Peer Reviewed Unpublished Report Commissioned and Submitted
on Behalf of Spirit of the Sage Council and
the National Endangered Species Network.
Science Missing in the "No Surprises" Policy


K. Shawn Smallwood, PhD1
Consulting in the Public Interest,
53 Clinton Street, Lambertville, New Jersey 08530

1 Correspondence to Davis Office: 109 Luz Place, Davis, CA 95616
Phone/FAX: (916) 756-4598 email: puma@davis.com

Commissioned by

The Spirit of the Sage Council, 30 N. Raymond Avenue, Suite 303, Pasadena, CA 91107
and
National Endangered Species Network, 915 L St., Suite C-347, Sacramento, CA 95814


The federal government recently agreed to full public review of the No Surprises policy, as applied to the issuance of Habitat Conservation Plans and Incidental Take-permits. As part of a settlement agreement of a federal lawsuit, (Spirit of the Sage Council, et al., v. Bruce Babbitt, Secretary of the Interior, et al.), the No Surprises policy is being proposed as a rule through the Federal Register, which will ensure public comment. The public comment period will end July 28, 1997. It is critical for scientists to take this opportunity to comment on how the proposed No Surprises and Safe Harbor rules and Candidate Conservation Agreements including the permit-shield provision will influence the use of scientific information and scientific methods in responding to declines in the distribution and abundance of threatened and endangered species. It is also critical for scientists to explain how the No Surprises policy and other rules providing assurances to take-permit holders will weaken the Endangered Species Act, and what impacts this weakening will likely have on the threatened and endangered species.

The Spirit of the Sage Council and the National Endangered Species Network commissioned me to write this paper, which is intended to support the efforts of comment preparation by scientists, small land holders, and the general public. My role is to first describe the modern scientific standards that are applicable to biological assessments prior to, and following, the issuance of incidental take-permits. These standards are defined in light of the wording and intent of the Endangered Species Act and current scientific principles and methods. Second, I will describe how the proposed No Surprises rule will affect the application of these scientific standards. This paper will present the reasons the proposed No Surprises Rule will further degrade the Endangered Species Act by preventing the use of scientific information and management practices in response to surprises, which are inherent to nature, science, and the mitigation plans typically developed by environmental consultants for incidental take-permit applicants. This paper is intended to introduce scientists and Services administrators to the implications and likely ramifications of the No Surprises policy as they relate to the application of science to conservation required by the Endangered Species Act. Walley (1996) summarized the legal implications of the No Surprises policy.

Science Applied to the ESA

According to Section 10 of the Endangered Species Act (ESA), Habitat Conservation Plans (HCPs) and Incidental Take-permits (ITPs), the incidental taking of threatened and endangered species shall not appreciably reduce the likelihood of the survival and recovery of the species in the wild. To this end, the best scientific information is to be used (Section 7(a)(2)), meaning that project impact and risk assessments must be made. By modern standards, these assessments must be spatially referenced (Gilpin 1996), and must include known or potential error rates or some estimate of uncertainty associated with the estimate of risk (Schulze et al. 1994). Such uncertainty analysis is required for the admissibility of scientific expert testimony in US courts -- scientific testimony cannot consist of unsubstantiated speculation (William Daubert, et ux., et al., Petitioners v. Merrell Dow Pharmaceuticals, Inc., No. 92-102, Supreme Court of the United States). The risk assessments must not be limited to the defined project area, because Section 4(a)(3)(A) requires the designation of critical habitat inside and outside the species range of distribution at the time of listing, and Section 2(b) requires conservation of the ecosystem upon which the threatened and endangered species depend. Furthermore, an incidental take-permit cannot be issued by the Services without the permit applicant assuring that adequate funding will be made available to implement minimization and mitigation of the taking to the maximum extent practicable.

The No Surprises policy, announced in Federal Register vol. 62 (103:29091-98), proposes as a rule that the Services (US Fish and Wildlife Service and National Marine Fisheries Service) will provide "general assurances" to HCP/ITP permit-holders that only the Services will need to respond to unforeseen circumstances that jeopardize species conservation pursuant to the ESA, so long as the take-permit holder has adhered to the terms of a properly functioning HCP. In other words, the No Surprises policy would relieve incidental take-permit holders from ESA obligations when the threatened and endangered species are discovered to be less likely to survive and recover due to the permitted incidental take. Instead, the Services would have the obligation of providing mitigation measures for these species. The use of public funds would therefore increase as surprises arise, although there would be no guarantee the new environmental conditions following the incidental take would allow for the Services to halt and reverse the declining status of the species and their critical habitat. As surprises arise, land holders who have not yet applied for HCP/ITP permits will face greater conservation challenges and costs.

According to the ESA, a properly functioning HCP is one that does not jeopardize the survival and recovery of the legally rare species. An HCP is more likely to be functional when it is based on application of the best scientific information along with appropriate adaptive management practices and all the other required administrative steps described in the first paragraph. (Note that the Federal Register 62 (103: 29094) announcement defined proper function on an HCP as fully implemented and in full compliance with original terms, rather than achieving conservation goals in full compliance with the ESA.) Whether any HCP can function properly is debatable (Beatly 1994, Shilling 1997), and will be the subject of much of this paper. The No Surprises policy assures only that appropriate scientific methods will not be adequately considered in the implementation of an HCP/ITP, for reasons that will be described in the following paragraphs. These reasons are inter-related, and deserve in-depth, comprehensive examination.

