Environmental Management

, Volume 37, Issue 3, pp 367–379

An Evaluation of the Influence of Natural Science in Regional-Scale Restoration Projects

Authors

  • F. Brie Van Cleve
    • School of Marine AffairsUniversity of Washington
    • School of Marine AffairsUniversity of Washington
  • Terrie Klinger
    • School of Marine AffairsUniversity of Washington
  • Charles Simenstad
    • School of Aquatic and Fishery SciencesUniversity of Washington
Article

DOI: 10.1007/s00267-005-0014-8

Cite this article as:
Van Cleve, F.B., Leschine, T., Klinger, T. et al. Environmental Management (2006) 37: 367. doi:10.1007/s00267-005-0014-8

Abstract

Regional-scale restoration is a tool of growing importance in environmental management, and the number, scope, and complexity of restoration programs is increasing. Although the importance of natural science to the success of such projects generally is recognized, the actual use of natural science in these programs rarely has been evaluated. We used techniques of program evaluation to examine the use of natural science in six American and three Western European regional-scale restoration programs. Our results suggest that ensuring the technical rigor and directed application of the science is important to program development and delivery. However, the influence of science may be constrained if strategies for its integration into the broader program are lacking. Consequently, the influence of natural science in restoration programs is greatest when formal mechanisms exist for incorporating science into programs, for example, via a framework for integration of science and policy. Our evaluation proposes a model that can be used to enhance the influence of natural science in regional-scale restoration programs in the United States and elsewhere.

Keywords

Environmental restorationScience utilizationEnvironmental decision-makingProgram evaluation

The majority of landscapes on the planet have been altered by human actions that have, in many cases, reduced the goods and services provided by these systems and thereby threatened the integrity of ecosystems necessary for human well-being (Higgs 2003, Karr 1997). Environmental restoration offers a means to recover damaged elements of ecosystems and to restore ecosystem function. Consequently, environmental restoration is a growing field of academic inquiry and practical application that has the potential for substantial societal benefit (Allen and others 1997, Dobson and others 1997, Higgs 1994, MacMahon 1997, National Research Council 1992).

The performance of restoration programs is sensitive to variables arrayed along a natural–social science continuum. Attributes of the physical and biological environment play a large role in determining the success of restoration efforts, as do historical, cultural, social, political, aesthetic, and moral attributes (Higgs 1997). Here we restrict our focus to the role of natural science in restoration, and use the term “science” to refer to natural science only. Although we acknowledge the important role played by social science, we do not address its role in program performance.

Natural science is central to determining what is realistically achievable in environmental restoration and in providing overall direction as restoration efforts move forward (e.g., Anderson and others 2003, Bradshaw 1983, MacMahon 1998, Winterhalder and others 2004). Science is the primary form of knowledge in the modern era; its status has been highly regarded for centuries (Higgs 1994). The discipline of restoration ecology has emerged from a need to improve restoration efforts by incorporating contributions from the natural sciences and engineering, and science now is perceived as necessary but not sufficient for successful restoration (Gaudet and others 1997, Higgs 1994, Houck 2003, Lackey 1998, 2001, MacMahon 1997).

The restoration of damaged ecosystems has not always been based on science (MacMahon 1998), and even now natural science is not effectively incorporated into regional-scale restoration efforts (National Research Council 1992). The importance of natural science to environmental management efforts and restoration performance and the processes by which science can be incorporated into such projects both have received inadequate attention in project planning and implementation (National Research Council 1992, 1995). Consequently, there exist few analyses of the influence of science on development of natural resource policy (Healey and Hennessey 1994) or environmental restoration. However, a growing recognition of environmental problems of global scale, especially climate change, has reinvigorated the study of factors and approaches that enhance the influence of science in environmental policy (Clark and others 2002, Schnellnburger 1999).

Here we evaluate the influence of natural science in the development and implementation of nine regional-scale restoration programs. Our goal is to provide a framework to optimize the influence of natural science in coastal restoration, in order to provide maximum benefits to society and the environment. We used program evaluation techniques (Palumbo 1987, Wholey and others 1994) to test the hypothesis that the way in which programs integrate science into their organizational structure (the effective use of science) has a stronger influence on program performance than the content or attributes of the science itself (effective science). A corollary of this hypothesis is that the organizational structure of a program will dictate the effective use of science within the program. Consequently, an objective of this study was to determine the institutional mechanisms by which science is incorporated into the organizational structure of restoration programs in the United States and Western Europe.

