You are on page 1of 14

Impact Assessment and Project Appraisal

ISSN: 1461-5517 (Print) 1471-5465 (Online) Journal homepage: http://www.tandfonline.com/loi/tiap20

A case study on visual impact assessment for wind


energy development

Robert C. Corry

To cite this article: Robert C. Corry (2011) A case study on visual impact assessment for wind
energy development, Impact Assessment and Project Appraisal, 29:4, 303-315

To link to this article: https://doi.org/10.3152/146155111X12959673796047

Published online: 20 Feb 2012.

Submit your article to this journal

Article views: 861

View related articles

Citing articles: 1 View citing articles

Full Terms & Conditions of access and use can be found at


http://www.tandfonline.com/action/journalInformation?journalCode=tiap20
Impact Assessment and Project Appraisal, 29(4), December 2011, pages 303–315
DOI: 10.3152/146155111X12959673796047; http://www.ingentaconnect.com/content/beech/iapa

A case study on visual impact assessment


for wind energy development

Robert C Corry

Wind energy development is expanding. Preparations for wind farms include environmental impact
assessments, specifically with visual impact assessments (VIAs). Often using simulated photographs,
VIAs depict the post-development landscape appearance with new wind energy structures. This
Ontario case study compares simulated photographs submitted as part of a VIA with post-development
conditions. Simulations were judged for accuracy in turbine number, height, diameter, and location,
and adequacy in representation of built conditions. Simulations were partially representative of post-
development conditions, yet commonly under-represented turbine number and size in different
locations than they were built. Simulation frames were too narrow to adequately represent human
vision and simulations under-estimated how many wind turbines were visible from a single landscape
position. Outcomes suggest that visual simulations can lack accuracy or representativeness, misleading
the public. Adopting province-wide VIA criteria and increasing adherence to visual simulation
guidance may be necessary to improve the accuracy and representativeness of VIA content.

Keywords: visual simulation, environmental screening, wind turbines, transmission lines, green
energy, Ontario, Canada

R ENEWABLE ENERGY DEVELOPMENT


IS INCREASING, driven by rising energy
costs, concerns about greenhouse gas emis-
sions, government support (e.g., Ontario, Canada’s
Green Energy and Green Economy Act; Bill 150,
established electricity grids to locations of greatest
demand (Lewis, 2008).
The changing appearance of the landscape is of
concern to stakeholders who seek to anticipate the
direction and magnitude of landscape change (Klick
Statute of Ontario, 2009 C.12), and conventional en- and Smith, 2010). Visual impact assessment is usu-
ergy exploration consequences like the Deepwater ally (not always) a requirement for permitting, and
Horizon drilling accident and subsequent oil leak in- documentation is submitted as part of development
to the Gulf of Mexico, USA (April 2010). Land- applications (Gipe, 2002; County of Bruce Planning
based renewable energy development (compared to & Economic Development Department, 2005;
off-shore) has commonly generated controversy Taylor et al, 2006). The quality of visual assess-
around the changing appearance of the landscape ments, therefore, should be strictly maintained in or-
(Krohn and Damborg, 1999) and aesthetic responses der that the pre-development predictions reliably,
remain important to the public (Klick and Smith, accurately, and reasonably completely depict the
2010). Yet land-based wind energy production has nature and extent of landscape change resulting from
advantages for transmission of electricity through renewable energy development (Smardon and Karp,
1993; Sheppard, 1989).
Guidance documents for wind energy visual as-
sessment vary in terms of expectations and regulato-
Robert C Corry is Associate Professor, Landscape Architecture, ry approach (Pasqualetti et al, 2002; Horner +
at the University of Guelph, Guelph, Ontario, Canada N1G
2W1; Email: rcorry@uoguelph.ca; Tel: +1 519 824 4120 ext. Maclennan and Envision, 2005; University of
58034. Newcastle, 2002). Photographic simulations are the
For acknowledgements see page 314. most common way of illustrating likely landscape

Impact Assessment and Project Appraisal December 2011 1461-5517/11/040303-13 US$12.00  IAIA 2011 303
A case study on visual impact assessment for wind energy development