No Surprises

The No Surprises policy is antithetical to the ESA because it precludes application of the best scientific information, adaptive management, and proper function of the HCP/ITP, especially in light of the take that inevitably resulted from the permitted project. By implementing this policy, the Services would:

• conclude that the HCP/ITP is functioning properly;

• rely on faith that today’s available scientific information, as applied to the HCP/ITP agreement leading to the issuance of an ITP, will remain the best into the future;

• forego adaptive management, which appears to be regarded as a critical tool by the Services (Federal Register 62(60):14940), and which has some legal foundation (ESA Section 2(b));

• disregard a large body of scientific evidence, along with the professional opinions of many scientists, that surprises are inherent in human interpretation of the distribution and abundance of both common and rare species (Connell 1978, Grossman et al. 1982, Belovsky et al. 1994), and indeed, in nature generally (Prigogine and Stengers 1984, Davies 1988).

Science is a structured process by which humans gain understanding of nature (Popper 1969, Kuhn 1970). It involves testing hypotheses in an attempt to refute conjectures generated by previous experience and theory. This effort typically leads to more hypotheses, paradigm shifts, and hopefully to stronger theory, based on those hypotheses most difficult to refute. Science is challenging because:

• sensory perception of the investigator is superficial with respect to the space-time continuum and complexity of nature that needs to be measured (Whitehead 1938);

• scientists are split between two world-views, one of "natural order," and one of "chaotic nature" (Smallwood 1993); and

• events and observations are not independent from one another (Einstein et al. 1923).

Ecologists are therefore often required to violate the assumptions of their most important statistical tests used to measure ecological pattern and its significance (Smallwood 1993). The theoretical meanings of ecological patterns continue to change as methods of pattern measurement and analysis also change.

Our scientific understanding is therefore always changing. For example, most scientists no longer ascribe to the theory of Lamarckian evolution acting alone, which holds that organisms can acquire their morphological characters through inheritance based on levels of use by parents responding to a changing environment. Most scientists now believe that evolution of morphologies requires natural selection. When rare species lack the generations needed to adapt to a quickly changing environment, they go extinct (of course, many other factors also influence extinction). Land cover (vegetation) and land uses around the world are changing quickly (Meyer and Turner 1992) and extensively (Hannah et al. 1994), and species are going extinct as a result (Hester 1967). That is why the ESA was written and authorized by the US Congress.

Environmental science is also a social process (Soule 1991, Norgaard 1992). Many scientists study topics that can be funded by private and government sources. They study what others want them to study, lest they rely on their personal funds and risk of obscurity. The Congress provided a strongly written ESA. The ESA, like the Constitution, was based upon the best science and philosophy of the time and had, like the Constitution, at least some ability to accommodate change. However, the Congress has not provided the funding for scientists to adequately study threatened and endangered species, nor the ecosystems upon which these species depend. There might be a day in the future when society does decide to fund research that can substantially improve our scientific understanding of rare species and their supporting ecosystems. Our better understanding will lead to new management prescriptions better able to assure survival and recovery of legally rare species in the wild. The science applied to ESA compliance for HCP/ITP-holders during the next 20 to 100 years should not be restricted to today’s scientific body of knowledge -- it must be allowed to change as more research is funded and paradigms shift.

Surprises in Science

Surprises are common for those who study or manage environmental elements (Holling 1986, Watt 1992, Levin 1992, Costanza and Folke 1994), such as rare species and aspects of their supporting ecosystems. Ecologists typically attribute surprises to the following origins.

(1) Environmental Stochasticity

Studies of coral reefs, tropical rain forests, and midwest stream fish assemblages have indicated that ecological diversity and redundancy are at least in part evolutionary responses to natural catastrophes, which occasionally or even periodically or routinely clear ecological space or change the distribution of species (Connell 1978, Grossman et al. 1982). Although every species is unique and contribute uniquely to ecological communities, many legally rare species are redundant taxonomically and functionally within ecological communities subject to such catastrophes (Walker 1992), which is a means for such systems to achieve stability (Watt and Craig 1986). Maintaining redundancy in these communities increases the odds the communities will remain resilient to catastrophes. Other forms of environmental stochasticity are internal to the species, such as demographic stochasticity, genetic drift and mutations (Shaffer 1981, Wilcox 1984). Natural catastrophes are not only common, but species and ecological communities organize around them. Scientists require long time periods and regional scales to reduce levels of indeterminacy in biology due to "randomness of an event with respect to the significance of the event" (Mayr 1961). Catastrophes are surprises to unprepared scientists and resource managers for reasons explained throughout this paper.

(2) Inadequate Scale

Ecological investigations are often conducted at spatial and temporal scales that are too small to accurately characterize important ecological patterns and interactions (Connell and Sousa 1983, Ricklefs 1987, Wiens 1989, Levin 1992, Brown 1995). The spatial and landscape perceptions of species vary (Holling 1992, With 1994), but study reports rarely acknowledge this variation or its implications for interpretation of the results. Patterns of distribution that might appear random at conventional spatial scales of observation often appear highly aggregated and predictable at larger scales of observation (Greig-Smith 1983, Smallwood et al. 1996b). Population dynamics and spatial patterns often fail to reveal their true magnitudes until multigenerational studies are conducted (den Boer 1981, Taylor and Taylor 1977, 1979), which are rare (Connell and Sousa 1983). Without studies conducted at very large spatial scales and for multiple generations, surprises will be inevitable.

