Read an Excerpt
CHAPTER 1
Monitoring birds
K. SHAWN SMALLWOOD
In memory of Robert Anderson: a champion of the Golden Eagle.
Summary
Studies are usually performed before the development of a wind farm to determine whether the study area, or specific locations within the study area, will pose an unreasonable avian collision risk. It has long been suspected that the terrain presents an uneven collision risk by channelling bird activity, and preconstruction surveys can inform sound decisions over wind turbine layout. Post-construction surveys are also needed to measure displacement and barrier effects, improve understanding of wind turbine collision mechanisms and test the efficacy of mitigation measures. Visual scans have been favoured for obtaining use rates to be compared to use and fatality rates elsewhere and to predict fatality rates at a proposed wind farm. This chapter reviews those survey methods for wind farm assessment. It then examines use rates in North America for sources of bias and uncertainty and whether use rates are predictive. Use rates are compared from 82 wind farms, including 43 based on pre-construction surveys, 32 with post-construction surveys (21 from the Altamont Pass Wind Resource Area in California, USA), 7 with both pre- and post-construction surveys and 54 with accompanying fatality rate estimates. Potential biases are serious and sources of uncertainty are many, meaning that use rates are often poor predictors of fatality rates. The assumptions used to justify survey methods need testing and efforts to standardise methods need more effective direction. Use surveys can be improved and telemetry expanded, and radar, thermal imaging and behaviour surveys should be developed for predicting and minimising avian collision impacts at wind farms.
Introduction
Environmental studies, typically referred to as 'baseline studies' in North America, often precede construction of new wind farms to assess potential impacts on birds, bats and other wildlife. This process is analogous to Environmental Impact Assessment (EIA) adopted in Europe. Wind farms can adversely affect birds in several ways: (1) modifying or destroying habitat by the construction of wind turbine pads and access roads; (2) displacing some birds that avoid the new facilities (Leddy et al. 1999; Whitfield & Madders 2006; Pearce-Higgins et al. 2009; Garvin et al. 2011; Langston 2013); and (3) injuring and killing birds when they collide with wind turbines (Smallwood & Thelander 2004; 2005; de Lucas et al. 2007; 2008; 2012a; Smallwood 2007; 2013; Dahl et al. 2012). Less obvious impacts include collision of birds with turbine towers, birds being struck by automobiles on the access roads, electrocution of birds along electric circuit lines, on transformers or other electrical infrastructure, oiling of feathers of birds entering the nacelles, and entrapment of birds within nacelle and tower spaces. Despite the variety of wind farm impacts, most of the focus of pre-construction environmental assessments has been on the risks of collision with wind turbines.
Baseline studies typically include on-site surveys to assess utilisation as an indicator of relative abundance. Utilisation rates, also known as use rates, and usually expressed by the metric 'birds seen per unit time', are typically assessed by comparing them to use rates at other wind farms where impacts had also been estimated following post-construction fatality surveys. Surveys to estimate use rates are typically intended to predict impacts so that an informed decision can be made whether the wind farm should be developed. According to the objectives typically appearing in baseline study reports, a long-established secondary objective has often been to guide wind turbine siting, also known as 'micro-siting', to minimise collisions with flying birds [Morrison 1998; Anderson et al. 1999; California Energy Commission California Department of Fish and Game (CEC CDFG) 2007; see Chapters 16 and 17 in this volume].
An important element of baseline studies and EIAs is learning from the impacts already experienced at existing wind farms (New et al. 2015). Studying causal factors of fatalities at existing wind farms can help to predict impacts at new wind farms. It is important to learn how eagles and other birds react to wind turbines (Osborn et al. 1998; Leddy et al. 1999; Hoover & Morrison 2005; Smallwood et al. 2009a; 2009c; May et al. 2010; Dahl et al. 2013; Hull & Muir 2013; Kitano & Shiraki 2013). Also needed are more accurate fatality rate estimates for comparison with predictor variables such as use rates, so that investigators can develop predictive tools and formulate mitigation strategies.
The prediction of wind turbine impacts benefits from knowing the sensitivity of local species to wind energy development, where sensitivity links to the conservation importance or relative rarity of the species (Percival 2007; Desholm 2009). It also benefits from understanding the collision susceptibility of bird species due to flight behaviours, relative abundance and reactions to wind turbines. It benefits further from understanding vulnerability to collisions posed by the planned arrangement and design of wind turbines, including the wind farm's size (rated capacity), spatial extent, tower heights, rotor diameters, rotational speed, inter-turbine spacing and locations of turbines on the landscape. That is, susceptibility is a species' predisposition to being harmed by wind turbines due to morphology, ecology and environmental perception, whereas vulnerability is the likelihood of individuals being harmed once wind turbines are installed (Smallwood & Thelander 2004; 2005).