change — more common than models and three- Landscape Institute: Institute of Environmental
dimensional virtual depictions. Suggestions for how Management and Assessment, 2002; Committee on
photographic simulations are constructed, presented, Environmental Impacts of Wind Energy Projects —
and viewed are intended to help standardize ap- National Research Council, 2007). Under recent sus-
proaches and yield highly representative simulations tainable energy legislation in Ontario (Bill 150,
(Horner + Maclennan and Envision, 2005; 2009) sustainable energy developers (e.g., solar,
University of Newcastle, 2002; Sheppard, 1989), yet wind, biomass) are not required to produce visual
in many cases the simulations submitted for permit- impact assessments, and no guidance is given. Court
ting and development approval are not methodically challenges to the ‘Green Energy Act’ (Bill 150,
evaluated against the post-development reality (see 2009) are demonstrating that concern remains over
Wood, 2000, as a rare exception). visual and other impacts of large wind energy devel-
Principles for visual simulation preparation and opment in rural Ontario so the statute may be subject
presentation include the representativeness and accu- to revision.
racy of the visual simulation (Sheppard, 1989). This paper is a case study of the visual impact
Sheppard (1989) defines the representativeness prin- analysis contained in preparatory materials for a sin-
ciple as, ‘a set of simulations is representative if it gle development of 45 wind turbines (2004–2006) as
shows important views of the project and shows the compared to the post-development conditions of the
project in typical views and conditions’ (p. 65). The wind energy landscape. The development occurred
principle of accuracy (Sheppard, 1989) is sometimes in two phases and this study includes only the first: a
considered the truthfulness or honesty of a simulated second phase of 85 additional turbines used modi-
view and is stated as, ‘a fully accurate simulation fied visual simulations that are included here for
shows a view of the project that is not significantly illustration purposes only. The setting is open
different in appearance from the real view when seen agricultural land with extensive farms, dispersed
from the same viewpoint’ (p. 76). Accuracy is com- farmsteads, and rural housing. The land is substan-
plicated by the nature of the subject matter shown tially flat and visibility of tall structures (e.g., tele-
(objects or elements of landscape change) and the communications towers) can be several kilometres
elements within the picture image (i.e., the visual in extent.
setting). Sheppard notes that, ‘A simulation may For comparison purposes environmental screening
omit parts of the project or scene that would, in real reports for other wind farm developments in Ontario
life, be seen’ or that occluding or mitigating objects were reviewed (e.g., Manitoulin Island, Greenwich
such as trees may be simulated but not actually Lake, Coldwell, Grand Valley, Harrow). There is no
present (p. 76). standard accepted methodology for preparing visual
This paper reports on a single case study where assessments, not all reports included photographic
simulations (in the ‘environmental screening report’ material, and details for photographic locations and
part of a visual impact assessment) have been com- camera focal length were highly variable. For the
pared to the post-development landscape and as- Manitoulin Island project the consultant reported
sessed qualitatively and quantitatively for accuracy that, ‘there are no regulations or policies regarding
and representativeness of the nature and extent of the visual impacts of wind turbines in Ontario’ and
landscape change. Note that this study does not iden- provided no images of visual impact (Dillon
tify the location, municipal authority, developer, or Consulting Limited, 2009: 124). The case presented
sub-contractors to the developer: the intent is limited here is similar to others in Ontario.
to appraising how an example visual impact assess- This case study is divided into descriptions of the
ment compares to the post-development conditions accuracy of components of the visual impact analy-
and commenting on the accuracy and representa- sis and representativeness of simulations compared
tiveness by comparing simulations to the built pro- to the post-development condition and landscape
ject as a way to inform future visual assessments and experience. Assessments of accuracy are provided
preparation of guiding documents. for the location, height and diameter, number, and
ancillary infrastructure associated with wind energy
turbines. Graphic illustrations are used to demon-
Ontario case study strate the accuracy of the visual impact assessments.

Renewable energy development in Ontario is grow-


ing rapidly (about 200 megawatts per year since Methods
2003) (Ontario Ministry of Energy and Infrastructure,
2010, mostly in the higher electricity demand areas of Procurement of visual simulations
southern Ontario. Wind energy development for a
rural southern Ontario municipality was preceded by Visual simulations were prepared by the developer’s
preparatory materials including environmental as- sub-consultant, a multi-disciplinary practice includ-
sessments submitted by the commercial energy de- ing architecture, landscape architecture, and engi-
veloper. These assessments included analysis of the neering, and the best-available representations were
likely visual impacts of the development (The used for comparison purposes. The administrative