Effective Science Versus Effective Use of Science

The use of program evaluation techniques to examine the role of natural science in selected regional-scale restoration efforts required that we develop attributes reflective of the overall effectiveness of the science as it is applied in the context of these restoration programs—in effect, unpacking the notion of “good” restoration science as described by Higgs (1994). Generally speaking, “good” restoration science (sensu Higgs 1994) is that which positively influences the development and implementation of restoration goals and policy. In the view of Clark and others (2002), influential science is perceived to be salient (addressing policy-relevant questions), credible (meeting standards of scientific rigor, technical adequacy, and truthfulness), and legitimate (fair and politically unbiased) by a broad array of actors in the relevant policy arenas (Clark and others 2002).

Others have argued that the content and the quality of the scientific information itself is important, as are the institutional processes by which it is generated, evaluated, and applied (Imperial 1999b, Lee 1993, Sabatier and Jenkins-Smith 1993). In this study, we make a similar distinction, characterizing these two dimensions of scientific assessment in support of environmental restoration as effective science and effective use of science, respectively. This characterization necessitated that we develop a template for decomposing these two dimensions of science—effective science and effective use of science—into component parts.

Somewhat akin to Lee’s (1993) “compass and gyroscope,” we view science as effective when it has attributes, such as peer review, that confer dynamic stability (the gyroscope) in addition to attributes, such as the use of program-specific endpoints and historic and baseline data, that assure its directedness toward goals appropriate for the particular restoration context (the compass). As we use it here, effective science has technical rigor and includes features that increase its utility for the intended purpose. Science is effectively used when it is influential (per Clark and others 2002) in program direction and outcome. Thus, attributes such as early use of science and transparency of science to the public become relevant because they enhance the likelihood of acceptance by the public and decision-makers, and provide opportunities for such groups to influence and assist the design and conduct of the restoration actions performed.

In this study, the allocation of attributes into separate categories that correspond to the dual objectives of assuring effective science on the one hand and effective use of science on the other is intended to highlight the importance of both content and process aspects of the use of science in restoration planning and implementation. Our emphasis on “effectiveness” reflects an important focus in program evaluation and the implicit assumption that the causes of ineffective application of science in restoration can be diagnosed and corrected via systematic “lessons learned” studies (e.g., Day and others 2004, Van Cleve and others 2004). Effectiveness can be thought of as both an attribute of policy or program alternatives (the extent to which a policy objective is achieved by an alternative (Quade 1975)) and a measure of technical efficiency (“a ratio measure relating observed output to planned output over some time period” (Brewer and deLeon 1983, p. 338)).

We developed a subjective, quantifiable scoring approach derived from the summative approach of Palumbo (1987) that permitted us to evaluate subjectively derived relative measures of effectiveness of the science component of each program against a scale whose upper bound represents an ideal of perfectly effective science or use of science. In order to do this, we identified factors that promote or prevent the effective use of science in environmental restoration programs, formulated criteria to measure the effectiveness and effective use of the science employed in these programs, identified factors common to programs judged to be successful, and identified desirable attributes that can contribute to the more effective use of natural science in other restoration programs.

Methods

We selected nine programs that were regional in scale, organizationally complex, and ecosystem focused. To the extent possible, we selected programs with estuarine components. This choice was driven by our collective knowledge of estuarine and marine systems. Nine programs met our criteria: the Chesapeake Bay Program, the Kissimmee River Restoration Project, the Comprehensive Everglades Restoration Project, the California Bay-Delta Authority, the Glen Canyon Adaptive Management Program, the Louisiana Coastal Areas, the International Commission for the Protection of the Rhine, the Skjern River Restoration Project, and the Salisbury Plain LIFE Project (Figure 1 and Table 1). Five of these programs had already been considered by the Puget Sound Nearshore Ecosystem Restoration Project (PSNERP) for a different purpose ( Van Cleve and others 2004).
https://static-content.springer.com/image/art%3A10.1007%2Fs00267-005-0014-8/MediaObjects/267_2005_14_f1.gif
Figure 1

(A) Locations of European programs studied. (B) Locations of U.S. programs studied.

Table 1

Summary of programs

 

Program purpose and area

Program structure

Science team established?

External science team?

Chesapeake Bay Program (CBP) 1983 http://www.chesapeakebay.net

To manage the Chesapeake Bay basin (166,000 km2) as an integrated ecosystem and to restore and protect it

State partnership with federal involvement and strong citizen input

STAC does not conduct or fund research, but advises and delegates to subcommittees

STAC provides peer review and advice

Kissimmee River Restoration Project (KRRP) 1992 http://www.sfwmd. gov/org/erd/krr/index.html

Return much of the Kissimmee River to its historic riverbed and floodplain (104 km2) and potentially reestablish pre-1960s fauna and flora

Partnership between SFWMD and the USACE, rather than stand-alone project

No science team. The KRREP is a science center run by the SFWMD

The Scientific Advisory Panel provides independent peer review and advice

Comprehensive Everglades Restoration Project (CERP) 1992 http://www.evergladesplan.org/