Furthermore, modeling landscape spatial patterns with increasing resolution tends to provide worse prediction (Costanza and Maxwell 1994). Information meaningful for ecosystem assessments and broad patterns of species distribution and abundance tends to be discernible at larger scales of observation and analysis (O’Neill et al. 1986, Kotlier and Wiens 1990, Levin 1992), and is the justification for development of the ecosystem indicators approach (Karr 1994, Schulze et al. 1994, Hammond et al. 1995). Surprises can occur just because studies were conducted with different levels of resolution, even at the same location and using the same data set.

(3) Improper Experimental Design

Ecological investigations lacking principles of experimental design can fail to identify critical habitat and landscape associations, as well as other important ecological patterns or factors (Green 1979, Hurlbert 1984, Smallwood 1993). Even the establishment of reserves or protected areas, in whatever configuration or design is in vogue, is not an "experiment." Reserves lack experimental controls, replication and interspersion of treatments, nor do they provide any means for testing the effects of observation scale (Green 1979, Connell and Sousa 1983, Smallwood 1993). A mitigation involving habitat protection cannot be experimental without identifying the critical habitat and population parameters and the ways in which they need to be studied so as to reduce their corresponding levels of uncertainty in risk assessment. That is, the relevant hypotheses need to be formulated and tested, lest the mitigation be based on faith in numerous unsubstantiated opinions. Once the environment is changed, such as typical of projects leading to HCP/ITPs, then the National Research Council recommends the implementation of an experimental design so that we can learn as much as possible from the manipulation (National Research Council 1986). The National Research Council (1986) and Green (1979) also recommend collection of base-line data prior to the manipulation, but these data need to be collected with clear understanding of the research hypotheses to be tested.

(4) Weak Theory

Ecologists study complex systems rich in feedbacks, homeostatic devices and potential multiple, dynamic pathways (Hairston et al. 1960, Mayr 1961, Roughgarden et al. 1989). Thus, ecological concepts and theories are often vaguely described with many different definitions for the same operational terms (Peters 1991), thereby allowing individuals to choose whichever definition suits their purpose. Density, e.g., the number of organisms per unit area, was presented by Peters (1991) as the most frequently used and most important operational term in ecology. However, density was inadequately defined by Peters (1991) as well as by most other ecologists. Density must be defined to spatial scale (Smallwood 1995, Blackburn and Gaston 1996, Smallwood and Schonewald 1996), and once defined in such a manner, Peter’s (1991) example of the best ecological theory, the allometry of density, actually turns out to have resulted from circular logic when applied to species of mammalian carnivores. Investigators of mammalian carnivores chose to study increasingly larger geographic areas for species with larger body mass. Theoreticians then related density to body mass without accounting for the fact that density decreases with increasing study area (Smallwood et al. 1996a). Therefore, ecological theories and even the most important operational terms are subject to surprise and revision.

Weak theory has always been a problem in ecology because the discipline has suffered from:

• data derived from a small group of species, risking bias due to lack of representation;

• lack of inter-specific comparison;

• short-duration data collection;

• focus on high density populations, neglecting the much more common cases of rarity;

• fragmented focus on numbers, energy flow, diversity, stability, ecological space, etc.; and,

• lack of synthesis of data, analytical results, and ideas in ecology (Watt 1971).

Ecological investigations are often uncoordinated, the terms and concepts are weakly described, and investigators often begin studies with poorly defined research hypotheses (Green 1979). An increasing number of ecological theories stand without foundation in data, despite being increasingly more complex (MacIntosh 1985, Peters 1991). Given such a state of affairs, it should be no surprise that ecologists and managers have been, and will continue to be, surprised.

Despite the challenges to ecology and the surprises they predispose, this scientific discipline offers the most useful information for formulating conservation and mitigation plans. Like any other scientific discipline, theory and method in ecology is always changing as our knowledge increases. The principles of ecology provide the most reliable prediction of how biota will respond to environmental changes.

Summary.--Prediction generally is poor in ecology (Peters 1991), thereby providing for many surprises. Ecology is a fairly young discipline, experiencing significant paradigm shifts (MacIntosh 1985) and lack of funding. The discipline of ecology will experience tremendous theoretical development into the future, as well as practical and heuristic development. Its application to environmental problem-solving would most likely be restricted by the proposed No Surprises rule.

Surprises from Incidental Take and Mitigation

Science and environmental stochasticity are not the only sources of surprises that can jeopardize the survival and recovery of legally rare species. Human activities are changing landscapes and ecosystems in substantial ways, which leads to greater risk of extinction and sometimes requires ESA compliance. The implementation of an HCP/ITP agreement or other mitigation plans can also lead to surprises, due to the following circumstances.

(1) Habitat Fragmentation and Cumulative Impact

The most certain environmental impact of any development project is the reduction and increased insularization of available habitat, otherwise known as habitat fragmentation (Wilcox 1984, Wilcox and Murphy 1985). Habitat fragmentation has become the principal anthropogenic impact with which the scientific discipline of conservation biology is concerned. It is obvious and well accepted among ecologists that "habitat fragmentation is the most serious threat to biological diversity and is the primary cause of the present extinction crisis" (Wilcox and Murphy 1985). No mitigation application is likely to eliminate the impact of any development project on habitat fragmentation, whether that strategy includes minimization of take, translocation, habitat "enhancement" in a reserve, acreage replacement in the form of easement or fee title purchase, nor the establishment of "preserves" interconnected by habitat "corridors." Development projects almost always reduce the spatial areas within which the legally rare species can survive, and they reduce the capacity of the landscape to provide for the dispersal and interchange of individuals from metapopulations. Fragmentation also increases habitat edges and their associated edge effects (Yahner 1988). Reduced spatial extent and increased edge-to-interior ratio correlate with more frequent intrusion and establishment of exotic species populations and other effects (Smallwood 1994). Land conversions extending right up to the reserve boundary will reduce the effectiveness of the boundary and of the reserve function well beyond the physical/political boundary of the reserve (Schonewald-Cox and Bayless 1986).

