COMMENTARY

Electronic copy available at: http://ssrn.com/abstract=2265478

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

151

“Not Supported by Current Science”: The

National Forest Management Act and the

Lessons of Environmental Monitoring for

the Future of Public Resources

Management

Ryan P. Kelly*

Margaret R. Caldwell

I. INTRODUCTION ...........................................................................

152

II. I

NDICATORS GENERALLY ............................................................. 154

III. T

HE RISE OF ENVIRONMENTAL MONITORING AND INDICATOR

PECIES ........................................................................................ 157

A. Environmental Monitoring and the Need to Simplify a

Complex System ............................................................... 157

B. Indicator Species as a Type of Environmental

Monitoring ........................................................................ 161

IV. NFMA

AND “MANAGEMENT INDICATOR SPECIES” (MIS) .......... 163

A. The Statutory Planning Scheme ...................................... 163

B. NFMA’s Implementing Regulations ................................ 164

C. The 2012 Planning Rule Revision, and the Fate of Previous

Revisions ........................................................................... 167

V. P

OST-1982 SCIENCE OF ENVIRONMENTAL MONITORING ............ 170

A. Problems with the Indicator Species Concept(s) ........... 170

* Analyst for Science, Law, and Policy, Center for Ocean Solutions, Stanford Univer-

sity. J.D., University of California, Berkeley, School of Law (Boalt Hall), Ph.D., Columbia

University. Email: rpk@stanford.edu.

** Executive Director, Center for Ocean Solutions, and Director, Environmental and

Natural Resources Law & Policy Program, Stanford University. J.D., Stanford University.

Email: megc@law.stanford.edu.

Electronic copy available at: http://ssrn.com/abstract=2265478

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

152 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

B. Toward Holistic Measures of Ecosystem Structure and

Function ............................................................................ 174

C. Best Practices of Environmental Indicators .................... 176

VI. E

VALUATING THE MANAGEMENT INDICATOR SPECIES (MIS)

ROVISION IN LIGHT OF SUBSEQUENT SCIENCE ......................... 179

A. Evaluation in Light of Best Practices Developed for

Indicators of Environmental State ................................... 183

B. Evaluation in Light of Implicit Purposes......................... 190

VII. L

ESSONS OF MIS FOR FUTURE PUBLIC RESOURCES MANAGEMENT,

AND FOR COASTAL AND OCEAN PLANNING IN PARTICULAR ........ 196

A. Marine Spatial Planning as Closely Analogous to National

Forest Planning ................................................................. 197

B. Lessons from the Experience of NFMA’s Management

Indicator Species .............................................................. 201

VIII. C

ONCLUSION .............................................................................. 211

INTRODUCTION

Environmental monitoring presents a fundamental challenge:

how should we measure a complex system with enough specificity

to be accurate, but with enough generality to be useful? From this

central challenge flow related difficulties; for example, tracking

the environmental impacts of particular development or manage-

ment decisions, or embedding methods of analysis in a regulatory

framework. Thirty-plus years of science have yielded enormous in-

sight into the structure and function of the ecosystems in which

human activities take place, but higher-resolution views of the sys-

tem’s complexity do not automatically point to the best means of

measuring our effects on the state of the environment.

Nevertheless, a wide (and growing) variety of laws and regula-

tions relies on environmental monitoring as a critical tool to assess

the impact of particular human activities on the ecosystems on

which we depend.

1. The authors wish foremost to acknowledge the significant contribution to this pa-

per from Matt Armsby. Further valuable input—variously through comments, conversa-

tion, and research—also came from Erin Prahler, Ashley Erickson, Jack Kittinger, Debbie

Sivas, Melissa Foley, David Weiskopf, Eric Biber, Nell Green Nylen, Kai Lee, and Kate

Wing.

As a legal proposition, this makes eminent

2. See generally Eric Biber, The Problem of Environmental Monitoring, 83 U. COLO. L. REV.

1 (2011) (describing the rise of ambient monitoring in environmental regulation, and the

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 153

sense: how else might we measure the success or failure of laws

meant to safeguard environmental public health and safety, or es-

timate the potential impacts from a proposed change in land use?

But as a scientific question, environmental monitoring is far less

straightforward than it might seem.

Legal and regulatory regimes routinely assume a level of scien-

tific certainty (or the existence of particular scientific tools) that is

unwarranted, leaving the implementing agencies scrambling to

fulfill regulatory requirements. In the case of environmental as-

sessment—required in one form or another by NEPA,

NFMA,

FLPMA,

CAA,

CWA,

CZMA,

It isn’t that such metrics don’t exist. Rather, the problem is that

too many such metrics exist. There is a nearly infinite number of

ways to describe the state of an ecosystem. Further, such methods

require data that is often expensive and difficult to collect. And

because science tends to change much faster than do statutes or

regulations, the scientific challenges compound the longstanding

problem of how to enshrine an appropriate level of specificity in

environmental regulation. This risks leaving agencies with an on-

erous top-down mandate (where regulation is too specific, requir-

ing a particular monitoring methodology), an overgenerous

amount of regulatory discretion (regulation is too broad, leaving

agencies vulnerable to agency capture and/or leading to erratic

outcomes), or in some cases, both.

and other major environmental

statutes—a core legal assumption is the existence of metrics of en-

vironmental “health” or state.

The case of the National Forest Management Act provides a viv-

id example of these issues: a legal mandate to include biological

diversity among Forest management priorities, a regulatory im-

plementation creating an expensive and unproductive mismatch

with the science of monitoring, a seemingly clear agency mandate

that nevertheless provided broad discretion, and a regulatory

scheme without a “best available science” provision that left sci-

ence frozen in time.

increasing importance of monitoring generally).

3. National Environmental Policy Act of 1969, 42 U.S.C. §§ 4321-4370(h) (2012).

4. National Forest Management Act of 1976, 16 U.S.C. §§ 1600-1687 (2012).

5. Federal Land Policy and Management Act of 1976, 43 U.S.C. §§ 1700-1786 (2012).

6. Clean Air Act of 1963, 42 U.S.C. §§ 7401-7671(q) (2012).

7. Clean Water Act of 1972, 33 U.S.C. §§ 1251-1387 (2012).

8. Coastal Zone Management Act of 1972, 16 U.S.C. §§ 1451-1466 (2012).

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

154 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

A review of the science of monitoring and its legal applications

is particularly appropriate now, both as the Forest Service’s 2012

NFMA planning rule

takes effect and as the federal government

considers embarking on an analogous comprehensive public-

resource management endeavor: ocean planning. The nation’s

first-ever National Ocean Policy

—which was announced in 2010

and which is just now beginning implementation—contains many

elements that move management of the ocean as a natural re-

source into the 21

century. Among these are ecosystem-based

management and coastal and marine spatial planning (CMSP)

In this Article, after discussing some general properties of indi-

cators, we review the Management Indicator Species (MIS) provi-

sion of the 1982 NFMA regulations, one of the statute’s relevant

environmental monitoring requirements. We then place this re-

quirement in the research context from which it arose, applied

ecology, and discuss the subsequent development of the science of

environmental monitoring. We go on to evaluate the National

Forests’ MIS in light of guidelines for good environmental indica-

tors developed in this subsequent literature. Finally, we introduce

ocean and coastal management as a context relevant for applying

the lessons of NFMA’s monitoring provisions, and suggest some

concrete steps that the managing agencies might take to avoid the

mistakes of past environmental monitoring efforts in public re-

sources management. In particular, these steps require intense fo-

cus on the what, why, and how of environmental monitoring.

unify the disparate regulatory regimes governing the United

States’ vast territorial waters. These waters are in many ways analo-

gous to the National Forests, being sparsely populated, multiple-

use public resources that are difficult to monitor due to their vast

sizes, and so the lessons of forest planning are especially salient as

we move toward improved management of the nation’s ocean re-

sources.

II.

INDICATORS GENERALLY

An environmental indicator is a “measurable surrogate[] for

9. 36 C.F.R. § 219 (2012).

10. Exec. Order No. 13547, 75 Fed. Reg. 43023 (Jul. 22, 2010).

11. N

AT’L OCEAN COUNCIL, DRAFT NATIONAL OCEAN POLICY IMPLEMENTATION PLAN

8 (2012), available at http://www.whitehouse.gov/sites/default/files/microsites/ceq/ na-

tional_ocean_policy_draft_implementation_plan_01-12-12.pdf (National Priority Objec-

tives 1 and 9).

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 155

[an] environmental end point[],”

or more generally, “a parame-

ter, or a value derived from parameters, which . . . describes the

state of a phenomenon/environment/area, with a significance ex-

tending beyond that directly associated with a parameter value.”

Body temperature is a familiar example that illustrates the con-

cept nicely in the context of human (rather than environmental)

health. A person’s body temperature is an indicator of her state of

health, because a fever is correlated with an infection. Although we

are actually interested in the patient’s state of health (and not her

temperature for its own sake), we take her temperature because it

is much more easily measured than viral load or any more direct

measure of infection. In order for body temperature to be an ef-

fective indicator of health, we must first understand something

about the relationship between temperature and infection, and

about the background variation of body temperature in healthy

individuals.

Thus, as we use the term here, an indicator is not itself the meas-

urement of interest, but rather is an easier-to-measure stand-in for

the actual measurement target.

In the context of environmental monitoring, indicators are

measurement tools that allow researchers to track environmental

state, structure, or function. These provide critical feedback for

public resources management, making it possible to track the

managed natural resource as it changes in response to manage-

ment actions. Indicators may take the forms of direct measures of

physical, chemical, or biological parameters. Particularly relevant

in the case of the National Forests, discussed at length below, are

indicator species—biological entities monitored because their pres-

ence indicates something about the state of the larger ecosystem.

There is no best way to summarize the state of the environ-

12. Reed F. Noss, Indicators for Monitoring Biodiversity: A Hierarchical Approach, 4

CONSERVATION BIOLOGY 355, 357 (1990).

13. O

RG. FOR ECON. CO-OPERATION AND DEV., OECD ENVIRONMENTAL INDICATORS:

DEVELOPMENT, MEASUREMENT AND USE 5 (2003), available at http://www.oecd.org/env/.

14. Note that this is an imperfect metaphor for ecological indicators because human

body temperature is an equilibrial value. There is a narrow range of body temperatures

that a human body will sustainably maintain; excursions from that range tend to be brief,

and body temperature quickly returns to its equilibrial (that is, homeostatic) value. By con-

trast, environmental variables may have multiple stable states, or none at all.

15. Indicator species have also been used as proxies for other (harder to count) spe-

cies or communities, although this use is disfavored among ecologists. See Gerald F. Niemi

& Michael E. McDonald, Application of Ecological Indicators, 35 A

NN. REV. OF ECOLOGY,

EVOLUTION, & SYSTEMATICS 89 (2004); see also discussion infra p. 120.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

156 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

ment. The cacophony of interactions among species, nutrients,

climate, and geography that occurs within even the smallest forest,

grassland, or estuary calls into question the desirability of develop-

ing an index of environmental “quality.” An index of any kind re-

quires summarizing and simplifying—indeed, that is the point—

which loses information in the process. An optimal index strikes a

balance between increased comprehension (gained through sim-

plifying a complex system) and the attendant loss of detail. But

precisely where the optimal balance lies depends upon at least two

things: what we want to know, and why we want to know it.

Compare, for example, the Dow Jones Industrial Average with

the S&P 500. The two indices have similar purposes, measuring the

state of the American economy and acting as yardsticks for chang-

es to that state.

The Dow represents a weighted average stock

price of 30 companies

whose stocks are listed on the New York

Stock exchange, while the S&P is a weighted index of the stock

prices of 500 such companies.

The S&P therefore represents a

broader swath of the market, but changes to any one stock price

are likely to be swamped out in the process of weighting and aver-

aging the values of all 500 companies.

The Dow, conversely, is

more sensitive to changes in any one of its constituent companies’

stock prices. Neither can be said to be the “better” index; they

simply have different strengths. If one cares about the prices of

blue chip companies to the exclusion of all else, the Dow is proba-

bly of more interest, while the S&P is likely to be a better snapshot

of market-wide changes. Note, however, that neither provides any

information about any individual stock: such resolution is neces-

sarily lost by creating an index.

16. Note, too, that the stock price of a company is itself a kind of index, a summary

of the company’s worth or future prospects in the eyes of the stock-buying public. We use

the analogy to economic indicators with apologies to Jameal Samhouri and coauthors, who

seem to have beaten us to the punch on this point. See J.F. Samhouri et al., Using Existing

Scientific Capacity to Set Targets for Ecosystem-Based Management: A Puget Sound Case Study, 35

MARINE POL’Y 508, 509 (2011). At the time of writing, we had not yet become aware of

their paper, in which they use the same comparison.

17. See Overview, S&P DOW JONES INDICES,

http://www.djaverages.com/?go=industrial-overview (last visited Jan. 13, 2013).

18. See S&P

DOW JONES INDICES, http://us.spindices.com/ (last visited Jan. 13, 2013).

19. Note that both the Dow and the S&P are weighted such that changes to the stock

price of companies with a higher market capitalization will influence the index to a greater

degree. Nevertheless, the broader point remains true.

20. Just as when measuring an index of species diversity, information about the indi-

vidual constituent species is necessarily lost.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 157

The creation of indices is thus a value-laden and purpose-

bound process. Selecting which variables to include and how to

weight those variables will determine the index’s output, and re-

quires a value judgment regarding which variables are most im-

portant. Such judgment, whether implicit or explicit, goes to the

critical question: what do we want to know? In the case of the stock

market, including some stocks in a particular index will necessarily

force us to pay less attention changes to others. Put another way,

we remain willfully ignorant about some aspects of the market, fil-

tering them out as background “noise” in order to focus on the

signal coming from a subset of particularly interesting stocks.

Closely related is the fact that the utility of an index’s output de-

pends strongly upon the particular study question at hand. Using

the economic example again, if one is interested in the health of

the steel industry, a tech-heavy index is unlikely to be the most in-

formative barometer.

In an ecosystem context, indices carry the same set of limita-

tions. Like an economy, an ecosystem is a complex of interactions

among a vast number of interacting units at different hierarchical

levels of organization. Perhaps we are interested in the physical

structures of an ecosystem as a way of summarizing such complexi-

ty; or maybe quantifying nutrient flows into and out of a circum-

scribed area would be more helpful for a given purpose. But be-

cause tradeoffs are inherent in crafting an indicator of any type,

durable and effective sets of indicators may only be selected after

an explicit assessment of the purpose and goals of that particular

indicator set.

III. THE RISE OF ENVIRONMENTAL MONITORING AND INDICATOR

SPECIES

A. Environmental Monitoring and the Need to Simplify a Complex System

The late 1960s saw the rise of environmental awareness in the

United States leading to the landmark federal environmental stat-

utes of the early 1970s: NEPA,

ESA,

CWA,

CAA.

21. The lack of explicit goals and purposes for indicators is perhaps the most com-

mon criticism in the academic literature regarding the use of indicator species in ecology

and conservation biology, and has remained so for decades. NFMA’s regulations requiring

MIS are a prime example of such an omission.

These stat-

22. 42 U.S.C. §§ 4321-4370(h) (2012).

23. Endangered Species Act of 1973, 16 U.S.C. §§ 1531-1544 (2012).

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

158 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

utes and their implementing regulations demanded ways of mak-

ing abstract concepts—such as environmental health and integri-

ty—into concrete and measurable quantities.

This changing legal landscape likely accelerated the growth of

applied ecology as a discipline, as researchers began to formalize

ways of summarizing nature’s complexity, simplifying the dense

tangle of organismal interactions in just the same way the Dow

Jones Industrial Average provides a window into the workings of

the American economy. An analogous yardstick to measure chang-

es in environmental condition was an attractive goal for academics

seeking field data, and had immediate policy relevance for agen-

cies newly charged with environmental responsibilities.

We can understand most of the resulting environmental meas-

urement techniques in terms of a black box model: the addition of

some input (or stressor) results in a change to some environmental

outcome. We need not understand the mechanism by which the

input results in the outcome, so long as we can predict a given out-

come from a particular level of input. All methods of ecosystem as-

sessment focus on inputs, outcomes, or both, with environmental

indicators being cheap or easy ways of measuring inputs or out-

comes.

The simplest examples of biological monitoring to measure an

ecosystem parameter are bioassays, EPA-mandated animal-based

tests for acute and ambient toxicity of effluents. The EPA sought to

establish dose-response curves, calculating the effect of an input

(say, selenium in effluent) on an outcome (the detrimental effect

on a particular species of fish, for example). Bioassays are straight-

forward tests of both input and outcome, and establish a clear rela-

tionship between the two.

In the wake of the environmental legislation of the 1970s,

applied ecologists developed a variety of competing and comple-

mentary environmental metrics, a scientific give-and-take that can

be seen as a struggle for the appropriate hierarchical level of focus.

24. 33 U.S.C. §§ 1251-1387 (2012).

However, they focus on the narrow

25. 42 U.S.C. §§ 7401-7671(q) (2012).

26. Note that a danger of selecting such indicators is that it tends to focus attention

on the indicator rather than the indicated, losing the larger point of environmental quali-

ty/health/function, by instead fixating on particular species or metrics.

27. A more familiar example of the same principle is the canary in a coal mine. See

infra p. 112. The survival of the canary (the outcome) indicates the absence of toxic levels

of methane or carbon monoxide in the mine (the input). As above with the human body

temperature example, we must first understand the relationship between toxic gas levels

and the canary’s survival before we may treat the canary as an indicator of toxic gases. Fur-

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 159

questions of determining single-stressor, single-species toxicity lev-

els, rather than more complex ecosystem-level assessments.

Such narrow focus is problematic, as it fails to address many of

the underlying motivations for environmental regulation.

For ex-

ample, the Clean Water Act (CWA) defines water quality with re-

spect to particular designated uses of water bodies,

measuring

individual chemical parameters against national criteria. Hence,

water “quality” is defined in large part by the absence of high lev-

els of pollutants.

Following the passage of the CWA, ecologists therefore criti-

cized its means of water quality assessment as overly reductive and,

moreover, ineffective. An ecosystem is the entire set of interactions

among species and nonliving components of an environment

(such as temperature or sunlight), and therefore merely tracking

pollutant levels is no measure of ecosystem integrity because it

misses much of what defines an ecosystem. At the same time, how-

ever, a catchall measure of the health of a complex ecosystem was

(and is) elusive for the same reasons as described above for eco-

nomic systems: any index is a tradeoff between loss of information

and increased simplicity, and there is no all-purpose “best” meas-

ure. Moreover, there is no one equilibrial state to an ecosystem (in

contrast to the temperature of the human body), such that even if

there were a “best” measure of ecosystem state, it would remain

unclear what value that measure should take in any given case. The

seemingly-simple questions of what do we want to know? and why do

This approach is equivalent to measuring some

of the inputs to the system, but not measuring the outcome: we

track levels of stressors, but no resulting measure of “quality” or

environmental state itself. Many water bodies may be appropriate

for a particular designated use—swimming pools are “swimma-

ble,” for example—but they are as far from a functioning ecosys-

tem as one might imagine.

ther, the canary is not an indicator of “mine safety” more generally until we understand

the relationship between the canary’s requirements for life and our own.

28. For example, the purpose of the Clean Water Act is “to restore and maintain the

chemical, physical, and biological integrity of the Nation’s waters.” 33 U.S.C. § 1251(a)

(2012). The idea of environmental “integrity” necessarily entails a broad focus on the

structure and function of the environment in question.

29. Examples of beneficial uses include recreation, fishing, shellfish culture, and

many others.

30. A water body’s acceptability for particular beneficial uses is also a measure of wa-

ter quality, insofar as those uses are quality-dependent. This may be seen as an ecosystem

“output,” albeit one with only tenuous connections to ecosystem composition, structure,

or function.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

160 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

we want to know it? are surprisingly slippery when applied in a real-

world context.