To restore, protect, and preserve the water resources of Central and S. Florida (47,000 km2); mitigate flood risk and meet water supply needs to 2050

Partnership b/t SFWMD and USACE; numerous projects coordinated with state efforts

Not centralized; science occurs on the RECOVER team and the Science Coordination Team

CROGEE provides peer review

California Bay-Delta Project (CALFED) 1994 http://www.calwater.ca.gov

To improve water supplies in CA and the health of the Bay- Delta watershed (3000 km2)

Agreement between state and federal partners, run by The Authority

The Science Program is the nexus of science

The Independent Science Board advises tech. panels, oversees peer reviews

Glen Canyon Dam Adaptive Management Program (GCDAMP) 1996 http://www.gcmrc.go/

To measure the effects of the Glen Canyon Dam operations on the Colorado River from GCD to Lake Mead (473 km)

Directed experimental management aiming at restoring habitat. Mostly federal partners

The AMWG is the central scientific body

The AMWG also provides advice and reviews the program semiannually

Louisiana Coastal Area Study (LCA) 1999 http://www.coast2050.gov/lca.htm

To restore and/or mimic the natural processes that built and maintained coastal LA from MS to TX

Federal and state partner- ship.The program is very directed or top-down

No formal science team

The NTRC advises the delivery team and provides independent peer review

International Commission for the Protection of the Rhine (ICPR) 1950 http://www.iksr.org/

To keep the Rhine ecosystem alive and in good health and to restore species that have disappeared in the basin (185,000 km2)

Multilateral partnership among riparian countries; agreed-upon goals are implemented by nation-states

No central science team. Each nation provides its own science, which is reported back to ICPR working groups

Not coordinated in multilateral agreement. Peer review provided, if at all, by nation-states

Skjern River Restoration Project (SRRP) 1998 http://www.sns.dk

Promote integrated catchment management using restoration to achieve improvements in environmental quality, flood control, and nutrient reduction (22 km2)

Partnership between national and EU LIFE. The national control is directed by state and federal agencies

No standing science team; science is hired out to consultants or generated/ provided by governmental agencies

No formal peer review or outside advice

Salisbury Plain LIFE Project (SPLP) 2001 http://www.english-nature. org.uk/Salisbury/

Protect and restore the largest unbroken expanse of calcareous grassland in north- west Europe (197 km2)

EU LIFE and governmental partnership

No science team. Defense Estates manages the area and provides science

No formal peer review or outside advice

STAC (Science and Technical Advisory Committee), SFWMD (South Florida Water Management District), USACE (U.S. Army Corps of Engineers), KRREP (Kissimmee River Restoration Evaluation Project), RECOVER team (REstoration, COordination, and VERification), CROGEE (Committee for the Restoration of the Greater Everglades Ecosystem), The Authority (California Bay-Delta Authority), RFPs (Requests for Proposals), AMWG (Adaptive Management Work Group), NTRC (National Technical Review Committee), EU LIFE (European Union LIFE project).

We used summative merit and worth program evaluation techniques to evaluate the role and influence of natural science in each of these programs. As described by Guba and Lincoln (1987), “merit” evaluation focuses on the intrinsic character of the object being evaluated, in this case the effectiveness of the science employed in the restoration program. “Worth” evaluation focuses on extrinsic characteristics, in this instance the effectiveness of the use made of the science in these restoration efforts. The main components of this evaluation were, first, the development of attributes that comprise the intrinsic and extrinsic roles of science as described above, and second, an evaluation of the science component of each of the nine programs in terms of those attributes.

Program evaluation employs measures that allow the researcher to ask how well the program achieves the intended objectives (Guba and Lincoln 1987, Putt and Syringer 1989). In a similar way, we defined a number of attributes that collectively could be used to describe the influence of science in the programs we evaluated. We developed these attributes by using publications and reports (National Research Council 1992, 1995, Van Cleve and others 2004) to identify techniques and tools used at the interface of science and policy within the fields of environmental management and, specifically, restoration ecology. Because restoration ecology is a developing discipline, some attributes used here may be too new to have received much attention in the published literature. As a result, an implicit bias may undervalue some attributes in our analysis.

We classified the selected attributes according to our interpretation of their contribution to effective science versus effective use of science (Table 2). A priori, we defined attributes of effective science as those that contribute to the content of science used in the program, and we defined attributes of effective use of science as those that relate to the position and prominence of science within the organizational and decision-making framework of each program. In the end, we assigned 15 attributes to each category, although it was neither necessary nor intentional to assign equal numbers of attributes to each. Finally, we validated our choice and classification of attributes through consultation with experts in restoration science, program evaluation, and marine science. We made minor changes to our definitions and classifications based on their comments.
Table 2

List of attributes and definitions (n= 30)a

Effective science

  Peer review in use† — rigorous and anonymous review of products by outside technical experts.