Furthermore, recent research results indicate a consistent pattern of dynamic spatial distributions among species populations. That is, populations usually are clustered spatially (Greig-Smith 1983, Smallwood 1995, Smallwood and Schonewald 1996), and these clusters shift locations through time (Taylor and Taylor 1977, 1979; den Boer 1981; Hanski 1994). Once the population density has been defined, its life span is limited, because nature is always achieving some balance between dispersive and congregatory behaviors (Taylor and Taylor 1977). By constraining populations to "preserves" that are some fraction of the spatial area of existing habitat, and by not guaranteeing contiguity of the mitigation preserves, the HCP/ITP planners deny the listed species and other species the ability to naturally shift to new locations. Contiguity has been shown to be important for population sizes among habitat patches for amphibians (Laan and Verboom 1990) and small mammals (La Polla and Barrett 1993), and its role in ecosystem functionality is a well accepted principle among ecologists and conservation biologists (Foreman 1981, Wilcox and Murphy 1985, Turner 1989). The typical HCP/ITP preserves will be mere fragments of previously contiguous habitat, which may very well fail to protect the listed species, and they certainly will not provide for recovery.

Clusters of natural populations shift locations through time, but ecologists remain unclear as to the reasons for this pattern of dynamic distribution. Several hypotheses have been proposed (Taylor and Taylor 1979):

• population clusters must move once they deplete their most limiting resources;

• the individuals in a cluster shift locations innately so as to prevent the exhaustion of resources;

• dispersal and territory establishment of the next generation also establishes the location of the next cluster, while the previous cluster senesces; and,

• a combination of the other three.

Whichever hypotheses are true, natural populations clearly require more space than someone might calculate with an existing density estimate and its unadjusted extrapolation to the geographic space needed to support a certain population size (e.g., a result of a population viability analysis). Estimating adequate space should become more problematic as space is restricted due to habitat fragmentation caused by development. Information in living matter (sensu Rothstein 1951) will also become more difficult to interpret.

Besides habitat fragmentation per se, development projects also fragment the landscape in ways that disrupt ecosystem function. Residential, commercial and industrial developments disrupt ecosystem functions by interfering with the movement of energy, important materials, and biota across the landscape, and by fouling ecosystem elements such as streams and downwind soils with contaminants. Development projects reduce water percolation through underlying soils, thereby reducing groundwater recharge. The associated vehicle traffic destroys many animals, and plant dispersal is halted at the boundaries of the developments. Reduced ecosystem function (Dynesius and Nilsson 1994) and increased habitat fragmentation set the stage for unfortunate surprises in the distribution and abundance of legally rare species.

(2) Failed Translocation

Translocation of threatened, endangered, or sensitive animal species have achieved poor success (Griffith et al. 1989, Dodd et al. 1991). The translocation of various plant communities has provided little if any evidence of success (Howald 1993), and abundant evidence of failures (Fahselt 1988). There have been few known or peer-reviewed, published studies confirming successful translocation of plant communities, nor has there been much evidence that plant communities can be successfully recreated as a scientifically acceptable mitigation method (Peter Bowler, personal communication, February 16, 1997; Read et al. 1996). Plant communities such as coastal sage scrub are much too complex ecologically to be translocated or created de novo somewhere else with any confidence of success. Such attempts do not qualify as "experiments," but only as trials based on faith in multiple assumptions and anecdotes. Scientists lack the understanding necessary to translocate plant populations that will survive longer than the first generation. For most rare plant species, scientists lack knowledge of the nutritional and moisture requirements and the allelopathic and other complex interactions that determine species distribution (Fahselt 1988). Scientists are nowhere near prepared to translocate or recreate natural ecosystems (Maycock 1985). Therefore, failed mitigation based on translocation should be expected, and if it is not, then it will be a surprise.

The Safe Harbor policy applied to translocation further threatens complete failure of conservation pursuant to the ESA. According to Safe Harbor assurances, the holder of a take-permit can translocate the legally rare individuals to another location and not be held responsible for the fate of individuals that return to the place from which they were removed. This policy provides no new habitat area, and risks causing harm to the biota in the area receiving the translocation. The proposed Safe Harbor rule will protect the permit-holder from the consequences of a scientifically unacceptable mitigation practice. Safe harbor is the No Surprises policy applied to translocation.

(3) Inadequate Species Accounts

The accounting protocol of legally rare species occurring on the project site usually involves reconnaissance-level surveys for species thought likely to occur there based on the examination of species’ distribution maps. Sometimes, there are no surveys whatsoever, e.g., Yolo County HCP (EIP Associates 1996) and often they are grossly inadequate. However, species distribution maps are not always accurate, and species’ populations are dynamically distributed (Taylor and Taylor 1979, den Boer 1981). Species range and other ecological boundaries should not be thought of as hard, but rather "fuzzy" (Rejesky 1993).

Sightings records are often used to assess presence/absence and habitat associations, even though Joseph Grinnell (1928) long ago served notice that sightings records are unreliable and should not be used for such purpose. Grinnell, who was a great naturalist and ecologist, stipulated that occurrence records are reliable only when they consist of specimens collected using museum standards. Although there are more modern techniques for assuring reliability of occurrence records, Grinnell’s recommendation should have been regarded more seriously than it has been among environmental consultants. Most scientists will agree today that occurrence records based on sightings can be used for legal compliance issues, so long as the sightings were made by trained scientists during research studies. Sightings records used to conclude absence are bound to produce surprises, to the detriment of the species.