304 Impact Assessment and Project Appraisal December 2011


A case study on visual impact assessment for wind energy development

office of the municipality was visited to view the difference in vertical or horizontal scale. All as-built
environmental screening report (ESR) — a printed photos were clipped to the width of the simulations
document submitted by the developer — and the de- based on horizon-line objects (keeping photo extents
veloper’s world-wide web site was scoured for doc- consistent with the simulations).
umentation, preferably in digital format. While the Several comparisons were made using digital
ESR for the second phase of development was avail- photo overlays and photographic assessment for
able on-line (it has since been removed), the ESR each vantage point simulation and the as-built photo-
for the first 45 turbines was not available on-line. graphs. The number of wind turbines in the field-of-
A municipal officer was contacted to see if an elec- view were counted for both simulations and as-built
tronic version had been submitted and was available, conditions. The height and rotor diameter of wind
but only the print copy was available. The best- turbines were compared between simulations and
quality printed simulations were scanned directly built conditions. The location of simulated and built
from the ESR (Phase One) that was provided to wind turbines were compared (photographically and
municipality. using construction document co-ordinates). Finally,
In the ESR the developer submitted simulated the ancillary structures associated with wind energy,
photographs from six ‘vantage points’, all from pub- as represented in as-built conditions photographed at
licly accessible roadsides. Hard-copies of the simu- the vantage points, were assessed.
lated photographs were scanned at high resolution
(600 dots-per-inch) in a flatbed scanner and convert- Assessment of simulation accuracy
ed to digital files (the basis for Figure 1). Simula-
tions were colour-adjusted only for increasing clarity The first objective was to learn if the simulations ac-
of the scanned images (permissible legibility curately portrayed the as-built conditions. Accuracy
improvements in best practice documents, see is assessed based on comparisons of as-built condi-
Horner + Maclennan and Envision, 2005), and no tions of turbines in the landscape compared to visual
other characteristic or attribute of the simulations simulations provided by the project development
was changed. Given the nature of print media and team. Here ‘accuracy’ refers to a near-complete
scanning of print documentation it is likely that the overlap of position between proposed locations
simulations have been slightly distorted in the con- (from simulations and planned co-ordinates) and as-
version of hard-copy to digital file. The computer built locations (from field photographs and built co-
screen image of the scanned photographs was visually ordinates), sensu Sheppard’s (1989) definition (see
compared to the original paper copy to qualitatively the start of the paper). Accuracy is quantified where
assess distortion; any change appears inconsequential possible, and qualitatively assessed where photo-
to the attributes assessed in this report. based measurement (like depth and position) is
complicated. Accuracy is judged on attributes of po-
Overlay of as-built conditions sition, height, and diameter, and number of wind
turbines in simulations and as-built conditions.
The rural municipality was visited and each vantage
point used to create simulations was identified Number of wind turbines For the number of wind
(using maps from the developer’s ESR). At each turbines, in-field counts were used and compared
vantage point several high-resolution digital photo- with counts from as-built photographs. At each
graphs were taken of the as-built conditions. Photos vantage point the numbers of turbines visible were
were taken with a digital camera (23 May 2008) counted and recorded as visible at the nacelle, to the
with a 35mm lens to approximate the date, frame blades at the height of rotation, and as a sum of both.
width, and vantage point position and direction of This count focused on the frame limits of a digital
the simulations. Photos were opened in software, re- simulation. In digital overlays the simulations and
viewed, and compared to simulations for the degree as-built photos were reviewed at the same pixel
of similarity of location and focal point. Best matches resolution and counts of wind turbines in each layer
were chosen for further comparison. were made.
Digital overlays were created by making the simu-
lations a background layer and putting the as-built Height and diameter of wind turbines The height
photos on top of simulations. Horizontal and vertical and diameter of wind turbines in the photos were
alignment was corrected with software so that horizon complicated by distance from viewer (the problem
lines and features such as buildings and vegetation of a shrinking perspective), camera focal length (not
were matched. In some cases foreground features reported and unknown for simulations), and changes
like fencelines, powerlines, or road verges were not in rotor angle to viewer between simulations and as-
identically matched (and not manipulated to match built photos. Quantifying height and diameter was
in order to maintain scale and proportion — and difficult because comparing a wind turbine simulat-
avoid stretching — of the digital photographs). To ed for a particular location yet built at a different
maintain consistency with the simulations, middle- location made such comparison questionable. Quali-
ground and background elements (usually trees) tative assessment and visualization through photo
were matched for location and height – ensuring no comparison better represented height/diameter fit.

Impact Assessment and Project Appraisal December 2011 305


A case study on visual impact assessment for wind energy development

Location of wind turbines The location of wind In panoramic photos the vantage point simulation
turbines changed from proposal stage to built condi- frame extent was gauged by comparing the simula-
tions. The locations of turbines in the proposal were tion with the panorama and matching horizon-line
transcribed from the ESR and built location co- points or objects (a box was illustrated on the photo to
ordinates were provided by the wind energy devel- show the simulation frame). The number of turbines
oper to the municipal office. Changes between pro- both within this frame and outside it were counted.
posed and as-built locations were precisely To assess the adequacy of visual simulations, typ-
quantified as measured shifts in east–west or north– ical human field-of-vision is compared to the view
south directions. Challenges in quantifying location frame of visual simulations to gauge how well the
in photos are similar to height and diameter meas- simulations match human experience in a static
urement. While locations obviously differ in simula- landscape view. If human field-of-vision is more ex-
tions and as-built photographs, it is difficult to pansive than the view frame of visual simulations,
discern which (or if a) built turbine was shown in a panoramic as-built photos are used to count the
simulation photo (i.e., the simulation intent; the ESR number of turbines within the width of human field-
did not specify if simulated turbines were located at of-vision. No assessment of height or locational ac-
proposed co-ordinates). It would be questionable to curacy is made using the merged panorama photos,
try to quantify measured changes in location for a but numbers of turbines partially or fully visible can
turbine shown in both photographs. Like for height be counted from as-built high-resolution panoramic
and diameter, photo comparison was chosen to rep- photos.
resent the fit between wind turbines as-built with
those simulated at vantage points.
Results
Ancillary attributes As-built photos showed some
changes in attributes other than wind turbines. In Accuracy: wind turbine number
photographic overlays the images are compared for
ancillary attributes like access roads, powerlines, or The numbers of wind turbines are shown in Table 1.
power substations. The comparison is visual only, Each vantage point is assessed for numbers of tur-
with ancillary attributes highlighted for selected bines visible to different heights (at the nacelle or
vantage points. generator housing, or to the tips of the blades) in the