Researchers of the 1970s focused on different hierarchical lev-

els of ecosystem organization to attack this problem: some zoomed

in (looking at species or other component parts of an ecosystem),

some zoomed out (focusing on system-wide variables, such as the

change in species composition over space). Those that zoomed out

suggested holistic measures best captured a portion of ecosystem

complexity, or sought to measure fundamental processes as indices

of ecosystem function. For example, measures of energy or nutri-

ent flow through a lake would provide a view of the “function” of

that lake in the context of global energy or nutrient cycles.

The ecologists that zoomed in used smaller-scale, more field-

friendly methods, such as monitoring individual species thought to

be surrogates for larger ecosystem processes.

But

quantifying and measuring these properties can be costly and diffi-

cult, and the remote sensing data that today makes these meas-

urements easier was not yet widely available in the 1970s and early

1980s, making these holistic measures less useful in practice at the

time.

But of course, any single-species measure necessarily failed to

capture much of what we think of as important about an ecosys-

tem. Single-species indicators are akin to choosing a single compa-

ny’s stock to measure the state of the New York Stock Exchange.

The ecological solution to this problem mirrored the economic

approach: creating indices that combined multiple species’ trends

into a single, trackable, number. One particular index of note was

the Index of Benthic Invertebrates (IBI),

A species that is es-

pecially sensitive to change in habitat, for example, might be a

good stand-in for change to that habitat. This approach had the

additional benefit of integrating the effects of stressors over

time—a change in the sensitive species would reflect changes to

the habitat now or at any time in the recent past, unlike periodic

chemical monitoring, which is likely to miss discrete events that

impact habitat (such as the sudden release of a pollutant).

31. See, e.g., William H. Schlesinger, Community Structure, Dynamics and Nutrient Cycling

in the Okefenokee Cypress Swamp-Forest, 48

ECOLOGICAL MONOGRAPHS 43 (1978).

which occupied a kind

32. See infra p. 112 and note 83.

33. James R. Karr, Assessment of Biotic Integrity Using Fish Communities, 6

FISHERIES 21

(1981) (providing an early and widely-cited example of the biotic community approach to

monitoring freshwater environments).

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 161

of compromise position between species-focused and ecosystem-

focused metrics, combining elements of species diversity and larg-

er ecosystem processes into a single metric. But this and other

multimetric indices shared many of the drawbacks of species-based

approaches, insofar as they were place-specific and required a high

level of expertise in a particular ecosystem to implement.

It was into this area of active ecological research that the NFMA

regulations were born, in which monitoring individual indicator

species became a mandatory feature of forest management in the

United States.

B. Indicator Species as a Type of Environmental Monitoring

By the mid-1970s the indicator species concept—that is, the

idea of monitoring one or a few species as an indicator of some

larger process or state—was appearing in the academic literature

as a method of environmental monitoring. This was an attractive

monitoring tool because it was cheaper and easier than many of

the alternatives, and because monitoring species captured at least

some level of information about the larger ecosystem: the occur-

rence of particular species demonstrates the existence of the eco-

system services necessary to support that species, as well as the on-

going absence of lethal conditions.

But important details of using species as indicators remained

fuzzy, details that would be necessary to make the concept of indi-

cator species useful in practice. For example, authors differed

widely regarding exactly which ecosystem states indicator species

might depict, how the connections between indicators and ecosys-

tem states might be verified, and how to go about selecting indica-

tors. One 1974 paper on the use of indicator species to measure

pollution is a helpful illustration of early ideas about the use of in-

dicator species as ecological assessment tools.

The author—John

Cairns, Jr., a future member of the National Academy of Scienc-

es—begins with a broad conception of what an indicator species is:

The idea that certain species can be used to indicate certain types

of environmental conditions is well established. Trout are usually

associated with cold water, game birds are usually associated with

a particular kind of habitat, gardeners know that plants have cer-

tain preferences regarding soil, amount of sunlight, temperature,

34. John Cairns, Jr., Indicator Species vs. the Concept of Community Structure as an Index of

Pollution, 10

WATER RESOURCES BULL. 338 (1974).

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

162 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

and the like . . . their presence indicates something about the na-

ture of the environment in which they are found.

By the next paragraph, though, Cairns has narrowed the indicator

concept to one sometimes known as a subtype of indicators called

“sentinel” species: “[t]he presence of a species furnishes assurance

that certain minimal conditions have been met.”

The classic sen-

tinel species is the canary in the coal mine.

So long as the canary

is alive, it is clear that sufficient oxygen is present for the bird’s re-

quirements and that, conversely, toxic levels of methane, carbon

monoxide, or other gases are not present.

Cairns then goes on to warn of the wide range of interpreta-

tions for an absence of such species, and ultimately reaches a con-

clusion that community-level assessments (rather than species-level

analyses) might be preferable.

Cairns’s usage empha-

sizes the presence of necessary environmental conditions, rather

than the absence of toxic conditions, but the point remains the

same.

35. Id. at 338.

This single example of ecological

scholarship in the mid-1970s thus struggles with many of the

themes that would be seen repeatedly in the ensuing decades: the

tension between simpler and more complex environmental indi-

ces, the tradeoffs between the desire for holistic measures and the

36. Id. at 346. The article goes on to state:

There are very few data supporting the indicator species assessment of pollution,

and the results are difficult if not impossible to quantify . . . and require much

information about the responses of organisms to various types of pollutional

stress that is not now in the literature. That is, most of the species likely to be

found in North America and other areas outside of Europe and even many of

the areas in Europe are not adequately characterized in terms of their response

to various pollutants. About the only alternative for assessing the biological con-

sequences of pollution is the use of information involving entire communities in

the receiving system. Bioassays involving one or more individual species are ex-

tremely useful but do not yet furnish sufficient evidence to predict what will

happen to a complex community with multiple interlocking cause-effect path-

ways exposed to the same waste discharges. Id.

37. See, e.g., supra note 27; see also William H. van der Schalie et al., Animals as Senti-

nels of Human Health Hazards of Environmental Chemicals, 107

ENVT’L HEALTH PERSP. 309,

309 (1999).

38. Importantly, canaries are more sensitive to these gases than are humans, such

that the bird’s continued vitality suggests a safe working environment for human miners,

leaving some margin for error.

39. Cairns, supra note 34, at 346.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 163

costs of obtaining the information necessary to make those

measures robust, ambiguity about the definition and uses of spe-

cies-as-indicators altogether, a desire for higher quality quantitative

data, and a call for more data overall.

The idea of using indicator species was seductive, promising a

means of ecosystem-level assessment with simple and field-friendly

tools, but in the 1970s the idea was still coalescing. Nevertheless, it

was about to be implemented in the NFMA regulations as a tool of

environmental management on a grand scale.

IV.

NFMA AND “MANAGEMENT INDICATOR SPECIES” (MIS)

A. The Statutory Planning Scheme

Congress passed the National Forest Management Act (NFMA)

in 1976 to address over-exploitation concerns and to resolve in-

tense conflict about the “correct” balance of industrial use and

conservation in the national forests.

NFMA requires the U.S. For-

est Service (USFS) to develop Land and Resource Management

Plans (LRMPs) for each management unit.

The USFS must also

issue a layer of national regulations to flesh out the statutory re-

quirements and provide further guidance to resource managers as

they develop individual LRMPs.

With respect to biodiversity management, the Act specified that

USFS regulations for land use management plans must “provide

for diversity of plant and animal communities” among other mul-

tiple-use objectives.

40. 16 U.S.C. §§ 1600-1687 (2012); see also Oliver Houck, The Water, the Trees, and the

Land: Three Nearly Forgotten Cases That Changed the American Landscape, 70

TUL. L. REV. 2279,

2291-309 (1996) (discussing the litigation and politics resulting in the Act).

More specifically, NFMA required the Forest

Service to specify “guidelines for land management plans . . .

which insure consideration of the economic and environmental

aspects of various systems of renewable resource management . . .

to provide for outdoor recreation (including wilderness), range,

timber, watershed, wildlife, and fish.” Together, these provisions

41. 16 U.S.C. § 1604(a) (2012). In theory, these plans are revised at least every fif-

teen years. Id. § 1604(f)(5).

42. The NFMA regulations are currently codified at 36 C.F.R. § 219 (2012).

43. 16 U.S.C. § 1604(g)(3)(B) (2012). But see The Lands Council v. Powell, 395 F.3d

1019, 1025 n.2 (9th Cir. 2005) (“The Forest Service is obligated to balance competing de-

mands on national forests, including timber harvesting, recreational use, and environmen-

tal preservation . . . . The national forests, unlike national parks, are not wholly dedicated

to recreational and environmental values.”) (citations and internal quotations omitted).

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

164 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

form the core of the NFMA biodiversity protections; land use plans

must make some accommodation for fish and wildlife.

B. NFMA’s Implementing Regulations

The Forest Service was then left to create the rules by which

more than 225 million acres of National Forest

would be man-

aged via land use management plans. Fulfilling a statutory re-

quirement, the Service appointed a Committee of Scientists in

1977,

This Committee saw biological diversity as compelled by

NFMA’s mandate to provide guidelines for management plans that

would “provide for diversity of plant and animal communities.”

to help draft this first set of forestry regulations.

However, the Committee regarded diversity per se as insufficient for

evaluating management effects on the forest, and no practical

strategy for measuring ecological outcomes had emerged as a clear

winner out of the foment of academic ideas.

Management indica-

tor species

(MIS) were the Committee’s solution, multiple forest

species to be selected by individual forest managers, to serve as

field-friendly metrics for a variety of interrelated ecological goals.

44. As of 2011. FS Directive 383, Land Areas of the National Forest System 1

(U.S.D.A. 2012).

45. Congress required the Forest Service to convene a Committee of Scientists

(COS) to assist the Forest Service in developing regulations to implement the new law:

“[T]he Secretary of Agriculture shall appoint a committee of scientists who are not officers

or employees of the Forest Service. The committee shall provide scientific and technical

advice and counsel on proposed guidelines and procedures to assure that an effective in-

terdisciplinary approach is proposed and adopted.” 16 U.S.C. § 1604(h)(1) (2012). The

Carter administration assembled the first COS in 1977, with advice from the National

Academy of Sciences. This group of scientists would have a profound impact on the result-

ing regulations, creating the “Management Indicator Species” nomenclature and estab-

lishing the contours of this new regulatory requirement. For a short history of the COS,

see Steven E. Daniels & Karren Merrill, The Committee of Scientists: A Forgotten Link in Nation-

al Forest Planning History, 36

FOREST & CONSERVATION HIST. 108 (1992). The Committee

consisted of eight scientists (seven at a time, with one substitution), headed by Arthur W.

Cooper of the School of Forest Resources at North Carolina State University.

46. 16 U.S.C. § 1604(g)(3)(B) (2012).

47. As evidence of this foment, see James R. Karr, Biological Monitoring and Environ-

mental Assessment: A Conceptual Framework, 11 E

NVTL. MGMT. 249 (1987) (“Some direct ap-

proaches for biological monitoring have been developed but a lack of consensus among

biologists, fueled by bureaucratic inertia, tends to favor established procedures.”).

48. C

OMM. OF SCIENTISTS, FINAL REPORT OF THE COMMITTEE OF SCIENTISTS 112-13

(1979); see also 36 C.F.R. § 219.12(g)(2)(1979).

49. The Committee’s product proved sturdy: despite the Reagan administration’s

encouragement to cut back the forest planning rules, the COS’s Carter-era product re-

mained largely intact when reissued as the 1982 NFMA regulations. Compare 44 Fed. Reg.

53928 et seq. (Sept. 17, 1979) with 47 Fed. Reg. 43026 et seq. (Sept. 30, 1982).

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 165

The original 1979 NFMA land use planning regulations and

the subsequent 1982 revision had largely identical biodiversity pro-

visions. First, land use plans were to ensure that habitat would be

managed to maintain “viable populations of existing native and

desired non-native vertebrate species” (the “vertebrate viability”

requirement).

Second, forest units were to select MIS as tools to

aid in evaluating alternative management actions:

In order to estimate the effects of each alternative on fish and

wildlife populations, certain vertebrate and/or invertebrate spe-

cies present in the area shall be identified and selected as man-

agement indicator species and the reasons for their selection will

be stated. These species shall be selected because their popula-

tion changes are believed to indicate the effects of management

activities.

The regulations went on to enumerate categories of species that

“shall be represented, where appropriate,” including federally

threatened or endangered species, game species, and “plant or an-

imal species selected because their population changes are be-

lieved to indicate the effects of management activities on other

species of selected major biological communities or on water quali-

ty.”

A final layer of guidance for forest managers comes from the

Forest Service Manual and the Forest Service Handbook,

MIS would be used in evaluating planning alternatives, and

their population trends monitored.

50. 36 C.F.R. § 219.19 (1982). The vertebrate viability requirement largely merges

into the MIS requirement, as forest units have tended to select vertebrates as MIS, thus

fulfilling both provisions simultaneously.

field-

51. Id. at (a)(1), (a). Notably, Forest Service proposals to remove these provisions

through “updates” to the 1982 regulations have been the subject of intense litigation for

the last decade. For an overview, see Citizens for Better Forestry v. U.S. Dep’t of Agric., 632

F. Supp. 2d 968, 982 (N.D. Cal. 2009), discussed infra p. 117.

52. 36 C.F.R. § 219.19(a)(1) (1982).

53. For a treatise discussion on the Manual and Handbook, see 1 P

UB. NAT.

RESOURCES L. § 7:16 (2d ed. 2012), explaining that:

[M]anual provisions continue to govern many procedural and some substantive

matters. Forest Service regulations, for example, indicate that procedures for the

conduct of agency activities are issued as directives, which include the Forest Ser-

vice Manual and related Handbooks . . . . The Forest Service Manual and hand-

book are published by the Office of the Chief, supplemented as necessary for

field office use by Regional Foresters and others, while guidance issued through

letters and memoranda must be issued in accordance with signing authorities

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

166 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

level instructions that are more frequently updated than regula-

tions. Since 1982, these guidelines have periodically changed to re-

flect an evolving view of MIS, and have attempted to increase the

scientific rigor with which the MIS provision is carried out.

How-

ever, neither the Handbook nor the Manual has the “independent

force and effect of law,” and consequently do not bind the Forest

Service.

This Article focuses on the regulations themselves, while

recognizing the existence of these lower-level guidelines more re-

sponsive to dynamic science.

In sum, the 1982 regulations bind forest managers, requiring

they select particular species and monitor them as an aspect of

evaluating management alternatives under their Land and Re-

source Management Plans (thus locking managers into how they

should monitor). The first Plans to incorporate MIS were due in

1985, to be revised at least every 15 years.

The 1982 NFMA regulations were ambiguous on their face, fail-

ing to address the what and the why of environmental monitoring

by providing little instruction to forest managers as to how best to

identify a suite of MIS and by implying a variety of regulatory pur-

poses for their use. For example, the fact that MIS “shall be select-

ed to indicate the effects of management activities” suggests MIS

are a generic outcome variable for environmental health, but the

enumerated categories of MIS species suggest multiple purposes,

including ESA compliance and maintaining hunting opportuni-

The regulations do not

tie MIS explicitly to action, such that so long as foresters disclose

why they selected MIS, they need not use the selected species in

any particular way.

delegated through issuances to the Forest Service Directive System. While the

agency makes available for public inspection and copying all unpublished direc-

tives, it is obviously more difficult to procure these materials than more formal

sources of ‘law,’ such as agency regulations. Indeed, the Forest Service does not

even publish the indices of much of these informal documents.

54. See, e.g., U.S.

FOREST SERV., FOREST SERV. HANDBOOK 1909.12 ch. 40 (2006) (up-

dating provisions on science and sustainability).

55. W. Radio Services Co. v. Espy, 79 F.3d 896, 901 (9th Cir. 1996) (noting that nei-

ther the Handbook nor the Manual are subject to notice-and-comment periods consistent

with the Administrative Procedures Act, that neither is routinely published in the Federal

and Handbook do not have the independent force and effect of law”).

56. We also discuss the Manual and Handbook briefly infra Part 6, in evaluating the

MIS provision in light of best practices.

57. 16 U.S.C. § 1604(f)(5)(A) (2012).

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 167

ties. At the same time, the original underlying logic for creating

the MIS was to ensure compliance with NFMA’s mandate to pro-

vide for the “diversity of plant and animal communities,” seeming-

ly a different goal entirely. This profusion of regulatory goals

would enshrine a significant level of agency discretion in fulfilling

the national forests’ biodiversity requirements.

C. The 2012 Planning Rule Revision, and the Fate of Previous Revisions

NFMA has proved highly controversial since 1976. Local, indus-

trial, and conservation interests have fought intense battles over

the degree to which NFMA’s substantive and procedural require-

ments should constrain the discretion of regional and unit-level

forest managers in the face of evolving science, local economic

demands, and competing interests. Within the last decade courts

have repeatedly struck down several Forest Service attempts to pro-

vide managers with additional flexibility on issues such as biodiver-

sity conservation, public process, and environmental review.

The biodiversity regulations have been a particular administra-

tive battleground over the past decades, as different presidential

administrations have fought to ensconce their preferred manage-

ment policies governing land-use planning in the National Forests.

The first complete set of planning regulations under NFMA came

into force in 1982 during the Reagan Administration,

creating a

baseline for both the process and substance of planning docu-

ments. The MIS provision focused on single species as indicators of

management activities’ effect on forest health. Late-Clinton-era

regulations

then sought a broad philosophical change, rooting

USFS decisionmaking in principles of ecosystem-based manage-

ment and sustainability, but were quickly replaced by a new set of

rules early in the George W. Bush administration.

58. See Citizens for Better Forestry v. U.S. Dep’t of Agric. (Citizens II), 481 F. Supp. 2d

1059, 1063 (N.D. Cal. 2007) (summarizing this procedural history); see also Citizens for

Better Forestry v. U.S. Dep’t of Agric. (Citizens III), 632 F. Supp. 2d 968, 982 (N.D. Cal.

2009); Citizens for Better Forestry v. U.S. Dep’t of Agric. (Citizens I), 341 F.3d 961 (9th Cir.

2003); George Hoberg, Science, Politics, and U.S. Forest Service Law: The Battle over the Forest

Service Planning Rule, 44

NAT. RESOURCES J. 1 (2004).

The 2005 Bush

59. National Forest System Land and Resource Management Planning, 47 Fed. Reg.

43037 (Sept. 30, 1982) (codified at 36 C.F.R. pt. 219).

60. National Forest System Land and Resource Management Planning, 65 Fed. Reg.

67568 (Nov. 9, 2000) (codified at 36 C.F.R. pt. 217, 219).

61. National Forest System Land and Resource Management Planning, 67 Fed. Reg.

72770 (Dec. 6, 2002) (codified at 36 C.F.R. pt. 219).

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

168 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

rules,

aimed to streamline forest management by weakening en-

vironmental review and biodiversity management requirements for

LRMPs, but were enjoined for lack of proper environmental as-

sessment.

A subsequent similar set, the 2008 Bush regulations,

met a nearly identical fate.

In Citizens for Better Forestry v. United

States Dept. of Agriculture,

the court vacated the 2008 regulations,

giving USFS the option of continuing under either the 1982 or the

2000 regulations because of the Service’s view that the 2000 ver-

sion was unworkable in practice.

The Forest Service then prom-

ulgated a rule in December 2009 reinstating the 2000 regulations

and leaving open the option for forest managers to use the 1982

rule under existing transitional provisions.

The politicization of the NFMA planning regulations and re-

sulting litigation created years of administrative uncertainty. In the

end, Forest Service planning has tended to rely on the Reagan-era

1982 regulations, which have remained on firm legal footing and

which the 2009 rulemaking explicitly validated.

Perhaps as a re-

sult of the legal battles, however, forest planning has fallen badly

behind schedule; as of 2012, “[o]f the 127 land management plans

for National Forest System lands, 68 are now more than fifteen

years old and are past due for revision.”