  Monitoring in use† — includes baseline, trend, implementation, effectiveness, compliance, impact, and validation.

  Pilot studies used/in use† — demonstration projects test alternative restoration treatments against controls.

  Indicators in use† — an indicator is a proxy for a larger process, structure, or state and shows environmental change.

  Conceptual models in use — models that summarize conceptual understanding with diagrams or text.

  Numerical models in use — models that summarize current understanding with parameters based on empirical data.

  Transparency of science in use — visibility and understandability of scientific process and methods used.

  Independence of science — science operating free of political pressures.

  Collaboration with outside science — working with external scientists or practitioners.

  Multidisciplinary science in use — functional specialization and creative collaboration.

  Baseline and historic information in use† — necessary for evaluating change and restoration possibilities.

  Reference sites in use† — experimental controls comparable in structure and function to restoration sites.

  Performance measures in use† — quantifiable metrics of processes or functions that can be used to judge “success.”

  Scientific achievability and relevancy of goals† — tangible and relevant restoration goals stated in scientific terms.

  Program-specific endpoints in use† — objectives should be expressed in measurable and attainable quantities.

Effective use of science

  Time elapsed before science was incorporated (early science)† — science should be incorporated early for optimal contribution.

  Defined science team — a designated source of scientific knowledge facilitates integration, creativity, and innovation.

  Likelihood of spatial interaction — increases chances for informal interaction, facilitating integration.

  Likelihood of temporal interaction — enhanced by frequent interactions and meetings among individuals and groups.

  Identifiable science leadership — leaders, individuals or groups, enhance accountability and improve program coordination.

  Vertical integration pathway in use† — transmission of scientific products and science representation on policy panels.

  Combined directed + discovery approach — combined “top-down” and “bottom-up” approach to directing program science.

  Mechanism for incorporating fresh perspective — turnover avoids burnout and ensures new ideas are incorporated.

  Accessibility of program science — how easily the public and external practitioners can access program information.

  Mechanism for communicating science to public — outreach is a critical step in assuring public buy-in.

  Methods of incorporating public input into science goals — restoration goals must incorporate public values to ensure buy-in.

  Adaptive management in use† — applies results from scientific experiments to management decisions.

  Programmatic review in use† — captures the perspective and advice from outside experts and enhances program credibility.

  Use of science in goal setting — the goal-setting process can be as important as the goals themselves.

  Process-focused† — holistic approach to cumulative environmental degradation by addressing underlying physical processes.

aAbbreviations of attributes appear in bold. The symbol “†” indicates attributes frequently identified as important.

We collected information about the role of science in individual programs by conducting interviews with natural scientists and decision-makers associated with each program. Initial interviews with representatives of five U.S. restoration programs were conducted in 2001 by members of the PSNERP Science Team ( Van Cleve and others 2004). In 2003, Van Cleve re-interviewed at least one representative from each of the five programs, and added a sixth U.S. program, the Kissimmee River Restoration Program, by conducting similar interviews. Van Cleve traveled to Europe in the summer of 2003 and conducted interviews in person and over the phone with representatives of three European programs. Subsequently, Van Cleve interviewed five independent experts familiar with regional-scale restoration efforts in the United States and Europe to develop additional perspective regarding the focal programs.

Interviewing methods consisted of semistructured, elite, organizational interviews with key individuals (Dexter 1970, Kvale 1996, March 1965). Most inter- viewees were lead scientists or persons familiar with the role of science and the interface between science and policy in their respective program. At least one representative from each program was interviewed and in several cases two or more were interviewed. Each interview lasted approximately 1 h and was structured around seven research questions, each further divided into interview questions designed to elicit the role of science in programs (see Appendix 2 in Van Cleve 2004).

We combined information gathered in these interviews and site visits with information from the peer-reviewed literature, from program and other websites, and from unpublished documents to determine the role of each attribute in each program. We did this by developing a rating scheme to quantify levels of performance for each attribute (see Appendix 1 in Van Cleve 2004) and assigning a score of 0–3 to each attribute for each program. We tallied the scores separately by category (effective science and effective use of science). According to our rating scheme, a score of 0 indicates that the attribute was not incorporated or utilized in the program; a score of 1 indicates that the attribute was only minimally incorporated; a score of 2 indicates an intermediate or moderate level of use; and a score of 3 indicates that the attribute was fully incorporated into the program. We used discrete values only (0, 1, 2, 3), rendering our data categorical rather than continuous, and did not a priori establish a median value for each attribute. Our rating scheme is coarse, but reflects the maximum precision of our data and our confidence in discriminating four (but not more) levels of performance. The scheme is necessarily subjective, and although we were internally consistent in our application of scores, they remain subjective measures developed for comparative purposes only ( von Winterfeldt and Edwards 1986).