Assessment of simulation representativeness Table 1. Accuracy of wind turbine numbers through


comparisons of visual simulations and as-built
Representativeness of simulations was assessed conditions
by comparing as-built conditions from the same
viewpoints as the vantage point locations using Basis of Visible turbines Totals
on-ground experience and photography with litera- comparison
ture on dimensions of human field-of-vision. The in- To nacelle To tips of
tent was to appraise how well the simulations height blade
represent the on-ground experience of the turbine Vantage point 1
development. The methodological approach follows Simulation 1 1 2
As-built 0 1 1
that of Sheppard (1989) for post-construction ap- Accuracy +1 = +1
praisal of simulations with a notable difference (a Vantage point 2
wider field-of-view; documented in the section on Simulation 5 0 5
‘Representativeness of as-built conditions’ below). As-built 13 3 16
Accuracy -8 -3 -11
As-built photography from the six vantage points
demonstrated that more turbines are visible than Vantage point 3
Simulation 7 0 7
were shown in a narrow frame of a simulation. To As-built 11 1 12
assess the adequacy of simulations as representa- Accuracy -4 -1 -5
tions of as-built conditions two vantage points were Vantage point 4
arbitrarily selected for further analysis. On 26 July Simulation 9 0 9
2008, additional as-built photos were taken with a As-built 10 1 11
Accuracy -1 -1 -2
digital camera for vantage points 3 and 4. The photo-
Vantage point 5
graphs were merged using software to create a 360- Simulation 1 0 1
degree visual field panoramic photo, with the As-built 1 0 1
horizon line approximating the mid-point of the ver- Accuracy = = =
tical view field. Merging photographs yields a con- Vantage point 6
cave horizon line, so the merged panoramic photo Simulation 1 2 3
As-built 2 1 3
was adjusted in software to make the horizon line Accuracy -1 +1 =
approximately level. The adjustments alter the verti-
Cumulative -13 -4 -17
cal profile of wind turbines but have no effect on the summary
number of turbines.

306 Impact Assessment and Project Appraisal December 2011


A case study on visual impact assessment for wind energy development

simulation frame view. Differences in turbine num- Table 2. Differences in proposed and built locations for wind
energy turbines. Easting and northing values
ber from simulation to as-built condition is derived (metres, from Universal Transverse Mercator co-
as a plus/minus ratio (showing either an over- or an ordinate projection system) were used to calculate
under-estimate of visible turbines, respectively). change. Change is shown as absolute values
The vantage point #1 simulation over-estimates
number of turbines. For vantage points #5 and 6 the Turbine no. Change easting Change northing
simulations approximately matched the as-built
numbers of turbines. Under-estimates characterize 1 11 18
comparisons for vantage points #2, 3, and 4 (see 2 10 16
3 9 14
Table 1). 4 7 13
5 0 0
Accuracy: location of wind turbines 6 0 0
7 6 31
8 29 24
Location accuracy is quantified from the proposed 9 2 1
and as-built co-ordinates for turbines. Table 2 shows 10 124 39
the arithmetic differences of proposed and built tur- 11 2 5
12 53 60
bine locations. Change is quantified in east–west and 13 1 0
north–south directions using absolute values (that is, 14 33 40
the direction of either east or west is not considered 15 0 1
16 2 0
critical, but rather how many metres of change 17 0 0
occurred). 18 0 0
The proposed location co-ordinates (metres, in 19 1 1
20 0 2
Universal Transverse Mercator projection) were re- 21 1 1
ported as integers, while built co-ordinates were 22 1 2
reported to three decimal places. The minimum 23 0 0
change reported by the developer from proposal to 24 0 0
25 0 0
built conditions is 0.01 and 0.03 metres in north– 26 39 32
south and east–west directions, respectively, and 27 17 16
within reason should be considered to be zero. The 28 28 15
29 5 5
maximum change is 123.62 and 76.16 metres (N–S 30 24 76
and E–W), with an average change of 13.32 and 31 6 10
14.00 metres. Because the proposed locations were 32 1 1
33 3 1
integer values, all change values are rounded to the 34 0 0
nearest metre. 35 0 0
Photographic comparisons appear to show more 36 Not built Not built
37 0 0
change in location than the numeric values in Table 38 Not built Not built
2 imply. Photographic comparisons are shown in 39 Not built Not built
Figure 1. These photos are used to qualitatively 40 Not built Not built
compare the location, but begin to hint at differences 41 Not built Not built
42 Not built Not built
in height and rotor diameter dimensions. A notable 43 1 0
challenge in the comparisons (Figure 1) is that it is 44 3 3
difficult to know which built turbines have moved 45 62 51
46 23 18
from planned locations. Horizontal lead-lines are 47 28 34
used to give the best approximation of ‘matched’ 48 21 47
turbines between simulations and as-built photos. 49 32 39
50 2 1
This may introduce some error as turbines that have
changed location were generally assumed to have Mean change (13.321, 13.996)
done so approximately laterally to the viewpoint ax- Maximum change (123.622, 76.162)
is; that is, that turbines moved visibly left or right in Minimum change (0.014, 0.027)
the frame, not strictly closer to or farther from the
viewer in a straight line. This assumption allowed
for more direct comparisons of heights and rotor dia-
meters, is plausible for the vantage point photos, and change, i.e., built precisely where planned), the sim-
is not challenged by the simulations. ulations appear to show only one turbine that is built
Photograph overlay figures should be compared where planned (Figure 1A). Using the turbine towers
with the tables of turbine location differences (Table as relative measures (they are 80 metres tall), many
2). It is understandable that location adjustments of the turbine locations appear to have changed by
would be required in the field during construction as more than the average change distance of 14 metres.
indicated by co-ordinates (Table 2). While the The turbines in Figure 1E and 1I appear to be more
changes in location of turbines in tables is common- than the maximum change distance listed in Table 2
ly quite small (many distance values are zero metres as 124 metres (approximately 11/2 tower heights). In

