influence the limnological attributes of reservoirs and the

expression of trophic responses. In addition to their importance to the development of nutrient criteria,

these relationships also describe potential management opportunities unique to reservoirs.

3-6

Nutrient Criteria—Lakes and Reservoirs

Hydraulic Retention Time

Hydraulic retention time, defined as lake or reservoir volume divided by outflow rate and expressed

as days or years, strongly influences limnological processes in lakes and reservoirs (e.g., Straškraba et al.,

1993). These influences include changes in material retention rates (Straškraba et al., 1995; Kennedy,

1998), modifications to thermal structure, and impacts on the size and composition of planktonic

communities (Straškraba and Straškrabova, 1975; Soballe and Threlkeld, 1985; Soballe and Bachmann,

1984).

Residence times vary widely between natural lakes and reservoirs and among reservoirs. Thornton et

al. (1980) evaluated data for selected U.S. Army Corps of Engineers reservoirs and lakes contained in the

National Eutrophication Survey database, and the authors reported significantly higher geometric mean

values for lakes (270 days) than for reservoirs (135 days). A similar assessment of data included in the

National Inventory of Dams (U.S. Army Corps of Engineers, 1998) indicates a broad range in water

residence time for reservoirs impounded by U.S. dams (Kennedy, 1999b). Values for the nearly 65,000

reservoirs ranged from less than 1 day to more than 750 days; a similar range was observed for those

operated by the Corps of Engineers. Many of those with short residence times are operated as

run-of-the-river reservoirs for the purpose of navigation.

Ryding and Rast (1989) suggest that impoundment-related changes in water quality will occur when

doubling times for algae are less than water residence times. Since Reynolds (1997) suggests that algal

doubling rates are in the range of ½ to 1½ per day, it is clear that nutrient-related influences on trophic

state are possible at relatively short water residence times. For reservoirs with longer residence times,

anticipated differences in trophic responses between natural lakes and reservoirs will be minimized. In

such cases, it may be possible to include both natural lakes and reservoirs with similar water residence

times, assuming broad similarities in other attributes, in the same group when establishing reference

conditions or developing nutrient criteria. In regions with a limited number of reservoirs, this will allow

increased sample size for statistical treatments of the data.

On the basis of the above considerations, a reasonable and defendable approach to the identification

of appropriate groups of reservoirs would employ multiple descriptors based on operational and physical

attributes. Suggested measurement variables for each attribute are presented in Table 3.1. Taken

together, this suite of physical and operational characteristics attempts to define factors influencing the

expression of trophic responses to changing nutrient levels relative to reservoirs.

Therefore, reservoirs having short detention times should be considered separately from natural and

man-made lakes because of their different origins, morphometry, and hydrodynamics. Reservoir studies

have shown that the nutrient loading paradigm fits with some modifications (Canfield and Bachmann,

1981). The rapid flushing rates and longitudinal gradients that typify most mainstem reservoirs require

modifications of the models to account for down-reservoir changes in water from sedimentation and

dilution with passage through the system. Also, nutrient loading models that work well to explain in-lake

concentrations in natural lakes overestimate values measured in reservoirs (Jones and Bachmann, 1978);

for a given external load, reservoirs appeared to have lower in-lake phosphorus values than natural lakes.

Differences between reservoirs and natural lakes were thought to be tied to the fact that reservoirs are

constructed in erosional topography and receive much larger inputs of suspended solids than most natural

lakes. With greater sediment input, it follows that reservoirs would have greater sedimentation rates and

that more phosphorus would be lost from the water column as compared with natural lakes.

Nutrient Criteria—Lakes and Reservoirs

3-7

Table 3.1. Categories and attributes for a Composite Classification Approach for CE Reservoirs

and Suggested Measurement Variables (Kennedy, 1999b)

Category Attribute Measured Variable

Location and size Watershed dimension

Reservoir dimension

Drainage area

Location of dam in hydrologic continuum

Surface area

Volume

Len

gth

Mean and maximum depth

Shoreline development ratio

Hydrology

Hydraulic loading

Storage dynamics

Inflow and outflow rates

Annual/seasonal hydraulic retention time

Pool elevation/volume

Chan

ge in pool elevation

Structure and

operation

Dam design

Dam operation

Dam height

Outlet depth (relative to water column depth)

