MEETING SUMMARY
This Meeting Summary is a synopsis of the main points discussed. The majority of reviewers of the draft Proceedings considered that a summary was necessary and helpful. However, some reviewers were concerned that readers would not obtain a full understanding of the nature of the discussions from a summary. Readers are encouraged to review and reference the full text.
The overall goal of the meeting, as set out in the invitation, was to define a research program that addresses wind power-related avian mortality issues. This research program should investigate both individual site impacts and national cumulative impacts. To reach this goal, the meeting should (1) identify and prioritize key issues with respect to bird-wind turbine interactions, (2) define a research agenda to resolve scientific and technical issues, while (3) insuring transferability of results, (4) avoid duplication and inadequate science, and (5) build consensus on approaches to the research needed to address the issues."
At the start of the meeting, a more specific list of meeting objectives was agreed upon by meeting participants:
 |
to help all parties understand the principal interests and concerns of one another; |
|
to identify and where possible prioritize the key scientific and technical questions regarding avian-wind interactions at wind power plants; |
|
to define and where possible prioritize research projects to address the questions identified; |
|
to identify research study requirements (e.g. time frame, resources and challenges associated with particular research proposals); |
|
if there is time, to develop consensus on a national research plan and establish priorities were possible; and |
|
define possible next steps. |
For the purposes of this meeting, consensus was taken to mean that "all participants at the table can live with a decision being considered".
Wind Technology Overview
Electricity generation from wind power is becoming a commercial reality in many parts of the United States, expanding across the country from the established base in California. In California, installed capacity is about 1,600 MW, with an additional 500-1,000 MW under discussion. In the U.S. outside California, about 50 MW has been installed, about 250 MW is at the permitting or construction stages, and at least 500 MW of further capacity has been proposed.
At present, about 15,000 wind turbines are installed in the U.S.A. To reach 10,000 MW of capacity would require about 16,000 new turbines of present design. If the upper limit on the economically exploitable wind resource in the U.S.A. is 100,000-200,000 MW, an upper limit of about 200,000 to 400,000 wind turbines would be needed, assuming no further increase in average power per turbine. However, power generated per turbine has increased ten-fold in the past decade. Also, expansion to this level would require many decades.
The U.S. wind turbine supplier industry is expanding, in part because of technical and financial support from the EPRI/DoE/NREL/Utility wind turbine development and testing programs. These programs provide risk-shared funding to conduct operational tests of about 20 turbines of a given type.
Continued expansion of wind power in the U.S. will require successful resolution of a number of key issues, including concerns about collisions between wind turbines and birds.
Resolution of these issues will require careful, reasoned discussion and teamwork among the major sectors of society that will be affected by the growth and use of wind power. The National Wind Coordinating Committee (NWCC) has been formed to provide a forum for discussion and resolution of such issues.
Some meeting participants noted that plans for rapid expansion of the wind power industry provide a unique opportunity to do carefully planned tests of wind plant effects on birds, including well-controlled pre- vs. post-construction comparisons. This might be done in an Adaptive Resource Management framework.
Avian Mortality Questions at Wind Power Plants
Meeting participants reviewed and discussed a lengthy list of questions about bird-wind plant issues that had been compiled in advance of the meeting. It was agreed that many of these questions were related, and that essentially all of them could be taken into account by a much shorter list of general questions about bird-wind plant issues. The initial list was condensed first to a 14-point list of questions (p.
*), then to a 7-point "distilled" list (Table 3A) and then to a 5-point "sequenced" list (Table 3B).Some attendees preferred the five-point formulation and others preferred the seven-point list. There was insufficient time during the meeting to reach consensus on a single list. However, there was general agreement that the 5-, 7- and 14-point lists of potential research areas, whichever one prefers, provide a good indication of the research topics that the meeting attendees collectively considered important.
Table 3. Condensed lists of major research areas.
A. Seven-Point "Distilled" List, not in any logical or priority sequence:
What are the population effects of avian mortality at wind plants, including cumulative effects?