Impact Assessment and Project Appraisal December 2011 307


A case study on visual impact assessment for wind energy development

fact it is difficult to identify an alternative built tur- turbines turned in different directions in photo-
bine in the field that helps to explain the discrepancy, graphs, yet blade length accuracy provides a good
suggesting that either the turbines shown in simu- surrogate. Vantage point #6 simulations both over-
lations do not correspond with planned locations or and under-estimated turbine height for a net neutral
that as-built location co-ordinates are incorrect. effect (Figure 1L). Vantage points #2, 4, and 5
consistently show simulated turbines slightly shorter
Accuracy: height and diameter of wind turbines than in the as-built conditions (Figure 1D, 1H, 1J).
While the distance from viewer to object and camera
As suggested in the assessment of location accuracy, focal length certainly affect the appearance of the
photographic overlays provided qualitative evidence object’s height, for half of the simulations the tur-
as to the accuracy of height of wind turbine towers bines’ heights are under-represented, with only one
and rotor diameter in simulations. In most cases the case of over-estimation (Figure 1L) and one case of
rotor diameter could be observed by the heights of a turbine simulated that could not be matched with
turbine blade tips and distance to nacelle for turbines an as-built turbine (Figure 1D). To assess camera
of similar nacelle heights. Figure 1 (right column) focal length effects other vertical elements in the
show the comparison for turbine heights and rotor photographs (e.g., utility poles, trees) were com-
diameters. pared across the depth-of-field and close matches or
For vantage points #1 and 3 the simulations close- only minor differences were observed. When the
ly match built conditions for turbine heights of simulations err for illustrating turbine height, they
nacelle and blades at top of rotation (Figure 1B and err most consistently by showing turbines that are
1F). Rotor diameter can be difficult to compare for shorter than they are built.

Figure 1. Simulation overlays showing as-built conditions (grey tone) and simulated turbines (white) for six vantage points.
Lead-lines help to compare turbine locations (left column) and heights (right column) (continued

308 Impact Assessment and Project Appraisal December 2011


A case study on visual impact assessment for wind energy development

Figure 1 (continued)

Accuracy: ancillary attributes were noticeable and in one the changes were sub-
stantial (Figure 2). The changes most vividly show
Simulations illustrate wind energy development access roads and powerlines. The vantage points
structures with a focus on wind turbines. However, did not allow for opportunities to view bases of
other elements are added to the landscape to sup- turbine towers where additional ancillary structures
port wind energy processing and delivery, including exist.
roadways and access lanes, electricity regulation
structures, and transmission lines and poles. In field Representativeness of as-built conditions
visits some marked changes were noted between
the simulated landscape view and the built view The adequacy of simulations compared to as-built
with respect to these ancillary attributes. For most conditions is assessed by comparing human field-of-
vantage points (four of the six), the changes were vision and 360-degree fields-of-view to the simula-
not easily visible, but in two vantage points changes tion frames (Horner + Maclennan and Envision,

Impact Assessment and Project Appraisal December 2011 309


A case study on visual impact assessment for wind energy development

Figure 2. Ancillary attributes (grey tone) that were not simulated for vantage point #4

2005). This approximates the experience of someone Because field work in this study demonstrated the
embedded in the landscape with a fixed or sweep- vividness of the turbines against the sky and keeping
ing view (e.g., anyone whose vision is not fixed by a fixed, directed view is unlikely within a surround-
a windshield frame). The figures compare the por- ing envelope of turbines, the lower values for field-
tion of visibility of the simulation frame to human of-view were considered to be narrow and unlikely
binocular vision and the 360-degree landscape in experience. Instead a field-of-vision of 120 de-
view, and count the number of turbines visible grees was applied (10 degrees less than suggested by
within the simulation frame compared to the other Smardon and Karp, 1993, but 70 degrees more than
ways of viewing. Because additional turbines were suggested by Sheppard, 1989) and diagrammed on
under construction at the time of photography, a the panoramic photos (Figure 3), compared to the
map figure from developer documentation was used simulation frame, with turbine numbers counted
to distinguish between the locations of first and across different extents of visibility and reported in
second construction phase turbines in the photo- Table 3.
graphs and limit the comparison to only the first For both vantage points (Figure 3), the simula-
phase (clear geographic zones for first- and second- tion frame was substantially narrower than field-of-
phase development helped eliminate any errors or vision values. The conservative field-of-vision val-
confusion). ue of 120 degrees was applied, yet simulation
Field-of-vision of a ‘typical’ human (binocular frames were measured to be 49 and 48 degrees (for
vision) ranges from 120 degrees to 200 degrees vantage points #3 and 4, respectively) or approxi-
(Smardon and Karp, 1993; Horner + Maclennan and mately 41% of typical binocular vision. Best man-
Envision, 2005), yet in that breadth of vision only a agement practices suggest that narrow frames can
very limited six to ten degrees is in a person’s focus be adequate: ‘For some views, it may be possible
(Horner + Maclennan and Envision, 2005). For con- to incorporate the desired field of view within a
sideration of typical views Sheppard (1989) notes single frame (under 39 degrees with a 50 mm lens);
that human binocular sight is almost 180 degrees in but usually this will require a panorama’ (p. 63,
width, but focal vision is usually limited to the cen- Horner + Maclennan and Envision, 2005). Because
tral 50 degrees of a field-of-view (his recommenda- of the differences between simulation frame width
tion for simulations is 40–45 degrees, p. 127). The and field-of-vision width, turbines shown in the
narrowest values (6 to 50 degrees) for field-of-view simulation are fewer than would be seen in the
implies fixed eyes and immobile heads (i.e., looking as-built situation by a typical person and many
at a landscape but without moving either eyes or fewer than a person standing in the landscape and
head). In this setting the 80 metre towers with 40 turning their head. The discrepancy between turbine
metre blades makes turbines highly visible beyond visibility in the simulation frame, human field-of-
the landscape’s existing high points (tops of trees or vision frame, and 360-degree view is reported in
tower silos, both commonly < 20 metres tall). Table 3.