62. National Forest System Land and Resource Management Planning, 70 Fed. Reg.

1023 (Jan. 5, 2005) (codified at 36 C.F.R. pt. 219).

Given the constant state

of change in the nation’s forest ecosystems—for example, 5 out of

the twenty five largest wildland fires in the past fifteen years have

63. Citizens II, 481 F. Supp. 2d at 1100.

64. National Forest System Land and Resource Management Planning, 73 Fed. Reg.

21468 (Apr. 21, 2008) (codified at 36 C.F.R. pt. 219).

65. Citizens III, 632 F. Supp. 2d 968, 982-83 (N.D. Cal. 2009).

66. Id.

67. Id. (“The effect of invalidating an agency rule is to reinstate the rule previously

in force. It appears that the 2000 Rule was in force before the 2008 Rule was promulgated.

However, the USDA has expressed in the past its view that the 2000 Rule is unworkable in

practice. Accordingly, the agency may choose whether to reinstate the 2000 Rule or, in-

stead, to reinstate the 1982 Rule.”) (citation, quotation marks, and footnote omitted).

68. See National Forest System Land and Resource Management Planning, 74 Fed.

Reg. 67059, 67060 (Dec. 18, 2009) (codified at 36 C.F.R. pt. 219) (“[R]esponsible officials

may continue to revise or amend land management plans under either the 1982 rule pro-

visions or the 2000 rule provisions.”).

69. Id. This is especially the case because forest planning can take many years to

complete. As such, completing an LRMP under an invalid set of regulations can result in a

court invalidating the entire Plan, an extremely expensive prospect.

70. F

OREST SERVICE FACT SHEET, HOW IS THE PREFERRED ALTERNATIVE DIFFERENT

FROM THE

1982 RULE PROCEDURES? at 1 (2012), available at

http://www.fs.usda.gov/Internet/FSE_DOCUMENTS/stelprdb5349609.pdf.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 169

been in National Forests

The 2012 revision to the planning rule, which went into effect

on May 9, 2012,

—this state of chronic delay hampers ef-

fective resource management.

is the result of a years-long process that aimed to

avoid a similar fate.

In response to widespread criticism of the

1982 MIS provision, the new rule includes a requirement to moni-

tor “focal species”

as “one of many ways to gauge progress to-

ward achieving desired conditions in the plan.”

The new regula-

tions then define “focal species” broadly:

A small subset of species whose status permits inference to the in-

tegrity of the larger ecological system to which it belongs and

provides meaningful information regarding the effectiveness of

the plan in maintaining or restoring the ecological conditions to

maintain the diversity of plant and animal communities in the

plan area. Focal species would be commonly selected on the basis

of their functional role in ecosystems.

Because no forest units have yet revised their plans under the 2012

rule, it remains to be seen how this new set of rules will fare by ei-

71. See NATIONAL INTERAGENCY FIRE CENTER, http://www.nifc.gov/fireInfo/

fireInfo_stats_lgFires.html (last visited: Jul. 2, 2012).

72. National Forest System Land and Resource Management Planning, 77 Fed. Reg.

21162, 21162 (Apr. 9, 2012) (to be codified at 36 C.F.R. pt. 219) (listing effective date as

May 9, 2012).

73. For the Forest Service’s account of this process, see N

ATIONAL FOREST

MANAGEMENT ACT (NFMA) / PLANNING, http://www.fs.fed.us/emc/nfma/index.htm (fol-

low “USDA Forest Service Launches Collaborative Process for New Planning Rule”) (last

visited Mar. 23, 2012).

74. 36 C.F.R. § 219.12(a)(5)(iii) (2012); see also U.S.

DEP’T OF AGRIC., FOREST SERV.,

FINAL PROGRAMMATIC ENVIRONMENTAL IMPACT STATEMENT 129 (“The Committee of Sci-

entists (1999) advanced the term ‘focal species’ to allow for a variety of approaches to se-

lecting species whose status and trends provide insights to the integrity of the larger eco-

logical system to which it belongs. Their use of the term focal includes several existing

categories of species used to assess ecological integrity, such as indicator species, keystone

species, ecological engineers, umbrella species, link species, strong interactors, and species

of concern. Focal species would commonly be selected on the basis of their functional role

in ecosystems, for example: species that act as ecosystem engineers by modulating the

availability of resources to other species through changes in biotic or abiotic materials,

thus creating or maintaining habitats; ecological indicators that indicate the action or con-

sequences of key environmental stressors; or strongly interactive species that are dispro-

portionately significant to the survival of other native species and ecosystems, such as

plants that provide critical resources, insect pollinators, and carnivores.”(citations omit-

ted)).

75. U.S.

DEP’T OF AGRIC., FOREST SERV., FINAL PROGRAMMATIC ENVIRONMENTAL

IMPACT STATEMENT, APPENDIX A, PROPOSED PLANNING RULE A26-27 (2012).

76. Id. at A

PPENDIX I, MODIFIED ALTERNATIVE A, § 219.19, I29-30 (2012).

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

170 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

ther environmental or legal standards.

Although the proposed 2012 NFMA regulations phase out the

1982 regulations’ MIS provision, the older rule contains important

lessons about the interaction between science and law that are im-

portant for the future of public resources management. The 1982

rule, not quite state-of-the-art when implemented, quickly became

outdated as the science of environmental monitoring moved on.

The Forest Service remained bound by this aging methodology—

frustrating foresters, environmentalists, and courts alike—even as

the Service expanded its environmental monitoring and assess-

ment program far beyond what the regulations required. This con-

flict between dynamic science and static law

frequently occurs

when statutes or regulations must identify a particular scientific

means to a policy end.

V. POST-1982 SCIENCE OF ENVIRONMENTAL MONITORING

Below, we review the scientific progress

since the 1982 regulations locked the MIS requirement in place,

and then evaluate specifically what made the MIS rule a misfire.

We then suggest how public resources agencies can do better in

the future, and in particular, how federal ocean governance

should benefit from the mistakes of NFMA’s MIS provision. The

fate of the new 2012 NFMA planning rules remains unclear, and

will likely be settled only after litigation. As the Forest Service

launches this most recent revision, it seems a particularly appro-

priate time to evaluate the way in which the 1982 regulations set

up a conflict between dynamic science and static law.

A. Problems with the Indicator Species Concept(s)

The MIS requirement began to appear dated almost immedi-

ately, as applied ecological research continued apace following the

Forest Service’s adoption of the 1982 NFMA regulations. In con-

trast to the ESA and other major environmental statutes, NFMA

has neither a “best available science” mandate

77. We use this phrase with apologies to Holly Doremus, author of The Endangered

Species Act: Static Law Meets Dynamic World, 32

WASH. U. J.L. & POL’Y 175 (2010).

nor any analogous

78. See Emily Hammond Meazell, Super Deference, the Science Obsession, and Judicial Re-

view as Translation of Agency Science, 109

MICH. L. REV. 733 (2011) (discussing the related

issue of courts’ deference to agencies regarding scientific issues, ultimately diminishing

agency incentives to use the best available science).

79. Note that the 2012 NFMA planning regulations do require the Service to “take

into account the best available scientific information throughout the planning process.”

36 C.F.R. § 219.3 (2012).

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 171

provision for updating its indicator scheme as new techniques

came to the fore. The result was a bright-line requirement to use

MIS in every national forest, a requirement that remained in place

even as ecological researchers began a marked shift away from us-

ing individual species as tools of ecosystem assessment.

As academic and agency scientists gained experience with envi-

ronmental monitoring techniques, a trend towards formalization

began to emerge. Between roughly 1970 and 1982, monitoring had

gone from ad-hoc to experimental to routine, and routine de-

manded reliability. It was as if coal miners required not just the ca-

nary, but had to determine the canary’s precise tolerance for me-

thane, or the distribution of such tolerance among all canaries.

With respect to species-as-indicators, this drive for increased rigor

led academic authors to make increasingly fine distinctions among

categories of indicator species based upon specifically what ecolog-

ical outcome variables were being measured (that is, based upon

just what was indicated.)

By the early 1990s, the result was a proliferation of indicator

species sub-types, often with overlapping purposes, among them

“focal species, umbrella species, flagship species, [and] guilds,”

where

[t]he term focal species has been used in many ways in the litera-

ture. . .[and t]here is not a consistent definition of a focal spe-

cies, except when they are selected by various means as the “fo-

cus of a study.” Umbrella species [are] those whose conservation

confers a protective umbrella to numerous co-occurring species .

. . Flagship species are those that have large public appeal, such

as charismatic megafauna like bears and tigers. . . Guilds [are

groups] of species that exploit the same class of environmental

resources in a similar way.

Review articles proliferated, attempting to wrangle the fragmented

80. BIOSIS Previews, an online database of academic publications in the life scienc-

es, shows an exponential increase in publications with the topic “indicator species” begin-

ning around 1976. This is a very robust trend, with a significant regression (R

=0.92). The

database lists 16 publications in 1980 with the topic “indicator species,” compared to 188

in 2010. See http://www.webofknowledge.com/ (search last performed Nov 30, 2012).

Note, however, that this trend in part reflects a tremendous increase in the overall number

of academic articles published since 1980.

81. Niemi & McDonald, supra note 15, at 97 (citing various authors as proponents of

each sub-concept).

82. Id. (citations and quotation marks omitted).

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

172 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

indicator concepts into meaningful categories, each arriving at a

different organizational scheme.

Given such a diversity of views

on the state of the indicator species approach to monitoring, it is

hardly surprising that frustration has ensued. There came to be as

many types of indicators as there were motivations for measuring

environmental state or change, and such specificity came at a cost:

the panoply of closely related—but not identical—ideas led to

suspicion among some ecologists regarding the indicator species

concept altogether.

In short, “[t]he term ‘indicator’ has been de-

fined in many different ways, exacerbating confusion about how to

use them.”

An authoritative 2004 paper covers the waterfront of indicator

critiques, and provides perhaps the most useful organizational

framework for the present purposes. Defining “indicator” as “a

general term to refer to approaches that use one or a few species

to ‘indicate’ condition or a response to stress that may apply to

other species with similar ecological requirements,”

83. See Erik A. Beever, Monitoring Biological Diversity: Strategies, Tools, Limitations, and

Challenges, 87 N.W.

NATURALIST 66 (2006); Vincent Carignan & Marc-André Villard, Select-

ing Indicator Species to Monitor Ecological Integrity: A Review, 78 E

NVTL. MONITORING AND

ASSESSMENT 45 (2002); Robert J. Lambeck, Focal Species: A Multi-Species Umbrella for Nature

Conservation, 11 C

ONSERVATION BIOLOGY 849 (1997); David B. Lindenmayer, Chris R.

Margules & Daniel B. Botkin, Indicators of Biodiversity for Ecologically Sustainable Forest Man-

agement, 14 C

ONSERVATION BIOLOGY 941, 943 (2000); Niemi & McDonald, supra note

the authors

cite three categories of indicator species: (a) those that reflect the

15, at

97-99 (discussing each concept separately); Reed F. Noss, Assessing and Monitoring Forest

Biodiversity: A Suggested Framework and Indicators, 115 F

OREST ECOLOGY AND MGMT. 135

(1999); Daniel Simberloff, Flagships, Umbrellas, and Keystones: Is Single-Species Management

Passé in the Landscape Era?, 83 B

IOLOGICAL CONSERVATION 247 (1998); see also T.M. Caro &

Gillian O’Dogherty, On the Use of Surrogate Species in Conservation Biology, 13

CONSERVATION

BIOLOGY 805 (1999) (discussing similar material, though using the term “surrogate spe-

cies” as a general term for all of the abovementioned applications). Note also that spatial

scale plays an important role in selecting and assessing indicator species, such that the use-

fulness of a particular set of species will vary both for different purposes and over different

spatial scales. See, e.g., Jan C. Weaver, Indicator Species and Scale of Observation, 9

CONSERVATION BIOLOGY 939, 939 (1995) (noting that species richness varies with spatial

scale and focusing attention on the scale dependence of indicator species).

84. For example, Daniel Simberloff was moved to title one well-cited 1998 paper

Flagships, Umbrellas, and Keystones: Is Single-Species Management Passé in the Landscape Era?,

supra note 83, and Sandy J. Andelman & William F. Fagan expressed their frustration with

the title Umbrellas and flagships: Efficient conservation surrogates or expensive mistakes?, 97

PROC.

NAT’L ACAD. SCI. 5954 (2000).

85. Andrew A. Whitman & John M. Hagan, F

INAL REPORT TO THE NATIONAL

COMMISSION ON SCIENCE FOR SUSTAINABLE FORESTRY, A8, BIODIVERSITY INDICATORS FOR

SUSTAINABLE FORESTRY 2 (2003).

86. Niemi & McDonald, supra note 15, at 96.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 173

state of the environment,

(b) those that serve as markers for en-

vironmental change, and (c) those that stand in as surrogates for

other “species, taxa, or communities within an area.”

NFMA’s MIS requirement pre-dates this mushrooming of spe-

cialized indicator designations, seemingly encompassing many or

all of the particular goals later given different names. Against this

diversifying backdrop, the nonspecific MIS appears ever more

vague and dated, reflecting an earlier untested assumption about

the strength of links between MIS and environmental outcome var-

iables.

It is this

third category, the authors argue, that is the cause of ontological

confusion, in part due to a lack of data supporting the use of one

species as a surrogate for others, and that this uncertainty has led

to the proliferation of narrow subtypes of indicators.

Despite the abundance of ideas and naming schemes for indi-

cator species as a basic tool of conservation, biological or ecologi-

cal assessment—and regardless of where MIS fit into this universe

of ideas—as of the late 1990s there was little in the way of data to

substantiate claims that indicator species were effective in the

field.

One major study evaluated a wide variety of different indi-

cator species schemes across different ecosystems, datasets, and

spatial scales.

87. In the marine context, an example of this kind of “state” indicator is benthic in-

vertebrate species composition. See

S. CAL. COASTAL WATER RESEARCH PROJECT, SCCWRP,

2012-13 RESEARCH PLAN, available at http://www.sccwrp.org/Documents/

ResearchPlan.aspx#a._Development_of_Benthic_Macrofauna_as_Indicators_for_Sediment

_Quality_Assessment. Conservation International’s Ocean Health index encompasses both

status and trends monitoring. C

ONSERVATION INTERNATIONAL, Ocean Health Index, available

at http:// www.conservation.org/global/marine/initiatives/ocean_health_index/pages/

ocean_health_index.aspx.

This broad sampling was intended to be a rigorous

test of the species-as-indicator concept, and the authors found that

none of the schemes performed significantly better than a random

88. Niemi & McDonald, supra note 15, at 96-97 (citing JH Lawton & KJ Gaston, Indi-

cator Species, in Encyclopedia of Biodiversity 437 (S. Levin ed., 3d ed, 2001)). Note that

“taxon” (plural = “taxa”) is a general term referring to species or coherent groups of spe-

cies that share an evolutionary history.

89. Conversely, MIS may be seen as a sui generis indicator type. Lindenmayer, supra

note 83, at 943, for example, refers to MIS as a distinct category of indicator.

90. Perhaps the plural “tools’” would be more appropriate here, but as the focus

here is on the rise, diversification, and assessment of indicator species generally, we will

treat the complex of techniques as a singular entity for simplicity.

91. Andelman & Fagan, supra note 84, at 5955. The authors evaluated up to 14

schemes for selecting indicator species, in three ecosystems. Each of these three made use

of a different dataset and occurred at a different spatial scale.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

174 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

selection of species from the relevant database.

Another analysis

demonstrated that areas of high diversity for one taxonomic group

rarely overlap with those for other taxonomic groups, and that rare

species (often of conservation interest) often do not occur within

the most species-rich geographic areas.

B. Toward Holistic Measures of Ecosystem Structure and Function

Taken together, these da-

ta undermine the very concept that one or several species can

stand as a measure of many others, or of a larger idea of biodiversi-

ty.

The fragmentation of indicator species concepts and the lack

of data supporting their general use contributed to a shift away

from indicator species and toward more synthetic methods of envi-

ronmental monitoring. However, the move away from focusing on

individual species (as individual components of a larger ecosys-

tem) and towards more holistic measures of ecosystem structure

and function has also reflected a larger conceptual trend in envi-

ronmental management: from single-species management to eco-

system-based management.

In this new light, monitoring individ-

ual species seemed to miss the larger point.

The shift away from

individual-species-based management was apparent even within

the Forest Service; the Forest Health Monitoring Program (begun

in 1990) used forest crown cover, chemistry, morphology, and spe-

cies diversity as metrics of forest health.

92. Id. at 5954. See id. at 5955-56 for a description of the different surrogate (that is,

indicator) species schemes the authors evaluated, and for the evaluation criteria for each.

93. See J.R. Prendergast et al., Rare Species, the Coincidence of Diversity Hotspots, and Con-

servation Strategies, 365

NATURE 335, 335 (1993).

94. “Ecosystem-based management is an integrated approach to management that

considers the entire ecosystem, including humans. The goal of ecosystem-based manage-

ment is to maintain an ecosystem in a healthy, productive and resilient condition so that it

can provide the services humans want and need. Ecosystem-based management differs

from current approaches that usually focus on a single species, sector, activity or concern;

it considers the cumulative impacts of different sectors.” K

AREN L. MCLEOD ET AL.,

CIENTIFIC CONSENSUS STATEMENT ON MARINE ECOSYSTEM-BASED MANAGEMENT,

COMMUNICATION PARTNERSHIP FOR SCIENCE AND THE SEA 1 (2005), available at

http://www.compassonline.org/science/EBM_CMSP/EBMconsensus. This influential

statement was signed by hundreds of scientists and policy experts with relevant expertise,

among them the now-Administrator of the National Oceanographic and Atmospheric

Administration, Jane Lubchenco.

95. Measuring the trees and not the forest, as it were.

96. See Samuel A. Alexander & Craig J. Palmer, Forest Health Monitoring in the United

States: First Four Years, 55

ENVTL. MONITORING AND ASSESSMENT 267 (1997); Dayle D. Ben-

nett & Borys M. Tkacz, Forest Health Monitoring in the United States: A Program Overview, 71

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 175

In 2000, the National Research Council issued a report entitled

“Ecological Indicators for the Nation,” in an effort to synthesize

the thinking about the use of indicators in environmental monitor-

ing, and to recommend particular indicators of broad applicabil-

ity.

Indicator species were not among the Council’s recommen-

dations. Rather, the report centered on a small handful of

fundamental ecosystem structures and processes, reflecting the

improved status of these metrics in environmental monitoring. In-

dicators of ecosystem structure included land cover and related

variables, as well as species diversity—an index of the biological

components of an ecosystem that avoids tracking individual spe-

cies.

The report recommended parameters such as carbon stor-

age, net primary productivity, and nutrient flows as indicators of

ecosystem function,

and echoed many other efforts to develop

similar metrics.

100

Further evidence of the trend toward holism comes from

emerging efforts to measure and manage coupled social-ecological

systems, reflecting a formalization of the basic observation that the

sphere of human activities is not somehow distinct from a separate

sphere of the “environment.”

It became possible to measure more important

variables directly, rather than relying heavily on indicators of those

same underlying variables. These and similar measures represent

the present state of the art in environmental monitoring.

101

AUSTL. FORESTRY 223 (2008).

Such integrative monitoring at-

tempts to place environmental data into an appropriate social con-

97. N

ATIONAL RESEARCH COUNCIL, ECOLOGICAL INDICATORS FOR THE NATION

(2000). (“[T]he Indicators Committee decided that its main task was to identify and char-

acterize general ecological indicators capable of informing the public and decision makers

about the overall state of the nation’s ecosystems and how those ecosystems may be chang-

ing.”).

