HomeMy WebLinkAboutPhase III Interim WRA report
City of Unalaska Wind Power Development
and Integration Assessment Project,
Interim Wind Resource Assessment Report
Douglas Vaught photo
November 19, 2020
Douglas Vaught, P.E.
V3 Energy LLC
Anchorage, Alaska
City of Unalaska Wind Power Phase III Report Page | 1
Contents
List of Figures ................................................................................................................................................ 3
List of Tables ................................................................................................................................................. 4
List of Equations ............................................................................................................................................ 5
Introduction .................................................................................................................................................. 6
City of Unalaska Power System ..................................................................................................................... 6
Wind‐Diesel Concepts ............................................................................................................................... 6
Wind Resource Assessment .......................................................................................................................... 8
Site Selection ............................................................................................................................................. 8
Pyramid (lower Pyramid Valley) .................................................................................................................. 11
Pyramid Site and Met Tower Information .............................................................................................. 11
Pyramid Data Quality Control ................................................................................................................. 13
Pyramid Environmental Measurements ................................................................................................. 14
Pyramid Wind Speed and Data Synthesis ............................................................................................... 15
Pyramid Wind Speed Adjustment Against Airport Reference Data .................................................... 17
Pyramid Vertical Wind Flow .................................................................................................................... 20
Pyramid Wind Distribution ..................................................................................................................... 21
Pyramid Wind Shear and Roughness ...................................................................................................... 22
Pyramid Extreme Wind Behavior ............................................................................................................ 23
Periodic Maxima ................................................................................................................................. 23
Method of Independent Storms ......................................................................................................... 24
European Wind Turbine Standards II (EWTS II) .................................................................................. 24
Turbulence .............................................................................................................................................. 24
Pyramid Wind Direction .......................................................................................................................... 25
Pyramid IEC Classification ....................................................................................................................... 25
Hog Island ................................................................................................................................................... 26
Hog Island Site and Met Tower Information ........................................................................................... 27
Hog Island Data Quality Control ............................................................................................................. 28
Hog Island Environmental Measurements ............................................................................................. 29
Hog Island Wind Speed ........................................................................................................................... 30
Hog Island Wind Distribution .................................................................................................................. 31
Hog Island Wind Shear and Roughness .................................................................................................. 32
City of Unalaska Wind Power Phase III Report Page | 2
Hog Island Turbulence ............................................................................................................................ 32
Hog Island Wind Direction ...................................................................................................................... 33
Hog Island and Pyramid Comparison ...................................................................................................... 33
Icy Creek Reservoir (upper Pyramid Valley) ................................................................................................ 34
Icy Creek Reservoir Site and Met Tower Information ............................................................................. 35
Icy Creek Reservoir Data Quality Control ................................................................................................ 37
Icing Data ............................................................................................................................................ 37
Icy Creek Reservoir Wind Speed and Data Synthesis .............................................................................. 37
Icy Creek Reservoir Wind Direction ........................................................................................................ 38
Icy Creek Reservoir and Pyramid Comparison ........................................................................................ 38
Bunker Hill (aka Little South America) ........................................................................................................ 39
Bunker Hill Site and Met Tower Information .......................................................................................... 40
Bunker Hill Data Quality Control ............................................................................................................. 42
Bunker Hill Wind Speed and Data Synthesis ........................................................................................... 42
Bunker Hill Wind Direction ..................................................................................................................... 43
Solar Irradiance ....................................................................................................................................... 45
Other Wind Power Site Options .................................................................................................................. 45
Ballyhoo (east summit area of Amaknak Island) .................................................................................... 46
Ptarmigan Road (eastern flank of Iliuliak Valley) .................................................................................... 46
Wind Flow Modeling ................................................................................................................................... 46
WAsP Software Methodology ................................................................................................................. 46
Lower Pyramid Valley Wind Model ......................................................................................................... 47
Wind Turbines ............................................................................................................................................. 48
Recommended Wind Turbine ................................................................................................................. 48
Utility‐scale Wind Turbines ..................................................................................................................... 49
Estimated Wind Energy Production, Pyramid Met Tower Site ............................................................... 50
Estimated Wind Energy Production, Pyramid Valley Alternative Sites ................................................... 51
Wind Power System Capacity ................................................................................................................. 53
Appendix A – Wind Resource Technical Information ................................................................................. 54
Wind Speed and Power ....................................................................................................................... 54
Direction .............................................................................................................................................. 54
Temperature ....................................................................................................................................... 54
Pressure .............................................................................................................................................. 55
City of Unalaska Wind Power Phase III Report Page | 3
Wind Shear .......................................................................................................................................... 55
Turbulence .......................................................................................................................................... 55
Extreme Wind ..................................................................................................................................... 55
IEC Classification ................................................................................................................................. 55
Icing ..................................................................................................................................................... 56
Appendix B – Pyramid Valley detailed met tower information .................................................................. 57
Appendix C – Hog Island detailed met tower information ......................................................................... 59
Appendix D – Icy Creek Reservoir detailed met tower information ........................................................... 60
Appendix E – Bunker Hill detailed met tower information ......................................................................... 61
List of Figures
Figure 1: Met tower locations and heights (map from Topozone.com) ..................................................... 10
Figure 2: Pyramid 60‐meter met tower (Andy Dietrich aerial photo) ........................................................ 11
Figure 3: Pyramid met tower location (orange line shows underground power distribution routing, 3
phase to the water house/tank, continuing at single phase to Icy Creek Reservoir), view north; Google
Earth image ................................................................................................................................................. 12
Figure 4: Pyramid met tower data recovery rate graphic (tower shading filtering excluded) ................... 14
Figure 5: Pyramid met tower temperature, relative humidity, and air density boxplots .......................... 15
Figure 6: Pyramid mean (mean of monthly means) wind speeds, all anemometers, reconstructed (gap‐