The spatial distribution of each species is not static, so planning based on static maps of species range or habitat distribution is fundamentally flawed. Potential occurrence of legally rare species is more appropriate for planning purposes than documented occurrence, so long as the potential occurrence is based on critical habitat designation. Potential occurrence can be assessed using range maps and general vegetation, soil and land-use types. The benefit of the doubt on lands proposed for management under ITPs always should be given to the species, because we know that populations tend to be dynamically distributed, and rare species, by their nature, are usually difficult to detect in the first place.

(4) Impact Assessment Lacked Uncertainty Analysis

Environmental consultants and the Service’s Section 10 support staff biologists usually assess future project impacts qualitatively. That is, their frequent conclusions of ‘less than significant impact’ due to the proposed project is based on sightings records and purported habitat occurrences on project areas. Population estimates are rarely made specifically for impact assessments. Estimates are made for listing packages, often using outdated or inadequate spatial data, and then various periods of time elapse between these estimates for listing and the HCP/ITP impact assessments. Most impact assessments are therefore inadequate for assessing risk of extinction and ability to recover in the wild. When estimates are made, they lack any confidence intervals or error rates. Surprises in nature led to the scientific standard in the courtroom of uncertainty analysis (William Daubert, et ux., et al., Petitioners v. Merrell Dow Pharmaceuticals, Inc., No. 92-102, Supreme Court of the United States). Scientific evidence is admissible in court when it has been assessed for known or potential error rates. Admissible knowledge in the court requires more than subjective belief or unsubstantiated speculation, and scientific knowledge is admissible only when the knowledge is derived from an inference or assertion as part of the scientific method.

A relevant risk assessment method for environmental consultants is Population Viability Analysis (PVA). PVA is a flexible approach to estimating time to extinction, probability of extinction by a given date or period of time, probability of persistence, and minimum viable population size for persistence (Boyce 1992). All of these estimates have corresponding error rates, or uncertainty ranges, which meet one of the National Research Council’s (1986) standards for broad ecological studies used for environmental problem-solving. The negative version of PVA is population vulnerability analysis, which is more appropriate for assessing whether mitigation will comply with the ESA recovery standard. Until 1992, PVA has been applied to 35 species (Boyce 1992), and has been applied numerous times since then. It is a widely accepted method among ecologists and conservation biologists, and is intended for use with rare, vulnerable species. Connell and Sousa (1983) recommended that the minimum area be estimated for population or community persistence, and Schonewald and Buechner (1991) furthered this recommendation by providing the methodology. Soule (1991) recommended a viability analysis be performed for nature reserves. All these recommended variations of PVA would provide useful risk assessments with legally required error rates.

The PVA parameter values are especially relevant to assessing extinction risk in the face of declining space available for the listed species (Shaffer 1981). Based on the theoretical foundations of ecology and conservation biology (Gleason 1922, MacArthur and Wilson 1967, Soule and Wilcox 1980, Meffe and Carroll 1994), ecological space is one of the most important resources for all the listed species. To conclude the listed species are not in jeopardy of extinction due to the level of take proposed in many HCP/ITPs is to defy scientific knowledge and common sense. Such conclusions are often driven by political expediency and profit margins, with no empirical evidence to support them. The existing uncertainty in the parameter values used for PVA will be exacerbated by the loss of ecological space and by the possible loss of spatial contiguity, because space influences many aspects of species’ natural history, population dynamics, genetics, as well as the impacts of environmental stochasticity.

Areal reduction resulting from project take can translate to reductions in species distributions and population sizes within the planning area. Wide geographic distributions appear to be critical for the persistence of rare species (Goodman 1987), so reduced distributions of rare species will contribute significantly to cumulative impacts. Without careful analysis of the nature of the areal reduction, the planners would be prudent to give the benefit of the doubt to the species by assuming at least proportional reductions in distribution and abundance along with aerial reduction. Time to extinction decreases with smaller spatial areas on which the population can occur (Schoener and Schoener 1983, Pimm et al. 1988). That smaller populations are more vulnerable to extinction is fundamental to Population Viability Analysis (Boyce 1992).

In a number of cases in which I and a number of scientists have been directly involved, the conventional assessment of risk conducted by environmental consultants and the Services biological opinions all too often appears to be based on unsupported speculation, at best. Many of these risk assessments may lack adequate scientific rigor for institutional and programmatic reasons, which will be described in a later section. Whatever the reasons, however, these conventional, qualitative risk assessments provide no means to prevent future surprises due to the HCP, and cannot possibly suffice in court.

Summary.--In the experience of myself and others, environmental consultants often dismiss development projects as having less than significant impacts to the legally rare species in question, yet habitat fragmentation has resulted from these projects and constitute the greatest threat to the survival and recovery of these species in the wild. Mitigation plans based on translocation and Safe Harbor assurances are scientifically flawed and are usually indefensible on anything other than a subjective level, as are impact assessments based on species’ sightings records and lack of quantitative estimates and error rates. The establishment of reserves would be functional, so long as the management focus is on sufficient spatial extent and contiguity of habitat necessary for species recovery and measurable net benefits in perpetuity. A functional reserve needs to be established by scientist from start to end of planning and beyond with monitoring.

Scientific Standards for ESA Compliance

If the requirements of the ESA are to be met, and if the best science is to be used, and if the court standards for the admissibility of scientific expert testimony are to be met, and if the recommendations of the National Research Council (1986) are to be met, HCP/ITP and other project planners must perform impact assessments that meet the following conditions.