98. Id. at 7.

99. Id.

100. For example, the Heinz Center’s 2008 State of the Nation’s Ecosystems Report

included a large number of metrics meant to express ecosystem health in a pluralistic way

(including indicators of nutrient cycling, biological productivity, and other metrics). T

HEINZ CENTER, HIGHLIGHTS: THE STATE OF THE NATION’S ECOSYSTEMS 6 (William C. Clark

et al. eds., 2008). The focus on fundamental biological and chemical variables underscores

the conceptual shift away from monitoring individual ecosystem components (such as par-

ticular species) and towards more a more holistic, ecosystem-level view of environmental

management. Improved technology probably also played a role in promoting these fun-

damental indicators, as more widely available remote sensing, GIS, and modeling tools

made such measurements more feasible.

101. Note that the lack of separation between these spheres is made clear by the very

existence of environmental statutes, all of which mediate the interactions between humans

and the natural resources on which we depend.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

176 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

text, and therefore requires a conceptual model that more fully in-

corporates human dimensions.

102

For example, some conservation

projects have begun to measure social outcome variables rather

than solely ecological ones.

103

Although certain social factors are

likely to be difficult to quantify, the future of environmental moni-

toring will likely entail improved means of incorporating social

and ecosystem variables—a far broader view of environmental

health than the single-species monitoring of the 1970s captured.

104

C. Best Practices of Environmental Indicators

The move toward ecosystem-based monitoring does not allevi-

ate the need for well-designed metrics to reflect ecosystem state.

On the contrary, because more holistic measures tend to be com-

posite statistics, ensuring transparency in the values that underlie

these statistics is more salient than ever, as the metrics themselves

are more removed from everyday experience. The trade-offs be-

tween generality and specificity (inherent in any indicator or in-

dex, as discussed supra in Part 2) remain, and striking a reasonable

balance among a metric’s assets and liabilities requires significant

up-front time investment to determine the purposes and practicali-

ties of a particular monitoring regime. A summary of best practices

102. One such conceptual framework that integrates social and ecological variables is

the five-part Driver-Pressure-State-Impact-Response (DPSIR) model developed by the Eu-

ropean Environment Agency. In this case, the two-part ecological system (P

ressure and

tate) occurs within a social context (social and demographic Drivers leading to environ-

mental P

ressure, the magnitude of social Impact resulting from a change in ecosystem

tate, and the social/political Response to that Impact). (This is a more elaborate version

of the earlier Pressure-State-Response model). These kinds of models form a framework

for developing social indicators of environmental pressures, as well as for identifying the

least socially-disruptive paths to changes in environmental management. See R.E. Bowen &

C. Riley, Socio-Economic Indicators and Integrated Coastal Management, 46

OCEAN & COASTAL

MGMT 299 (2003); Hanne Svarstad et al., Discursive Biases of the Environmental Research

Framework DPSIR, 25

LAND USE POL’Y 116, 116 (2008) (arguing that the framework is bi-

ased by the “discursive positions the applicant brings to it”).

103. See, e.g., Helen E. Fox, Reexamining the Science Of Marine Protected Areas: Linking

Knowledge to Action, 5 C

ONSERVATION LETTERS 1 (2011); Patrick Christie, Marine Protected

Areas as Biological Successes and Social Failures in Southeast Asia, in A

QUATIC PROTECTED

AREAS AS FISHERIES MANAGEMENT TOOLS: DESIGN, USE, AND EVALUATION OF THESE FULLY

PROTECTED AREAS 155 (J. B. Shipley, ed., 2004); Patrick B. Christie et al., Toward Developing

a Complete Understanding: A Social Science Research Agenda for Marine Protected Areas, 28

FISHERIES 22 (2003).

104. Note, too, that the social context of environmental variables blurs the line be-

tween what we want to know and why we might want to know it: for example, the incen-

tives leading to increased population density are as relevant to understanding environ-

mental stressors as they are to finding ways of ameliorating those stressors.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 177

for designing and selecting environmental indicators is therefore

helpful.

Adequately summarizing thirty years of academic literature on

environmental indicators and monitoring is neither within the

scope of this article nor particularly desirable for the present pur-

poses. However, a broad consensus has emerged out of this body

of work as to the necessary or desirable properties of environmen-

tal indicators generally, whether such indicators are narrow and

single-species-focused or broad and holistic. We can then use these

properties to evaluate the NFMA MIS provision in light of the best

practices of the present. Environmental monitoring regimes using

indicators should have:

105

1. Clear purposes for which the indicator is being used

2. Explicit criteria used to select the indicator

3. Known, robust, and reliable relationship between indicator

and the indicated environmental or biological variables that in-

forms purposes

4. Appropriate spatial scale of analysis, given purposes

5. Clear baseline or reference condition against which to

measure change or state

6. Appropriate statistical power, precision, and accuracy of the

indicator set, given purposes

7. Logistical, financial, and social feasibility

8. Explicit monitoring standards

9. Explicitly-evaluated sources of error, including sampling er-

ror and intra-annual, inter-annual, and spatial variability in the in-

dicator

10. A clear plan for information management over lifetime of

data collection

This is not a trivial list: each of these ten requirements requires

substantial analysis and engagement across sectors, and most pre-

105. These suggestions represent a synthesis of the following reviews and reports:

U.S.

ENVTL. PROT. AGENCY, EPA/620/R-99/005, EVALUATION GUIDELINES FOR

ECOLOGICAL INDICATORS 1-1 to 1-5 (2000); NATIONAL RESEARCH COUNCIL REPORT, supra

note 97; N

AT’L ACAD. OF ENGINEERING, MEASURES OF ENVIRONMENTAL PERFORMANCE AND

ECOSYSTEM CONDITION (1999); Carignan & Villard, supra note 83; Caro & O’Dougherty,

supra note 83, at 805; Lindenmayer, supra note 83; David Niemeijer & Rudolf S. de Groot,

A Conceptual Framework For Selecting Environmental Indicator Sets, 8 E

COLOGICAL INDICATORS

14 (2008); Niemi & McDonald, supra note 15; Simberloff, supra note 83.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

178 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

suppose the existence of datasets relevant to the management

question at hand.

Note that the use of individual species-as-environmental-

indicators is not necessarily inconsistent with these best practices,

so long as that use is supported by sufficient background analysis

and information.

106

For example, one relative bright spot for indi-

cator species has been the sustained use of benthic

macroinvertebrates—that is, easily-visible species without back-

bones that live on the bottoms of rivers and lakes, which are gen-

erally insect larvae from various taxonomic groups—to determine

and monitor the environmental health of freshwater environ-

ments.

107

The success of the biomonitoring technique in this con-

text is in large part due to the sheer bulk of relevant data. A review

of selection criteria for indicators, published in 2000, found 84%

of invertebrates used as indicators had documented tolerance lev-

els to stressors, necessary information for a valid use of indicators.

This was in stark contrast to a mere 8% of vertebrates with the

same available data.

108

As a result, many benthic

macroinvertebrates have a substantial basis supporting their use of

indicators, unlike most vertebrates.

109

The common use of explicit

reference conditions

110

—nearby rivers against which to measure

the condition of the focal river—also makes freshwater indicators

more useful and rigorous in practice than most MIS in the Nation-

al Forests.

111

106. Note especially the use of benthic macroinvertebrates as indicator species that

reflect water quality in freshwater streams.

In fact, monitoring programs using benthic

107. For a discussion of this suite of indicators, see FRESHWATER BIOMONITORING

AND

BENTHIC MACROINVERTEBRATES (D.M. Rosenberg & V.H. Resh eds., 1993).

108. Jodi Hilty & Adina Merenlender, Faunal Indicator Taxa Selection For Monitoring

Ecosystem Health, 92 B

IOLOGICAL CONSERVATION 185, 190 (2000). Note, however, the au-

thors found only 1% (vertebrates) and 3% (invertebrates) of indicator taxa were tied to

data “correlating changes in the indicator status with changes to the ecosystem.”

109. See, e.g., Thomas F. Cuffney et al., Responses of Physical, Chemical, and Biological

Indicators of Water Quality to a Gradient of Agricultural Land Use in the Yakima River Basin,

Washington, 64 E

NVTL. MONITORING & ASSESSMENT 259, 267 (2000) (providing correlation

data for agricultural intensity and indicator response).

110. T.B. Reynoldson et al., The Reference Condition: A Comparison of Multimetric and

Multivariate Approaches to Assess Water-Quality Impairment Using Benthic Macroinvertebrates, 16

N. AM. BENTHOLOGICAL SOC’Y 833, 833 (1997). Note that the reference condition pro-

vides a critical directionality to indicators that is otherwise lacking from environmental

monitoring: given a reference condition, we know what we’re aiming at. Without a refer-

ence condition, measures of ecosystem states are descriptive, rather than normative.

111. Note that, although most MIS in National Forests are vertebrates, some Forests

list benthic macroinvertebrates as MIS—see the Kaibab National Forest discussion infra

page 142, for an example.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 179

macroinvertebrates often meet nearly all of the Best Practices for

indicators, supra.

112

The NFMA regulations missed all of this, from the wholesale

movement of applied ecology toward ecosystem-based manage-

ment to the development of best practices. The regulations were

stuck in 1982, with species-based monitoring and ambiguous moni-

toring goals that spawned costly litigation, did little to illustrate the

state of the National Forests, and included common uses of indica-

tors that are “not supported by current science.”

113

VI. EVALUATING THE MANAGEMENT INDICATOR SPECIES (MIS)

PROVISION IN LIGHT OF SUBSEQUENT SCIENCE

Having reviewed some of the science of environmental moni-

toring and set of best practices for developing environmental indi-

cators, we can evaluate NFMA’s MIS provision in light of these sub-

sequent developments in order to distill lessons for future public

resources management efforts. Despite having evolved subsequent

to the 1982 regulations, present-day scientific standards are the

relevant bases for measuring the MIS provisions’ effectiveness be-

cause most National Forest units continue to use these indicator

species under current Land and Resource Management Plans, and

because future resources management regimes—such as the new-

ly-enacted 2012 NFMA regulations, or the analogous effort in the

coastal oceans that we discuss further below—must meet the best

available scientific standards.

114

112. See Hilty & Merenlender, supra note

108, at 20(comparing invertebrates to ver-

tebrates); see also A

NDREW A. WHITMAN & JOHN M. HAGAN, FINAL REPORT TO THE

NATIONAL COMMISSION ON SCIENCE FOR SUSTAINABLE FORESTRY, A8: BIODIVERSITY

INDICATORS FOR SUSTAINABLE FORESTRY 2 (2003) (listing many more desirable criteria,

many of which benthic macroinvertebrates meet).

113. W

AYNE OWEN, BEST PRACTICES FOR SELECTING AND USING MANAGEMENT

INDICATOR SPECIES, USFS TECHNICAL GUIDANCE MEMO 4 (2010) (describing the use of

indicator species as proxies for other species, and citing 36 C.F.R. § 219.19(a)(1) [1982] as

motivating this use of MIS). Additionally, an internal Forest Service review of the MIS

concept concluded that species may be used as indicators of environmental quality (such

as water quality), but are not reliable measures of the effects of management decisions on

other (non-MIS) species. C

HRISTINA VOJTA, A REVIEW OF THE MANAGEMENT INDICATOR

SPECIES CONCEPT AS USED BY THE FOREST SERVICE FOR PLANNING AND MONITORING (Aug.

2009) (graciously provided by Wayne Owen, USFS). Notably, this language made it into

the Federal Register to accompany the final planning rule. 77 Fed. Reg. 21162, 21175

(Apr. 9, 2012) (to be codified at 36 C.F.R. pt. 219).

114. The 2012 NFMA planning regulations do require forest managers use the best

available scientific information to inform the planning process. 36 C.F.R. § 219.3 (2012).

Moreover, a failure to meet modern scientific standards could leave any management re-

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

180 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

A threshold difficulty in evaluating MIS performance in the

National Forests is that the 1982 regulations and subsequent guid-

ance documents are nonspecific in establishing purposes for moni-

toring using MIS. As noted above, the regulations provide a list of

candidates for MIS: federally threatened and endangered species,

“species with special habitat needs that may be influenced signifi-

cantly by planned management programs; species commonly

hunted, fished, or trapped; non-game species of special interest;

and additional plant or animal species selected because their pop-

ulation changes are believed to indicate the effects of management

activities on other species of selected major biological communi-

ties or on water quality.”

115

As a result of this multiplicity of implied goals, it is impossible

to assess the success of the MIS provision as a whole in terms of its

particular aims. With neither a clear what or why, we therefore as-

sess the 1982 MIS requirement using other available criteria: the

best practices for environmental indicators listed above, internal

evaluations by Forest Service employees themselves, and success in

meeting the implicit purposes of MIS in particular forests. The

MIS program fares poorly by any metric.

This mandate implies—but does not

state—at least five different goals: 1) safeguarding endan-

gered/threatened species, 2) species sensitive to changes in par-

ticular habitats, 3) maximizing game species, 4) maximizing un-

specified non-game species of undefined “special interest,” 5) and

species selected to indicate the effects of logging or other man-

agement activities on unmonitored species or on the state of water

quality.

It bears noting that one measure of ineffectiveness is extensive

litigation. Where limited resources are dedicated to defending

agency decisions in court, funding for environmental monitoring

probably suffers. Uncertainty caused by litigation is also unlikely to

result in a robust and continuous dataset of the type most useful in

environmental management. The Forest Service’s MIS provision

has been the subject of repeated and acrimonious litigation over

the course of more than two decades, not least because of the am-

biguities inherent in the regulation.

116

gime vulnerable to challenge as arbitrary and capricious under the Administrative Proce-

dure Act. 5 U.S.C. § 706 (2)(A) (2012).

MIS are bright-line regula-

115. 36 C.F.R. § 219.19 (2012).

116. See discussion of NFMA planning regulations, supra; see also Utah Envtl. Confer-

ence v. Bosworth, 372 F.3d 1219 (10th Cir. 2004) (finding forest service monitoring of MIS

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 181

tory requirements; forest managers must designate a number of

species as MIS, and such clarity limited litigation on this particular

issue. But because the regulations provide broad discretion as to

how forest managers might select, use, and monitor MIS, decades

of litigation resulted in court-mediated biology and forestry. Fu-

ture efforts at biodiversity management would do well to avoid

such an outcome.

Litigation has ultimately focused on the question of how Na-

tional Forests must monitor MIS once it has selected them. For ex-

ample, an important question has been whether a Forest manager

must monitor the populations of MIS directly, or if the manager

may use models that predict the effects of management actions on

the habitats associated with those species. Because an indicator

species is already a proxy—it substitutes for a more comprehensive

accounting of ecosystem structure and function—monitoring the

MIS habitat rather than populations is a technique that the Ninth

Circuit has described as a “proxy on a proxy.”

117

For instance, despite its somewhat pejorative characterization

of the Forest Service’s “proxy-on-proxy” monitoring technique,

the Ninth Circuit has sporadically approved the practice.

The practice of

monitoring habitat rather than populations has been controversial

perhaps because it distills a number of legal, scientific, and policy

questions into a single issue: to what extent must the Forest Service

ensure that MIS adequately incorporate available science and ac-

count for uncertainty? Because of the lack of specificity in the

NFMA diversity regulations, federal courts have reached conflict-

ing decisions on the question.

118

insufficient); Idaho Sporting Cong., Inc. v. Rittenhouse, 305 F.3d 957 (9th Cir. 2002)

(finding Forest Service use of habitat data as proxy for MIS arbitrary and capricious); Sier-

ra Club v. Martin, 168 F.3d 1 (11th Cir. 1999) (finding habitat viability analysis insufficient

as monitoring to satisfy MIS requirement); Neighbors of Cuddy Mountain v. U.S. Forest

Serv., 137 F.3d 1372 (9th Cir. 1998) (holding environmental impact analysis inadequate

due to insufficient treatment of MIS impact); Inland Empire Pub. Lands Council v. U.S.

Forest Serv., 88 F.3d 754 (9th Cir. 1996) (finding Forest Service analysis of project alterna-

tives using evaluation of MIS prior to timber sale adequate).

In Ida-

ho Sporting Congress v. Thomas, the court held that the Forest Service

could use habitat as a proxy for MIS population measurements,

but only “if it demonstrate[d] no appreciable habitat disturbance”

117. Idaho Sporting Cong. , 305 F.3d at 962 ; Lands Council v. Powell, 395 F.3d 1019,

1036-37 (9th Cir. 2004).

118. See Inland Empire Pub. Lands Council, 88 F.3d at 762-63 (finding no actual popu-

lation counts required); Idaho Sporting Cong., 305 F.3d at 971-73 (finding habitat availability

acceptable as a proxy for population data).

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

182 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

from proposed industrial activities.

119

However, in Ecology Center v.

Austin, the court later held that the Forest Service had violated the

NFMA diversity requirements by failing to use on-the-ground ob-

servations verify its assumptions about the effects of timber harvest

on dependent MIS.

120

In a 2008 en banc decision, The Lands Council v. McNair,

121

the

Ninth Circuit then overruled both Idaho Sporting Congress v.

Thomas

122

and Ecology Center v. Austin,

123

holding that the Forest

Service need not empirically verify its estimates of the effects of

projects on MIS. Instead, the Service may model the effects of pro-

posed actions on habitat as a proxy for their effects on MIS, even if

it knows that a proposed project will subject that habitat to appre-

ciable disturbance.

124

The Ninth Circuit’s waffling on this issue exemplifies the judi-

ciary’s inability to impose scientifically-informed checks on the

Forest Service’s use of MIS. The Seventh Circuit appears to agree

with the Ninth in giving the Forest Service broad discretion to

choose and monitor MIS.

125

By contrast, the Eleventh Circuit has

invalidated project approval decisions because the Service failed to

gather population data on indicator species.

126

District Courts in

the Tenth Circuit have similarly required the Forest Service to

conduct species-specific monitoring and data collection to validate

its management models,

127

119. Idaho Sporting Cong. v. Thomas, 137 F.3d 1146, 1154 (9th Cir. 1996); accord

Native Ecosystems Council v. U.S. Forest Serv., 428 F.3d 1233, 1250 (9th Cir. 2005).

and some Circuits have failed to settle

120. Ecology Ctr., Inc. v. Austin, 430 F.3d 1057, 1063-65 (9th Cir. 2005); see also Ore-

gon Natural Res. Council Fund v. Goodman, 505 F.3d 884, 890 (9th Cir. 2007) (“[W]e

find that in this instance the Forest Service’s use of habitat as a proxy for population vio-

lated the NFMA.”).

121. Lands Council v. McNair, 537 F.3d 981 (9th Cir. 2008) (en banc).

122. Id. at 997.

123. Id. at 990.

124. Id. at 998. Note, of course, that a court could still find that any particular use of

habitat-as-proxy to be arbitrary and capricious where the facts on the ground do not sup-

port the use of the technique. (“We will defer to its decision to use habitat as a proxy un-

less the Forest Service makes a ‘clear error of judgment’ that renders its decision arbitrary

and capricious.” Id.).

125. See, e.g., Ind. Forest Alliance, Inc. v. U.S. Forest Serv., 325 F.3d 851, 865 (7th Cir.

2003) (“We find that the Forest Service reasonably relied on habitat and survey infor-

mation about management indicator species to monitor the effects of the forest openings

management project on those species. Because this method was reasonable, the Forest

Service did not act arbitrarily or capriciously in proceeding with the action.”).

126. See, e.g., Sierra Club v. Martin, 168 F.3d 1, 4 -5 (11th Cir. 1999).

127. Utah Envtl. Cong. v. Zieroth, 190 F. Supp. 2d 1265, 1271 (D. Utah 2002).

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 183

on any answer whatsoever to this question. The Fifth Circuit, for

example, held in Sierra Club v. Peterson (1999) that “NFMA requires

on-the-ground inventorying and monitoring and is not simply a

planning statute.”

128

It then vacated its own decision in a rehearing

en banc.