filled) data ................................................................................................................................................... 16
Figure 7: Pyramid diurnal wind speed profile ............................................................................................. 17
Figure 8: Pyramid raw, filtered, and reconstructed (gap‐filled) wind speed data comparison .................. 17
Figure 9: Dutch Harbor Airport wind speed comparison, Pyramid test period vs. 32.5‐year average ....... 18
Figure 10: Pyramid raw, filtered, reconstructed, and adjusted against Dutch Airport weather station
mean wind speeds ...................................................................................................................................... 20
Figure 11: Pyramid vertical wind flow scatterplot, all direction sectors, best fit: y=0.0422x (2.4° average
up flow angle) ............................................................................................................................................. 20
Figure 12: Pyramid vertical wind flow rose, combined 60‐meter anemometers ....................................... 21
Figure 13: Pyramid wind speed probability distribution histogram ........................................................... 21
Figure 14: Pyramid vertical wind shear profile (calculated 0.079 power law exponent) ........................... 22
Figure 15: Pyramid vertical wind shear rose (0.12 power law exponent, outer ring) ................................ 23
Figure 16: Pyramid turbulence intensity graph .......................................................................................... 24
Figure 17: Pyramid wind energy rose, 60‐meter level combined anemometers and 50‐meter wind vane
.................................................................................................................................................................... 25
Figure 18: Hog Island met tower (D. Vaught photo) ................................................................................... 26
Figure 19: Hog Island met tower location, view north; Google Earth image ............................................. 27
Figure 20: Hog Island met tower data recovery graphic (tower shading filtering excluded) ..................... 29
Figure 21: Hog Island barometric pressure boxplot ................................................................................... 30
Figure 22: Hog Island barometric pressure vs. 60 m level wind gust scatterplot (color code indicates wind
direction) with linear trend ......................................................................................................................... 30
City of Unalaska Wind Power Phase III Report Page | 4
Figure 23: Hog Island mean wind speeds, filtered (but not gap‐filled) data, complete months only ........ 31
Figure 24: Hog Island wind speed probability distribution histogram ........................................................ 31
Figure 25: Hog Island vertical wind shear profile (calculated 0.191 power law exponent) ....................... 32
Figure 26: Hog Island vertical wind shear rose (0.40 power law exponent, outer ring) ............................ 32
Figure 27: Hog Island wind energy rose, 60‐meter west anemometer and 60‐meter wind vane ............. 33
Figure 28: Hog Island vs. Pyramid wind speed comparison, 60 m W anemometers, overlap period only 34
Figure 29: Icy Creek Reservoir 34‐meter met tower (D. Vaught photo) ..................................................... 35
Figure 30: Icy Creek Reservoir met tower location, view north, Google Earth image ................................ 36
Figure 31: Icy Creek Reservoir met tower data recovery graphic (tower shading filtering not employed)37
Figure 32: Icy Creek Reservoir wind energy rose ........................................................................................ 38
Figure 33: Icy Creek Reservoir vs. Pyramid wind speed comparison, overlap period only ........................ 39
Figure 34: Icy Creek Reservoir vs. Pyramid wind direction comparison ..................................................... 39
Figure 35: Bunker Hill 10‐meter met tower (K. Arduser photo) ................................................................. 40
Figure 36: Bunker Hill met tower location, view north, Google Earth image ............................................. 41
Figure 37: Bunker Hill met tower data recovery graphic ............................................................................ 42
Figure 38: Bunker Hill wind energy rose, 10‐meter NE anemometer ........................................................ 44
Figure 39: Cold Bay upper air (4500 ft. level) wind rose (from Phase II report) ......................................... 44
Figure 40: Bunker Hill solar irradiance boxplot, units of Watts/meter2 ..................................................... 45
Figure 41: Bunker Hill solar irradiance Dmap, units of Watts/meter2 on right‐hand scale ........................ 45
Figure 42: WAsP software wind flow model of lower Pyramid valley; purple/blue color ~ 5.2 m/s, red
color ~ 9.7 m/s ............................................................................................................................................ 48
Figure 43: EWT DW52‐900 wind turbines in Kotzebue, Alaska .................................................................. 49
Figure 44: Lower Pyramid valley sites of potential interest for wind power, WTG = wind turbine
generator .................................................................................................................................................... 52
List of Tables
Table 1: AEA penetration categories of wind‐diesel system configuration .................................................. 7
Table 2: Revised penetration categories of wind‐diesel system configuration ............................................ 8
Table 3: Pyramid met tower summary information ................................................................................... 11
Table 4: Pyramid met tower sensors .......................................................................................................... 13
Table 5: Pyramid met tower data recovery rate table (tower shading filtering excluded) ........................ 14
Table 6: Pyramid wind speeds with reconstructed (gap‐filled) data .......................................................... 16
Table 7: Pyramid wind speed adjustment against Dutch Harbor Airport long‐term weather station
reference ..................................................................................................................................................... 19
Table 8: Pyramid wind speed distribution table ......................................................................................... 22
Table 9: Pyramid turbulence intensity table and IEC categories ................................................................ 25
Table 10: Hog Island met tower summary information .............................................................................. 27
Table 11: Hog Island met tower sensors ..................................................................................................... 28
Table 12: Hog Island met tower data recovery rate table (tower shading filtering excluded)................... 29
Table 13: Hog Island wind speeds with filtered (but not gap‐filled) data, 10 months data, note poor DRR
for 60m E and 50m W anemometers .......................................................................................................... 31
Table 14: Icy Creek Reservoir met tower summary information ................................................................ 35
Table 15: Icy Creek Reservoir met tower sensors ....................................................................................... 36
City of Unalaska Wind Power Phase III Report Page | 5
Table 16: Icy Creek Reservoir met tower data recovery rate table (tower shading filtering not employed)
.................................................................................................................................................................... 37
Table 17: Icy Creek Reservoir wind speeds with reconstructed (gap‐filled) data ...................................... 38
Table 18: Bunker Hill met tower summary information ............................................................................. 40
Table 19: Bunker Hill met tower sensors .................................................................................................... 42
Table 20: Bunker Hill met tower data recovery data (tower shading filtering not employed) .................. 42
Table 21: Bunker Hill wind speeds with filtered data ................................................................................. 43
Table 22: Wind turbine annual energy production estimate, Pyramid met tower site, includes 10% AEP
loss; NCF = net capacity factor .................................................................................................................... 51
Table 23: WAsP wind flow modeled EWT DW58‐1000 wind turbine AEP at six selected sites in lower
Pyramid valley ............................................................................................................................................. 52
Table 24: Wind Power Penetration Categories at Increasing Wind Power Capacities ............................... 53
Table 25: IEC 61400‐1, 3rd edition basic parameters for wind turbine classification ................................. 56
Table 26: Pyramid met tower complete sensor installation information ................................................... 57
Table 27: Pyramid met tower monthly anemometer data (gap‐filled data set) ........................................ 58
Table 28: Hog Island met tower complete sensor installation information ............................................... 59
Table 29: ICR met tower complete sensor installation information ........................................................... 60
Table 30: Bunker Hill met tower complete sensor installation information .............................................. 61
List of Equations
Equation 1: Wind power density equation (P=power, A= rotor swept area, ρ=air density, V=wind speed;
units Watts/m2) ........................................................................................................................................... 18
Equation 2: Power and wind speed relationship ........................................................................................ 54
Equation 3: Density, pressure, and temperature relationship ................................................................... 54
Equation 4: Wind shear relationship .......................................................................................................... 55
Equation 5: Turbulence intensity ................................................................................................................ 55
City of Unalaska Wind Power Phase III Report Page | 6
Introduction
With high sustained winds, Unalaska Island, home of City of Unalaska and Dutch Harbor, has long been
considered an optimal location for wind energy, but prior to 2017, only a few preliminary wind studies
and analyses were previously completed and none that included the collection of high quality wind data.
The purpose of this report is present and discuss the data collected in Phase III of the City of Unalaska’s
wind power development efforts.
City of Unalaska Power System
The City of Unalaska uses high efficiency diesel generators for power generation. These are two 5.2 MW
Wartsila units and two 4.4 MW Caterpillar units. Wartsila and Caterpillar generators provide good
response to power quality (frequency and voltage) requirements, but the Caterpillar units are sensitive
to prolonged operation below 50% load, which City of Unalaska tries to avoid because of maintenance
and environmental concerns.