Understanding susceptibility and vulnerability has been hampered by five major problems. First, the distinction between susceptibility and vulnerability can be muddled by changes in behaviour or relative abundance caused by wind turbine installations. Patterns of behaviour and abundance seen before construction may not resemble the patterns seen after construction, or they may differ in small but significant ways. Experimental designs and more thorough investigations are needed to discern how and to what degree susceptibility translates to vulnerability following a wind farm development (Dahl et al. 2012). Secondly, comparing observational studies among wind farms by opportunity rather than by sampling design can lead to pseudoreplication (Huso & Dalthorp 2014). Thirdly, European studies found poor prediction of fatality rates based on pre-construction use rates (de Lucas et al. 2008; Ferrer et al. 2012), but no serious verification has been attempted elsewhere. Fourthly, the degree to which relative abundance and behaviours vary within and between species is often unknown and can confound comparisons of fatality rates. Relative abundance measured at a certain place and time may differ from the relative abundance measured after wind farm construction for reasons independent of the site's construction. By not surveying far enough beyond the boundaries of wind farms, investigators often miss measures of aggregation within wind farms that would be more predictive of collision impacts (Carrete et al. 2012). Behaviours can also vary temporally and spatially in response to wind, topography, food and social conditions. Finally, high uncertainty and large biases are likely in the methods used to estimate use rates. Whereas uncertainty and biases in fatality rate estimates have been debated, little debate has been directed towards sources and magnitudes of error and bias in use rates other than potential biases pointed out by Madders and Whitfield (2006).
The development of North American baseline study methods focused on standardisation and comparability of metrics (Gauthreaux 1995). Many guidelines documents have been prepared, usually emphasising methodological standardisation (Table 1.1). Despite the emphasis on standardisation, guidelines have varied considerably in their goals and objectives and level of detail (Table 1.1). Guidelines in Japan, England and California also collectively recommended recording certain details associated with use surveys, including the observer's name, station number, survey date, start time, temperature, wind speed (average, maximum), wind direction, weather (cloud cover, precipitation), visibility, time of each bird observation, species, number of birds composing an observation, social context (single, pair, flock), behaviour (perching, nesting, flying, type of flight), mapped location or distance from the observer, slope aspect or habitat, flight direction and distance of the bird from the nearest wind turbine (if post-construction). Guidance conspicuously absent from the guidance documents reviewed in Table 1.1 includes the assessment of barrier effects or avian energetic costs associated with birds having to fly around wind turbines or wind farms (Drewitt & Langston 2006), and the use of peer review in formulating survey plans and in reporting results.
It is also immediately clear from the guidance documents that a range of methods may be recommended in particular circumstances and according to survey goals and objectives. The metrics potentially derived from the different methods include: (1) use rates, including use rates within heights above ground equal to the anticipated low and high reaches of the turbine blades, for use in a collision risk index and modelling collision risk (New et al. 2015); (2) passage rates, which are at least theoretically the basis of collision risk models based on avoidance rates (Band et al. 2007; Smales et al. 2013; see Chapter 13 in this volume); (3) behaviour rates, such as hovering time or aggressive encounters per hour, which are used for spatially explicit collision hazard models (see Chapter 17 in this volume); and (4) nesting densities, which are also used for spatially explicit collision hazard models (Smallwood et al. 2009a).
The same survey methods can also be used during post-construction monitoring to estimate barrier effects and displacement (Garvin et al. 2011; Loesch et al. 2012), to test mitigation measures or to compare use rates and fatality rates directly for sensitive species. In at least one case, post-construction monitoring also included nocturnal surveys to quantify behaviour rates, passage rates and avoidance rates of owls and nocturnal migrants (Smallwood, unpublished data 2015). Post-construction monitoring was used by de Lucas et al. (2012a) to shut down select turbines as soaring Griffon Vultures Gyps fulvus approached during the migratory months of October and November, hence reducing Griffon Vulture fatalities by 50% with negligible impacts on wind farm energy generation.