(1) Identify and Designate Critical Habitat

Habitats are defined by the species’ use of the environment, and therefore use and availability of environmental elements must be considered (Johnson 1980, Smallwood 1993, Hall et al. 1996, Morrison et al. 1992, Verner et al. 1992). Vegetation and soil types are not habitat unless they can be directly linked as required elements of the environment by a particular species (Hall et al. 1997). Spatial scale of observation can also influence interpretation of habitat associations because as the scale of observation changes, the availability likely also changes, and possibly use as well (Smallwood 1995, Smallwood and Fitzhugh 1995, Riitters et al. 1997). Therefore, the multiple spatial and temporal scales used to estimate population distribution and abundance also need consideration for identifying and designating critical habitat.

The experiments used to identify critical habitat can be mensurative or manipulative (sensu Hurlbert 1984). Mensurative experiments involve counts of species’ individuals or their signs along with each possible habitat type. These counts are then related to the availability of habitat types in the sampled landscape. Manipulative experiments involve replication and interspersion of multiple possible conditions for each presumed critical habitat element. Knowledge of habitat for legally rare species is usually limited to unscientific, natural history observations, which are indeed important, but inadequate for reducing the uncertainty as to which condition is preferable among those claimed by multiple naturalists. Smallwood and Morrison (1997) synthesized opinions from multiple naturalists to design an experiment for giant garter snake (Thamnophis gigas), in which water channels were arranged hierarchically for water depth management and in parallel for replicating and interspersing treatments representing the various opinions of critical habitat conditions. This and other types of appropriately and rigorously designed experiments can be used by the Services as one tool to develop and implement recovery plans.

Many GIS maps of vegetation types and other environmental variables have been presented to the public in support of HCP/ITPs. These maps usually look impressive, but are very often flawed with inappropriate categorization (e.g., habitat types), data aggregation at scales too coarse for the intended analysis, inaccuracy, and inappropriately hard ecological boundaries (Rejesky 1993). It is also important not to carelessly replace the terms "vegetation" and "vegetation type" with "habitat" or "habitat type," because such replacement is likely to be inappropriate (Hall et al. 1997), and the wording has ramifications for land-use decisions and policy (Rjesky 1993). The "habitat" depicted in a GIS map may not be the critical habitat that still needs to be identified and designated for ESA compliance.

The spatial data in GIS maps need to be analyzed for error rates at least in part by conducting ground-truth surveys (verification analysis) and iterative re-assignment of derived values per land unit (sensitivity analysis). Using state-of-the-art research methods, an accuracy assessment with error rates can be applied to predicted species’ ranges based on habitat designations and vegetation and landform maps, thereby meeting court standards, ESA requirements, and biological reality and adequacy.

(2) Ecosystem Assessment

The ecosystem upon which the threatened, endangered, and other rare species depend must be described and assessed for project impacts. Again, multiple scales must be considered because ecosystems are hierarchically organized (Klijn and Udo de Haes 1994), are best examined from the top-down of the hierarchy (O’Neill et al. 1986), and are not predisposed to convenient description within project boundaries, even though they are conceptual and arbitrary assignments of the environmental elements into compartments. Assessments of ecosystem condition, sensitivity, vulnerability, and impacts due to project activities require scientists who are trained in ecological theory and method with an emphasis on ecosystems. Assessments cannot be made from all the detail in many environmental elements comprising the ecosystem. Ecosystems are more reliably assessed using indicators of landscape structures and biological inventory influencing biological and physical transport and storage of material and energy.

Assessments of ecosystem function should be made at multiple spatial scales, ranging from the project area to the region (Klijn and Udo de Haes 1994). They also require uncertainty analysis of indicator values expressing sensitivity, vulnerability and impact, where sensitivity is the predisposition of the system to degrade due to changes in the conditions, vulnerability is the likelihood of degradation when particular anthropogenic activities pressure the sensitive parts of the system, and impact is the consequence of the pressure to the system. Risk assessment and uncertainty analysis of project impacts on ecosystem function are now possible (at least in part) using GIS and landscape ecology (Foreman 1981, Turner 1989, Graham et al. 1991), and the ecosystem indicators approach (Adriaanse 1993, Battaglin and Goolsby 1995, Bedford and Preston 1988, Cairns and McCormick 1992, Hammond et al. 1995, Karr et al. 1986, Karr 1994, Rapport et al. 1985, Rotmans et al. 1994, Schulze et al. 1994, USDA 1994). These modern assessment methods must be used to conserve the ecosystems upon which the legally rare and other sensitive species depend. Just stating that the ecosystem will be conserved by reserve establishment or other types of mitigation is inadequate.

(3) Population Estimates for PVA

Estimates of distribution and abundance of threatened, endangered, and other rare species must be made at multiple and spatial and temporal scales, the minimum scale being the area encompassing a persistent, natural population or community. Smaller scales are unlikely to reveal spatial requirements. The largest scale considered should include the species’ recent and current geographic range of distribution, so as to assess cumulative impact and collateral take (losses). Estimates of population size and project impact due to foreseeable take must represent at least several generations of each species (for estimates of variance), and must account for dynamic spatial and temporal patterns. Error rates must accompany the estimates of population size and distribution, as well as of risk to survival and recovery, and these error rates must be attributed to sources such as variability in the data (uncertainty analysis) and measurement error. An honest description of all the sources of uncertainty and model limitations must be represented (Rejesky 1993). Using these modern scientific standards, a PVA can be accomplished to satisfy court standards for expert testimony and level of assessment rigor needed to qualify it as the best scientific information, pursuant to the ESA.