Specifically, the City of Unalaska power system is comprised of the following generation units:
Unit 10: Wartsila 12V32 – 5.2 MW
Unit 11: Wartsila 12V32 – 5.2 MW
Unit 12: Caterpillar C280‐16 – 4.4 MW
Unit 13: Caterpillar C280‐16 – 4.4 MW
Organic Rankine Cycle (ORC)1 generators – 3 units – 50 kW each
Unalaska differs from much of rural Alaska. As the largest fisheries port in the United States in terms of
fish volume processed, City of Unalaska’s Department of Public Utilities Administration Division (DPUA)
serves many large industrial customers. At present, electric load demand averages 6.1 MW with a peak
load of approximately 12.5 MW and a minimum load of approximately 4.0 MW. Generation fuel
efficiency is an admirable 15.5 kWh/gallon.2 Average system load demand has been relatively stable
over the past several years, but it does not reflect Unalaska’s total electric load demand as not all local
shore‐based seafood processors buy power from the City of Unalaska; instead, they generate their own.
Unalaska’s electric load profile, because of its seafood processing‐related industrial base, peaks in
February and March and again in August, as opposed to the seasonally cyclic load demand (winter high,
summer low) in most of rural Alaska.
Wind‐Diesel Concepts
Wind‐diesel power system configurations are categorized by Alaska Energy Authority based on their
average penetration levels, or the overall proportion of wind‐generated energy compared to the total
amount of electrical energy generated. AEA’s categories of wind‐diesel penetration levels are very low,
low, medium, and high and are roughly equivalent to the amount of diesel fuel displaced by wind power.
Refer to Table 1 for a detailed explanation.
Related to wind‐diesel system design, secondary loads refer to non‐electric demand such as thermal
loads, which can include district heat loop‐connected (recovered heat system) and remote node
1 See https://en.wikipedia.org/wiki/Organic_Rankine_cycle for an explanation of the organic Rankine cycle
2 Alaska Energy Authority Power Cost Equalization Program Statistical Report, FY2019
City of Unalaska Wind Power Phase III Report Page | 7
hydronic systems. A typical isolated grid district heat loop carries jacket water heat from the diesel
generators to off‐powerplant receivers before routing back to the cooling radiators. A remote node
hydronic system is a hot water heat system that is not connected to the district heat loop. For use as a
secondary thermal load for wind‐diesel operations, a remote node typically will be a large structure such
as a school or hospital.
During periods where the electric load demand is met but winds are such that excess wind power is
available, excess energy can be diverted to secondary loads to partially or fully replace lost heat from
lower diesel generator loading in a district heat loop and/or heating fuel usage in remote node hydronic
systems. If secondary loads are not connected, not available, or do not require heat, wind turbines can
be curtailed (secured or power output reduced through pitch control) and/or excess energy can be
dissipated through an electric boiler in the diesel generator jacket water cooling system and to the
atmosphere via the radiators.
Table 1: AEA penetration categories of wind‐diesel system configuration
Penetration
Category
Wind Penetration Level
Operating Characteristics and System RequirementsInstantaneous Average
Very Low <60% <8% Diesel generators run full time
Wind power reduces net load on diesel
All wind energy serves primary load
No supervisory control system
Low 60 to 120% 8 to 20% Diesel generators run full time
Secondary loads or wind turbine curtailment to
ensure sufficient diesel loading
Relatively simple supervisory control system
Medium 120 to 300% 20 to 50% Diesel generators run full time
Secondary loads or wind turbine curtailment to
ensure sufficient diesel loading
At high wind power levels, complex secondary load
control system to avoid over‐saturation of heat loads
Sophisticated supervisory control system
High
(Diesels‐off
Capable)
300+% 50 to
150+%
Diesels‐off capability
Auxiliary components required to regulate voltage
and frequency
Energy storage (typically)
Highly sophisticated supervisory control system
Medium penetration is often considered a compromise between fuel use offset and relatively minimal
system complexity and is the system configuration for most Alaska village wind‐diesel systems. This
choice though, while initially attractive in achieving high wind penetration for relatively minimal cost,
has sometimes proven difficult to manage in practice as it combines high instantaneous wind input with
an occasionally insufficient system control strategy.
Many wind‐diesel experts though, recognizing the limitations of the medium penetration configuration
design, collapse the wind‐diesel categories to just two: low and high (see Table 2). This reflects the
essential nature of wind‐diesel power quality and system management. At low penetration,
City of Unalaska Wind Power Phase III Report Page | 8
instantaneous wind turbine power input is too low to significantly influence diesel loading and power
quality, hence control measures are minimal. At high penetration, instantaneous wind turbine input is
sufficient to significantly impact diesel engine loading and power quality, hence control measures must
be sophisticated and robust. To avoid turbine curtailment and/or excessive diesel generator spinning
reserve during periods of high wind penetration (i.e., a second diesel generator online to provide
reserve capacity in event of turbine fault), large secondary loads are needed, battery electrical storage
may be necessary, a flywheel or ultracapacitors could be required, or all could be employed.
Table 2: Revised penetration categories of wind‐diesel system configuration
Penetration
Category
Wind Penetration Level
Operating Characteristics and System RequirementsInstantaneous Average
Low 60 to 120% 8 to 20% Diesel generators run full time
Secondary loads or wind turbine curtailment to
ensure sufficient diesel loading
Relatively simple supervisory control system
High
(Diesels‐off
Capable)
120 to
300+%
20 to
150+%
Diesels‐off capability
Auxiliary components required to regulate voltage
and frequency
Energy storage (typically)
Highly sophisticated supervisory control system
Wind Resource Assessment
The August 2017 Request for Proposals, Analysis of the City of Unalaska Wind Power Development and
Integration Assessment Project Phases II to IV noted a wind power project goal of a low penetration
system with approximately 500 kW wind turbines (refer to the City of Unalaska Power System section of
this report for a definition and discussion of wind power penetration). Subsequent discussions though
with Department of Public Works (DPW) personnel managing the wind project broadened the goals to
possible consideration of medium to high penetration and consideration of larger, approximately 1,000
kW capacity wind turbines. Advantages of higher penetration include lower project cost per kilowatt of
installed wind power capacity and more reduction of diesel fuel usage. Disadvantages include increased
system complexity and higher initial capital cost.
With these matters in mind, the highest met tower towers recommended in the Phase II report were 60
meters in lieu of possible 50 meters or less if only considering smaller capacity wind turbine models.
Even 60 meters though is lower than the hub height of some 1,000 kW wind turbine variants, but it is
standard wind power industry practice to equip met towers of this height with anemometers at three
levels and extrapolate to higher with a wind shear power law exponent calculated from the data.