129

A. Evaluation in Light of Best Practices Developed for Indicators of

Environmental State

The best practices for ecological indicators, developed above,

provide a means of assessing the 1982 regulations’ MIS require-

ment, incorporating subsequent work on environmental monitor-

ing. We address these in turn, evaluating the regulations them-

selves. We note at the outset that because the regulations provided

broad authority to individual forest units, those forest units

could—and in some cases, did—implement the regulations in a

way that more closely aligned with best practices than the federal

regulations required. However, our focus remains identifying the

regulatory floor, rather than ceiling: what did the 1982 NFMA reg-

ulations require in their MIS provision? Where applicable, we note

Forest Service guidance documents—the Forest Service Manual

and Handbook—that inform MIS implementation, but because

these documents are not binding on the Service, they do not

change NFMA’s regulatory floor.

1. Clear purposes for which the indicator is being used

The 1982 NFMA regulations fall well short of this most funda-

mental practice for environmental monitoring and indicators. As

discussed above, the enumerated categories of MIS imply a multi-

plicity of sometimes-inconsistent goals, and do not explicitly set

out any particular purpose as paramount. A lack of clarity about

the purposes of the MIS program at the federal level made overall

assessment difficult or impossible; it became unclear how to meas-

ure the success or failure of this environmental monitoring tech-

nique. The Forest Service Manual provided little further guidance

on the purposes of MIS, instructing individual foresters only to

“[u]se management indicators to address issues, concerns and op-

portunities for plants, wildlife, fish, and sensitive species habitats

through all planning levels.”

130

128. Sierra Club v. Peterson, 185 F. 3d 349, 372 (5th Cir. 1999).

This guidance simply restated the

129. Sierra Club v. Peterson, 228 F. 3d 559 (5th Cir. 2000).

130. U.S.D.A.

FOREST SERVICE: FOREST SERVICE MANUAL § 2620.3(1) (effective Jul.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

184 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

instruction to use MIS as a tool, without clarifying why the tool

might be helpful. However, another section of the Manual did in-

struct foresters to identify goals and objectives relating to MIS spe-

cifically.

131

Because the 1982 regulations did not direct foresters to use

MIS for any specific purpose, forest unit managers were free to use

indicators for any number of permissible aims, leading to confu-

sion and frustration among managers.

132

For example, in amend-

ing the list of MIS for three national forests in 2005, Forest Super-

visor Charles Richmond encapsulated much of this frustration:

The concept and application of MIS have come under critical re-

view. . . . Identifying species which are well suited as MIS, and

which meet the intent and letter of the 1982 regulation has prov-

en to be a challenge . . . Adjoining National Forests have gone

through similar selection processes, applying the best science and

reasonable judgment, and have come up with different species

lists. It appears that there, in fact, is no set of species which

meet[s] the theoretical intent of the regulations

133

. . .

Identifying species which truly meet the intent of MIS, to indicate

some change in environment or condition caused by manage-

19, 1991).

131. U.S.D.A.

FOREST SERVICE: FOREST SERVICE MANUAL § 2621.4 (1991) (“The for-

est plan must identify habitat components required by management indicators; determine

goals and objectives for management indicators; specify standards, guidelines, and pre-

scriptions needed to meet management requirements, goals, and objectives for manage-

ment indicators. Prescribe mitigation measures, as appropriate, to ensure that require-

ments, goals, and objectives for each management indicator will be sufficiently met during

plan implementation at the project level.”), available at

http://www.fs.fed.us/im/directives/fsm/2600/2620.txt. We note also that the Forest Ser-

vice Handbook was revised in 2006 to provide greater guidance surrounding planning,

MIS, and monitoring. U.S.D.A.

FOREST SERVICE: FOREST SERVICE HANDBOOK § 1909.12

(2006). However, because most Forests have not undergone planning since this update, we

focus the analysis below on the NFMA regulations themselves and the 1991 Forest Service

Manual guidance that informed the majority of Forests’ Land and Resource Management

Plans now in effect. U.S.

FOREST SERVICE: SCHEDULE OF FOREST SERVICE LAND

MANAGEMENT PLAN REVISIONS & NEW PLANS (2010), available at

http://www.fs.fed.us/emc/nfma/includes/LRMPschedule.pdf.

132. Note that this ambiguity also led to great discretion on the part of forest man-

agers, which was no doubt valuable in shaping the overall thrust of land use planning in

different forests.

133. C

HARLES S. RICHMOND, U.S.D.A. FOREST SERVICE: DECISION NOTICE & FINDING

NO SIGNIFICANT IMPACT, MANAGEMENT INDICATOR SPECIES FOREST PLAN AMENDMENT

TO THE

LAND AND RESOURCE MANAGEMENT PLAN FOR THE GRAND MESA, UNCOMPAHGRE

AND

GUNNISON NATIONAL FORESTS 3 (May 11, 2005) (emphasis added), available at

http://www.fs.usda.gov/Internet/FSE_DOCUMENTS/fsbdev7_003183.pdf.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 185

ment has proven to be far less straight forward than was thought

at the time the regulation was promulgated. Very few species

meet all criteria for being a good MIS. . . . [C]ollectively, the bur-

den of monitoring the large number of species suggested ex-

ceeds the usefulness of the information. It becomes instead a

barrier to efficient planning and decision-making.”

134

Data-driven policy requires that testable hypotheses provide the

foundation for management decisions.

135

2. Explicit criteria used to select the indicator; 3. Known rela-

tionship between indicator and the indicated environmental or bio-

logical variable

If there is no obvious way

of testing the success or failure of a policy, its continued imple-

mentation lies in the realm of faith, not science. Although it may

be politically desirable to institute policy whose effectiveness can-

not be publicly disproved, such a justification for opaque policy-

making is normatively undesirable and contravenes the public-

participation rationale of both the APA and NFMA.

The mingled purposes of MIS necessarily obscured the criteria

by which those MIS might be selected: one cannot derive meaning-

ful criteria for aspects of any environmental monitoring regime

without a clear purpose against which to judge those criteria.

However, one MIS purpose seems clear enough to infer a selec-

tion criterion: those species “selected because their population

changes are believed to indicate the effects of management activi-

ties on other species of selected major biological communities or

on water quality.”

136

134. Id. at 6 (responding to public comment requesting a more extensive list of MIS)

(citation removed and emphasis added). Nevertheless, a core scientific concern does seem

to have reached forestry decision makers: Richmond removed the mule deer from the

three forests’ MIS lists because it is an ecological generalist, insensitive to changes in habi-

tat. “I find that mule deer, while of great economic, and public, interest, is such a habitat

generalist that it would serve as a poor indicator of management effects on the Forest.” Id.

at 4.

Only species with documented associations

between the species’ population on the one hand and the effects

of some management action, un-monitored species, or water quali-

ty on the other hand would fulfill this purpose. Although the

NFMA regulations themselves fail to require such a link, the 1991

MIS guidance in the Forest Service Manual remedies this shortfall,

135. See, e.g., KAI N. LEE, COMPASS AND GYROSCOPE 51 (1993) (discussing “adaptive

management” as a means of deriving policy through testable hypotheses).

136. National Forest System Land and Resource Management Planning, 47 Fed. Reg.

43048 (Sept. 30, 1982) (to be codified at 36 C.F.R. § 219.19(a)(1)).

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

186 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

instructing managers to “[s]elect ecological indicators (species or

groups) only if scientific evidence exists confirming that measura-

ble changes in these species or groups would indicate trends in the

abundance of other species or conditions of biological communi-

ties they are selected to represent.”

137

In practice, however, the adequacy and performance of MIS of-

ten went unassessed. In the words of one study co-authored by

Forest Service employees, “[d]espite this increased use [of indica-

tor species], the conceptual bases, assumptions, and published

guidelines for using ecological indicators have not been adequate-

ly examined.”

Thus at least for one particu-

lar purpose of MIS, Forest Service guidance provides a baseline cri-

terion for identifying particular candidate MIS species.

138

4. Appropriate spatial scale of analysis, given purposes

The 1982 NFMA regulations do not mention spatial scale ex-

plicitly. But because each Forest unit was required to select its own

MIS, the provision implicitly required MIS appropriate to the scale

of the individual Forests. Where multiple Forests collaborated to

137. U.S.D.A.

FOREST SERVICE: FOREST SERVICE MANUAL § 2621.1(3) (1991), available

at http://www.fs.fed.us/im/directives/fsm/2600/2620.txt. Note, however, that interpret-

ing this guidance is complicated by the fact that the Forest Service Manual defined “man-

agement indicators” without mentioning the NFMA regulations’ MIS provision specifical-

ly: “Management Indicators. Plant and animal species, communities, or special habitats

selected for emphasis in planning, and which are monitored during forest plan implemen-

tation in order to assess the effects of management activities on their populations and the

populations of other species with similar habitat needs which they may represent.” Id. §

2620.5(1). Note that this type of indicator falls into the problematic third category of

Niemi and McDonald, supra note 81, at 96-97, but that the problems with the use of such

surrogate species are eased if sufficient data exist to demonstrate the link between indica-

tor and indicated.

138. Peter B. Landres et al., Ecological Uses of Vertebrate Indicator Species: A Critique, 2

ONSERVATION BIOLOGY 316, 317 (1988) (discussing the use of vertebrates, specifically, as

indicators). Both Verner and Thomas were listed as Forest Service affiliates on the paper.

In this well-cited critique, the authors suggest best practices including setting explicit as-

sessment goals and criteria for indicators, explicit analysis of sources of subjectivity, peer

review, and the incorporation of variability. See id. at 316-17. These remain among the most

common suggestions for improving the MIS requirement and the use of indicator species

generally. The lack of data on indicator effectiveness points out a catch-22: indicators are

useful as a management tool insofar as they represent cost-effective and labor-saving

means of accurately assessing an environmental state or change. But in order to test the

indicators’ efficacy and validity, one must have sufficient baseline data describing the indi-

cated environment. Because avoiding such a resource-intensive, comprehensive census is

often the very impetus for using indicator species in environmental management, manag-

ers are likely to be forced to decide between using unverified indicator species or none at

all. Under the 1982 NFMA regulations, foresters had to develop MIS, but without a means

or motive to ground-truth the indicators, the requirement became a mere hurdle. See, e.g.,

ICHMOND, , supra note 133.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 187

select MIS, as in the case of the ten Sierra Nevada National Forests,

their evaluation of candidate MIS occurred over the scale of the

combined Forest units. But because the regulations appear to re-

quire (albeit implicitly) congruent spatial scales of MIS selection

and MIS function, they tend to satisfy this Best Practice.

139

5. Clear baseline or reference condition against which to meas-

ure change or state

Again, both the regulations and the guidance documents avail-

able when most Plans were completed were silent as to establishing

a baseline or reference condition, a necessary condition for effec-

tive ecosystem management.

140

6. Appropriate statistical power, precision, and accuracy of the

indicator set, given purposes

Although the regulations contain no provision that speak to

the statistical adequacy of the selected MIS, the 1991 Forest Service

Manual directs Forests to “[i]nvolve Research Stations, universities,

and other research entities in monitoring to ensure that appropri-

ate sampling methods are employed and statistically valid results

are obtained.”

141

7. Logistical, financial, and social feasibility; 8. Explicit moni-

toring standards

It is not clear why external entities—and not the

Forest Service itself—were necessary to ensure the statistical validi-

ty of sampling results, but nevertheless the guidance document

does incorporate a level of statistical awareness into the MIS moni-

toring process.

Because Forest units selected their own MIS, the 1982 regula-

tions built in feasibility to some degree. Foresters are presumably

more likely to select MIS that are easy to monitor than those that

would be pose more substantial time- or resource-burdens.

142

139. The Forest Service Handbook contains directives from 2006 — well after most

of the current forest Plans were completed — that do explicitly reference spatial scale.

“Select characteristics for evaluation that are appropriately matched to the scale of plan-

ning.” U.S.D.A.

FOREST SERVICE: FOREST SERVICE HANDBOOK § 1909.12, ch. 43.12 (2006).

140. Again, the later Forest Service Handbook amendments do include guidance on

this point, discussing the historical range of variation, which is clearly relevant to establish-

ing a baseline condition. Id. at ch. 43.13 (“The range of variation under historic disturb-

ance regimes is an important context to evaluate current and desired conditions.”). How-

ever, because these guidelines arose only after the most recent Plan revisions for nearly all

forest units, it is hard to know how influential the guidelines are.

141. U.S.D.A.

FOREST SERVICE: FOREST SERVICE MANUAL § 2621.5 (1991), available at

http://www.fs.fed.us/im/directives/fsm/2600/2620.txt.

142. Nevertheless, there are many instances of named MIS that are absent or difficult

to find within a forest unit, as in the cases of the Chequamegon and Kaibab Forests.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

188 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

However, the automatic inclusion of federally threatened or en-

dangered species as MIS introduced an additional difficulty; by

definition, these species are rare, and therefore likely to be hugely

burdensome to monitor. This led the Forest Service to seek new

ways of monitoring MIS, and in turn, resulted in extensive litiga-

tion over monitoring details.

143

More specific monitoring requirements in the regulations

might have decreased the likelihood of litigation on this particular

point, but the text of the regulations is reasonably specific:

“[p]opulation trends of the management indicator species will be

monitored and relationships to habitat changes determined.”

There were good policy reasons to

require Forest management take into account federally listed spe-

cies—among these reasons, ensuring that the Forest Service wasn’t

working at cross-purposes with the Department of Fish & Wildlife,

which had listed the species—but by including these among the

MIS, the regulations created a more onerous monitoring burden

for the Service.

144

Specifying how often monitoring should take place, or by what

methods, would have greatly limited the flexibility of individual

Forest units to monitor in a way that made sense for their particu-

lar species.

145

9. Explicitly-evaluated sources of error; 10. Plan for infor-

mation management

Finding the appropriate balance between regulations

that provide too much and too little of this kind of specificity is a

classic problem of administrative law, and we discuss some means

of doing so in the scientific monitoring context below in Part 7.

The regulations make no mention of evaluating the sources of

error associated with environmental monitoring. This is unsurpris-

ing: as with the monitoring details themselves, perhaps guidance

documents (being lower-level and more fluid than regulations) are

the more logical place for such details. The Forest Service Manual,

U.S.D.A. FOREST SERVICE, LAND AND RESOURCE MANAGEMENT PLAN FOR CHEQUAMEGON

NATIONAL FOREST (1986).

143. See Idaho Sporting Cong., Inc. v. Rittenhouse, 305 F.3d 957 (9th Cir. 2002).

144. National Forest System Land and Resource Management Planning, 47 Fed. Reg.

43048 (Sept. 30, 1982) (to be codified at 36 C.F.R. § 219.19(a)(6)).

145. The Forest Service Manual, too, references monitoring throughout the relevant

sections. For example, it instructs to “monitor management indicators to evaluate compli-

ance of management activities with plan direction, effectiveness of prescribed manage-

ment, and validity of information used in habitat evaluation and planning.” U.S.D.A.

FOREST SERVICE: FOREST SERVICE MANUAL § 2620.3(5) (1991). See also 53 Fed. Reg. 26807,

26812-13 (Jul. 15, 1988) (describing “implementation,” “effectiveness,” and “validating”

monitoring in Forest Service Manual 1922.7).

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 189

however, also neglects the topic.

Information management, too, is mentioned neither in the

regulations nor the guidance. This Best Practice is key for develop-

ing a robust time-series of data, which in turn is important for mak-

ing management decisions on ecologically relevant time scales.

In sum, the 1982 NFMA regulations embodied few of the Best

Practices for environmental monitoring, which is perhaps not sur-

prising given that it was primarily subsequent research—

conducted after 1982, and so after the regulations were in place—

that informed the development of the Best Practices. Although the

guidance documents available to forest managers (most notably

the Forest Service Manual and Handbook) were updated periodi-

cally and reflected some substantial improvements over the

Reagan-era rules, these generally did not have the force of law and

moreover, the improvements came too late to influence the land

use planning processes in many forests.

A 2007 revision of MIS for a set of Sierra Nevada forests echoes

these sentiments, seeming to lament the existence of the MIS re-

quirement while awaiting a revision of forest planning rules:

I want to acknowledge the problems with the MIS concept and

the associated difficulties with implementing this concept to meet

the continued requirement to use MIS until forest plans are re-

vised or new NFMA regulations permit otherwise. Until revision

occurs or new planning regulations permit otherwise, each of

these National Forests will be required to use MIS.

146

These are not isolated complaints. Taken together with further

statements below, and especially considering that the Forest Ser-

vice has instituted additional, parallel monitoring schemes over

the years that do not include MIS,

147

146. U.S.D.A. FOREST SERVICE, PACIFIC SOUTHWEST REGION, SIERRA NEVADA FORESTS

MANAGEMENT INDICATOR SPECIES AMENDMENT, RECORD OF DECISION 11 (Dec. 14, 2007).

it seems clear that many forest

147. Such monitoring efforts include at least the following: (1) Forest Health Moni-

toring Initiative, see Alexander & Palmer, supra note 96, at 267; (2) Forest Inventory Analy-

sis, see B.K.

SCHULZ ET AL., U.S.D.A. FOREST SERVICE, U.S.D.A. GEN. TECH. REPORT, PNW-

GTR-781, S

AMPLING AND ESTIMATION PROCEDURES FOR THE VEGETATION DIVERSITY AND

STRUCTURE INDICATOR (2009); U.S.D.A. FOREST SERVICE, FOREST INVENTORY AND

ANALYSIS NATIONAL CORE FIELD GUIDE (2006); SUSAN WILL-WOLF, U.S.D.A. FOREST

SERVICE, GEN. TECH. REPORT, PNW-GTR-818, ANALYZING LICHEN INDICATOR DATA IN THE

FOREST INVENTORY AND ANALYSIS PROGRAM (2010); and (3) Multiple Species Inventory

and Monitoring (an effort within the larger Forest Health Monitoring Initiative that in-

cludes repeated baseline monitoring of set plots), see P.N.

MANLEY ET AL., U.S.D.A. FOREST

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

190 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

managers came to view MIS as a “barrier to efficient planning”

148

B. Evaluation in Light of Implicit Purposes

rather than a useful management tool.

In practice, forest units have often used MIS as surrogates for

larger species assemblages—as a metric for all species using a par-

ticular habitat, for example, or for an index of species diversity in a

particular area.

149

These related purposes fall into Niemi &

McDonald’s

150

problematic third category, indicating neither envi-

ronmental state nor change to that state, but rather standing in for

some number of other species.

151

As surrogates for larger assemblages of species, MIS have not

performed well. One study directly on point examined the avian

MIS for the Chequamegon National Forest in detail.

152

There, the

USFS had designated twelve bird species as MIS and thirteen oth-

ers as “sensitive” species

153

If the designated MIS were to function effectively as indicators

of the presence of other species, the MIS would have to be associ-

important for monitoring. The authors

were able to census eighteen species out of this combined group,

assessing whether the selected MIS were effective stand-ins for

presence and health of other species in the forest.

SERVICE, U.S.D.A. GEN. TECH. REP. WO-73, MULTIPLE SPECIES INVENTORY AND

MONITORING GUIDE (2006), available at http://www.treesearch.fs.fed.us/pubs/24985.

148. R

ICHMOND, supra note 133, at 6.

149. See case studies, infra pp. 139-45. Note also that there is some evidence of Forest

Service intent to use MIS as indicators of ecosystem state or change to that state. See

NITED STATES GENERAL ACCOUNTING OFFICE, GAO/RCED-91-123, WILDLIFE

MANAGEMENT: PROBLEMS BEING EXPERIENCED WITH CURRENT MONITORING APPROACH 2

(1991) (“Population changes in the indicator species being monitored are interpreted as a

signal of changes in the health of the ecosystem.”). But we have not found examples of

this use in practice.