310 Impact Assessment and Project Appraisal December 2011


A case study on visual impact assessment for wind energy development

Figure 3. Comparison of simulation frame width with panoramic and 120-degree fields-of-vision
for two vantage points
(‘*’ symbols identify turbines)

Impact Assessment and Project Appraisal December 2011 311


A case study on visual impact assessment for wind energy development

Table 3. Comparison of Phase One turbine numbers in over-estimated and under-estimated in visual simula-
simulations, field-of-vision, and panoramic views for tions, but on the whole the number of turbines visi-
two vantage points. (Parenthetic values show
magnitude of difference from the simulations) ble in the simulations was an under-estimate of the
as-built conditions. In six simulated vantage points a
cumulative under-estimate of 13 turbines visible to
Simulation Field-of-vision 360-degree the generator/nacelle was observed, and four more
frame frame view
were visible to the tips of blades (an average 2.8 ad-
Vantage point 12 19 26 ditional turbines at each vantage point). Best prac-
#3 (58% more) (117% more) tices (for Scotland) state that ‘it should be ensured
Vantage point 11 28 43 that all the wind turbines that could potentially be
#4 (155% more) (291% more) seen from a viewpoint should be shown within the
Note: The figures in this table include only the Phase One image’ (Horner + Maclennan and Envision, 2005:
turbines (Phase Two turbines increase the above 92–93).
values) For turbine location accuracy, the planned co-
ordinates reasonably approximate the built co-
Discussion and conclusions ordinates provided by the developer with many cases
implying nearly complete locational accuracy (aver-
Environmental screening reports (by any name) are age of 14 metres different; maximum of 123 metres).
intended to demonstrate the likely visual effect of Yet when comparing the as-built photos to the visual
development (Smardon and Karp, 1993; University simulations, the locational differences appeared to
of Newcastle, 2002) and inform decision-making by be greater than the co-ordinates indicated. For many
residents and local authorities. It is important, there- photographic comparisons, the illustrated change in
fore, that the information is adequate and accurate. location looked much more than the average change
Visual simulation of new development has been a (and sometimes even more than the maximum
way to understand landscape change and a recom- change) indicated by the table of planned and built
mended part of impact assessment for years (Smardon co-ordinates (Table 2). While it is challenging to
and Karp, 1993; USDA Forest Service, 1995; ‘match’ or ‘pair’ the simulated turbine with a built
Sheppard, 1989; The Landscape Institute: Institute of turbine, the photographic comparison does not illus-
Environmental Management and Assessment, 2002). trate locational accuracy on the order of the planned
Simulations of wind energy development are a lo- and built co-ordinates reported. For people who
gical part of any impact assessment and have been relied upon simulations to approximate turbine loca-
advocated for jurisdictions in Ontario (Taylor et al, tions (and hence plan visual screening), the changes
2006) yet not adopted on a provincial basis. A could make screening ineffective.
commercial software for wind energy planning, Heights and rotor diameters of wind turbines in
‘WindPro’, includes a highly specific visual simula- simulations were commonly similar to the as-built
tion module, demonstrating the recognized im- conditions, with both over-estimates and under-
portance and value of visual impact assessment estimates of heights. As with locational accuracy it
for wind energy development (Windpro, 2008). is challenging to match turbines in the simulation to
Other reports have consistently and thoroughly the as-built condition. The majority of turbines
described the importance of visual impact assess- simulated, however, were simulated slightly shorter
ment for wind energy development (Gipe, 2002; than the as-built conditions. A number of reasons
Horner + Maclennan and Envision, 2005; University might explain this difference, not the least of which
of Newcastle, 2002; Gipe, 2003). It is logical and is the difficulty in simulating objects of a fixed
consistent that wind energy development in rural height at an estimated distance from the camera. Yet
Ontario includes visual simulations as part of on the whole the simulations under-estimated the
environmental screening reports. tower and blade heights (a surrogate for rotor diame-
This case study examined the accuracy and ade- ter) in the simulations. In the Scottish Best practices
quacy of visual simulations provided for 45 turbines document, authors note that simulations ‘should at
in a rural wind energy development project. The least have the correct hub height and rotor diameter’
second phase of the project is now complete and has (Horner + Maclennan and Envision, 2005: 92).
effected more visual landscape change, even for van- Ancillary attributes are noticeable in their absence
tage points reported on here (i.e., changes since mid- from visual simulations. Access lanes and electricity
2008). The description of the visual impact and the utilities (including transmission lines and equipment
comparison to simulations is therefore a conservative vaults) are common ancillary attributes with wind
estimate to the new rural landscape experience. energy development. In the simulations these attrib-
Results showed a number of discrepancies be- utes are typically not illustrated, and for two vantage
tween the visual simulations in the environmental points their visibility was noticeable, even substan-
screening report and the as-built situation. Location- tial (e.g., for vantage point #4, Figure 2). Multi-
al information was drawn from the environmental phase transmission lines and poles on road rights-of-
screening report and compared to built data provided way appear to have the greatest visibility of ancillary
by the developer. Turbine numbers were both attributes and have a visual impact that was not