Site Selection
There were several criteria to consider for wind prospecting in Unalaska (completed under Phase II of
the wind project), that commenced with an assessment of the regional wind climate (refer to pages 13
through 20 of the Phase II report). In short, developable locations for wind power in rural Alaska,
including Unalaska, are those where the following criteria are met:
Wind resource: high (but not too high) mean wind speed, normal or near normal Weibull
City of Unalaska Wind Power Phase III Report Page | 9
distribution, low‐to‐moderate turbulence (steady wind flow), acceptable extreme winds, and
unimodal or bimodal wind direction distribution.
Power distribution infrastructure: proximity to existing (or near‐term planned) distribution
lines with sufficient amperage capacity to accept input from planned wind farm capacity,
including expansion potential.
Roads/access: proximity to existing roads, or reasonable cost to develop or improve access.
Site area: large enough to host a wind turbine array that meets project wind power capacity
goals.
Land use: available for development (ownership, easement restrictions, lease rates, etc.).
Airspace: no insurmountable FAA restrictions for airport flight operations.
Terrestrial wildlife and avian species: no or minimal impacts to critical habitat, flyways, etc.
Wetlands, parks, and other high‐value environments: no insurmountable restrictions and/or
acceptable mitigation requirements are possible.
Noise, shadow flicker, and aesthetics: no or minimal impact to residents.
Rime icing environment and/or ice throw risk: no or minimal risk and/or acceptable
mitigating measures possible.
With these considerations, four locations were chosen for installation of meteorological (met) towers
for wind resource evaluation (see Figure 1):
1. Pyramid (Lower Pyramid Valley)
2. Hog Island
3. Icy Creek Reservoir
4. Bunker Hill (referenced in the Phase II report as Little South America)
City of Unalaska Wind Power Phase III Report Page | 10
Figure 1: Met tower locations and heights (map from Topozone.com)
There are two primary uses of wind data for wind power development. First is classification of site(s) as
that determines suitable turbine models3. Wind turbine manufacturers require IEC classification of a site
to ensure that the proposed turbine model is suitable, and that warranty coverage is valid. Financial
institutions and/or partners require proper classification to ensure their wind turbine investment is
appropriate and can be expected to perform throughout the planned service life
The second use of wind data is calculation of annual energy production (AEP) for wind turbines of
interest with reasonable deductions for wake, electrical, O&M, soiling, and other losses. Net AEP data is
used to model economic benefit of a wind power project.
3 See IEC Classification discussion in Appendix A
Hog Island,
60 meters
Bunker Hill,
10 meters
Pyramid,
60 meters
Icy Creek
Reservoir,
34 meters
City of Unalaska Wind Power Phase III Report Page | 11
Pyramid (lower Pyramid Valley)
Pyramid Valley, source of Unalaska’s water supply, was considered at the outset to be the most
promising location in Unalaska for a wind power project. The plateau area that comprises the lower
valley is large enough to contain several megawatts of wind power capacity; a wide, well‐maintained
gravel road provides access; the area devoid of housing and other community‐use development other
than the water plant; and of considerable importance, the valley is served by an underground high
capacity, three‐phase power distribution line (3 phase power routes to the water plant with single phase
continuing to Icy Creek Reservoir) that is minimally loaded at present. Additionally, Pyramid Valley is
relatively distant from Dutch Harbor Airport and displaced from established landing patterns and normal
air traffic routing.
Figure 2: Pyramid 60‐meter met tower (Andy Dietrich aerial photo)
Pyramid Site and Met Tower Information
A 60‐meter height (197 ft.) NRG Systems, Inc. tubular, guyed met tower was installed4 in mid‐October
2018 on City of Unalaska land just south of Veronica Lake and remains operational at time of this report.
Table 3: Pyramid met tower summary information
Data dates 10/16/2018 to 6/16/2020 (20 months)
Datalogger information NRG Symphonie PRO, 26 channel, site no. 3550
Site coordinates 53.8496 North, 166.5625 West (WGS 84 datum)
Site elevation 103 meters (334 ft.)
Wind speed, mean annual, 60 m level 6.37 m/s (7.08 m/s when corrected to Dutch Harbor
Airport long‐term weather station data)
Wind power density, mean annual, 60 m 443 W/m
2 (616 W/m2 when corrected to Dutch Harbor
Airport long‐term weather station data)
4 Met tower installation accomplished by V3 Energy LLC with contracted assistance from Bering Straits
Development Company and Solstice Alaska Consulting. The considerable support provided by City of Unalaska
Dept. of Public Works personnel is much appreciated.
City of Unalaska Wind Power Phase III Report Page | 12
Wind power class 4 (good), of 7 defined classifications; 5 (excellent), when
corrected to Dutch Harbor Airport long‐term weather
station data)
Maximum 10‐min. avg wind speed 29.8 m/s
Maximum 2‐sec. gust wind speed 42.1 m/s (94 mph)
Wind shear power law exponent 0.079 (very low; 0.140 considered nominal)
Calm wind frequency (winds < 4 m/s) Approx. 33%
Extreme wind probability (50‐year period) 31.7 to 44.9 m/s
Turbulence intensity, 60 m level 0.12
IEC 61400‐1 3rd ed. classification Class IIB
Figure 3: Pyramid met tower location (orange line shows underground power distribution routing, 3 phase to the water
house/tank, continuing at single phase to Icy Creek Reservoir), view north; Google Earth image
Prior to installation of the met tower, a Federal Aviation Administration (FAA) obstruction evaluation
was requested. FAA issued Aeronautical Study No. (ASN) 2018‐WTW‐5350‐OE in July 2018 with a
determination of no hazard to air navigation. Obstruction lighting was not required although FAA
requested alternating bands of aviation orange and white paint on the met tower and orange high‐
visibility marker balls be attached near the top of the outer guy wires to improve visibility of the tower
for aviators. Both requirements were accomplished.
The Pyramid met tower is equipped with two anemometers each at 60 meters, 50 meters and 40
meters; one wind vane each at 60 meters and 50 meters; a vertical wind propeller anemometer at 55
meters; and temperature and relative humidity sensors at the tower base (refer to Table 4). Refer to
Appendix B for detailed sensor technical information.
Pyramid met tower
(south of Veronica Lake)
City of Unalaska Wind Power Phase III Report Page | 13
Table 4: Pyramid met tower sensors
Pyramid Data Quality Control
The met tower sensor data was manually filtered to remove compromised records. This included startup
sequencing, isolated periods of power supply problems, icing events, tower shading5, and poorly
functioning sensors. As indicated in Figure 4, anemometer data recovery from the Pyramid met tower
was outstanding. All six anemometers functioned admirably until May 2020 when the channel 1 (60
meter east) and channel 6 (40 meter west) anemometers began “dragging”, or behaving abnormally
compared to their companion anemometers. This behavior was detected in the datasets of the other
three met towers installed for this project but has not been routinely noted in datasets from scores of
met towers installed throughout Alaska over the past 20 years. There is no obvious correlation of this
“dragging” behavior to environmental factors other than observing during met tower installation and
several subsequent site visits that the considerable local bald eagle population found the sensor boom
arms to be convenient places to perch. This report author theorizes that eagles attempt to land on
anemometers, damaging their bearings. From the ground, a damaged anemometer appears to function
normally, but close observation – both visual and via the data record – indicates that it spins more
slowly than its companion and stops moving at slightly higher wind speeds.