We summed the attribute scores in two ways: (1) by program, totaling the scores across categories (effective science attributes and effective use of science attributes), and (2) by attribute, totaling the scores across programs. We summed scores for each program individually and for all U.S. programs combined and all European programs combined. We avoided adding scores from different categories (that is, we did not add scores for effective science with those for effective use of science) because the attributes contained in each category likely are not independent. When appropriate, we tested for significance using a two-sample t-test (α = 0.1) or a Spearman rank correlation test.

Results

Programs tended to score slightly higher in effective science than in their effective use of science, but this difference was not significant (P = 0.12; Figure 2 and Table 3). Only the Rhine and Kissimmee programs achieved significantly different scores for effective science versus effective use of science (two-tailed t-test, P > 0.01 and P = 0.003, respectively).
https://static-content.springer.com/image/art%3A10.1007%2Fs00267-005-0014-8/MediaObjects/267_2005_14_f2.gif
Figure 2

Total scores for all programs, plotted as effective science versus The straight line is drawn to show equally balanced effective science and effective use of science. If programs score better in effective science relative to effective use of science, they fall below and to the right of the diagonal line. If programs score better in effective use of science relative to effective science, they fall above and to the left of the diagonal line.

Table 3

Program and attribute scoresa

 

Chesapeake

Kissimmee

Everglades

Bay-Delta

Glen Canyon

Louisiana

Rhine

Skjern

Salisbury

Total

Effective science

          

  Peer review

2

3

2

3

3

2

2

0

0

17

  Monitoring

3

3

2

2

3

1

3

3

2

22

  Pilot studies

2

3

3

3

1

1

1

1

1

16

  Indicators

3

3

3

2

3

1

3

2

3

23

  Conceptual models

2

3

3

3

3

2

2

3

0

21

  Numerical models

3

3

3

3

3

2

1

3

0

21

  Transparency of science

2

3

3

3

2

0

0

1

1

15

  Independence

3

2

1

3

2

1

1

1

2

16

  Collaboration

2

2

1

3

2

0

3

2

2

17

  Multidisciplinary

2

3

3

3

2

3

3

2

3

24

  Baseline info

3

3

2

3

2

3

3

3

3

25

  Reference sites

3

2

0

3

0

0

1

1

2

12

  Performance measures

3

3

3

2

3

1

3

0

3

21

  Scientific achievability

3

3

2

3

3

2

1

1

3

21

  Endpoints

1

3

3

0

0

1

3

0

3

14

Total

37

42

34

39

32

20

30

23

28

 

Effective use of science

          

  Early science

1

3

2

3

3

1

3

3

3

22

  Science team

3

2

3

2

3

2

0

1

1

17

  Spatial interaction

3

2

2.

2

3

0

1

3

3

19

  Temporal interaction

3

2

3

3

1

0

1

1

1

15

  Leadership

2

3

3

3

3

2

0

3

2

21

  Vertical integration

3

1

3

3

3

2

3

3

2

23

  Directed + discovery

3

1

1

3

2

1

2

1

1

15

  Fresh perspective

2

0

1

1

0

1

0

0

0

5

  Accessibility

3

3

2

3

2

2

1

1

0

17

  Communicating to public

3

2

2

3

2

2

1

3

2

20

  Public input

0

0

1

3

0

1

0

0

1

6

  Adaptive management

0

1

3

3

3

1

1

1

1

14

  Program review

1

2

1

1

3

3

1

2

1

15

  Science in goal-setting

3

3

2

3

3

2

0

2

1

19

  Process-focused

1

3

2

3

2

2

2

3

3

21

Total

31

28

31

39

33

22

16

27

22

 

a0 = attribute not used/incorporated; 1 = attribute minimally used/incorporated; 2 = attribute moderately used/incorporated; 3 = attribute fully used/incorporated.

On average, U.S. programs scored higher for both effective science and effective use of science than did European programs. Combining ranks across both metrics yielded an overall rank order (from highest to lowest) of Bay-Delta, Chesapeake, Kissimmee, Glen Canyon, Everglades, Skjern, Salisbury, Rhine, and Louisiana (Table 4). Across programs, total scores for effective science were not correlated with those for effective use of science. Consistent with this result, the rank order of programs based on effective science differed significantly from that based on effective use of science (Spearman rank correlation test, 0.05 < P < 0.10), suggesting that program-specific differences exist in the contribution of science to program performance. However, both effective science and effective use of science were significantly positively correlated with overall rank order (Spearman rank correlation, P = 0.001 and 0.02 < P < 0.05, respectively), suggesting that both elements contributed to overall performance in these programs.
Table 4

Program ranksa

 

Effective science

Effective use of science

Sum of ranks

Overall rank

Bay-Delta

2

1

3

1

Chesapeake

3

3

6

2

Kissimmee

1

5

6

2

Glen Canyon

5

2

7

3

Everglades

4

4

8

4

Skjern

8

6

14

5

Salisbury

7

7

14

5

Rhine

6

9

15

6

Louisiana

9

8

17

7

aRank order of programs based on effective science attributes, effective use of science attributes, and overall rank (effective science rank + effective use of science rank). A Spearman rank correlation test showed that the rank order of programs for effective science and effective use of science differed significantly at a 10% confidence level (0.05 < P < 0.10).