Quantit

y and seasonality of release volumes

Depth of release

Other response

effects

Light regime

Mixing regime

Nonalgal turbidity

Photic depth to mixed depth ratio

Thermal stability

Mixed layer depth

Another factor contributing to apparent differences in water column nutrient values is that reservoirs

typically have large watersheds (Canfield and Bachmann, 1981). As a result, inflow enters from a parent

river that, during stratified periods, forms a density flow below the warm surface water but above the

colder bottom waters (tropholytic zone) that does not mix or contribute nutrients to the photic zone

(Ford, 1990). Timing of nutrient-laden inflows relative to seasonal stratification can be as important as

their volume in controlling nutrient values within the water column. Water bodies with density currents

do not always show a response to external inputs. In these water bodies, loading models need to take into

account the effects of inflow timing and flow stratification. The relative timing of flow and stratification

will vary from year to year and could make nutrient content of the surface layer unpredictable except as a

long-term average.

�

Water Chemistry (nonnutrient)

Intrinsic water chemistry (not including nutrients) can be used to classify lakes. The most likely

variables include acid-base chemistry (any of alkalinity, pH, conductivity) and dissolved organic matter

(water color). Color and pH are cheaply and easily measured in the field and are therefore highly cost-

effective.

Lake water chemistry is largely determined by the hydrologic pathways of water entering the lake

and the material the water contacts along its path. Lakes with large inputs of water from shallow ground

water, including wetlands, tend to be stained yellow or brown with dissolved humic compounds. Water

entering a lake as deeper ground water tends to be clear but will contain the cations of the soils and

3-8

Nutrient Criteria—Lakes and Reservoirs

aquifer. Highly colored lakes have been termed dystrophic because they often are observed to have low

productivity in spite of moderate to high nutrient concentrations (Wetzel, 1975). Colored water not only

reduces light penetration, but the dissolved organic matter also can chelate nutrients, making them

unavailable for algal uptake. Therefore, water color is an important classification variable (or covariate;

see below) for lake nutrient criteria.

Alkalinity also influences lake productivity, in part because alkaline soils are richer in several

nutrients (especially phosphorus and potassium) than are acid soils, and because the nutrients are more

readily available to plants. The world

�

s most productive agricultural regions are in alkaline soils.

Alkalinity, or its related variables pH and conductivity, are important classification variables for nutrient

criteria. As an example, Florida lakes were characterized as acidic or alkaline and as colored or clear,

resulting in four lake types (Shannon and Brezonik, 1972). Although pH and as color are continuous

variables, it was more convenient to cluster the Florida lakes into four groups because response to

nutrient enrichment and macroinvertebrate communities also clustered according to the four groups

(Gerritsen et al., 1999).

�

Nonalgal Turbidity (suspended sediments)

High concentrations of nonalgal suspended materials are prevalent in lakes in many regions of the

world and can inhibit growth of phytoplankton, causing light limitation. Nonalgal turbidity from

suspended clay or organic matter is also strongly regional, depending on soil characteristics, vegetation,

and hydrology. It is a prominent characteristic of many impoundments in Midwestern and arid Western

States. Nonalgal turbidity can produce low algal chlorophyll-to-nutrient ratios and cause a lack of

relationship between chlorophyll and phosphorus in some regions (Jones and Novak, 1981; Hoyer and

Jones, 1983; Carlson, 1991; Jones and Knowlton, 1993). Light limitation of algal biomass in the mixed

zone of lakes occurs when irradiance absorbed by the phytoplankton community is less than is required

for net growth of biomass over time. Light limitation extended over periods of a week or longer is

common in deep or turbid lakes during winter because of low incident light, but it is less common in

summer when incoming irradiance is maximal and when mixing depth is reduced by thermal

stratification.

The Carlson trophic state index (TSI) (1977) can be used to identify certain conditions in the lake or

reservoir in which algal biomass is not related to phosphorus or nitrogen. When more than one of the

three variables are measured, it is probable that different index values will be obtained. Because the

relationships between the variables were originally derived from regression relationships and the

correlations were not perfect, some variability between the index values is to be expected. However, in

some situations the variation is not random, and factors interfering with the empirical relationship can be

identified. These deviations of the total phosphorus or the Secchi depth index from the chlorophyll index

can be used to identify errors in collection or analysis of real deviations from the “standard” expected

values (Carlson, 1980b). Some possible interpretations of deviations of the index values are given in

Table 3.2 (Carlson, 1983, 1992).