On a positive note, infrequent icing events6 have been detected, indicating minimal concern for
atmospheric icing that can negatively impact wind turbine operations.
Note in Figure 4 two long periods of loss of function of the wind vanes and temperature sensor in
November 2018 and January 2019. This was due to a power supply problem that was corrected in
5 Tower shading results from airflow distortion by the met tower. Air decelerates slightly upwind of the tower,
accelerates as it goes around the tower (Bernoulli principle), and decelerates markedly in the lee of the tower
where a flow separation bubble may occur, resulting in disturbed airflow downwind (source: Windographer help
menu). Because of that, anemometers in a 30‐degree arc downwind are filtered from the dataset. Anemometers
are paired opposite each other and perpendicular to the prevailing winds to minimize the tower shading effects.
6 Icing is inferred in the dataset by observing stationary anemometers and/or wind vanes combined with
temperature near freezing or below and relative humidity at or near 100%, indicating the likelihood of snow or
freezing rain.
City of Unalaska Wind Power Phase III Report Page | 14
February 2019. During that site visit, a relative humidity (RH) sensor was installed to aid in the inference
of wintertime icing events.
Figure 4: Pyramid met tower data recovery rate graphic (tower shading filtering excluded)
Table 5: Pyramid met tower data recovery rate table (tower shading filtering excluded)
Pyramid Environmental Measurements
Unalaska experiences a cool, damp maritime climate, August 2019 notwithstanding, with a relatively
narrow range of temperatures, especially compared to northern and interior Alaska, with typically high
relative humidity. From the perspective of wind turbine operations, cool damp air is beneficial as it
yields higher air density than equivalent elevation in warmer climates. This directly improves the lift
force imparted to the rotor blade and increases the resulting turbine power output. Hence, standard air
density at 103 meters (334 ft.) elevation is 1.213 kg/m3, but the measured air density at Pyramid is 1.246
kg/m3, 2.7% higher than standard.
DRR
Data Channel Height (%)
Ch1_Speed_60m_E 59.7 m 88.8
Ch2_Speed_60m_W 59.3 m 96.0
Ch3_Speed_50m_E 50.2 m 95.8
Ch4_Speed_50m_W 49.7 m 95.6
Ch5_Speed_40m_E 38.9 m 94.6
Ch6_Speed_40m_W 38.4 m 92.6
Ch13_Direction_60m_NNE 57.3 m 96.5
Ch14_Direction_50m_NE 48.0 m 96.4
Ch20_Vert Wind_55m_NW 55.2 m 96.9
Ch16_Temperature_3m_N 3 m 96.7
Ch19_RH_2m_N 2 m 80.2
City of Unalaska Wind Power Phase III Report Page | 15
Figure 5: Pyramid met tower temperature, relative humidity, and air density boxplots
Pyramid Wind Speed and Data Synthesis
Wind speed data filtered from the data set as noted in Data Quality Control can be reconstructed to
yield a more accurate and more probable data set than raw (unfiltered) data and filtered data. This is
especially true for tower shading filtering where, depending on relative wind direction frequencies, one
of the paired anemometers will be filtered more often than the other. This can result in a data
divergence between paired anemometers. Wind analysis software allows one to remove filtered data
from the data set and fill the gaps with a pattern‐based reconstruction methodology that references
non‐filtered data from paired anemometers for tower shading filtering and pre and post‐wind speeds
for icing and power disruption events that affect all the anemometers.
City of Unalaska Wind Power Phase III Report Page | 16
With reference to reconstructed or gap‐filled data, mean wind speeds at the 60‐meter level were
measured at approximately 6.35 m/s (14.2 mph) with a mean wind power density of 440 Watts/m2 (see
Table 6). This classifies lower Pyramid Valley as a Class 4 (description: good) wind resource.7
Table 6: Pyramid wind speeds with reconstructed (gap‐filled) data
Pyramid’s monthly wind speed profile (see Figure 6) demonstrates a pronounced seasonal variation in
wind speeds with high winter winds and lower summer winds. This is a normal pattern and matches well
with typical seasonal power demands in a community. Figure 7 indicates a normal, though somewhat
muted, diurnal (daily) wind speed profile of stronger afternoon winds compared to night and morning.
Figure 6: Pyramid mean (mean of monthly means) wind speeds, all anemometers, reconstructed (gap‐filled) data
7 Wind power classification is a U.S. Department of Energy measurement based on wind power density at a defined
height above ground level, typically 50 meters (164 ft.). Class 1 on the low end is considered a poor wind resource
while Class 7 on the high end is considered superb.
Variable
Ch1_Speed_
60m_E
Ch2_Speed_
60m_W
Ch3_Speed_
50m_E
Ch4_Speed_
50m_W
Ch5_Speed_
40m_E
Ch6_Speed_
40m_W
Mean wind speed (m/s) 6.36 6.35 6.26 6.26 6.14 6.19
Mean wind speed (mph) 14.2 14.2 14.0 14.0 13.7 13.8
Max 10‐min wind speed (m/s)29.8 29.8 28.9 28.6 28.5 28.6
Max gust wind speed (m/s) 41.0 41.5 42.1 41.3 40.8 40.5
Max gust wind speed (mph) 91.7 92.8 94.2 92.4 91.3 90.6
Mean power density (W/m²) 443 439 410 411 395 391
Frequency of calms (%) 33.3 33.4 33.5 33.7 34.7 33.9
City of Unalaska Wind Power Phase III Report Page | 17
Figure 7: Pyramid diurnal wind speed profile
Pyramid Wind Speed Adjustment Against Airport Reference Data
With reference to the collected Pyramid wind speed data, it can be assessed in three ways: raw, filtered,
and reconstructed (gap‐filled), as displayed in Figure 8, with the reconstructed data typically higher than
the other two.
Figure 8: Pyramid raw, filtered, and reconstructed (gap‐filled) wind speed data comparison
But, because the met tower measurement period is relatively brief, it may misrepresent true site
climatology. For example, a one or two year met tower project may capture an unusually windy or
unusually calm winter season, which can skew, or bias, the results. At Pyramid, the measured mean
annual wind speed of 6.35 m/s at the 60‐meter level is 0.60 m/s less than that predicted by AWS
Truepower Windnavigator software,8 which warranted further investigation.
Nearby Dutch Harbor Airport, 5.6 km (3.5 miles) north‐northeast of the met tower, serves Unalaska.
Automated airport weather station data from January 1988 to June 2020 was obtained to provide 32.5
years of continuous comparative wind speed for the Pyramid met tower. Airport wind speed data was
converted from miles per hour to meters per second (for consistency with met tower data) and using
the pivot table feature of Excel software, sorted to yield average monthly wind speeds for the 32.5 year
8 See Table 4 on page 30 of the Unalaska Wind Assessment Phase II project report
5.95
6.00
6.05
6.10
6.15
6.20
6.25
6.30
6.35
6.40
Mean Wind Speed, m/sRaw Filtered Reconstructed
City of Unalaska Wind Power Phase III Report Page | 18
period. With that, wind speeds measured at the airport for the 19 complete months of overlap with
Pyramid met tower were compared. With reference to Figure 9, one can see that for the test overlap
period of November 2018 to May 2020 (assessing complete months only), recorded wind speeds at the
airport were less than the 32.5‐year average every month except December 2019 and January 2020.