150. Niemi & McDonald, supra note 81, at 96-97.

151. In large part, this is due to NFMA’s statutory mandate to provide for a “diversity

of plant and animal communities,” 16 U.S.C. § 1604(g)(3)(B), among the multiple uses of

National Forest land. The MIS provision was the way in which the regulations made this

requirement concrete, and so it followed that particular indicator species would stand in

for some larger suite (“diversity”) of species.

152. Gerald J. Niemi et al., A Critical Analysis on the Use of Indicator Species in Manage-

ment, 61 J. W

ILDLIFE MGMT. 1240, 1240 (1997) (“Here we focus on the overall question on

[sic] whether the MIS approach can be used to ensure the perpetuation and well-being for

many other species in a forest setting.”). The authors note that the “scientific basis for se-

lecting MIS in the plan is obscure.” Id.

153. The authors report a total of sixteen “sensitive” species, but three—the com-

mon loon, sharp-tailed grouse, and olive-sided woodpecker—appear on both the MIS and

sensitive species lists.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 191

ated with the presence of those others in a statistically significant

way. The 1986 Management Plan for the forest reiterated this aim,

stating that the MIS would “each represent all other game and

non-game species associated with similar habitat needs.”

154

The

study’s authors found a different reality, however: only seven of the

eighteen study species were even sufficiently abundant for analy-

sis.

155

Four of the eighteen appear to have been absent from the

forest altogether.

156

Of the seven analyzed bird species, only one

157

was consistently and significantly associated with a particular habi-

tat type. Moreover, the authors found only inconsistent associa-

tions between the monitored species and other, non-monitored

forest species,

158

The data therefore showed that most of the USFS’s designated

species were rare or absent, and of those abundant enough for

analysis, most were associated neither with particular habitat types

nor with particular species. As a group, the MIS indicated nothing.

And the Chequamegon National Forest MIS are not alone, espe-

cially insofar as they represent the shortfalls of the first generation

of MIS sets designated under the 1982 NFMA regulations.

the very parts of the forest assemblage the MIS

were supposed to reflect.

159

1991 Government Accountability Report reported a similar lack of

information in an anonymous forest:

At a national forest, the wildlife biologist said that the forest does

not have habitat-monitoring data for the eight management indi-

cator species specified in the forest plan. . . . The forest wildlife

biologist said that predicting species population levels from habi-

tat availability is risky because not all species/habitat relation-

ships have been defined. For example, in the case of the sage

154. U.S.D.A. FOREST SERVICE, LAND AND RESOURCE MANAGEMENT PLAN FOR

CHEQUAMEGON NATIONAL FOREST (1986) (cited in Niemi et al., supra note 15, at 1240-41).

Note that this statement is an ecological non-sequitur: on the one hand, each of the vari-

ous MIS cannot possibly represent all other species with “similar habitat needs” as every

species—indicator and indicated—occupies a unique niche described by the intersection

of its biotic and abiotic needs. On the other hand, the statement could be read as trivially

true: the MIS represent all of the species associated with them.

155. Niemi et al., supra note 15, at 1243.

156. Id.

157. The yellow-bellied flycatcher (Empidonax flaviventris). Id. at 1244. A second spe-

cies, the pine warbler (Dendroica pinus), was associated with a particular habitat for a subset

of the analyzed data. Id. at 1244.

158. Id. at 1245.

159. The Chequamegon National Forest Land and Resource Management Plan was

completed in 1986.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

192 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

grouse, more needs to be known about its use of habitat and

about the impacts of fire, fencing, water developments, and graz-

ing.

160

A second GAO report makes the same point: “even when planned

data collection efforts are completed using this monitoring ap-

proach, the data can have limited usefulness because observed

population changes in the species being monitored often cannot

be related to overall habitat conditions or the effects of Forest Ser-

vice management actions.”

161

Of course, the lack of monitoring in

this particular forest contravenes the NFMA regulations.

162

But the

lack of information about species/habitat relationships suggests

that even had the monitoring been done, its value as a manage-

ment tool would have been negligible because of the lack of

known relationship between indicator species and indicated envi-

ronmental change.

163

160. UNITED STATES GENERAL ACCOUNTING OFFICE, GAO/RCED 91-64, PUBLIC

LAND MANAGEMENT: ATTENTION TO WILDLIFE IS LIMITED 30 (1991) (emphasis added).

Note that a subsequent GAO report from 2004 encourages the development of “indicator

sets” and coordination for their use across agencies, but that this report refers to environ-

mental indicators generally, and not indicator species specifically. U

NITED STATES

GOVERNMENT ACCOUNTABILITY OFFICE, GAO-05-52, ENVIRONMENTAL INDICATORS: BETTER

COORDINATION IS NEEDED TO DEVELOP ENVIRONMENTAL INDICATOR SETS THAT INFORM

DECISIONS (2004). Notably, the term “indicator species” and “management indicator spe-

cies” do not appear in the 2004 document.

161. UNITED STATES GENERAL ACCOUNTING OFFICE, GAO/RCED-91-123, WILDLIFE

MANAGEMENT: PROBLEMS BEING EXPERIENCED WITH CURRENT MONITORING APPROACH 1

(1991). The report continues: “First, relationships between indicator species and the habi-

tat characteristics they are supposed to predict are often not known. Without a clear un-

derstanding of such relationships, an observed population decline in an indicator species

may or may not represent a change in overall habitat conditions or establish whether the

change was caused by Forest Service management actions or other reasons. Second, as

noted by Forest Service managers, changes in population that are detected could be due

to habitat changes beyond management control, or be part of a normal cycle requiring no

management action.” Id. at 3.

162. National Forest System Land and Resource Management Planning, 47 Fed. Reg.

43,048 (Sept. 30, 1982) (to be codified at 36 C.F.R. § 219.19(a)(1)) indicates that MIS

should be “selected because their population changes,” (emphasis added), and are useful

to assess the effects of management activities. Without monitoring, it is impossible to

measure changes to populations. § 219.19(a)(6) requires more explicitly that “population

trends of the management indicator species will be monitored.” Note that these citations

refer to the 1982 regulations, no longer in force as of this writing. Nevertheless these are

the relevant regulations because they were in force at the time of the GAO report, and

remain relevant because the vast majority (if not all) of the forest land use plans to date

have been completed under the 1982 regulations.

163. Note that even if an MIS is carefully selected and monitored, MIS can lead to

perverse consequences as forests manage exclusively in favor of that species. Designating

particular species to signal the health of a forest unit, perhaps inevitably, has often led for-

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 193

The absence of species/habitat relationships remains a prob-

lem even today, with the result that Forests often select indicators

without much knowledge about what (if anything) they might in-

dicate. The 2010 evaluation of Kaibab National Forest’s MIS,

164

for

example, presented data on eighteen species

165

of management

importance. Each species was intended to represent a larger suite

of species using the same habitat. However, the 256-page report

included almost no data supporting the idea that any one of the

MIS in fact shared habitat with other, non-MIS species.

166

esters to concentrate their attention on those few focal species to the exclusion of the rest

of the ecosystem; to lose the forest for the trees, so to speak. Simberloff crystallizes this cri-

tique as follows:

The ab-

sence of such necessary information is striking more than twenty-

five years after MIS were first required in Forest Service planning.

Moreover, many species on the Kaibab’s newly-revised list of MIS

The legal status of the owl as an indicator species under the National Forest

Management Act has led to an undue focus on this particular species to the ex-

clusion of all that it is supposed to indicate. For example, logging industry repre-

sentatives frequently suggest management procedures specifically targeted at

owls, like moving or feeding them, artificially enhancing their prey density, or

providing added shelter, in order to boost their populations so that logging quo-

tas can be raised. The Forest Service proposed moving owls from site to site. Lost

in such suggestions is the recognition that single-species management of an indi-

cator species is a self-contradiction. After all, if the species’ status is artificially im-

proved, it no longer indicates the status of all the species it is supposed to represent. Would

we also add food or shelter for other birds, mammals, amphibians, and insects of

the old-growth forest, and move them around?

Simberloff, supra note 83, at 249 (emphasis added) (citations omitted). Thus, where MIS

become a focus of management rather than a yardstick for that management, much of the

logic behind the use of species as indicators disappears. NFMA and its implementing regu-

lations nearly demand this kind of selectively myopic management by lumping in federally-

listed threatened and endangered species with other MIS. NFMA requires foresters to

track the populations of listed species, presumably to ensure the ongoing existence of

those species in the National Forests. The listed species are therefore being managed for

their own sake, not as representatives of anything larger than themselves, and they are thus

“indicators” only by legal designation, lumped into the MIS term-of-art.

164. Valerie Stein Foster et al., M

ANAGEMENT INDICATOR SPECIES OF THE KAIBAB

NATIONAL FOREST: AN EVALUATION OF POPULATION AND HABITAT TRENDS, VERSION 3.0

(2010), available at http://www.fs.usda.gov/Internet/FSE_DOCUMENTS/

stelprdb5114494.pdf.

165. The document treats aquatic macroinvertebrates as a unit for analysis, though

they are in fact a highly diverse group.

166. One bird species, the juniper titmouse, may forage with chickadee species un-

der some conditions. Foster, supra note 164, at 49. This is the only cited example of a MIS

significantly associated with another (non-MIS) species.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

194 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

have the most undesirable traits for indicators: exceedingly rare

species or those absent from the National Forest entirely,

167

ubiqui-

tous or ecological generalist species that are unlikely to reflect

changes to the Forest habitat,

168

and species that are only rarely

monitored.

169

Finally, the Service’s analysis of the scant MIS data

leaves much to be desired; for example the authors conclude

“[t]he data from the [aquatic environment] studies indicate stable

conditions,” one paragraph after noting that the “[l]ow numbers

of individuals sampled suggest an unstable ecosystem.”

170

The re-

cent Kaibab evaluation document underscores the many challeng-

es that remain before MIS might be effective management tools in

the National Forests.

171

Necessarily, if some species are to stand in for others, the po-

tential MIS must be vetted to ensure they will fulfill their intended

function.

172

Assessing the usefulness of a species of quail as a MIS

to represent a particular ecological guild,

173

167. For example, the Cinnamon Teal, Lincoln’s Sparrow, Mexican Spotted Owl,

Red-Naped Sapsuckers, and Yellow-breasted Chat occur rarely or not at all in the Kaibab

National Forest. These are five of the eighteen species analyzed, or nearly thirty percent of

the biological indicators for the Forest. See id. at 18-92 for individual species evaluations.

for example, one study

found that the quail’s habitat differed significantly from the habi-

168. Mule deer and red squirrels are common in many habitats; wild turkeys are

widespread in North America, but have been observed only once in each of the 2005 and

2006 landbird surveys of the Kaibab National Forest. Id. at 244.

169. Benthic macroinvertebrates, for example, have not been assessed since 1998;

the Arizona bugbane, an ESA candidate species, likewise was last counted in the Clinton

administration. Id. at 20, 92.

170. Id. at 21, 20.

171. Other Forests have the same problems. The 2007 MIS revision for Sierra Nevada

Forests noted that one reason the revision was necessary was that prior MIS were “inade-

quate” in part because “they [were] not strongly linked to habitats or ecosystem compo-

nents that are affected by National Forest management activities or have population

changes that have no known link to the effects of our management activities (for example,

Canada Goose, largemouth bass, Peregrine Falcon, rainbow trout.” U.S.D.A.

FOREST SERV.,

SIERRA NEVADA FORESTS MANAGEMENT INDICATOR SPECIES AMENDMENT FEIS 6 (2007).

172. In terms of the Administrative Procedure Act, 5 U.S.C. §§ 701-06, it would seem

arbitrary and capricious to declare some species to be “indicators” of others in the ab-

sence of data in support of that association. “While the scope of review under the ‘arbi-

trary and capricious’ standard is narrow and a court is not to substitute its judgment for

that of the agency. Nevertheless, the agency must examine the relevant data and articulate a satis-

factory explanation for its action, including a ‘rational connection between the facts found and the

choice made.’” Motor Vehicle Mfrs. Ass’n of the U.S., v. State Farm Mut. Auto. Ins. Co., 463

U.S. 29, 43 (1983) (citation omitted) (emphasis added).

173. The authors use a definition of “guild” that refers to “assemblages of species

that use a particular class of resources in a similar way.” William M. Block, Leonard A.

Brennan & R.J. Gutierrez, Evaluation of Guild-indicator Species for Use in Resource Management,

11 J. E

NVTL. MGMT. 265, 265 (1987).

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 195

tat of other species in its guild in fourteen out of fifteen compari-

sons.

174

Because of these habitat differences, the presence of quail

would not indicate the presence of the others, and it would there-

fore fail as a surrogate for the other species.

175

A recent revision of the list of MIS for ten California National

Forests offers an improved model for such large-scale prospective

analysis, ensuring that the selected species would meet a list of

suitability and feasibility requirements that parallel many of the

best practices discussed supra.

But over the quar-

ter-century history of MIS in practice in National Forests, this sort

of a priori evaluation seems the exception to Forest Service prac-

tice, rather than the rule.

176

There, the USFS compiled data

on a large number of candidate MIS and evaluated them relative

to five explicit selection criteria,

177

174. Id. at 268. Note that such habitat differences are to be expected under standard

ecological theory: species are thought to have unique niches, no two of which overlap en-

tirely. See, e.g., Jared Verner, The Guild Concept Applied to Management of Bird Populations, 8 J.

NVTL. MGMT. 1, 4 (1984) (“We should not be surprised to find few if any groups of spe-

cies in a community with patterns of habitat use so alike that one species could be used as

an indicator of the others in its group.”).

rejecting those candidates that

175. Worse, using such a species to indicate the other members of its guild would be

actively misleading, akin to depending on a car with a broken fuel gauge. The authors, if

anything, understate their case: “[T]here is little assurance that habitat suitability or popu-

lation status of a guild indicator will parallel those of other species in the guild. Moreover,

if the guild includes an uncommon species, the welfare of that species may be jeopardized

by indirectly monitoring its status with a guild indicator.” Block et al., supra note 173, at

268. See also Winston P. Smith, Scott M. Gende & Jeffrey V. Nichols, The Northern Flying

Squirrel as an Indicator Species of Temperate Rain Forest: Test of an Hypothesis, 15 E

COLOGICAL

APPLICATIONS 689 (2005) (finding that the focal species was inappropriate as a surrogate

for old-growth habitat).

176. U.S.D.A.

FOREST SERV., SIERRA NEVADA FORESTS MANAGEMENT INDICATOR

SPECIES AMENDMENT RECORD OF DECISION 1-6 (2007). Note, however, that this set of MIS

may not be useful to evaluate the impacts of any given management decision because of its

massive spatial scale. The Deputy Regional Forester emphasized in the Record of Decision

that “all MIS monitoring is at the planning area level, not at the project level. For this

Amendment, the planning area is the 10 Sierra Nevada National Forests. The regulations

require that ‘population trends of the management indicator species will be monitored

and relationships to habitat changes determined’; this monitoring, as with all actions iden-

tified in 1982: 36 CFR 219.19, are required at the planning area level. There are no MIS

monitoring requirements in the project area or at the project level.” Id. at 11 (citation

omitted).

177. U.S.D.A.

FOREST SERV., SIERRA NEVADA FORESTS MANAGEMENT INDICATOR

SPECIES AMENDMENT FEIS 22-23 (2007) (“I. Suitability Criteria: A. The species is linked to

a habitat or ecosystem component that is affected by Forest Service management activities.

. . . B. The population changes of the species are thought to primarily indicate the effects

of Forest Service land management activities versus indicating the effects of other factors. .

. . II. Feasibility Criteria: A. There is an available, tested methodology (either currently be-

ing implemented or readily available to implement) to monitor the population or habitat

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

196 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

failed to meet the criteria or had inappropriate geographic ranges.

The document is far from perfect. For example, it features a po-

tentially serious drawback in stretching the same set of MIS across

a vast geographic range. It also makes some questionable substan-

tive decisions—such as including the mule deer, an ecological

generalist, as a MIS linked to seasonal shrublands.

178

Nevertheless, the California Forests’ effort is a step in the right

direction towards data-based decisionmaking: the Service set out

clear goals for their monitoring program, guidelines for selecting

MIS, and made available the data used to evaluate the candidate

MIS.

179

By combining the MIS selection across multiple forest

units, the individual National Forests were able to use information

and resources more efficiently, probably resulting in a more rigor-

ous process than would have otherwise been possible. But the

tradeoff is having less location-specific MIS, because the selected

species represent a compromise across all ten forest units. Howev-

er, insofar as these compromise species function as anticipated,

continue to be supported by data, and help the individual forest

units manage biodiversity, the California Forests’ approach to des-

ignating MIS is an improvement on past practice.

180

VII. LESSONS OF MIS FOR FUTURE PUBLIC RESOURCES MANAGEMENT,

AND FOR

COASTAL AND OCEAN PLANNING IN PARTICULAR

The NFMA regulations faced a problem of how to make a

somewhat vague statutory mandate—“provide for the diversity of

plant and animal communities” —operational and practical. MIS

were the tools that the Forest Service chose to accomplish this aim,

in an attempt to balance a variety of overlapping and competing

objectives. Faced with the same problem today, the agency might

choose a different suite of techniques, but the problem itself is no

less daunting now than it was in 1976: in practical terms, just how

should an agency measure and manage the living resources under

of the species. . . . B. The methodology, including data analysis, can be implemented with-

in budget constraints. . . . C. The methodology gives information regarding population or

habitat status and change of the species that is useful to informing management deci-

sions.”)

178. Id. at 121.

179. See generally Sierra Nevada Forests Management Indicator Species Amendment

FEIS, app. B.

180. Note that the FEIS emphasized that the Sierra Nevada National Forests would

re-do this analysis under a new planning rule, and so it remains to be seen what the final

outcome will be under the 2012 regulations.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 197

its care?

The experience of NFMA’s MIS provision helps clarify this

challenge somewhat, identifying three distinct questions surround-

ing biodiversity management: what, why, and how. First, “diversity”

is not a single thing, and there is no best way to simplify the con-

tinuum of hierarchical components that we think of as constituting

biological diversity.

181

This means that natural resources agencies

must think more deeply about what it is they are managing, and

approach the rulemaking process accordingly. Second, “diversity”

(however defined) probably is not a fundamental goal of natural

resources management; more likely, the core goal is to incorporate

some sense of ecosystem “health,” stability, performance, or sus-

tainability into management.

182

Each of these three aspects of biodiversity management holds

lessons for the future of natural resources regulation. Below, we

apply these lessons in the concrete context of the next federal ma-

jor federal public resources challenge, the oceans.

That is, defining why one might

want to include some such ecosystem-based measure is a reasona-

ble starting point for any biodiversity-related regulation. Finally,

even given clarity on these first two points, there remains the chal-

lenge of how to encapsulate appropriate ecosystem monitoring

techniques into a regulatory framework—that is, how to balance

static law with dynamic science, uniform standards vs. require-

ments flexible enough to be applicable in individual management

units, and how to make all of the above financially feasible.

A. Marine Spatial Planning as Closely Analogous to National Forest

Planning

On July 19, 2010, President Obama issued Executive Order No.

13547,

183

establishing a National Ocean Policy and making com-

prehensive coastal and marine spatial planning (CMSP)

184

181. Dale D. Goble, What Are Slugs Good For? Ecosystem Services and the Conservation of

Biodiversity, 22

J. LAND USE 411, 414-17 (2007); John M. Hagan & Andrew A. Whitman, Bio-

diversity Indicators for Sustainable Forestry: Simplifying Complexity, J.

FORESTRY 203 (2006).

the

Federal Government’s primary approach for managing ocean,

182. Although, of course, NFMA’s statutory language does not demand this interpre-

tation, and neither is it necessarily what the Forest Service or other agencies might see as

their institutional mission.

183. Proclamation No. 13,547 75 Fed. Reg. 43,023 (July 22, 2010).

184. Note that the federal acronym is CMSP (Coastal and Marine Spatial Planning).

Where we discuss the use of marine spatial planning approaches generally, we use the

more common acronym “MSP” or else the more general “ocean planning.”