In turbid lakes, it is common to see a close relationship between the total phosphorus TSI and the

Secchi depth TSI, while the chlorophyll index falls 10 or 20 units below the others. Clay particles

contain phosphorus, and therefore, lakes with heavy clay turbidity will have the phosphorus correlated

with the clay turbidity while the algae may neither utilize all the phosphorus nor contribute significantly

to the light attenuation. This relationship of the variables does not necessarily mean that the algae is

limited by light, only that the measured phosphorus is not all being utilized by the algae.

Nutrient Criteria—Lakes and Reservoirs

3-9

3. Covariatesr

Several of the above factors have strong influences on lake trophic state and may be expected to vary

widely within nutrient regions. Whether a given factor needs to be considered separately in lake

classification within regions depends on its variability in the region and its regional relevance in affecting

trophic state. Additional classification factors may be treated as additional classes (e.g., as classes of

large and small lakes) or as continuous covariates (e.g., a regression model to predict natural trophic state

of lakes according to lake size). State and regional experts can use their knowledge of lake

characteristics to determine if any of the modifying factors should be considered as part of a State-level

classification scheme.

Lake morphometry and lake hydrology affect trophic state through the influence of water movement,

retention time, and stratification. In-lake phosphorus dynamics in mixed and stratified lakes can

complicate the relationship between external loading and measurements of lake trophic state, making

model-based predictions uncertain in some cases. In mixed lakes, phosphorus has been shown to

increase during the spring to summer period (Riley and Prepas, 1985), presumably because of recycling

due to mixing of the water column and internal loading from the sediments (Osgood, 1988; Welch and

Cooke, 1995). In contrast, it is typical for phosphorus to decrease somewhat in stratified lakes because

of sedimentation processes, with the metalimnion acting as a barrier to upward transport into the photic

zone.

Shallow lakes can efficiently cycle phosphorus and, under favorable light conditions, convert

phosphorus into phytoplankton biomass. As a consequence of these internal loading mechanisms,

shallow lakes do not always readily respond to reductions in external nutrient loading. Among large

stratified lakes, evidence exists that the efficiency of nutrient recycling increases with lake size; mixed

layers in large lakes are more turbulent and thicker than in small lakes, and these processes increase the

probability of nutrient regeneration within the mixed layer rather than loss to the sediments. The

response to greater nutrient regeneration is greater phytoplankton photosynthesis. These findings suggest

that external nutrient loads should be converted into biological production more efficiently in stratified

lakes than in small lakes.

Water residence time can have a significant effect on the amount of algae in the water. The water

must remain in the basin for a period longer than the doubling time of the algae or the algae will wash

Table 3.2. Conditions Associated with Various Trophic State Index Variable Relationships

Relationship Between TSI Variables Conditions

TSI (CHL) = TSI(CHL) = TSI(SD) Algae dominate light attenuation

TSI(CHL) > TSI(SD) Large particulates, such as Aphanizomenon flakes,

dominate

TSI(TP) = TSI(SD) > TSI(CHL) Nonalgal particulates or color dominate light attenuation

TSI(SD) = TSI(CHL) > TSI(TP) Phosphorus limits algal biomass (TN/TP ratio greater than

33:1)

TSI(TP) > TSI(CHL) = TSI(SD) Zooplankton grazing, nitrogen, or some factor other than

phosphorus limits algal biomass

3-10

Nutrient Criteria—Lakes and Reservoirs

out of the basin. This means that for faster growing algae such as Chlorella, the water residence time

would have to be at least 2 days. For slower growing species, the residence times would have to be much

longer. In reservoirs with short residence times, the algae would not necessarily reach densities in which

nutrients were limiting to their growth. However, residence time may vary seasonally, and there may be

times when the algae do become nutrient limited. Another consideration would be that in short residence

time situations, attached plants (both macrophytes and attached algae) may proliferate, and criteria would

have to be set in reference to attached rather than planktonic forms.