Avoidance rates for use in collision risk models can also be estimated from post-construction behaviour surveys (May et al. 2010), but obviously cannot be estimated during pre-construction surveys unless followed up by post-construction surveys to measure differences in passage rates through the wind farm or wind turbine rotors (e.g. Hull & Muir 2013; Johnston et al. 2014). For avoidance rates to be effective, however, explicit definitions are needed, such as whether a bird's avoidance action was directed towards the entire wind farm, individual wind turbines, rotors or blades; or, another way of looking at it, whether the bird's avoidance was measured while within the bounds of a wind farm, within a certain distance of a wind turbine, within the rotor plane of a turbine, or within a certain distance of a blade. Hull and Muir (2013) developed an explicit definition for avoidance at the wind turbine level. According to Cook et al. (2014), micro-avoidance is a last-second action taken within 10 m of a rotor to avoid collision, meso-avoidance is any behavioural response to individual turbines from the tower base to the outer reaches of the rotor, and macro-avoidance is any behavioural response to the presence of the wind farm, measured from the outermost turbines. Cook et al. (2014) assigned distance thresholds to meso- and macro-avoidance, but the distance could change with species and with the size of wind turbines or wind farms. In addition to the need for more explicit definitions, Chamberlain et al. (2006) pointed out that avoidance behaviours may be less frequent during conditions when visual scans are not undertaken, such as during cold weather, fog, rain or darkness. Upon reviewing collision risk models, Masden and Cook (2016) advocated testing assumptions underlying the models. Owing to variation in the meaning of avoidance and the need for testing assumptions related to how birds behave while flying near wind turbines, it may be prudent to regard most avoidance rates as overoptimistic until evidence suggests otherwise.
Scope
The first objective of this chapter is to present methods used to survey wind farms for birds, particularly in relation to their susceptibility to collision impacts. Avian survey methods assessed herein were identified from accumulated worldwide literature on wind farm assessment and monitoring, but this is not exhaustive. The second objective is to assess the value of each of these methods for their comparability of rates and predictability of impacts according to the principles of wind farm assessment. The third objective is to explore sources of bias and uncertainty in estimating use rates from use surveys, which are most often used to predict impacts and sometimes subsequently applied in post-construction monitoring to help measure or explain impacts, including fatalities and displacement.
For the last objective, all the publicly available monitoring data used were collected in North America, where standardisation of methods has been strongly advocated to achieve comparability of use rates and fatality rates (Gauthreaux 1995; Anderson et al. 1999); and the research and monitoring data used were collected in the Altamont Pass Wind Resource Area (APWRA). No attempt is made to identify sources of uncertainty and bias in fatality rate estimates because this was the target of previous publications (Smallwood 2007; 2013).
Themes
Survey methods
Survey methods used in wind farm assessment have included point counts or visual scans conducted both by day and increasingly at night, behaviour surveys, radar, telemetry, nest surveys and transect surveys. The basic methodology and a brief discussion of strengths and weaknesses are provided in the sections below. Every survey method used for pre-construction site assessment or post-construction impact assessment should be examined closely for the value of the data in meeting objectives or testing hypotheses outlined in Table 1.2. Wind turbines potentially affect many bird species in multiple ways and these species vary greatly in daily and seasonal activity periods, and in numbers, behaviours and detectability. As a result, no single method is available for assessing wind energy impacts on all bird species.
Diurnal use surveys
Circular (360-degree) visual scans, also known as variable distance circular point observations (Reynolds et al. 1980), are usually performed by observers stationed at designated locations in a proposed or existing wind farm, consistent with scenario A in Figure 1.1 (Osborn et al. 1998; Rugge 2001; Lekuona & Ursúa 2007; Smallwood et al. 2009b). Each station, vantage point (VP) or observation point (OP) is typically selected for maximum vantage over long distances. The maximum survey radius and session duration are established before the surveys begin, and the observers follow a survey schedule that varies according to wind farm size and available budgets. Detected birds are identified to species level, using binoculars if necessary, and counted, and the counts are divided by the session duration to arrive at use rates. By changing the location and restricting the area surveyed, diurnal use surveys can be easily adapted to meet the requirements of scenario B or C in Figure 1.1.
Nocturnal use surveys
Thermal and infrared imaging is available for nocturnal surveys of owls, nocturnal migrants and nocturnal activities of particular species that may be adversely affected by wind turbines, including energetics related to foraging. Thermal imaging was first used in wind farms to quantify behavioural responses of bats to wind turbines (Horn et al. 2008; Cryan et al. 2014), but has also been used to calculate passage rates of bats, foraging owls and nocturnal migrants through or near the rotors of wind turbines (Smallwood, unpublished data). Thermal cameras mounted on tripods can be panned for 360-degree scans or fixed on particular wind turbine rotors during timed sessions, as required for scenario C in Figure 1.1.
(Continues…)
Excerpted from "Wildlife and Wind Farms, Conflicts and Solutions Volume 2"
by .
Copyright © 2017 Pelagic Publishing.
Excerpted by permission of Pelagic Publishing.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