(4) Monitoring for Impacts

All of the conceivable take, mitigation, and conservation impacts need to be considered, including cumulative impacts. However, the conceivable impacts do not always match the realized impacts, so adequate, scientific monitoring (sensu Morris 1955, Watt 1968, Green 1979) needs to be implemented along with an adaptive management strategy that details adaptive management practices to be put in place when monitoring reveals particular impacts. Monitoring of the legally rare species and functionally important ecosystem conditions (indicators) needs to be conducted at a spatial scale large enough to detect meaningful patterns of change through time. Meaningful patterns of change will be those that inform of likely impact. The monitoring also should be adequate for conducting power analysis (Gerrodette 1987; Morrison et al., in press). Monitoring for impacts should rely more on preventing Type II errors than Type I errors. If the null hypothesis is that the population or ecological indicator has not changed through time, while the alternative is that the population has declined, then rejecting the null hypothesis when it is actually true will lead to the false but conservative conclusion that the population or ecological indicator is declining. On the other hand, by not rejecting the null hypothesis when it is actually false, action will likely not be taken to adapt management for halting the decline of the species or ecological indicator. Concluding lack of statistical significance based on >5% chance of committing a Type I error should not be carelessly translated into lack of impact.

Scientific monitoring for impacts due to implementation of HCP/ITPs and Agreements should be described in the planning and take authorization documents. Appropriate goals and standards should be detailed for implementation of adaptive management practices.

(5) Adaptive Management

A conservation plan must be based on the social conditions, and it must maintain a plural, flexible approach (Soule 1991) to solving problems as they are discovered, pursuant to the ESA. Conservation or reserve boundaries are political only (Schonewald-Cox and Bayless 1986), and are permeable to people, their pets, invasive associates, and industrial and residential air and water emissions such as smoke, nitrogen compounds (van der Voet et al. 1996) and other contaminants. Many wildlife biologists agree that rare species will always require active management; assigning rare species to reserves will never suffice (Belovsky et al. 1994). The No Surprises policy would prevent implementation of adaptive management practices, which would render adaptive management strategies written into the HCP/ITP as nothing more than promotional text.

(6) Referencing of Information Sources

To comply with the ESA’s requirement of using the best scientific information, reference to information source of scientifically based conclusions must always be included in planning and take-authorization documents, whether these documents be prepared by environmental consultants or the Service’s Section 10 support staff biologists. Document writers are not being scientific when they use phrases such as "it is believed that…" without any reference to the source of such belief. Also, referencing will be more scientifically defensible and useful when the following standards are met:

• preference given to empirically-based reports, reviews of empirical reports, and scientific principles;

• balanced or comprehensive use of data analyses, scientific ideas, and anecdotal evidence supporting different sides of an argument, rather than tactical, selective referencing;

• accurate representation of referenced scientific research reports or published opinions.

Standard protocols for referencing in scientific document preparation are described in numerous books, papers, and scientific journal guidelines to authors.

(7) Independent Scientific Review

Scientific research results are usually subjected to peer review. If not, then they are published in what scientists refer to as the "gray literature." Scientists find value in gray literature where expedience in publication is useful, or where the author is targeting a select audience. However, scientists prefer to rely on peer-reviewed research results for building their theory. Many scientific journals require reference only to peer-reviewed research results. Peer review is an important quality of scientific research that keeps the process credible (Woolf 1981, Heath 1989) and more effective.

Because the ESA requires use of the best scientific information in biological assessments, independent scientific review should be a standard step preceding the issuance of any take-permit. Public review periods do not constitute independent scientific review, just as scientists do not obtain independent peer review by making their draft manuscripts available to the public. Rather, scientists solicit peer review, and usually the process is administered by journal editors. Independent, scientific peer review would greatly improve the public’s confidence in assessments and HCP/ITPs developed by environmental consultants, who usually are hired by the take-permit applicant. The environmental consultants have a vested interest in pleasing the take-permit applicant, so these consultants are vulnerable to bias. The Services and local government agency biologists should not serve as independent scientific reviewers, because they must issue the permits and oversee the plan’s implementation -- they are not independent from the projects under consideration, and could conceivably be biased.

The need for independent scientific review was made all too clear by the reviews of the Yolo County HCP (EIP Associates 1996), which were solicited from scientists at the University of California at Davis by Fraser Schilling (Schilling 1997). Just as the HCP appeared to be nearly approved by the last of the city governments in the county, Schilling asked some of his colleagues at U.C. Davis to review the scientific foundations of the Final Draft Yolo County HCP document. All 12 of these scientists concluded the science was flawed and the HCP should not be approved.

Independent scientific review should be standard and mandatory of HCP/ITPs and other biological assessments used to justify issuance of take-permits. The "science" behind take-permits issued by the government is in greater need of independent review than is non-applied science. The stakes are high for science when it is used to justify issuance of take-permits. These stakes include biodiversity, ecosystem functionality, the law, allocation of public funds, the public trust, and the integrity of the environmental sciences. Environmental consultants should not only be required to obtain independent scientific review, but they should also be obliged to publish their assessments in professional, scientist-reviewed outlets (National Research Council 1986). Such publications would help prevent errors and scientific fraud in its several forms (Woolf 1981, Chubin 1985, Stewart and Feder 1987). Sufficient detail should be provided in the publication to facilitate replication of the research (Chubin 1985), and raw data should be kept available for independent scientific review. Perhaps a clearinghouse of HCP/ITPs and supporting data would provide the openness needed for the process to work properly. Such a clearinghouse should be funded by the take-permit applicants, much as scientists fund their own professional journals.