Any continuous variable, such as pH, color, lake depth, or lake area, can be treated as a covariate in a

classification. Often, it may be more convenient to treat them as discrete classes (large and small; acid

and alkaline). Whether to treat an important classification variable as discrete classes or as a continuous

covariate may depend on the size of the database, the distribution of lakes across the gradient, and

whether the trophic response to the gradient is linear. For example, a uniform or unimodal distribution

across a gradient would suggest treating the variable as a covariate (e.g., lake surface area), while a

bimodal distribution would suggest dividing into classes (e.g., pH classes of acidic and nonacidic lakes).

If a relationship is found between measures of lake size or hydrology and trophic state, then additional

classes reflecting these factors, or treating them as a covariate, must be considered. Because

relationships with morphometric variables are often linear, the covariate approach usually is preferred for

those variables (area, depth, retention time). Conversely, water chemistry variables (pH, color, hardness)

may cluster naturally due to geology, soils, and vegetation, so the class approach may be preferred for

those variables.

Nutrient Criteria—Lakes and Reservoirs

3-11

Case Study: Ecoregional Classification of Min esota Lakesn

Minnesota has over 12,000 lakes spread across diverse geographic areas. Previous studies had shown

distinct regional patterns in lake productivity associated with regional differences in geology,

vegetation, hydrology and land use (Heiskary and Wilson, 1989). Minnesota contains seven

ecoregions (Omernik, 1987), and four of the ecoregions contain 98 percent of the lakes. These four

ecoregions are the Northern Lakes and Forest (NLF), North Central Hardwood Forest (NCHF),

Northern Glaciated Plains (NGP), and Western Corn Belt Plains (WCBP) (Figure, p. B-4). Minnesota

uses these ecoregions as the framework for analyzing data, developing monitoring strategies,

assessing use patterns, and developing phosphorus goals and criteria for lakes (Heiskary, 1989).

The Minnesota Pollution Control Agency (MPCA) and several other groups collected data on

chlorophyll a concentrations and several water quality parameters (total phosphorus, total nitrogen,

and Secchi transparency) in 90 reference lakes between 1985 and 1987. Secchi transparency data

were collected mostly by volunteer participants in the Citizen Lake Monitoring Program. Reference

lakes were chosen to represent minimally impacted sites within each ecoregion. Criteria used in

selecting reference lakes included maximum depth, surface area, fishery classification, and

recommendations from the Minnesota Department of Natural Resources (Heiskary and Wilson, 1989).

Lake morphometry had previously been examined. In addition to the reference lake database, MPCA

examined a statewide database containing data collected by these same groups on approximately

1,400 lakes form 1977 to 1987.

Differences in morphology, chlorophyll a concentrations, total phosphorus, total nitrogen, and Secchi

transparency were found among lakes in the four ecoregions in both studies. Lakes in the two

forested ecoregions (NLF and NCHF) are deeper (median maximum depth of 11 meters) with slightly

smaller surface areas (40 to 280 ha) than those in the plains ecoregions (NGP and WCBP). Lakes in

the two plains ecoregions were typically shallow (median maximum depth of 3 meters) with larger

surface areas (60 to 300 ha).

Box-and-whisker plots for chlorophyll a and water quality measurements in the reference lake study

paralleled the morphological differences seen among the ecoregions (Heiskary and Wilson, 1989).

The two plains ecoregions had significantly higher chlorophyll a levels than either of the two forested

ecoregions. Results of the statewide database analysis showed these same trends. The results of these

two database analyses support the use of ecoregions in developing frameworks for data analysis,

monitoring strategies, assessing use patterns, and developing phosphorus goals and criteria for lakes.

3-12

Nutrient Criteria—Lakes and Reservoirs

CHAPTER 4

A. Introduction

B. Evaluating Existing Data

C. New Data Collection

Establishing an

D. Database Mana

gement

Appropriate Database

A. Introduction

The development of nationwide regional nutrient criteria requires the availability of an extensive

amount of data from across the country for evaluation. Data may come from existing sources or can be

collected from new sampling programs. Nutrient-related data for lakes and reservoirs, collected by

various agencies for many different purposes, exist in various databases and have the potential to provide

the basis for the development of nutrient criteria on a regional level. This chapter presents an overview

of existing nutrient criteria databases and presents a general discussion on the evaluation of such data in

terms of their use in the nutrient criteria development process. The chapter also provides a description of

the process undertaken by EPA to use existing data from STORET and perhaps other existing data sets

(e.g., U.S. Geological Survey [USGS] NAWQA) to generate preliminary nutrient criteria on an

ecoregional level. In addition to discussing the use of existing data, the chapter discusses new data

collection, including sampling design and the types of monitoring to be considered as part of data

collection activities. The chapter ends with a general discussion of data management issues that are

integral in the overall discussion of data storage and accessibility.