Figure 9: Dutch Harbor Airport wind speed comparison, Pyramid test period vs. 32.5‐year average
The implication of an observation of lower measured wind speeds at the airport during the Pyramid
study period compared to the long‐term reference is that the mean wind speed calculated at Pyramid is
lower than expected. Processing the Dutch Harbor Airport weather station data to yield measured winds
speeds versus 32.5‐year average winds speed, as shown in Figure 9, yields the data shown in the first
column group of Table 7. The center data group shows measured versus adjusted Pyramid 60‐meter
level anemometers with a percentage adjustment calculated from the first column group. Because the
data period does not comprise a multiple of a complete year, recurring months are combined to yield
mean of monthly means (MoMM) to determine the mean, or average, adjusted wind speed as shown in
the third column group. Hence, in place of a measured 6.37 m/s wind speed at the 60‐meter level, a
mean wind speed of 7.08 m/s can be expected. This is nearer to the AWS Truepower Windnavigator
mean wind speed at the Pyramid met tower site of 6.95 m/s at the 60‐meter level. This has enormous
implications for wind power as the power of the wind is a function of the velocity cubed, as noted in
Equation 1.
Equation 1: Wind power density equation (P=power, A= rotor swept area, ρ=air density, V=wind speed; units Watts/m2)
𝑃
𝐴ൌ 1
2 ∗𝜌∗𝑉ଷ
So, although the long‐term average MoMM predicted wind speed of 7.08 m/s is 11% higher than that
measured at Pyramid during the study period, the power of the wind would be 37% higher (7.083
divided by 6.373). This adjustment boosts the wind power class of the Pyramid site from Class 4 (good)
to Class 5 (excellent). Note that although wind turbines cannot extract all full cubic increase of additional
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May
2018201820192019201920192019201920192019201920192019201920202020202020202020DUT Airport avg wind speed, m/sMonthly average speed 1988‐2020, m/s Measured speed, m/s
City of Unalaska Wind Power Phase III Report Page | 19
energy produced from higher wind power density, the relationship nonetheless is non‐linear and the
advantage of more typical winds in Unalaska enormous.
Table 7: Pyramid wind speed adjustment against Dutch Harbor Airport long‐term weather station reference
Dutch Harbor Airport 60 m combined anemometers Mean of Monthly Means
Year Month
Monthly
average
speed
1988‐2020,
m/s
Measured
speed, m/s
60 m,
m/s
60 m
adjusted,
m/s
%
adjust Month
60 m
MoMM,
m/s
60 m
adjusted
MoMM,
m/s
2018 Nov 5.64 4.93 5.92 6.77 114% Jan 7.15 8.18
2018 Dec 5.70 5.65 7.37 7.44 101% Feb 8.26 8.33
2019 Jan 5.51 5.26 7.56 7.92 105% Mar 7.35 7.71
2019 Feb 5.78 5.61 8.74 9.01 103% Apr 5.58 5.75
2019 Mar 5.49 5.19 7.27 7.67 106% May 4.90 5.17
2019 Apr 5.22 4.98 5.66 5.92 105% Jun 4.53 4.75
2019 May 4.40 3.25 4.42 5.97 135% Jul 5.81 7.87
2019 Jun 3.82 3.43 4.43 4.93 111% Aug 5.67 6.31
2019 Jul 3.62 3.26 5.81 6.47 111% Sep 7.27 8.09
2019 Aug 3.84 3.31 5.67 6.57 116% Oct 6.44 7.46
2019 Sep 4.70 4.49 7.27 7.63 105% Nov 6.04 6.33
2019 Oct 5.55 4.58 6.46 7.83 121% Dec 7.40 8.97
2019 Nov 5.64 4.26 6.15 8.14 132%Annual 6.37 7.08
2019 Dec 5.70 5.82 7.42 7.26 98%
2020 Jan 5.51 5.50 6.75 6.77 100%
2020 Feb 5.78 5.69 7.78 7.91 102%
2020 Mar 5.49 5.24 7.44 7.79 105%
2020 Apr 5.22 4.15 5.50 6.90 126%
2020 May 4.40 4.10 5.38 5.77 107%
Data period 5.11 4.67 6.47 7.09 111%
With reference to Figure 8 and , inclusion of the adjusted mean wind speed with raw, filtered, and
reconstructed (gap‐filled) data demonstrates the considerable impact of adjusting corrected (via filtering
and gap‐filling) wind speed to remove the bias of an unusually windy or calm measurement period
(unusually calm in this instance) with reference to a long‐term data source.
City of Unalaska Wind Power Phase III Report Page | 20
Figure 10: Pyramid raw, filtered, reconstructed, and adjusted against Dutch Airport weather station mean wind speeds
Pyramid Vertical Wind Flow
A RM Young propeller vane anemometer was installed at the 55‐meter (180 ft.) level to enable
calculation of wind flow angle, an important engineering consideration with wind turbines that affects
main rotor shaft bearing loading. The 2.4 degree average up‐flow angle demonstrated in Figure 11 is
normal and reasonable, but relatively high up‐flow angles from specific wind sectors (see Figure 12) may
pose some concern and should be discussed in detail with wind turbine manufacturers.
Figure 11: Pyramid vertical wind flow scatterplot, all direction sectors, best fit: y=0.0422x (2.4° average up flow angle)
5.50
6.00
6.50
7.00
7.50
Mean Wind Speed, m/sRaw Filtered Reconstructed Adjusted against DUT data
City of Unalaska Wind Power Phase III Report Page | 21
Figure 12: Pyramid vertical wind flow rose, combined 60‐meter anemometers
Pyramid Wind Distribution
The probability distribution function, or histogram, of the Pyramid met tower 60 meter wind speed data
indicates a shape curve dominated by low‐to‐moderate wind speeds with an unusually high percentage
of calm winds (see Figure 13).
Figure 13: Pyramid wind speed probability distribution histogram
With reference to Figure 13, Table 8 includes the statistical information of the fitted shape curves for
the measured wind speed distribution. Note that a Weibull k for all four models are lower than 2.0,
City of Unalaska Wind Power Phase III Report Page | 22
which represents a “normal” shape curve in the wind power industry. This strongly demonstrates the
predominance of calm (0 to 0.5 m/s) and lower wind speeds in the data set.
Table 8: Pyramid wind speed distribution table
Pyramid Wind Shear and Roughness
Wind shear is defined as the change in wind velocity (wind and direction vector) with height above
ground level, though in practical usage only speed is considered. Low wind shear is desirable as the
marginal increase in power output at higher heights is minimal, leading to the possibily of lower height
wind turbine towers to significantly reduce project cost (refer to Appendix A for a detailed discussion).