Ø Determine whether wind plant mortality is additive or compensatory (see definitions above, p. *)
Ø Estimate the decrease, if any, in the average annual survival rates of species of interestDetermine avian mortality, including consideration of appropriate and comparable tools, methods, and techniques.
Identify ways to prevent or mitigate mortality or enhance avian viability:
Ø Develop ability to predict impacts
Ø Develop methods to reduce unnecessary mortalityWhat causes avian mortality from wind turbines or wind plants?
Ø bird behavior
Ø turbine design
Ø wind plant design
Ø location of wind plantAssess overall direct effects and indirect ecological effects of avian mortality at wind plants.
What are the indirect effects of wind development on avian populations, i.e. disturbance and habitat modification effects?
Agreement on research design protocols, including how to involve lay people in monitoring and assessment of wind plant effects.
B. Five-Point "Sequenced" List:
Assess mortality attributable to wind turbines at existing sites (including control data from "no turbine" sites).
Predict mortality at planned wind power sites, based in part on (1).
Predict population consequences.
Identify ways to reduce bird kills at wind plants.
Set values for off-site mitigation.
Relevant Past and Ongoing Research
The History of Wind-Related Avian Research in the U.S.A.
The first study of the potential impact of wind turbines on birds in the U.S.A. involved a mid-1970s study of the ERDA/NASA 100 kW Experimental Wind Turbine in Ohio. The wind turbine was not proven to be a high risk to nocturnal migrants or other birds.
Bird movements and collision mortality were studied for one year as part of the performance monitoring program for the large Boeing/PG&E MOD-2 Wind Turbine in Solano County, California. One American Kestrel was seen to fly into the turbine, and seven dead birds were found although not all of these necessarily collided with the turbine. Actual mortality, allowing for scavenging and detectability biases, might have been as high as 54 birds.
A survey and review by the California Energy Commission (CEC) of reported bird mortality at California wind plants in 1984-88 found evidence of bird mortality, mostly of raptors. A follow-on 2-year study conducted in the Altamont Wind Resource Area by BioSystems Analysis Inc. found 183 dead birds, of which 119 were raptors (mainly Red-tailed Hawks, American Kestrels and Golden Eagles). Fifty-five percent of the mortality was attributed to collisions with turbines, but no birds were directly observed flying into turbines.
Several studies are underway or planned to assess the potential impact of wind farm development on bird injury and mortality. Nearly all of these studies involve pre-construction monitoring of bird movements. These studies are in Maine, New York and Texas, and similar studies are planned for Vermont, Oregon, and Montana. The opportunity to acquire valuable pre-construction data is great, but to date the methodologies applied in different studies have not been standardized.
Industry Research: Kenetech Windpower
The Kenetech Avian Research Task Force feels that it is important to acquire precise quantitative data in sufficient quantities to allow reliable assessments and conclusions. The rate of accidents per turbine is very low, so direct human observations of collisions are impractical for determining how and why birds collide with turbines.
Studies initiated to date by Kenetech include controlled flights of homing pigeons near turbines to obtain quantifiable data on general bird/turbine interactions, development of automatic machine-recording systems such as video monitoring, development of a tracking system to provide precise 3-dimensional data on bird movements near turbines, studies on the visual and acoustic capacities of raptors, a study of perching behavior by raptors on turbines, and a population study of Golden Eagles in the Altamont area, as monitored via radio-telemetry.
The Kenetech Avian Task Force feels that adequate management of bird collisions at wind plants must take account of (1) initial siting, (2) size and layout of the wind plant, (3) design of turbines and towers, and (4) the possibility of off-site mitigation.
U.S. Federal Wind Energy Program Avian Research Projects
The population study of Golden Eagles in the Altamont region, mentioned above, has been initiated as a one-year pilot study with federal support.
Avian research was identified as a high priority in a 1994 request to universities for proposals to the Federal Wind Energy Program. A raptor perching study likely will be funded.
The cost-shared Utility Wind Turbine Verification Program mentioned above includes federal support.