B. Evaluating Existing Data

In many States, historical data on lakes are extensive and may be sufficient to identify lake classes

and reference lakes and to establish interim criteria. With societal interest in cultural eutrophication

during the 1970's—programs such as EPA

�

s Clean Lakes Program (CLP) and citizen

�

s organizations

such as LakeWatch— lakes have been extensively monitored and studied. Many historical data are

available and may suffice for developing lake nutrient criteria. Although existing data may be sufficient,

program managers should recognize that gathering, reducing, analyzing, and interpreting existing data

sets can be costly and time consuming.

Development of nutrient criteria using the reference site approach requires a database that can be

used to characterize (1) reference lakes and (2) the biological response of lakes to nutrient enrichment.

At a minimum, observations should include total nitrogen (TN) (or Kjeldahl nitrogen plus NO

-NO

-N),

total phosphorus (TP) (or total dissolved phosphorus plus particulate phosphorus), chlorophyll a, and

Secchi depth. For each lake in the database, there should be information or inference on the status of

anthropogenic nutrient loading to the lake. At a minimum, this information would be informed best

professional judgment on whether anthropogenic nutrient loading was negligible or substantial for a

given lake. The judgment could be based on personal observation, discharge information, land use

information, or historical information. Such information may not reside together with the water quality

observation and may need to be found and obtained separately.

As mentioned above, the water quality data should be sufficient to characterize a lake. All lakes in

the database should have been sampled the same way and should be characterized the same. Ideally,

lakes will have been sampled once during an index period that characterizes nutrient or trophic state (e.g.,

Nutrient Criteria—Lakes and Reservoirs

4-1

during spring overturn) and with sampling methods that are assumed to characterize the lake (e.g.,

pumped or composite sample of the entire water column). Some existing data sets may permit estimation

of annual or growing season average nutrient concentrations.

Appropriate data analysis will be determined by the data set. The basic procedure is to consider each

lake an independent sample unit and to estimate an annual characteristic value (annual average, median,

minimum, maximum) of each water quality observation for each lake. These annual characteristic values

are the information used to develop nutrient criteria and biological response of lakes to enrichment. This

procedure assumes that lakes are independent and that annual averages among years are independent.

The investigators and the Regional Technical Assistance Groups should decide whether the

independence assumptions are reasonable and whether any modifications should be made.

1. Potential Data Sources

Databases could include water quality monitoring data from water quality agencies (often stored in

STORET), national surveys such as the EPA Eastern Lakes Survey (Linthurst et al., 1986) or the

Environmental Monitoring and Assessment Program (EMAP) (Paulsen et al., 1991), limnological studies,

and volunteer monitoring information.



STORET

STORET is EPA

�

s national data warehouse for water quality data. All State and Federal water

quality monitoring agencies are required to submit their data to STORET. STORET is huge, covering all

50 States since the 1970's; it includes lakes, streams, rivers, and estuaries and physicochemical and

biological data. STORET has no quality control for accepting or rejecting data, but it does require

extensive metadata (data descriptors) that show how data were collected and what analytical methods

were used. All sampling sites are referenced by latitude and longitude, by the EPA Reach File 3 (RF3),

and by USGS hydrologic unit codes. Extracting data from STORET requires familiarity with the system

as well as selection criteria (date ranges, location ranges, specified measured variables, collecting agency

with known quality control) to keep from being overwhelmed with irrelevant data.



National Eutrophication Survey (NES)

EPA conducted NES in the early 1970's. Several hundred lakes were sampled and nutrient budgets

were estimated. The lakes selected for the survey received discharges from municipal sewage treatment

or the States requested they be included. NES contains a broad but incomplete sample of lakes, and

therefore, the data from NES are not sufficient in themselves for developing reference conditions to

support regional nutrient criteria. The NES data may be used for determining biological responses to

enrichment and for developing site-specific criteria.



National Surface Water Survey (NSWS)

EPA conducted NSWS in the mid 1980's under the National Acid Precipitation Assessment Program.