Summary.-- Modeling genetic parameters will probably be less important than modeling demographic and ecological processes in conducting PVA (Shaffer 1987, Lande 1988, Boyce 1992). Loss and degradation of habitats are recognized by ecologists and conservation biologists as the principal causes and threats of extinction (Wilcox and Murphy 1985, Boyce 1992). Ecological indicators and associated methods (Bedford and Preston 1988, Graham et al. 1991, Cairns and McCormick 1992, Schulze et al. 1994, Hammond et al. 1995) can further improve the effectiveness and practicality of PVA, and in fact, can likely improve the reliability and validity of results because it provides measurement of ecological parameters at a more appropriate scale for assessing ecosystem functionality (O’Neill et al. 1986). Ecological indicators useful for assessing ecosystem functionality and PVA ought to be of great interest to the Services, which claim to prefer adaptive management and ecosystem strategies (Federal Register 62(60):14940). The indicators approach has emerged as the state-of-the-art scientific method for characterizing ecosystem condition, sensitivity, and vulnerability. These qualities are important for species survival and recovery, and they are spatially-dependent. The indicators approach should be built into PVA and applied to every region within which every HCP/ITP is considered, although the indicators approach must be used carefully as it is subject to abuse. Where uncertainty remains, mitigation should be implemented in the form of an experiment and should involve scientists (National Research Council 1986) prior to take-permit authorization.

Why are the Scientific Standards Missing from ESA Compliance?

Environmental consultants often lack incentives to apply academic and research scientific standards to their biological assessments of project impacts. Their employers are often the developers who want to reduce or obviate the regulatory requirements of the ESA. The scientific standards required for ESA compliance are the hurdles that take-permit applicants want their environmental consultants to overcome. Therefore, take-permit applicants are more likely to seek repeat business with environmental consultants who do not apply scientific standards to their biological assessments. Additionally, independent scientific review is not required by the Services nor other government agencies, even though the ESA requires use of the best scientific information. Conserving ecosystems, let alone legally rare species, not only requires high skill-levels among the environmental consultants and the Services, but also the benefit of reviews from well-trained scientists who are without conflicts of interest related to the project and have nothing to personally gain by promoting or halting the project.

Academic scientists are usually not engaged in ESA compliance and agency policies because such activities are unlikely to provide funding or research results needed by academics for publication and career advancement. In my experience, when academics are involved in biological assessments, they are usually involved only to lend post-facto or pseudo credibility to the assessments, and they are not the lead on the contract. Academic scientists are too busy conducting the scientific investigations that should serve as the new standards by which environmental consultants advise take-permit applicants in compliance with the ESA. However, some of these scientists are unaware that environmental consultants and the Services often fail to adhere to the scientific standards. They are also unaware these consultants often use scientific research results in whatever fashion is convenient to avoid ESA compliance. Environmental consultants and the Services need to be held to modern scientific standards either through independent peer-review by academic scientists or through legal actions by the Services enforcement branch and citizen lawsuit provisions.

Conclusion

Scientists need to be alert to the implications of the No Surprises policy (as already implemented and proposed as a rule). This policy will deny the appropriate application of scientific information to HCP/ITP mitigation when so-called unforeseen or extraordinary circumstances arise. These circumstances are bound to arise because surprises are inherent to nature and science, and because they are forced by poor mitigation planning that appears to be inherent to the role of environmental consultants in HCP/ITP preparation. Currently, HCP/ITPs are unlikely to contribute to the conservation of legally rare species and habitats, because the Services have not required permit holders and applicants to use the best scientific information, including the modern scientific standards and methods. Such scientific deficiencies are explained as follows:

• timelines for scientific data gathering and analysis are unrealistically short;

• most mitigation ratios and mitigation banking have no bases in science and contribute net losses in habitat space and contiguity;

• Safe Harbor assurances applied to translocations as mitigation for project and plan effects on the species are scientifically flawed, and risk harmful ecological interactions among biota at the location receiving translocated individuals;

• Congressional appropriations and private funding are inadequate for acquisition of threatened and endangered species habitats;

• Congressional appropriations and private funding are inadequate for recovery planning and reserve management;

• scientific research methods are not adequately applied by the Services to designation of critical habitat;

• scientifically defensible ecosystem assessments are not conducted by the Services nor the permit applicants and their consultants;

• risk assessments and uncertainty analyses are not conducted to estimate likelihood of species survival and recovery in the wild;

• scientific monitoring is not adequately conducted prior to or following the approval of plans, agreements and issuance of take permits, or are they adequately enforced by the Services; and

• no independent scientific review is required of HCP/ITP planning documents developed by environmental consultants.

Solutions to all these problems facing ESA compliance in the development of HCP/ITPs will require the implementation of explicit scientific standards as explained in this paper and referenced literature. Also, environmental scientists should rally in opposition to the No Surprises policy, lest their body of knowledge be made a mockery by those who will hasten the biological simplification of our country and increased dysfunctional behavior of our ecosystems.

Acknowledgements.--I thank Spirit of the Sage Council and National Endangered Species Network for financial support of this study, and Shu Geng, Chris Druxler and the Pacific Rim Science Center for use of their web site. I also appreciate the useful comments on the manuscript made by Dan Holland, Jan Beyea, Michael Morrison, Andrea Erichsen, Marc Commandatore, Gary Meffe, Joy Belsky, Verna Jigour, and Dean P. Keddy-Hector.