Lessons from Utility Structure Environmental Impacts
Based on almost 25 years of studies of bird collisions with powerlines and other utility structures, various biological generalizations can be formulated: Ø Sooner or later, birds collide with any tall structure. Ø Collision vulnerability varies with species, age, sex, habitat, weather, human disturbance, and location. Ø Not all dead birds found near a tall structure were killed by colliding with it; cause of death can be difficult to determine. Ø Knowledge of bird behavior is critical to finding quick and cost-effective solutions. Ø The significance of mortality depends on the population affected.
Procedural generalizations based on utility experience include the following: Ø It is difficult to get good estimates of collision mortality. Ø Necropsies of dead birds are needed. Ø Good scientific studies are needed in order to develop solutions; these studies are time consuming and expensive. Ø The perceived significance of mortality depends on many non-biological factors.
The Avian Powerline Interaction Committee (APLIC) has been successful because it has been a broadly-based cooperative effort with a narrow and specific focus, nurtured through periods of limited funding by dedicated agency and utility personnel. Other important factors have been that APLIC has Ø emphasized achievable objectives (reduction but not total elimination of bird deaths), Ø developed information about options for dealing with bird collision issues, but left decisions as to how to apply this information to other groups, and Ø emphasized standardized study methodologies.
Bird/Wind Turbine Investigations in Europe
Studies of bird-wind plant interactions in Europe have dealt with collision frequency, disturbance/habitat loss effects, and avian flight behavior.
Bird collision studies in Europe have found that determination of search bias is essential. Even with careful searching, the proportion of the bird bodies found can be low, especially for small birds in high vegetation.
Estimated average mortality in two Dutch wind parks was 0.04-0.09 birds per turbine per day. These are very high figures relative to values reported for California. Most bird victims were found after nights with both poor flight conditions and visibility.
Mean mortality per kilometer of wind park is similar to that per km of highway, and comparable to or somewhat lower than that per km of power line in risky situations. Total numbers killed per 1,000 MW of wind power capacity are low relative to other human-related causes of death.
Why is the kill rate so much higher in the Netherlands than in the Altamont? The Altamont, with a high density of Golden Eagles and other raptors, is not comparable to areas of Europe and North America with fewer raptors but concentrated nocturnal migration. The types and numbers of birds that would collide with turbines in some parts of North America may differ widely from those documented in the Altamont.
Even with the higher collision rate in the Netherlands, collisions are infrequent events. They can be observed directly, but this is very labor intensive.
Habitat loss/disturbance effects were demonstrated in the Netherlands up to 250-500 m from the nearest turbines. Numbers there were reduced by up to 95%, depending on species, site, season, tide, and whether the wind park was in operation. In Europe, disturbance/habitat loss effects are thought to be much more important than direct collision mortality. However, Spain may be an exception.
Flight behavior near Dutch wind turbines differed between day and night. By day, most reactions of migrating birds to turbines were calm and gradual. Few birds needed more than one passing attempt before crossing the wind park. Reaction frequency was higher when turbines were 150 m apart and/or operating than when 300 m apart and/or inoperative. Few birds were seen within 20 m of a rotor during daylight.
Nocturnal migrants were more commonly seen within 20 m of the rotors, especially with headwinds. By day, 15% of the closely-approaching birds altered their flight paths calmly to avoid crossing the rotor, whilst at night 36% did so. The other birds all crossed or tried to cross between the rotor blades. While doing so, most of them either flapped their wings powerfully or fluttered.
Bird/Wind Turbine Investigations in Southern Spain
There is an ongoing study of bird mortality and behavior in the Tarifa area of southern Spain—one of the three main routes of concentrated migration of soaring raptors and storks en route between Europe and Africa. Also, large numbers of soaring birds sometimes land in the area while waiting for weather conditions good for crossing the sea. The international significance of the Tarifa area for migratory birds has been recognized in several ways.