312 Impact Assessment and Project Appraisal December 2011


A case study on visual impact assessment for wind energy development

simulated as part of the environmental screening re- for a typical person with either a fixed or sweeping
port. Best practices for simulations include require- view, and simulations thus show lower numbers of
ments for associated infrastructure, saying that ‘if turbines than would be seen by a person at a vantage
these elements are likely to result in significant point in this rural Ontario landscape. While heights
impacts, either individually or collectively, they of turbine towers and blade lengths were similar to
should be included in photomontages’ (Horner + development outcomes, for most cases these attrib-
Maclennan and Envision, 2005: 110). utes are modestly under-represented. In this case the
The final assessment was for adequacy of simula- visual simulations consistently imply less visual im-
tion materials as compared to the built experience. pact than the developed reality. Best practices
The visual simulations were compared to the experi- suggest that simulations ‘can help illustrate the visu-
ence of viewing the landscape from a fixed position al impacts…to an audience that is less familiar with
using human field-of-view parameters (for binocular windfarm developments, the particular landscape in
vision) and to the experience of living within the question, and/or how windfarms typically appear in
landscape with the freedom to view the horizon in a landscape’ (Horner + Maclennan and Envision,
all directions. The simulation frame width was 2005: 105). Given that wind energy development in
smaller than all human field-of-vision parameters for Ontario remains relatively novel and was provincially
binocular vision found in the literature (and larger unique at the time of the proposal documents re-
than some focused fields-of-view), yet acceptable viewed here, the simulations should have better
for focused vision by a best practices report for wind communicated how the wind turbines would have
energy simulations (Horner + Maclennan and appeared in the breadth of the landscape setting.
Envision, 2005). Though a compromise width of More recent environmental screening reports still do
human field-of-vision was used the simulation not always include visual simulations of likely post-
frames approximated about 41% of what a person development outcomes (e.g., Dillon Consulting
standing at that point in the landscape could see. The Limited, 2009).
number of turbines in the simulation frame therefore If the accuracy and representativeness of the sim-
was less than would be in a typical person’s field-of- ulations for the first phase of development are indic-
view: in one vantage point the number of turbines ative, the second-phase simulations might have also
visible to an average person was 50% more than in under-estimated the number of turbines, heights and
the simulated view; in the other vantage point it was rotor diameters, and ancillary attributes associated
more than one-and-a-half times the simulated view’s with development. The same multi-disciplinary con-
number. The total number of turbines visible from sulting practice provided both sets of simulations,
the vantage points is larger still, so the simulations though the individual technician(s) and technique(s)
yield a poor estimate of the visibility of turbines may have changed. The Phase Two simulations
from either a human field-of-vision perspective or a (Figure 4) are wider fields-of-view than Phase One
landscape immersion perspective (i.e., poor repre- simulations, but this case study did not assess these
sentativeness). In this location the wind turbines sur- Phase Two simulations against built conditions.
round residents — they are not on a shoreline or Some photographic qualities of these simulations,
ridgeline, distant from viewers or seen typically such as the season, colour of the sky and clouds, and
from one side — so the immersion perspective is direction of rotor may compromise their legibility
most representative. (Horner + Maclennan and Envision, 2005; Bishop
In sum the visual simulations provided by the de- and Miller, 2007).
veloper and their consultants in the environmental Best practice documents and published research
screening report may mislead audiences with respect provide substantial direction to guiding visual simu-
to turbine number, location, and ancillary attributes lations to yield accurate and adequate simulations
and, to a lesser degree, height. The simulations do that fully represent the likely landscape change
not adequately represent the development situation (Horner + Maclennan and Envision, 2005; Uni-
versity of Newcastle, 2002; Bishop, 2002; Bishop
and Miller, 2007). Explicitly, visual simulations
should match the number, height, diameter and visu-
al experience of the development, even if perfect
In sum the visual simulations provided accuracy is unachievable (Horner + Maclennan and
by the developer and their consultants Envision, 2005; University of Newcastle, 2002). De-
tailed methods for constructing visual simulations
in the environmental screening report ensure replicability while removing potential bias
may mislead audiences with respect to (real or perceived) and improving neutrality. There
turbine number, location, and is no evidence to suggest that the consulting profes-
sionals are affected by their commercial relationship
ancillary attributes and, to a lesser to the client, yet an impartial consultant who creates
degree, height the simulation based on developer-supplied data
would diminish any perception of potential conflicts-
of-interest. Organizations such as the International