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

198 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

coastal, and Great Lakes waters.

185

The Policy also enshrines a suite

of other important features for natural resources management, in-

cluding science-based decisionmaking,

186

ecosystem-based man-

agement,

187

and a precautionary approach to environmental stew-

ardship.

188

The history of biodiversity management in the national forests

is a chance to derive concrete lessons that give meaning to these

new environmental policy goals. Ocean governance is closely anal-

ogous to forestry and public lands management, in which vast,

sparsely-populated areas contain valuable public resources subject

to overexploitation in the absence of responsible management.

As a result, it offers an important opportunity for

improved stewardship of our national public resources.

189

185. The Obama Administration has defined CMSP as “a comprehensive, adaptive,

integrated, ecosystem-based, and transparent spatial planning process, based on sound

science, for analyzing current and anticipated uses of ocean, coastal, and Great Lakes are-

as. Coastal and marine spatial planning identifies areas most suitable for various types or

classes of activities in order to reduce conflicts among uses, reduce environmental impacts,

facilitate compatible uses, and preserve critical ecosystem services to meet economic, envi-

ronmental, security, and social objectives. In practical terms, coastal and marine spatial

planning provides a public policy process for society to better determine how the ocean,

our coasts, and Great Lakes are sustainably used and protected—now and for future gen-

erations.” Proclamation No. 13,547 75 Fed. Reg. 43,023 (July 22, 2010). The Executive

Order adopted the Final Recommendations of the Interagency Ocean Policy Task Force, which

provides a blueprint for the future of ocean and coastal management in the United States.

WHITE HOUSE COUNCIL ON ENVTL. QUALITY, FINAL RECOMMENDATIONS OF THE

INTERAGENCY OCEAN POLICY TASK FORCE (2010), available at

http://www.whitehouse.gov/files/documents/ OPTF_FinalRecs.pdf. The Intergovern-

mental Oceanographic Commission, which has analyzed MSP approaches in a number of

countries, defines MSP more generically as “a public process of analyzing and allocating

the spatial and temporal distribution of human activities in marine areas to achieve eco-

logical, economic, and social objectives that are usually specified through a political pro-

cess.” UNESCO,

MANUAL AND GUIDES NO. 53, ICAM DOSSIER NO. 6, INTERGOVERNMENTAL

OCEANOGRAPHIC COMM’N, MSP: A STEP-BY-STEP APPROACH TOWARD ECOSYSTEM-BASED

MANAGEMENT 18 (2009), available at http://www.unesco-ioc-marinesp.be/msp_guide.

186. Proclamation No. 13,547 75 Fed. Reg. 43,023 (July 22, 2010).

187. Id.

188. N

AT’L OCEAN COUNCIL, DRAFT NATIONAL OCEAN POLICY IMPLEMENTATION

PLAN 97 (2012) (“One of the Policy’s guiding stewardship principles provides that deci-

sion-making will be guided by a precautionary approach as reflected in the Rio Declaration

of 1992.”).

189. Both terrestrial and marine systems generate ecosystem goods and services and

support local economies through their support of tourism and recreation; both sets of

ecosystems also support “extractive” economic activities and sectors that, when governed

responsibly, can serve as valuable sources of jobs and raw materials for the indefinite fu-

ture. At the same time, like the timber industry in the National Forests, ocean-based indus-

tries such as commercial fishing have suffered from overexploitation as users’ extractive

activity has degraded the underlying ecosystems and surpassed managers’ abilities to con-

trol extractive use.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 199

The large spatial scales of both terrestrial and ocean ecosystems

encompass a complex mix of habitats. This complexity poses simi-

lar challenges of management and monitoring in the terrestrial

public lands and in the federal jurisdictional oceans.

Pursuant to the Executive Order federal agencies will enlist

state and tribal partners to develop spatially-specific use and re-

source management plans on a regional basis,

190

similar to Land

and Resource Management Planning process in the National For-

ests under NFMA. For the last decade, scholars and policy experts

have advocated to make multi-resource, multi-sector spatial plan-

ning and management approaches more central to ocean govern-

ance. With the Executive Order, this approach is now at the fore-

front of U.S. ocean policy.

191

As with the origin of NFMA in the 1970s, ocean planning

comes to the fore amidst a growing recognition that our oceans

and inland seas and the resources they contain are finite assets. A

long list of stressors – including widespread habitat loss, pollution,

over-exploitation, invasive species, and the effects of climate

change and ocean acidification

192

– threatens the viability of our

marine ecosystems, in turn jeopardizing the billions of dollars and

millions of jobs that the ocean’s goods and services provide.

193

Some states have already taken action on marine spatial plan-

ning, independent of the emerging federal initiative. The state of

Massachusetts has been the leader with respect to adopting and

implementing an MSP framework. The Massachusetts legislature

passed the Massachusetts Oceans Act in 2008

194

; the state pub-

lished its first Ocean Plan in December 2009.

195

190. Proclamation No. 13,547, 75 Fed. Reg. 43,023.

Other states are

191. Though the Executive Order and the Final Recommendations apply to the

Great Lakes as well, this article focuses on ocean management specifically.

192. See, e.g., CTR. FOR OCEAN SOLUTIONS, PACIFIC OCEAN SYNTHESIS, SCIENTIFIC

LITERATURE REVIEW OF COASTAL AND OCEAN THREATS, IMPACTS, AND SOLUTIONS (2009),

available at

http://centerforoceansolutions.org/sites/default/files/pdf/PacificSynthesis.pdf; Melissa

M. Foley et al., Guiding Ecological Principles for Marine Spatial Planning, 34

MARINE POL’Y 955

(2010).

193. See generally Boris Worm et al., Impacts of Biodiversity Loss on Ocean Ecosystem Ser-

vices, 314

SCIENCE 787 (2006); Gretchen Daily et al., Ecosystem Services: Benefits Supplied to

Human Societies by Natural Ecosystems, 2

ISSUES IN ECOLOGY 1 (1997).

194. Massachusetts Oceans Act of 2008 (codified at M

ASS. GEN. LAWS CH. 114, §

35HH).

195. See M

ASS. OFFICE OF ENERGY AND ENVTL. AFFAIRS, MASSACHUSETTS OCEAN

MANAGEMENT PLAN (2009), available at http://www.env.state.ma.us/eea/mop/final-v1/v1-

complete.pdf. The Plan designates two areas for commercial-scale renewable energy de-

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

200 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

poised to follow suit: Washington passed marine spatial planning

legislation in early 2010

196

; Rhode Island and Oregon are conduct-

ing MSP under existing legal mandates; and New York has initiated

MSP as part of its coastal management program.

197

Meanwhile,

California has nearly completed a network of marine protected ar-

eas that could function as a starting point for more comprehensive

planning, and both the California Legislature and the California

Ocean Protection Council have expressed interest in MSP.

198

The-

se ocean governance initiatives—along with other initiatives, such

as monitoring in marine protected areas—can benefit significantly

from the lessons of prior large-scale public resource management

efforts such as NFMA (and its public lands equivalent, the Federal

Land Policy and Management Act.)

199

Maintaining sustainable ecosystems is a key aspect of natural

resources management, in part because these ecosystems and their

constituent parts form the basis for generating the goods and ser-

vices on which we depend. As such, the lessons of NFMA’s MIS bi-

odiversity provisions—and the lessons of decades of experience

velopment and retains a “prohibited area” in which development is forbidden pursuant to

prior legislation. The majority of Massachusetts waters remain designated for general use,

although new siting maps and performance standards identify “special, sensitive, or

unique resources” and establish a mechanism for resolving use-ecosystem and use-use con-

flicts. Id. at 2-1 to 2-23.

196. Marine Waters Planning and Management, Substitute S.B. 6350, ch. 145 (Wash.

2010), available at http://www.ecy.wa.gov/programs/sea/msp/pdf/SB6350.pdf.

197. See N.Y.

DEP’T OF STATE, N.Y. STATE COASTAL MGMT. PROGRAM, ATLANTIC

OCEAN AMENDMENT (2010), available at http://www.nyswaterfronts.com/downloads/pdfs/

NYS_CMP_Amendment.pdf;

OR. DEP’T OF LAND CONSERVATION AND DEV., OR. COASTAL

MGMT PROGRAM, OREGON TERRITORIAL SEA PLAN, PART FIVE: USE OF THE TERRITORIAL SEA

FOR THE

DEVELOPMENT OF RENEWABLE ENERGY FACILITIES OR OTHER RELATED

STRUCTURES, EQUIPMENT, OR FACILITIES (2009), available at http://www.oregon.gov/

LCD/OCMP/docs/Ocean/otsp_5.pdf; O

R. DEP’T OF LAND CONSERVATION AND DEV., OR.

COASTAL MGMT PROGRAM, REQUEST FOR ROUTINE PROGRAM CHANGE TO THE OREGON

COASTAL MANAGEMENT PROGRAM, RPC.OR-2010-001, ITEM 4 (2010), available at

http://www.oregon.gov/LCD/OCMP/docs/Public_Notice/RPC.OR-2010-01_Combined

.pdf; R.I.

COASTAL RES. MGMT. COUNCIL, OCEAN SPECIAL AREA MANAGEMENT PLAN, DRAFT

CH. 1 (2009), available at http://seagrant.gso.uri.edu/oceansamp/samp.html.

198. See Cal. A.B. No. 2125 (as amended by Senate, July 15, 2010), available at

http://www.leginfo.ca.gov/pub/09-10/bill/asm/ab_2101-2150/ab_2125_bill_20100715_

amended_sen_v96.pdf; C

AL. OCEAN PROTECTION COUNCIL, OPC SUPPORT FOR

COLLABORATION ON MARINE SPATIAL PLANNING, (resolution amended, Sept. 17, 2009),

available at http://www.opc.ca.gov/2009/11/opc-support-for-collaboration-on-marine-

spatial-planning/; see also C

AL. OCEAN PROT. COUNCIL, STAFF MEMO RE: COORDINATING

GEOSPATIAL DATA TO MAP HUMAN USES AND CONDITIONS IN THE OCEAN ENVIRONMENT

(2009), available at http://www.opc.ca.gov/webmaster/ftp/pdf/agenda_items/20090917/

0909COPC_03_MSP.pdf.

199. 43 U.S.C. § 1701 (2012).

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 201

with environmental monitoring generally—are especially salient

for the emerging ocean planning effort.

B. Lessons from the Experience of NFMA’s Management Indicator Species

1) What to measure: Diversity is not just one thing

As many authors have noted, and as we discuss briefly above in

Parts 2 and 3, biological diversity is a slippery concept with nested

hierarchical levels. Accordingly NFMA’s mandate that Forest man-

agement provide for a “diversity of plant and animal communi-

ties” proved to be difficult for the Forest Service to translate into

practical terms. In the context of ocean policy, any large-scale ma-

rine spatial planning effort will have to balance a host of often-

competing uses for nearshore waters—from shipping lanes to

commercial fishing to marine protected areas—while maintaining

a level of biological complexity sufficient to ensure the ongoing vi-

ability of our marine ecosystems. Ocean planning regulations (or

statutes) therefore will very likely feature some language that ex-

plicitly addresses the conservation of biological diversity, with con-

servation listed among many other desirable uses of federal ocean

waters—just as NFMA did in the case of the National Forests.

One clear lesson from the NFMA experience is that such lan-

guage will have to be more specific in order to effectively guide

implementation. Rather than espousing the value of “diversity” or

“biodiversity,” a more effective mandate would include preserving

the particular elements and hierarchical levels of biological diversity

that together contribute to ecological complexity. For example,

one such hypothetical regulation might require that:

resource use plans for each identified habitat type within a man-

agement unit must use the best available science to ensure and

maintain sufficient genetic, population, species, and community

biological complexity to ensure 1) the continuing composition,

structure, and function of the ecosystem relative to an identified

reference condition, and 2) the continued viability, sustainability,

and evolution of the ecosystem’s constituent species.

Although perhaps less elegant than preserving “diversity,” this

more specific mandate would both better protect what we think of

as biological diversity,

200

200. That is, the proposed language would address the various hierarchical levels of

biological complexity, would link that complexity to measureable ecosystem characteris-

and provide clear guidance for the man-

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

202 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

agement agency

201

to avoid the expense and uncertainty of litiga-

tion. It is also a tenable goal: as the cost of biological assessment

and monitoring decreases, the intensity of monitoring can and

should increase concomitantly. Genetic sampling of seawater, for

example, could provide a wealth of information about community

composition (such as which species are present, and in what pro-

portion) while simultaneously speaking to the existing level of ge-

netic diversity within each species.

202

As the cost of genetic tech-

niques continues to plummet, environmental monitoring using

these methods becomes increasingly cost-effective, and in many

cases may be cheaper than traditional monitoring.

203

The language above punts on a critical scientific question: what

is “sufficient” complexity to meet the goal of the hypothetical reg-

ulation? Defining sufficiency is a key entry point for scientific data,

and this might usefully be made explicit rather than implicit. Any

natural resources statute or regulation requires scientific data to

give meaning to its terms—for example, “maximum sustainable

yield” is meaningless in the abstract, and requires field measure-

ments and ecological models to approach a credible estimate of

MSY for any particular species. Here, the implementing agency

would use primary scientific data to make the regulatory language

operational in the same way.

204

2) Why to measure it: Diversity is not the goal in and of itself

This interaction between science

and law provides the important secondary benefit, discussed fur-

ther below, of allowing scientific methods and norms to change

while remaining informative for the same point of law.

The shift from species management to ecosystem management

tics, and would seek to protect the processes (evolution and the raw genetic diversity that it

requires) resulting in the variety of biological entities we observe in the world.

201. Presumably NOAA would be the relevant agency, in the case of CMSP, although

in principle other public resources agencies could adopt similar requirements.

202. See, e.g., Philip Francis et al., Monitoring Endangered Freshwater Biodiversity Using

Environmental DNA, 21 M

OLECULAR ECOLOGY 2565 (2011) (implementing similar tech-

nique in freshwater lakes and streams to sample imperiled species, which are by nature

difficult to census by traditional, hand-counting methods).

203. Id.

204. Of course, no one definitive scientific answer would be forthcoming, but exist-

ing science allows us to estimate a “sufficient” population size to ensure ongoing viability

and to avoid the loss of genetic diversity that results from very small population sizes. Pop-

ulation viability analysis—familiar in the context of the Endangered Species Act—

performs a similar role. Note, too, that although the implementing agency would be due

deference in selecting the “identified reference condition,” the agency would be limited

by the regulatory purpose of ensuring and maintaining sustainably viable ecosystem com-

position, structure, and function.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 203

illustrates a larger point: diversity per se probably should not be a

core goal of natural resources management. Rather, preserving the

composition, structure, and function of the ecosystems under

management is the more fundamental mission.

Oceans—like the Forests—are multiple-use public resources,

and so both politics and practicality demand balancing these uses

in some rational way. But regardless of the particular balance of

uses one might wish to strike, maintaining a functioning ecosystem

is a necessary underlying precondition to any natural resources

management. Neither commercial nor recreational fishing, for ex-

ample, has a future in the absence of an intact trophic web sup-

porting the targeted fishery species. Tourism, the primary contrib-

utor to the ocean economy, depends in significant part on the

existence of intact coastal ecosystems—for instance, diving on cor-

al reefs requires the existence of coral reefs as well as the entire set

of ecological interactions that support them. Building ecosystem-

level monitoring into marine spatial planning is therefore neces-

sary to bring ocean management into the 21

Quantifying the composition, structure, and function of ecosys-

tems (and the baseline levels of temporal variability in these pa-

rameters) remains challenging, but some well-accepted measures

have existed for many years, and the experience of NFMA under-

scores the fact that identifying the real goals of a management

program is one key to its continued relevance in the face of evolv-

ing science.

century and to en-

sure its ongoing effectiveness.

A set of critical ecosystem measures, such as mean trophic level

and the ratio of net primary productivity to biomass, would provide

agencies with relevant management feedback when paired with

explicit baseline or reference conditions. In the context of the

oceans, community-wide genetic sampling (described above) offers

a high-resolution means of assessing the membership of species as-

semblages, and therefore of determining ecosystem measures like

mean trophic level. Net primary productivity can be measured rou-

tinely by satellite, and much of this data is already publicly availa-

ble.

205

205. NOAA and NASA satellites, for example, provide such data.

Incorporating these kinds of measurements would be a sub-

stantial step toward ecosystem-based ocean management, avoiding

some of the pitfalls of single-species-based biodiversity regulations

in the National Forests by speaking to the more fundamental goals

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

204 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

of natural resource management. As further measures of ecosys-

tem composition, structure, and function become commonly ac-

cepted in the scientific community,

206

California’s Marine Protected Areas Monitoring Enterprise

the implementing agency

could employ such measures to fulfill the existing regulatory re-

quirement, allowing science to evolve within federal ocean policy.

207

offers one example of the process by which agencies or public-

private partnerships might work towards implementing ecosystem

measures within a regulatory scheme in the marine environment.

The Monitoring Enterprise builds upon baseline environmental

data to develop monitoring plans that capture a core set of ecosys-

tem attributes within California’s network of marine protected ar-

eas. It does so by engaging a broad cross-section of stakeholders—

including scientists with significant expertise in the specific geo-

graphic area—evolve region-specific plans.

208

3) How to incorporate dynamic science into regulation

Perhaps the clearest lesson of the NFMA MIS experience is that

science will continue to evolve, and therefore regulation that re-

quires a particular monitoring technique risks becoming quickly

anachronistic.

209

206. As the scientific process generates new methods of measurement, the question

of how to distinguish between legitimate/useful advances and pseudo-science arises. The

implementing agency can establish quality control for evolving scientific standards by

wielding its “best available science” mandate (coupled with a court’s deference to the

agency in reasonably defining that science). Where this is insufficient, environmental and

administrative law could borrow from evidence law, applying the Daubert or similar stand-

ard familiar for expert witnesses.

The case of MIS is analogous to requiring the

207. M

ARINE PROTECTED AREAS MONITORING ENTER.,

http://monitoringenterprise.org (last visited Jan. 26, 2013).

208. See M

ARINE PROTECTED AREAS MONITORING ENTER., NORTH CENTRAL COAST

MARINE PROTECTED AREAS MONITORING PLAN 4 (2010), available at

http://monitoringenterprise.org/pdf/NCC_MPA_Monitoring_Plan.pdf.

209. Requiring such a particular monitoring technique would be an example of what

Prof. Seidenfeld calls an ex ante constraint on agency action. Mark Seidenfeld, Bending the

Rules: Flexible Regulation and Constraints on Agency Discretion, 51 A

DMIN. L. REV. 429, 433

(1999) (“Broad statutory delegations of power to an agency to regulate a general area of

the economy, characteristic of much legislation adopted under the New Deal belief in

agency expertise, impose few ex ante constraints on the agency. In contrast, statutes that

provide rigorous formulae may preclude reasonable regulation that balances social costs

against benefits.”) The present discussion of how to balance dynamism with stasis in law is

closely related to the larger and older body of work on the proper degree of agency discre-

tion, and regulations’ generality vs. specificity, in administrative law generally. See generally

Richard J. Pierce, Jr., Political Accountability and Delegated Power: A Response to Professor Lowi,

36 A

M. U. L. REV. 391, 404 (1987); Edward L. Rubin, Law and Legislation in the Administra-

tive State, 89 C

OLUM. L. REV. 369 (1989); Peter L. Strauss, Legislative Theory and the Rule of

Law: Some Comments on Rubin, 89 C

OLUM. L. REV. 427 (1989). Richard Stewart adds:

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 205

Forest Service to conduct all land-use planning using an Apple II

computer: arguably the best of several options at the time the reg-

ulations were written, but very quickly surpassed as technology im-

proved over time. To better integrate science into a regulatory

framework, it therefore makes sense to focus on the process-based

(rather than product-based) lessons of NFMA, seeking to define

enforceable standards of management while also incorporating

constantly-advancing science.