Several wind parks have been established within the area traversed by soaring birds on migration. Most turbine strings are aligned roughly parallel to the main migration direction, but a few strings cross that axis.
Preliminary results show that, during the first few months of study, a number of birds were killed by collisions with turbine blades. These included 14 protected species. The majority were Griffon Vultures. This mortality has raised concern in Spain.
Designs for Avian-Wind Power Research
Standardized Assessment and Monitoring Protocols
Monitoring studies of birds in relation to wind power development can involve Ø preliminary site selection surveys; Ø pre-construction surveys at specific proposed sites; Ø post-construction monitoring; Ø development of mitigation methods; and Ø monitoring of decommissioned wind plants.
The objectives of pre- and post-construction monitoring are to gather data that can be used to assess the impact of wind plant development on avian populations. Pre-construction data are needed Ø to predict the impact of a proposed wind plant, and Ø as a baseline for measuring impact.
There is a need for agreed units of measurement and, where practical, standardized methods. Meeting participants agreed that data need to be comparable within and among studies, but had varying opinions about the importance and practicality of requiring the use of the same specific survey methods in different areas.
Preliminary site selection surveys should use existing information on bird species, densities, habitat dependencies, and flight patterns in potential project areas to help select the locations of potential wind turbine projects. If general information about birds and habitats is not available, some monitoring of bird populations and movements at potential wind turbine sites may be needed
Pre- and post-construction surveys should include surveys of bird populations, movements, and mortality. Ideally, surveys should be done on both the wind plant site and otherwise-similar comparison sites both before and after the wind plant is constructed. This BACI (Before-After-Control-Impact) design, provides both temporal and spatial control data. Both types of control data are needed to determine whether apparent differences are truly attributable to a wind plant.
Population surveys should be based on existing standard methods.
Bird movement studies should use standardized methodologies to document low-altitude movements, migratory and local, and to distinguish birds that are at risk because of their low flight altitudes from those flying higher and not at risk.
Searches for dead and injured birds should quantify mortality related to wind turbines, meteorological towers, and powerlines. Collision rates should be expressed in terms of the percentage of birds passing through the "envelope of risk". Methodological aspects requiring careful attention include area, timing and frequency of searches; data to be recorded; quantification of biases; provision for necropsies to determine cause of death; and methods for estimating total collisions and collision rates.
Attendees had varying views about the importance of mortality searches during the pre-construction phase, or in a nearby comparison area during the post-construction phase. In many areas, few dead or injured birds are found in the absence of turbines. However, some attendees felt strongly that it is essential to collect these data.
Conceptual Framework: Adaptive Resource Managementand Integration of Diverse Studies
The Adaptive Resource Management Approach: "Real world" management actions must often be taken in the absence of organized scientific knowledge. These actions often have some of the characteristics of a large-scale experiment. Adaptive Resource Management (ARM) is an iterative approach in which policies are treated as hypotheses and management actions as experiments. Based upon the outcome of such experiments, policies may be changed or refined and further tested by another stage of management action implemented as an experiment.
Many questions about bird-wind power interactions might best be addressed in an ARM context while wind energy was being produced, thus recouping some of the costs. The results could be used to refine the design of future expansions or new wind plants or, if necessary, to cancel a planned expansion or new development.
This approach would involve some costs and complications over and above the basic cost of establishing and operating a wind plant. However, the incremental costs are expected to be low when compared with the cost to obtain equally reliable scientific results in some other manner.
Integration of Diverse Studies: To determine the direct and indirect effects of wind development on birds, two main classes of studies are needed: (1) siting studies, and (2) existing site studies.
The main objective of siting studies is problem avoidance, through development of an ability to predict effects of future wind plants on birds, and avoidance of sites where serious bird-wind plant problems are predicted. A GIS (Geographic Information System) with good analytical abilities would provide a useful basis for integration and analysis of data.
Post-construction monitoring should be done to assess the accuracy of predictions. The results can then be used to refine future prediction abilities—a simple example of Adaptive Resource Management. A more elaborate variation would be to identify two similar sites, conduct pre-construction surveys on both, build a wind plant on one site, and then monitor birds on both sites (BACI design).