We performed a rough test of the resolution of the model by comparing the results of a full analysis that included all attribute scores with the results of an intentionally constrained analysis that included fewer scores. To do this, we identified 14 attributes that are frequently cited in the literature, for example, National Research Council (1992, 1995), as important to program performance and repeated the analysis using this smaller set of data. The rank order of programs differed significantly depending on the number of attributes included (Spearman rank correlation, 0.01 < P < 0.02). Consequently, we determined that the full set of attributes used here provides a different, and likely more robust, ranking than would be obtained if only a subset of attributes were considered. In other words, even though some attributes are most frequently cited as important to program performance, those attributes alone do not appear to fully explain the role of science in restoration programs.

We next considered whether program maturity (or program age) explained program ranks. To evaluate the impact of program maturity on the role of science in program performance, we first assigned each program a maturity score between 1 and 7. This scale captures the general stages of evolution of the programs we studied, namely (1) problem identification, (2) program formulation, (3) planning, (4) implementation, (5) feedback, (6) revised implementation, and (7) program termination. We used correlation analysis to test whether the role of science tended to increase with program maturity (Figure 3A and 3B). Although program maturity explained only 3% of the variation in effective use of science, program maturity explained 38% of the variation in effective science. The latter relationship was significant at a 10% confidence level. Thus, it appears that programs amass more science as they age but they do not necessarily use this science more effectively as they mature.
https://static-content.springer.com/image/art%3A10.1007%2Fs00267-005-0014-8/MediaObjects/267_2005_14_f3.gif
Figure 3

(A) Total score for effective science plotted against program maturity fitted with least-squares regression line. The straight-line relationship between the total score for effective science attributes and program maturity explains 38% of the variation in total score for effective science attributes. This relationship is significant at a 10% confidence level (P = 0.08). (B) Total score for effective use of science plotted against program maturity fitted with least-squares regression line. The straight-line relationship between the total score for effective use of science attributes and program maturity explains only 3% of the variation in total score for effective use of science attributes.

High-scoring attributes were more frequent among those relating to effective science than those relating to the effective use of science (Table 3), and the mean and median attribute scores, summed across programs, were greater for those attributes relating to effective science than effective use of science (mean values = 19.0 and 16.6, respectively; median values = 21 and 17, respectively).

Discussion

Our results indicate that 30% of attributes identified were generally well incorporated into the science aspect of these programs, suggesting that programs express their scientific components through a relatively small proportion of attributes. Among the attributes most often incorporated at a high level were the use of baseline information, the use of multidisciplinary science, and the use of indicators.

The contributions of other science-based attributes were program-specific or idiosyncratic. Programs were clearly distinguishable based on specific combinations of attributes. This may reflect the fact that programs excel in different ways, leading to the development of flagship attributes such as the adaptive management focus in the Glen Canyon Program. That idiosyncratic or flagship attributes appear to play a large role in program performance suggests that the attributes used to evaluate these studies should not be considered prescriptive; other attributes not identified here could be important to the performance of other programs. Collectively, our results indicate that the utilization and influence of science within programs is variable, context dependent, and (in most cases) still suboptimal.

Three generalizations emerge from our results. First, the effective use of science in regional-scale restoration projects is paramount to the integration of science within programs. Second, U.S. and European programs differ in their use of science. Third, both political and geographic context influence the role of science in regional-scale restoration programs. We discuss each of these in turn.

Effective Use of Science is Paramount

Attributes contributing to effective science were more frequently cited in the literature and more often incorporated into programs than those contributing to effective use of science. Consequently, both European and U.S. programs tended to achieve higher scores for attributes representing effective science than for those representing effective use of science. This suggests that it is easier for programs to generate or access scientific information than to actually employ science in an effective way.

Many attributes cited as important in the literature were incorporated at moderate levels in the restoration programs we studied, as indicated by moderate scores. Other important attributes were not. This may reflect difficulties inherent in application (e.g., the use of pilot studies and reference sites is expensive and not always feasible), or could reflect their emerging importance (i.e., they are too “new” to have been implemented in programs). For example, a process-based approach remains a relatively new and challenging concept to integrate into large-scale restoration (National Research Council 1994, 2004).