An important structural starting point is ensuring that the

management agency has appropriate institutional incentives; the

agency has to want to change, or at least must have an institutional

structure that can adapt in the face of new information.

210

In the

case of the National Forests, the Forest Service had long been

geared toward timber and game production rather than toward

maintaining a healthy balance of uses. Moreover, timber sales

helped to finance the Service, adding a financial incentive to err

on the side of timber production in land use planning.

211

Game

animals were important to a vocal subset of the Forest’s public, fur-

ther politicizing MIS selection and management decisions.

212

In many government endeavors it may be impossible in the nature of the subject

matter to specify with particularity the course to be followed. This is most obvi-

ous when a new field of regulation is undertaken. Administration is an exercise

in experiment. If the subject is politically and economically volatile — such as

wage and price regulation — constant changes in the basic parameters of the

problem may preclude the development of a detailed policy that can consistently

be pursued for any length of time. These limitations are likely to be encountered

with increasing frequency as the federal government assumes greater responsibil-

ity for managing the economy.

Richard B. Stewart, The Reformation of American Administrative Law, 88 H

ARV. L. REV. 1669,

1695 (1975). Note that if one reads “environment” in place of “economy” in Stewart’s

piece, his point retains at least equal force.

210. See, e.g., Biber, supra note 2, at 76 (discussing the evolution of the U.S. Geologi-

cal Survey as a monitoring agency).

211. See, e.g., Austin D. Saylor, The Quick and the Dead: Earth Island v. Forest Service and

the Risk of Forest Service Financial Bias in Post-Fire Logging Adjudications, 37 E

NVTL. L. 847

(2007) (discussing a particular case of financial conflict-of-interest, post-fire timber sales,

in the context of Earth Island v. U.S. Forest Service, 442 F.3d 1147 (9th Cir. 2006)). Note

that timber sales apparently represent net financial losses to the USFS, and that account-

ing within the Service (at least as of 2001) left the Forest Service’s “cost information totally

unreliable.” U.S.

GOV’T ACCOUNTABILITY OFFICE, GAO-01-1101R, FINANCIAL

MANAGEMENT: ANNUAL COST OF FOREST SERVICE’S TIMBER SALES PROGRAM ARE NOT

DETERMINABLE (2001).

212. Telephone interview with anonymous Forest Service employee integrally in-

volved with implementing biodiversity provisions in a suite of National Forests (Feb. 21,

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

206 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

It is critical that the agency implementing marine spatial plan-

ning on a national scale be financed independently of the re-

source-extraction rights they may confer. This was made clear, for

example, by the reorganization of the former Minerals Manage-

ment Service in the wake of BP’s Deepwater Horizon oil spill.

213

Ocean planning statutes or regulations could help ensure

structural adaptability by defining the relationship of the imple-

menting agency to the managed marine resources as that of a trus-

tee to a beneficiary. This would formalize the agency’s attendant

duty to act in the best interests of the beneficiary, the American

public, presumably including present and future generations. Be-

cause a trustee must act as a reasonably prudent person in manag-

ing trust property, it seems likely the agency would then be bound

to consider the best available information in making management

decisions regarding the public resources under its control.

Moreover, ocean planning would benefit from an agency culture

accustomed to balancing uses for long-term sustainability—

perhaps looking more to the Federal Reserve Board or OIRA ra-

ther than to the Forest Service or Bureau of Land Management as

a model for agency culture. This structural setting would improve

the odds of implementing testable, economically sustainable, and

data-based resource management policies.

214

Moreover, the trust relationship would impose the duties to inven-

tory and maintain the trust property, to account for actions regard-

ing the trust property, to be impartial with respect to beneficiaries

of the trust, and not to profit from trust property.

215

2012) (notes on file with the author).

213. MMS became the Bureau of Ocean Energy Management, Regulation, and En-

forcement (BOEMRE), which was then split to separate revenue-generating from safety-

regulation and enforcement activities. See Reorganization of the Bureau of Ocean Energy Man-

agement, Regulation, and Enforcement, http://www.boem.gov/About-

BOEM/Reorganization/Reorganization.aspx (last visited Jul. 3, 2012) (“In the place of

the former MMS – and to replace BOEMRE – we are creating three strong, independent

agencies with clearly defined roles and missions. MMS – with its conflicting missions of

promoting resource development, enforcing safety regulations, and maximizing revenues

from offshore operations and lack of resources – could not keep pace with the challenges

of overseeing industry operating in U.S. waters. The reorganization of the former MMS is

designed to remove those conflicts by clarifying and separating missions across three agen-

cies and providing each of the new agencies with clear missions and additional resources

necessary to fulfill those missions.”).

214. This fiduciary duty would of course be in addition to (and probably stronger

than) the APA’s baseline “arbitrary and capricious” standard for agency action.

215. As an example of these commonly-held duties see, for example, Cal. Prob. Code

§ 16000 (1990) (providing general fiduciary duties of trustee under California probate

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 207

Requiring the responsible agency to employ best practices

(such as those discussed above for environmental indicators) when

developing an environmental monitoring program would be a

more targeted means of embedding a dynamic element into public

resources regulation. Such a “best practices” mandate would be

akin to a best-available-science requirement,

216

Technology-based standards are a special case of a general

strategy that separates detailed, technical specifications (which

may change comparatively rapidly) from the slower-moving pro-

cess of notice-and-comment rulemaking under the Administrative

Procedures Act. Guidance documents play a similar role, albeit

one that is less enforceable. Just as regulations (found in the Code

of Federal Regulations) implement sections of statutes (the U.S.

Code), sub-regulatory guidance often contains ground-level details

useful for carrying out regulatory mandates. For example, as noted

above, the Forest Service Manual and Handbook contain guidance

relevant to the NFMA regulations. These documents do not have

or to the technolo-

gy-based standards of other environmental laws. The Clean Water

Act and Clean Air Act, for example, have provisions that require

polluters use the Best Available Control Technology or the Best

Practicable Technology (or some similar standard) to reduce their

pollution levels. Technology-based standards both institute an en-

forceable regulatory provision and allow that standard to become

more stringent over time. Moreover, they are particularly appro-

priate where some heterogeneity in the regulatory landscape—in

the case of environmental monitoring, this is literal habitat heter-

ogeneity—makes implementing uniform requirements difficult or

impossible. Applying the same logic, best practices for environ-

mental monitoring form an evolvable, “technology”-based stand-

ard for future public resources management; identified reference

conditions could ensure each geographic area meets core goals in

a quantifiable way. Critically, these best practices should be con-

strained by, and guided by, a clear statement of the purpose of the

monitoring regime and a requirement the agency use the best

available science in implementing it.

law). See also Mary Turnipseed et al., The Silver Anniversary of the United States’ Exclusive Eco-

nomic Zone: Twenty-Five Years of Ocean Use and Abuse, and the Possibility of A Blue Water Public

Trust Doctrine, 36 E

COLOGY L. Q. 1, 10 (2009) (discussing public trust applied to oceans).

216. For a discussion of possible improvements to the Forest Service planning regu-

lations, including a best-available-science mandate, see Nell Green Nylen, To Achieve Biodi-

versity Goals, the New Forest Service Planning Rule Needs Effective Mandates for Best Available Sci-

ence and Adaptive Management, 38 E

COLOGY L. Q. 241 (2011).

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

208 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

the force of law,

217

consistent with the larger principle of adminis-

trative law that more enforceable or coercive rules require greater

procedural safeguards, such as notice-and-comment periods in

administrative rulemaking.

218

By embedding malleable details—

such as scientific techniques subject to change—into sub-

regulatory documents, the implementing agency can update them

regularly,

219

although the price of such flexibility is decreased en-

forceability because the regulations were not promulgated through

the APA rulemaking process.

220

Another way to encourage agency accountability and respon-

sive policymaking is to establish an external scientific advisory

committee, with its role defined by regulation. In the case of

NFMA, the statute itself instructed the Forest Service to convene a

Committee of Scientists, limiting its membership to people not

“officers or employees of the Forest Service.”

221

The Committee

was a positive step to review and incorporate ideas from outside

the Service. However, it was not a permanent institutional compo-

nent: NFMA provided that “the Committee shall terminate upon

promulgation of the regulations.”

222

A standing external commit-

tee might have provided ongoing and valuable technical input, es-

pecially insofar as its members were unaffiliated with the regulators

or the regulated parties.

223

217. W. Radio Serv. Co. v. Espy, 79 F. 3d 896 (9th Cir. 1996) (holding that the Forest

Service Manual and Forest Service Handbook do not have independent force of law).

In the case of ocean planning, such a

committee could be charged with deriving or honing best monitor-

218. See, e.g., City of Williams v. Dombeck, 151 F. Supp. 2d 9, 22 (D.D.C. 2001)

(“[T]he law in this circuit expressly rejects publication in the Federal Register as the hall-

mark of a regulation. Rather . . . the real dividing point between regulations and general

statements of policy is publication in the Code of Federal Regulations, which the statute

authorizes to contain only documents having general applicability and legal effect.“) (cit-

ing Brock v. Cathedral Bluffs Shale Oil Co., 796 F.2d 533, 539 (D.C. Cir. 1986) (quotations

and emphases omitted)).

219. Better yet, the agency could empower an external committee of scientists to up-

date the scientific details in guidance documents, which would likely be a cost-effective

means of identifying and maintaining the best available science.

220. The limits of this internal guidance are then presumably set by a statutory or

regulatory “best available science” mandate.

221. 16 U.S.C.A. § 1604(h)(1) (2012). The Service went beyond this mandate, ex-

cluding timber company employees as well.

222. Id. Another committee was formed in 1997, which influenced the 2000 regula-

tions. See U.S.D.A.

FOREST SERV., COMMITTEE OF SCIENTISTS REPORT, available at

http://www.fs.fed.us/news/news_archived/science/.

223. Other agencies have similar scientific advisory committees. See, e.g., 16 U.S.C.A.

§ 1852 (2012). The National Academy of Sciences also plays an analogous role in shaping

national-scale policy, albeit from a more removed standpoint.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 209

ing practices and renewing them on a regular basis—say, every five

to ten years, consistent with an adaptive regulatory regime. This

periodic review could evaluate the techniques and performance of

the monitoring scheme as well as its purposes.

A necessary precondition to data-based (as opposed to faith-

based) natural resources management is the use of ecosystem out-

come data to inform future decisions that will inevitably affect that

ecosystem. Ideally, regulations would function as working hypothe-

ses, to be tested by outcomes on the ground; regardless of whether

the hypothesis is supported or refuted, new information (the out-

come resulting from the regulation) would animate the decision to

change or maintain the existing rule.

224

This is adaptive manage-

ment, a much-discussed philosophy of governance, but one that is

difficult to implement in part as a result of the agency incentive to

avoid explicit and public tests of decisions.

225

Because adaptive

management requires a consistent flow of information from which

to assess the results of past actions, ambient monitoring is critical

to adaptive natural resources management.

226

However, just as crit-

ical as re-evaluation and improvement of regulatory decisions is re-

evaluation of the methods used to collect the underlying monitor-

ing data. One means of embedding dynamic science in static regu-

lation, then, is ensuring that the regulation provides not just for

“adaptive management” in the abstract, but in concrete terms that

include the frequency and thoroughness with which the imple-

menting agency should re-evaluate monitoring methods.

227

224. See Lee, supra note

135; see also J.B. Ruhl, Regulation by Adaptive Management—Is

It Possible?, 7

MINN. J.L. SCI. & TECH. 21, 27 (2006) (arguing that adaptive management is

necessarily implicated in second-generation environmental law tools for dealing with dif-

fuse, complex problems); A. Dan Tarlock, The Nonequilibrium Paradigm in Ecology and the

Partial Unraveling of Environmental Law, 27 L

OY. L.A. L. REV. 1121, 1139 (1994) (discussing

adaptive management as part of a larger shift in academic ecology and natural resources

management, in which there is no such thing as a “final” management decision).

225. Lee, supra note 135. Note also that adaptive management is easily confused with

ad hoc management—each requires dynamic decisionmaking, but while the former is a

structured, scientific approach to learning from the results of one’s actions, the latter is

the opposite: unstructured decisions that evince a lack of learning from existing infor-

mation. The seminal work on adaptive management in natural resources is C

ARL J.

WALTERS, ADAPTIVE MANAGEMENT OF RENEWABLE RESOURCES (1986), and many authors

trace the origin of modern scholarship surrounding adaptive management to A

DAPTIVE

ENVIRONMENTAL ASSESSMENT AND MANAGEMENT (C.S. Holling ed., 1978).

226. See generally Biber, supra note 2.

227. For an insightful discussion of adaptive management in the context of law and

natural resource management, see Holly Doremus, Precaution, Science, and Learning While

Doing in Natural Resource Management, 82

WASH. L. REV. 547 (2007). One may also see adap-

tive management as providing iterative feedback. One aspect of complex systems is that

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

210 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

Finally, it bears mentioning that an ideal law, regulation, or

policy would be neither over-inclusive nor under-inclusive. It

would limit the behavior of all of its target population and only its

target population. Because no such ideal regulation exists, a key

policy question is whether the emerging marine spatial planning

effort (or any other natural resources management regime)

should err on the side of under-constraining or over-constraining

agency action. Too much agency deference leads to heterogene-

ous application of law and variable outcomes; too little agency def-

erence leads to inefficiency, frustration, and a sclerotic bureaucra-

cy.

Rational natural resources regulation should be tied to the spa-

tial and temporal scales of the processes that generate the re-

sources themselves. Given this, and applying baseline risk man-

agement in which both the risk of error and the severity of the

potential harm due to that error are factored into decisionmaking,

it seems reasonable to err on the side of over-constraining an

agency’s power to allocate resources more rapidly than those re-

sources are generated. Such a policy tilted toward avoiding the

harms of over-exploitation should sound familiar: it is essentially

the precautionary principle, well-established in international law

following the Rio Declaration of 1992, and now formally part of

the Implementation Plan for the National Ocean Policy.

228

This is analogous to maintaining a personal bank account: be-

cause it is far easier to spend money than it is to accrue money,

sustainable fiscal policy errs on the side of over-constraining ex-

penditures. In the case of biological diversity, reductions in diversi-

ty (however measured) are essentially irreversible on human time-

they can evolve over time. In this context, administrative law functions as a complex, adap-

tive system whose properties may be understood with reference to complexity theory in

physics and biology. See Donald T. Hornstein, Complexity Theory, Adaptation, and Administra-

tive Law, 54

DUKE L.J. 913 (2005).

228. N

AT’L OCEAN COUNCIL, DRAFT NATIONAL OCEAN POLICY IMPLEMENTATION

PLAN 97 (2012) (“One of the Policy’s guiding stewardship principles provides that deci-

sion-making will be guided by a precautionary approach as reflected in the Rio Declaration

of 1992, which states in pertinent part, ‘[w]here there are threats of serious or irreversible

damage, lack of full scientific certainty shall not be used as a reason for postponing cost-

effective measures to prevent environmental degradation.’”). It is worth noting that one of

the nation’s key laws protecting marine biodiversity, the Marine Mammal Protection Act,

16 U.S.C. §§ 1361-1423(h) (2012), implements an essentially precautionary regulatory re-

gime by creating a presumption against activities that might negatively impact marine

mammals. See Cara Horowitz & Michael Jasny, Precautionary Management of Noise: Lessons

from the U.S. Marine Mammal Protection Act, 10

J. INT’L WILDLIFE L. & POL’Y 225, 228 (2007)

(discussing the precautionary principle embedded within the Act).

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

2013] LESSONS OF ENVIRONMENTAL MONITORING 211

scales: even genetic diversity, which accrues most rapidly, takes mil-

lennia to develop in the species we tend to exploit most heavily.

For marine spatial planning and other future public resources

management regimes—which themselves depend upon the exist-

ence of intact ecosystems for the goods and services they are

charged with managing—sustainability turns on limiting the re-

source agency’s ability to erode ecosystem function faster than it

builds up. Thus, even as a future ocean planning regime should

provide the implementing agency with the procedural flexibility to

select techniques that evolve along with best available science, it

should constrain the agency’s substantive ability to unsustainably

allocate natural resources.

229

VIII. CONCLUSION

Public resources management requires management; that is, it

requires responsible agencies to make decisions about whether

and how to use natural resources that are public or public trust

property. As with any ongoing management—be it in the financial

sector, forestry, or ocean resources—making responsible decisions

requires some kind of feedback by which to gauge the impact of

those decisions. Environmental monitoring is thus an integral as-

pect of public resources management, providing the feedback

necessary for managers and for the public to evaluate the state and

the trajectory of our shared resources.

The insidious challenge of environmental monitoring is that it

seems simple, but its ground-level details are maddeningly com-

plex. In the case of National Forest management, a small statutory

element and brief interpretive regulation spurred years of litiga-

tion, failed regulatory reforms, and frustrations on the part of for-

esters attempting to implement the MIS regulation. The story of

229. One means of implementing such asymmetrical discretion is to focus on the

uncertainty surrounding any given data point. CMSP’s animating law could require that

the agency allow a margin of error (for example, 2 SE) in making allocation decisions. If,

for example, a management plan called for permitting shipping lanes in a given area of

ocean, and the agency estimated it could sustainably use 100 +/- 10 mi

for this purpose,

the 2 SE margin of error would allow it to lease only 80 mi

. As uncertainty decreased, the

margin of error would decrease accordingly. Such a precautionary limiting principle

would minimize risk to ecosystem composition, structure, and function, while providing an

agency incentive to improve its statistical methods (which would, in turn, lower the re-

quired margin of error and maximize agency discretion within allowable bounds). Of

course, the efficacy of such a system depends on the transparency and validity of the

method used to generate the margin of error.

I KELLY (DO NOT DELETE) 3/18/2013 1:08 PM

212 STANFORD ENVIRONMENTAL LAW JOURNAL [Vol. 32:151

NFMA’s MIS regulation is an object lesson in the interaction of

dynamic science and static law, and in the struggle within adminis-

trative law to balance flexibility and discretion in implementation

against a need to ensure the statute’s purposes are uniformly met.

Future natural resources management, and in particular the

emerging federal CMSP initiative, must improve upon the experi-

ence of federal biodiversity management in the National Forests.

Sustainable, data-based policy is a normatively desirable goal—

indeed, one might argue it is the only rational means of managing

public resources. But of course, raw data do not generate automat-

ic answers to core ethical questions about resources management:

what is sustainable use, over what timescales? What are acceptable

environmental tradeoffs for economic development? Such value-

laden questions are always embedded in resource management,

unaddressed by even the most thorough environmental indicator

data.

230

Environmental monitoring requires both a priori value-

based decisions (for example, which variables are important to

track?) and a posteriori value-based decisions (given the data in

hand, what is an appropriate use of public natural resources?).

231

Despite considerable agency incentives to avoid addressing these

questions directly in regulation, transparent and responsive gov-

ernance requires that value-based questions not be buried in tech-

nical standards. Rather, the lessons of NFMA and its MIS provision

for managing biodiversity suggest that future large-scale public re-

source management regimes—such as the emerging federal

coastal and ocean planning effort—must face difficult, value-based

decisions head-on or else risk substantial ecological and legal un-

certainty if they fail to do so.

230. See Holly Doremus & A. Dan Tarlock, Science, Judgment, and Controversy in Natural

Resource Regulation, 26

PUB. LAND & RESOURCES L. REV. 1 (2005).

231. See M.L. Morrison & B.G. Marcot, An Evaluation of Resource Inventory and Monitor-

ing Program Used in National Forest Planning, 19 E

NVTL. MGMT. 147, 153-54 (1995) (noting

that “[t]he [Forest Service] would be better served by developing a system to answer spe-

cific, key questions about the environment by selecting ecosystem element indicators that

will answer those questions, rather than forcing environmental questions a posteriori into

the MIS or other systems”).