Existing site studies can be designed to look at direct effects on mortality or at indirect effects involving disturbance, habitat change, or change in food availability. Results should be used to predict the impact of proposed sites on birds. The accuracy of these predictions could be tested when new wind plants are constructed.
Components of wind plants responsible for direct bird mortality could also be addressed by a two-step process: (a) existing-site studies, followed by (b) tests at new or expanded wind plants. Step (b) is necessary to determine whether predictions are correct.
Generic Hypotheses to be Tested: Three generic null hypotheses amenable to testing in an ARM context during studies at existing and proposed wind plants were proposed. The specific wording would need to be adapted to the individual circumstances. The first two hypotheses apply to existing sites and the third to siting studies:
Ho: Population sizes, recruitment rates, survival rates, etc., are the same at sites with
Ho: Mortality rates are the same at sites with and without mitigative measures. Ho: Cumulative effects at new wind plants will not differ from those predicted.
and without wind plants.
Data on bird behavior near wind turbines can also be necessary. Specific questions about behavior near turbines should be identified at the start of research programs. Behavioral observations should be done as necessary to test specific predictions.
Experimental Design: The best design for environmental impact studies in the field is some variant of the BACI (Before-After-Control-Impact) design. This design has limitations, but is much better than designs lacking pre-construction and/or contemporaneous control data. The limitations of BACI designs are known and can be largely overcome.
Population Models, Uncertainty, and ARM: The value of population models and associated sensitivity analyses in focussing research on key data gaps was discussed. Opinions varied as to the value of this approach, given the imperfect knowledge about bird population dynamics.
Many attendees felt that gradual wind power development, in conjunction with Adaptive Resource Management (ARM) on a national scale, could resolve uncertainties, develop siting guidelines and mitigation measures where necessary, and do so without risk to bird populations. The ARM approach incorporates pre-planned alternatives and trigger points. Alternative actions that might be taken at a trigger point can include going ahead with development, adding mitigation measures, or ceasing development.
This raised several questions: Ø What is the minimum economical size of a wind plant? Ø Is this size consistent with avian research needs? Ø Would the avian mortality associated with this size of wind plant be tolerable? Ø If not, would it practical to remove the wind plant? Answers to these questions were beyond the scope of the meeting, but they relate to the scientific and economic practicality of the ARM approach and need to be addressed.
Defining an Integrated Plan for Avian-Wind Power Research
Principles for a National Avian-Wind Power Research Plan
Management Principles: An effective national plan should provide for Ø cost sharing by various stakeholders, Ø different combinations of stakeholders to support different projects, Ø adequate management support, Ø clearly stated program objectives, and Ø a system to measure the degree of success.
Establishing a Research Climate: This might best be done through adoption of a competitive, peer-reviewed approach for proposals; replication of key work by different groups; peer-review of results; and emphasis on hypothesis-based research.
Who Should Sponsor or Conduct Various Kinds of Research? Government or trade associations should sponsor and/or conduct research when single firms have no incentive to do so, e.g. when other firms can use the results without having paid for them.
Collaborative Research: When research is controversial, a collaborative approach toward sponsorship and management is more likely to produce results that will be accepted and used by all concerned.
Toward a National Research Agenda
Questions and guidance from this technical group may help define questions, e.g. on wind power economics, policy and legalities, that require attention by a group with broader representation and mandate.
There was general acceptance that a collaborative approach would be appropriate for management of a national research plan on avian-wind power issues, and that the National Wind Coordinating Committee's Avian Workgroup may, when formed, be appropriate for at least part of this role.
The overall goal might be to devise a process, incorporating scientific research as a major element, that would allow the wind industry to develop without being stopped by the occurrence of an unacceptable number of bird deaths.
It was agreed that, in general, scientific research is urgently needed to help resolve existing questions related to bird-wind power developments.