Impact Assessment and Project Appraisal December 2011 313


A case study on visual impact assessment for wind energy development

Figure 4. Simulations provided for the second phase of the wind turbine development

Association for Impact Assessment could lead ef- wind turbines: the influence of distance, contrast, movement
and social variables. Renewable Energy, 32(5), 814–831.
forts to share guidance for visual impact assessments Committee on Environmental Impacts of Wind Energy Projects —
while cataloguing examples of before-and-after sim- National Research Council 2007. Environmental Impacts of
ulations and best practice case studies. Wind-Energy Projects. Washington, DC: The National
Academies Press.
Best practice documents specifically for Ontario — County of Bruce Planning & Economic Development Depart-
a location with a substantially-favourable renewable ment 2005. Application Requirements for Commercial Wind
energy policy (2009) — could build from the experi- Energy (Turbine) Facilities. Walkerton, Ontario: County of
Bruce.
ences elsewhere and improve documentation and Dillon Consulting Limited 2009. McLean’s Mountain Wind Farm
elaborate explicit requirements for visual assessment. Environmental Screening Report/Environmental Impact
Future environmental screening reports (particularly Statement. Toronto, Ontario.
Gipe, P 2002. Design as if people matter: aesthetic guidelines for
the visual impact assessment) should continue to be a wind power future. In Wind Power in View: Energy
closely and independently scrutinized prior to being Landscapes in a Crowded World, ed. M J Pasqualetti, P Gipe
used to inform decision-making or public processes and R W Righter, pp. 173–212. San Diego, California:
Academic Press.
to yield unsurprising landscape changes. Gipe, P 2003. British lessons to avoid anti-wind conflict — a best
practice guide. WindStats, windstats.com, 16, 7–9.
Horner + Maclennan and Envision 2005. Visual Analysis of
Windfarms: Good Practice Guidance. Inverness, Scotland:
Acknowledgements Scottish Natural Heritage, The Scottish Renewables Forum,
Scottish Society of Directors of Planning.
This paper was funded in part by a research grant from a private Klick, H and E Smith 2010. Public understanding of and support
citizen in the area of study, and administered through the Univer- for wind power in the United States. Renewable Energy, 35(7),
sity of Guelph Office of Research. No limits apply to publication 1585–1591.
and dissemination of the research results. Krohn, S and S Damborg 1999. On public attitudes towards wind
M Erik Coleman provided research assistant support to this power. Renewable Energy, 16(1–4), 954–960.
study, and Professor James R Taylor collaborated on a full review Lewis, G M 2008. High value wind: a method to explore the
of wind energy development of which this is a constituent part. relationship between wind speed and electricity locational
The research received support and encouragement from a private marginal price. Renewable Energy, 33(8), 1843–1853.
citizen in the municipality and was aided by the accessibility of Ontario Ministry of Energy and Infrastructure, 24 November 2009.
municipal officials and the transparency with which they dis- Ontario’s Green Energy Act: Our Path to a Green Economy
charged their civil service. I thank the editors and two anonymous and a Cleaner Environment. Available at <http://
reviewers for comments on an early version of this manuscript. www.mei.gov.on.ca/en/energy/gea/>, last accessed 21
October 2010.
Sheppard, S R J 1989. Visual Simulation: A User’s Guide for
Architects, Engineers, and Planners. New York: Van Nostrand
Reinhold.
References Smardon, R C and J Karp 1993. The Legal Landscape:
Guidelines for Regulating Environmental and Aesthetic
Bishop, I D 2002. Determination of thresholds of visual impact: Quality. New York: Van Nostrand Reinhold.
the case of wind turbines. Environment and Planning B — Taylor, J R, R C Corry, S Hall and L Balcerek 2006. Landscape
Planning & Design, 29(5), 707–718. and Visual Assessment Guidance for Wind Energy Facilities.
Bishop, I D and D R Miller 2007. Visual assessment of off-shore Guelph, Ontario: University of Guelph: 83.

314 Impact Assessment and Project Appraisal December 2011


A case study on visual impact assessment for wind energy development

The Landscape Institute: Institute of Environmental Management Service.


and Assessment 2002. Guidelines for Landscape and Visual Windpro 2008. Windpro. Available at <http://www.windpro.com>,
Impact Assessment. New York: Spon Press. last accessed 18 June 2010.
University of Newcastle 2002. Visual Assessment of Windfarms: Wood, G 2000. Is what you see what you get? Post-development
Best Practice. Edinburgh: University of Newcastle. auditing of methods used for predicting the zone of visual
USDA Forest Service 1995. Landscape Aesthetics: A Handbook influence in EIA. Environmental Impact Assessment Review,
for Scenery Management. Washington, DC: USDA Forest 20(5), 537–336.

Impact Assessment and Project Appraisal December 2011 315

You might also like