Pyramid wind shear is very low by wind industry standards with a mean calculated power law exponent
(combined paired anemometers) of 0.079 from all wind direction sectors (see Figure 14 and Figure 15).
This compares to a default power law exponent of 0.14, where there is a greater change in wind speed
with height. The calculated surface roughness of 0.00029 meters is reflective of a very smooth surface,
such as a calm sea.
Figure 14: Pyramid vertical wind shear profile (calculated 0.079 power law exponent)
Weibull Weibull Mean Proportion Power R
k A Above Density Squared
Algorithm (‐) (m/s) (m/s) 6.421 m/s (W/m2) (‐)
Maximum likelihood 1.42 7.03 6.40 0.415 475.2 0.8941
Least squares 1.31 7.17 6.61 0.420 601.2 0.8925
WAsP 1.52 7.19 6.48 0.431 445.3 0.8926
Openwind 1.49 7.11 6.42 0.423 445.3 0.8932
Actual data (86,496 time steps) 6.421 0.431 445.3
City of Unalaska Wind Power Phase III Report Page | 23
Figure 15: Pyramid vertical wind shear rose (0.12 power law exponent, outer ring)
Pyramid Extreme Wind Behavior
Extreme wind is described by Vref, or reference velocity, in a 50‐year return period (refer to Table 25 in
Appendix A) as defined by International Electrotechnical Commission (IEC) 61400‐1, 3rd edition (2005)
standards. Reference velocity is the highest 10‐minute mean wind speed predicted to occur once every
50 years. Because very few wind measurements for wind power development are or near 50 years
where direct measurement would be possible, a Gumbel distribution analysis9 estimates the 50‐year
extreme wind probability using collected met tower data. There are three estimation methods: periodic
maxima, method of independent storms, and European Wind Turbine Standards II.
Periodic Maxima
One method to estimate Vref is a Gumbel distribution analysis modified for monthly maximum winds
versus annual maximum winds, which are typically used for this type calculation. Twenty months of
wind data are acceptable for this analysis, though minimally so. For this analysis, the 60‐meter channel 2
(west‐facing) anemometer is referenced because it recorded the highest wind speeds of the six
anemometers on the tower. With filtered and preconditioned data, the predicted Vref by this method is
42.7 m/s. With reference to Appendix A, this result just exceeds IEC Class II criteria, the middle‐defined
category of extreme wind probability.
9 In probability theory and statistics, the Gumbel distribution models the distribution of the maximum or minimum
of several samples of various distributions (Wikipedia); see https://en.wikipedia.org/wiki/Gumbel_distribution for
further explanation.
Vref (50 yr)
Method (m/s)
Periodic Maxima 42.7
Method of Independent Storms 31.7
EWTS II (Exact) 41.6
EWTS II (Gumbel) 42.1
EWTS II (Davenport) 44.9
City of Unalaska Wind Power Phase III Report Page | 24
Method of Independent Storms
A second extreme wind estimation method, method of independent storms, yields a Vref estimate of
31.7 m/s, which is significantly lower than that predicted by the periodic maxima method and would
classify the site as IEC 61400‐1 Class III.
European Wind Turbine Standards II (EWTS II)
The third estimation technique, EWTS II, ignores measured peak wind speeds and calculates Vref from
the Weibull k factor. There are three variants of this method – Exact, Gumbel, and Davenport – which
yield a Vref between 41.6 and 44.9 m/s at Pyramid. These results are like that of the periodic maxima
method and classify the site as IEC Class I or II.
Turbulence
Turbulence at the Pyramid met tower site is moderate with a mean turbulence intensity (TI) of 0.12 at
15 m/s (refer to Appendix A for an explanation of turbulence calculation). Considering the reputation of
the Aleutian Islands for extremely rough and turbulent wind conditions, this is highly desirable. Note in
Figure 16 the moderate TI for wind speeds up to approximately 23 m/s, at which point the TI increases
substantially, though curiously, decreases to Category C (low turbulence) levels at 28 m/s wind speeds.
Figure 16: Pyramid turbulence intensity graph
City of Unalaska Wind Power Phase III Report Page | 25
Table 9: Pyramid turbulence intensity table and IEC categories
Pyramid Wind Direction
The measured prevailing wind directions at Pyramid are broadly northerly, southeasterly, and
southwesterly, with southeasterly and southwesterly winds strongest (see Figure 17). The practical
interpretation of this graph for wind turbine operations is that power‐generating winds are generally
southerly and to a lesser extent, northerly. Hence, for the most part, the winds are bimodal, which is
advantageous for wind turbine array (multi‐turbine) layout.
Figure 17: Pyramid wind energy rose, 60‐meter level combined anemometers and 50‐meter wind vane
Pyramid IEC Classification
As noted in previous sections and discussed in greater detail in Appendix A, for the purposes of wind
turbine design and selection, IEC 61400‐1, 3rd edition (2005) standards classify a site by its extreme wind
and turbulence behavior. The Pyramid extreme wind probability indicates a probable Class II
Standard Represen‐ IEC 3 ed.
Height Data Mean Deviation tative Turbulence
Wind Speed Sensor (m) Points TI of TI TI Category
Ch1_Speed_60m_E 59.7 m 1,190 0.122 0.041 0.174 B
Ch2_Speed_60m_W 59.3 m 1,137 0.121 0.039 0.170 B
Ch3_Speed_50m_E 50.2 m 1,121 0.123 0.037 0.170 B
Ch4_Speed_50m_W 49.7 m 1,111 0.123 0.036 0.169 B
Ch5_Speed_40m_E 40.89 m 1,065 0.127 0.036 0.173 B
Ch6_Speed_40m_W 40.43 m 1,067 0.126 0.035 0.171 B
15 m/s Speed Bin
City of Unalaska Wind Power Phase III Report Page | 26
environment and calculate TI demonstrates Category B turbulence, hence a Class IIB (or possibly low
Class IB) site classification.
Hog Island
The August 2017 Request for Proposals, Analysis of the City of Unalaska Wind Power Development and
Integration Assessment Project Phases II to IV that initiated the wind resource study envisioned up to
five primary sites to be instrumented with met towers. Unalaska’s topography is complex however and
wind power site options are limited, as discussed in the Phase II report. Initially, only lower Pyramid
Valley was considered a primary site and recommended for a large, 60‐meter met tower. The 34‐meter
Icy Creek Reservoir met tower was intended as an auxiliary to the larger Pyramid met tower to serve as
a reference point for wind flow modeling, and the 10‐meter Bunker Hill met tower was installed as
higher elevation reference to validate climatology data derived from Cold Bay upper air monitoring data.
With that, a second primary site was desired as an alternative should the Lower Pyramid Valley wind
resource prove insufficient or unsuitable. With due consideration of the options, it was felt that only
Hog Island readily possessed the development characteristics necessary to host several wind turbines
and hence was added to the project. Meso‐scale wind resource models such as UL’s AWS Truepower
Windnavigator (discussed in the Phase II report) do not, however, include Hog Island and hence its
anticipated wind resource was uncertain. It was thought that Hog Island’s relative distance from high
elevation, shadowing terrain would prove beneficial, but its lower elevation may be disadvantageous
with respect to wind resource.