Lakes were surveyed only in those regions where they were initially thought to be at risk to acid

precipitation: New England, the Adirondacks, the mid-Atlantic highlands, the mid-Atlantic coastal plain,

the southeastern highlands (southern Appalachians and Ozark-Ouachitas), Florida, the upper Midwest,

and the montane West. NSWS sampled 2,300 lakes ranging in size from 4 to 2,000 hectares; thus, the

smallest and largest lakes were not represented. NSWS was a stratified random sample of lakes in the

4-2

Nutrient Criteria—Lakes and Reservoirs

selected region; therefore, inferences can be made to the populations defined by the regions. Sampling

took place in the fall. Nutrients were measured, so NSWS data could be used to help develop criteria.



Environmental Monitoring and Assessment Program

EPA

�

s EMAP sampled lakes in New England and the Adirondacks in 1991, with a probability-based

site selection procedure to obtain an unbiased sample (Paulsen et al., 1991). Further similar work is

planned in the western States. The EMAP lake sample should include potential reference lakes as well as

stressed lakes and could be used to identify reference lakes and characterize reference conditions.



Clean Lakes Program

The EPA CLP for restoring public lakes included a monitoring and assessment component. Lakes in

this program were selected because they were perceived to have water quality impairment. Like data in

the NES database, these data must be carefully scrutinized before being included in a database setting

regional reference conditions to support nutrient criteria.



Volunteer Monitoring Programs

• Individual State Lake Association Programs that may contain considerable information, may be

contacted through the National Directory of Environmental Monitoring Programs (contact:

Alice Mayo, U.S. EPA headquarters, Washington, DC).

• National data on Secchi disc transparency have been collected since 1994 by the “Secchi Dip-In”

(Carlson et al., 1997).

Elements of these databases could contribute to criteria development but, like the EMAP data, would

need to be screened to identify reference and nonreference lakes.



State Monitoring Programs

Most States monitor some subset of lakes and impoundments within their borders for eutrophication

and nutrient variables. Several of the more extensive lake monitoring programs (e.g., Minnesota,

Wisconsin, Maine, Florida) are profiled in this document as examples of using monitoring data to help

develop nutrient criteria. The purpose of the survey should be assessed before using the data. See

Representativeness, in Section 2 below.



U.S. Army Corps of Engineers

The U.S. Army Corps of Engineers is responsible for more than 750 reservoirs. Extensive

monitoring datahave been collected for many of these reservoirs that could contribute to the development

of nutrient criteria for reservoirs.



U.S. Department of the Interior, Bureau of Reclamation (BuRec)

The Bureau of Reclamation manages many irrigation and water supply reservoirs in the West. Data

from their operations may be available for some of these.

Nutrient Criteria—Lakes and Reservoirs

4-3



Electric Utilities

Many electric utilities own reservoirs for hydroelectric power generation, and the utilities are

required to monitor the reservoirs

�

water quality. The largest of these, the Tennessee Valley Authority,

has extensive chemical and biological monitoring data from most of its reservoirs from the early 1980's

to the present.

2. Quality of Historical Data

The quality of older historical data sets is a recurrent problem because the data quality is often

unknown. This is especially true of long-term repositories of data such as STORET and long-term State,

academic, commission, or municipal databases, where objectives, methods, and investigators may have

changed many times over the years. The most reliable data tend to be those collected by a single agency

using the same protocol for a limited number of years. Supporting documentation should be examined to

determine the consistency of sampling and analysis protocols.

When “mining” from large heterogeneous data repositories such as STORET, investigators must

screen data for acceptance considering a number of variables, as discussed below.



Location

STORET data are georeferenced with latitude, longitude, and RF3 codes. These can be used to

select specific locations or specific USGS hydrologic units. In addition, STORET often contains a site

description. If selecting, for example, all lake sites within a geographic region, it is also important to

know the rationale and methods of site selection by the original investigators. Such information may be

included in STORET metadata, if known.



Variables and Analytical Methods

Thousands of variables are recorded in STORET records. Each separate analytical method yields a

unique variable (called parameter in STORET); thus, five ways of measuring TP results in five unique

variables. Because methods differ in accuracy, precision, and detection limits, it is generally unwise to

mix methods in the same analysis. If there is one method that the investigator judges to be best, then

only observations using that particular method can be selected. Selection of a particular “best” method

may result