Although some attributes clearly are more important than others, we were unable to qualify this difference beyond attributes frequently versus infrequently identified as important in the literature. We acknowledge that the actual importance of attributes in restoration likely varies much more than these two categories can indicate. In addition, our coarse approach to scoring multiple attributes (i.e., 0–3 ratings) may favor programs that perform moderately well in many attributes because these programs may score higher overall than programs that excel in only one or two attributes of particular importance. A second consequence of the coarseness of our scoring approach is that it fails to adequately distinguish between degrees of implementation. For example, the Glen Canyon program has by far the best developed adaptive management program among the programs studied. However, the Everglades and Bay-Delta programs also have adaptive management programs that, according to our rating scheme, are of high quality. Therefore, all three programs were assigned scores of 3. The resulting bias is that programs that specialize in one or two important attributes may be undervalued in this analysis.

Effective use of science appears to be a better overall measure of science performance in programs than effective science. On its own, effective science is necessary but not sufficient to ensure the contribution of science to program performance. The Kissimmee program offers an example of this distinction. This program ranked first in effective science attributes, but fifth in effective use of science. Based on reports from inter- viewees, the rank based on effective use of science may be a more appropriate reflection of overall program performance: the program intensively collected data during feasibility studies and program planning but then stopped and is now proceeding haltingly with little ongoing research and relatively few actively involved scientists. Few project resources are devoted to continuing the historically prominent role of science in the program and vertical integration in the program is limited to the delivery of technical reports to decision-makers. This supports our hypothesis that the way in which programs integrate science throughout the organizational structure has a greater influence on the overall role of science than content of the science itself, as long as a fundamental level of technical excellence is attained.

In ecosystem-based management, developing effective institutional arrangements is at least as important as developing policies that ensure desired outcomes, and the evolution and improvement of ecosystem-based management strategies are dependent on improved institutional design and performance of management institutions (Imperial 1999a, 1999b). Incorporating science in restoration (itself an element of ecosystem-based management) appears to be no different, in the sense that development of effective institutional arrangements for incorporating science is at least as important as the science itself. An example of this might be the creation of a standing science team to address not only the basic science requirements of the project, but also to address emerging issues and to respond to low-frequency large-magnitude events (hurricanes, algal blooms, hypoxia). Institutional arrangements that provide for timely, appropriate responses to unforeseen events can substantially enhance the effective use of science in restoration, and can prevent unanticipated setbacks in program performance.

Effective science was more apparent in mature programs (e.g., Bay-Delta, Chesapeake, and Kissimmee) than in immature programs (e.g., Louisiana and Everglades). However, the effective use of science remained low across all programs, independent of program age or maturity. This again suggests that the absence of an effective programmatic framework for incorporation of science can hinder the influence of science throughout the life of a program, despite increasing access to relevant science.

Economic investment in science could be an important explanatory variable in analysis of the contribution of science in program performance. However, we were unable to compare levels of economic investment in science among programs because this information was not available in a form that could be accurately compared across programs. In addition to problems of access are problems of source and application of project funds. For example, some programs rely on direct or dedicated funding of science, whereas others rely largely on “in kind” contributions of expertise or labor. Further confounding the issue of investment in science are situational differences between pre-existing bodies of science (rendering scientific information relatively inexpensive) versus the generation of new, program-specific science (which can be very costly). Finally, the cost of science is not necessarily an indication of its utility, and the most expensive science may not be the most useful in achieving project goals. For all these reasons, we omitted economic considerations from our analysis.

U.S. and European Programs Differ in Their Use of Science

We found that the programmatic application of environmental restoration is less well developed in Europe than in the United States. We admit difficulty in finding examples of regional-scale restoration in Europe comparable to the U.S. programs we studied. Most examples in the peer-reviewed literature and elsewhere refer to restoration studies and programs within the United States, and all of our European contacts considered European environmental restoration efforts to be nascent. This may be due to differences in the duration of human alteration of nature in the U.S. versus Europe: in the United States, major human alterations (i.e., those following discovery and settlement by Europeans) have occurred over a few hundred years, whereas major human alterations of the environment in Europe have occurred over a few thousand years.

Consequently, in Europe, cumulative human impacts are likely to be greater, and memory of unmodified environments more remote, than in the United States. The relative speed of transition from natural to cultural and industrial landscapes likely was greater in the United States than in Europe, and was far more alarming to early environmental thinkers, such as George Perkins Marsh (Marsh 1864). Thus, the perceived social need for restoration in Europe may be less urgent and the technological challenges greater than they are in the United States.

Bradshaw (1996) argues that the goal of restoration is to return to a close approximation of pristine conditions. It is possible that the perceived existence of vast, open, and undeveloped landscapes, characteristic of the interior of the United States and Alaska, and even more so of Canada, correlates positively with the perceived need to restore impacted areas in the United States, as well as with the perceived possibility of doing so. In Europe, the bases for such perceptions and comparisons are likely to be different, and consequently a return to pristine conditions (or “wild nature”) may be perceived to be neither possible nor desirable.