Many attendees supported the Adaptive Research Management (ARM) concept. The ARM approach, combined with planned wind power development, may provide a "golden opportunity" to conduct well-designed scientific experiments on high-priority research topics at the scale necessary to obtain meaningful results. However, questions were raised relating to the phasing of wind plant development, and to the economics and politics of terminating a development if bird problems prove to be severe and unmitigable. Nonetheless, attendees agreed that well-designed avian-wind power research should be done in conjunction with new wind plant developments.
The merits of conceptual modeling of bird population processes and potential wind plant effects require further discussion if a consensus is to be reached.
Some attendees felt that detailed studies of population dynamics could be important in specific circumstances, but that the first priority at most existing and proposed wind plants would be systematic monitoring of bird numbers, movements, and mortality. These attendees tended to emphasize the need for standardized survey methodologies.
Approaches (5)-(7) are not mutually exclusive. Research might incorporate conceptual models to help identify key data gaps, an ARM approach to the experimental testing of hypotheses, and use of standardized field methods.
Wind power impacts on birds in other parts of the U.S.A. can be expected to be very different from those in California, where almost all U.S. studies to date have been done. Thus, there is a need to investigate the extent and nature of "the bird problem" in different parts of the country.
It was agreed unanimously that, for the foreseeable future, there should be some form of pre-development bird survey at each proposed wind plant. Pre-development studies should be divided into (a) preliminary site-selection surveys, and (b) more detailed pre-construction surveys of specific sites as summarized on p. *.
When calculating survival and mortality rates, the choice of denominator is critical, and the most appropriate denominator depends on the question being asked. Deaths must be distinguished from dispersal, and must be further partitioned into those that are and are not attributable to the wind plant. The most appropriate measures of mortality for present purposes deserve further consideration.
During a further discussion about the types of research needed, some attendees suggested that three main research questions need to be addressed: Ø Does the avian mortality caused by wind plants constitute a problem, from a population, ecological, or other perspective? Ø If mortality is a problem, how can the impact be reduced? Ø If mortality is a problem, how can future wind plants be sited to reduce mortality?
Other attendees felt that these three research categories are undesirably broad or incomplete, and preferred the previously-developed 7 or 5 point categorization schemes (p. *ff). It was agreed that, taken together, the closely-related lists of 7, 5 or 3 research topics provide a good sense of the types of research needed.
Attendees did not assign priorities to the categories of research identified on those "short-lists" of 3-7 general research topics. All categories on these lists are important.
Many attendees submitted specific research suggestions (Appendix 3, p. *ff). There was much interest in this list. However, in the absence of a more general research framework, the majority of attendees preferred to work toward consensus on that framework rather than discuss the specific research projects listed in Appendix 3.
The following key elements should be part of any scientific research on avian-wind power issues: Ø competitive proposals, Ø peer review of proposals and results, Ø clearly stated objective(s), Ø statement of hypotheses, Ø experimental design appropriate to test the hypotheses, Ø measurement protocols that will produce consistent data, and Ø specified measures of success.
There were different opinions about the importance of standardized methods (p. *). At the least, consistent and systematic methods should be used within a given study area, and the results must be comparable among regions whether or not the methods used are identical. Many attendees felt that standardized methods are very desirable and would be welcomed by many researchers.
Next Steps to be Taken
More specific recommendations for baseline studies should be formulated, starting from concepts discussed at this meeting and work being done under contract to EPRI by S.A. Gauthreaux. A process for this effort is to be worked out, coordinated by Resolve Inc.
A few participants in this meeting are expected to be members of the National Wind Coordinating Committee's soon-to-be-formed Avian Workgroup. Some other participants in this meeting could become a technical advisory group to the Avian Workgroup. In this way the activities of present group and of the NWCC Avian Workgroup could be merged onto a single track.
- Other topics that might be taken up by such a technical group for further consideration and consensus building include Ø mitigation/enhancement opportunities, Ø population level effects, modeling, etc., Ø experimental design approaches, and Ø peer review of proposals and results.