Although our findings suggest that the role of science in European restoration programs is less well developed than in U.S. restoration programs, we cannot exclude the possibility that this finding may be due to our U.S. perspective and our direct involvement with U.S. programs exclusively; the attributes that we chose to include may be more characteristic of U.S. than European programs and could reflect a U.S. bias in our approach. As we stated previously, there is no one “right way” to do regional-scale environmental restoration. However, there may be distinct U.S. and European approaches. If this is the case, the European programs we studied may have scored low because they were measured against standards created from a U.S. perspective. Extending this analysis to programs from Australia, New Zealand, and Canada could help to determine whether a U.S. bias influenced our results and to validate the applicability of the metrics we used. Furthermore, differences in economic investment could explain some of the differences we observed, and this we are unable to quantify.

Cultural differences between Europe and the United States relate not only to restoration, but also to the role of science in natural resource management and environmental policy decisions more generally. For example, Americans and Europeans have two distinct approaches to the accessibility to scientific information in environmental risk decision-making (Jasanoff 1996). “Americans demand full disclosure of all the facts, whereas Europeans are often contented with more targeted access to information” (Jasonoff 1996, p.66). Americans favor a “right to know” model, whereas Europeans prefer a “need to know” approach. The latter approach assumes it is the government’s responsibility to provide only as much technical information as is really necessary for citizens to make prudent decisions (Jasanoff 1996).

Science has been called the “American faith” because of the willingness of many Americans to unquestioningly accept the products and process of science (Doremus 1997). Science has played a prominent role in American society since colonial times, shaping the American legal structure, government institutions, and society as a whole. In the United States perhaps more than in Europe, science appeals to basic societal and moral values. Thus, the relative importance of science to Americans compared with Europeans may explain why the role of science in American programs was relatively advanced for their maturity compared with European programs.

Political and Geographic Context Matter

The political and geographic context of restoration programs must be considered in developing a scientific approach to regional-scale restoration because both may present unexpected limitations or opportunities for the use of science. The Rhine program is the only multinational program studied; it achieved the lowest combined score of any program. In this case, the political context may greatly influence the role of science in the program, for example, by subordinating the use of science to the more compelling political need to engage eight riparian and watershed nations in strategic restoration. Geographic context may similarly influence the role of science in restoration. The Louisiana program, which scored lower than any other U.S. program, is directly affected by the entire Mississippi watershed, or 41% of the contiguous United States. Not only does the spatial scale of the problem faced by the Louisiana program dwarf that of all other programs considered here, but this geography encompasses vastly different political landscapes, exemplified by a division within the scientific community regarding the relative importance of different sources of coastal Louisiana’s problems that require reconstructive action (Boesch 1996).

Science changes the way we understand the world (Jasanoff 1996). Consequently, the role of science in restoration affects understanding of the initial problem, formulation of appropriate goals, and ultimate performance of regional-scale, process-based, ecosystem restoration. Our results suggest that (1) ensuring the content and technical rigor of science is important but (2) lack of integration of science into programs can preclude optimal utilization of science, even when the content is relevant and the technical rigor is high. Therefore, science is most effective when policies and procedures, either formal or informal, exist for communication between policy and scientific aspects within specific programs. Once this integration framework exists, attributes relating to the content of the science become important. In the absence of an integration framework, the content of the science is less likely to influence project outcomes because science in any form is less likely to be incorporated into the program.

Finally, we acknowledge that natural science is not the sole predictor—or even the best predictor—of program performance, and that social dimensions (which we did not consider) can be determinative of program success. Even so, the effective use of natural science in regional-scale restoration is likely to have a direct positive effect on program performance (via increased success in physical and biological manipulations), which in turn can produce substantial indirect positive effects in the social dimension, resulting in better overall performance of regional-scale restoration programs.

Conclusion

We developed a framework for analysis of the role of natural science in regional-scale restoration programs. Our application of this analytical framework to selected programs in the United States and Western Europe suggests that effective science is important to program performance, but that the use of such science can be constrained by the absence of formal, integrated mechanisms for incorporating science into program execution. Program performance is optimized when effective, program-specific science is generated and explicitly incorporated into program action via established mechanisms.

Acknowledgments

Funding to FBVC was provided by the Puget Sound Nearshore Ecosystem Restoration Project, Washington State Sea Grant College Program, UW School of Marine Affairs, Smith College Frances Grace Scholarship, and Anchor Environmental, L.L.C. M. Logsdon and M. van Heeswijk were especially helpful in the conduct of this research. This study would not have been possible without the willing participation of independent experts and representatives from regional-scale restoration programs, and we are grateful for their assistance.

Copyright information

© Springer Science+Business Media, Inc. 2006