Figure 18: Hog Island met tower (D. Vaught photo)
Steep topography on the northern half of Hog Island and instrument approach area boundaries for
Runway 13 make it likely that only the southern half of the island could ever be developed for wind
power. Also, Hog Island is only accessible by boat or helicopter and has no existing power distribution.
But according to City of Unalaska Public Works personnel, Hog Island may be less expensive to develop
than the Ptarmigan Road site area in Iliuliuk Valley (refer to the Phase II report for site information and
discussion). This reflects the nature of power distribution supplying Iliuliuk Valley compared to a
City of Unalaska Wind Power Phase III Report Page | 27
relatively straight‐forward requirement to route approximately 1.25 miles of power distribution across
Unalaska Bay from a substation near the airport.
Hog Island Site and Met Tower Information
A 60‐meter height (197 ft.) NRG Systems, Inc. tubular, guyed met tower was installed in mid‐August
2019 on Ounalashka Corporation land on Hog Island and remains operational at time of this report.10
Table 10: Hog Island met tower summary information
Data dates 8/17/2019 to 6/24/2020 (10 months)
Datalogger information NRG Symphonie PRO, 26 channel, site no. 3550
Site coordinates 53.9029 North, 166.5755 West (WGS 84 datum)
Site elevation 30 meters (98 ft.)
Wind speed, mean annual, 60 m level 6.1 m/s (10‐month data period)
Wind power density, mean annual, 60 m 351 W/m
2 (10‐month data period)
Wind power class 3 (fair), of 7 defined classifications (partial year)
Maximum 10‐min. avg wind speed 26.6 m/s
Maximum 2‐sec. gust wind speed 40.7 m/s (91 mph)
Wind shear power law exponent 0.19
Calm wind frequency (winds < 4 m/s) 33% (10‐month data period)
Extreme wind probability (50‐year period) Not calculated
Turbulence intensity, 60 m level 0.13
IEC 61400‐1 3rd ed. classification Not determined
Figure 19: Hog Island met tower location, view north; Google Earth image
10 Met tower installation accomplished by V3 Energy LLC with contracted assistance from Bering Straits
Development Company and Solstice Alaska Consulting, and with the generous material and personnel support of
City of Unalaska Department of Public Works.
Hog Island met tower
City of Unalaska Wind Power Phase III Report Page | 28
Prior to installation of the met tower, a Federal Aviation Administration (FAA) obstruction evaluation
was requested. FAA issued Aeronautical Study No. (ASN) 2018‐WTW‐5353‐OE in September 2018 with a
determination of no hazard to air navigation. Obstruction lighting was required in addition to alternating
bands of aviation orange and white paint on the met tower and orange high‐visibility marker balls near
the top of the outer guy wires to improve visibility. Obstruction lighting was accomplished with a strobe
light kit from NRG Systems, Inc. and a 24 Volt custom designed and constructed battery power system
with a 3 kW wind turbine and 1,000 kW solar power capacity supplied by APRS World of Minnesota.
The Hog Island met tower is equipped with two anemometers each at 60 meters, 50 meters and 40
meters; one wind vane each at 60 meters and 50 meters; and temperature, relative humidity and
barometric pressure sensors at the tower base (refer to Table 11). Refer to Appendix C for detailed
sensor technical information.
Table 11: Hog Island met tower sensors
Hog Island Data Quality Control
As with data collected from the Pyramid met tower, Hog Island met tower data was manually filtered to
remove compromised records. This included startup sequencing, isolated periods of power supply
problems, icing events, tower shading, and poorly functioning sensors. Unlike the Pyramid met tower
though where all sensors have performed very well, several Hog Island anemometers have experienced
“dragging” problems (see Pyramid data quality control discussion) and as of May 2020 both wind vanes
have failed. NRG Systems anemometers and wind vanes are exceptionally reliable, and this rate of
failure can be considered unprecedented. A possible reason is the presence of an exceptionally high
population of bald eagles, which is a distinguishing aspect of Unalaska compared to locations of the
scores of met towers installed throughout Alaska over the past 20 years. During met tower installation
and subsequent site visits, bald eagles were often observed perched on the sensor boom arms. It is
probable that eagles occasionally attempt to land on the sensors themselves, damaging them.
City of Unalaska Wind Power Phase III Report Page | 29
Figure 20: Hog Island met tower data recovery graphic (tower shading filtering excluded)
Table 12: Hog Island met tower data recovery rate table (tower shading filtering excluded)
Hog Island Environmental Measurements
Environmental conditions at Hog Island do not differ substantially from those at Pyramid Valley, hence,
one may reference the previous section for temperature, humidity, and density information. Unlike
Pyramid though, Hog Island was equipped with a barometric pressure sensor (see Figure 21). The intent
of this sensor was ideally to record an extreme low‐pressure event (960 mb or lower) to document
possible accompanying extreme winds. While low pressure of this magnitude was measured on several
occasions, corresponding extreme winds were not (see Figure 22). Although a general trend of
increasing wind gust speeds with lower atmospheric pressure is documented, measured wind gusts
when pressure was less than 960 mB were not especially high. Interestingly, note that the highest wind
gusts generally were with southwesterly to westerly winds during low, but not extreme low pressure
weather events (again, see Figure 22).
DRR
Data Channel Height (%)
Ch1_Speed_60m_E 59.7 m 48.0
Ch2_Speed_60m_W 59.3 m 95.8
Ch3_Speed_50m_E 50.3 m 95.9
Ch4_Speed_50m_W 49.8 m 32.2
Ch5_Speed_40m_E 40.9 m 95.9
Ch6_Speed_40m_W 40.4 m 78.1
Ch13_Direction_60m_SE 57.3 m 88.7
Ch14_Direction_50m_SW 47.7 m 21.6
Ch16_Temperature_3m_N 3 m 99.9
Ch17_Pressure_2m_N 2 m 89.4
Ch18_RH_2m_N 2 m 89.6
City of Unalaska Wind Power Phase III Report Page | 30
Figure 21: Hog Island barometric pressure boxplot
Figure 22: Hog Island barometric pressure vs. 60 m level wind gust scatterplot (color code indicates wind direction) with linear
trend
Hog Island Wind Speed
For the three anemometers with high data recovery rates (60m W, 50m E, and 40m E), 10‐month data
period mean wind speeds were low (see Table 13) at between approximately 5.4 and 6.1 m/s. Because
comparison with Pyramid met tower (see succeeding discussion) demonstrates that Pyramid is the
preferred wind power site of the two locations, the gap‐filling subroutine as employed with Pyramid
data was not accomplished.
City of Unalaska Wind Power Phase III Report Page | 31
Table 13: Hog Island wind speeds with filtered (but not gap‐filled) data, 10 months data, note poor DRR for 60m E and 50m W
anemometers
Figure 23: Hog Island mean wind speeds, filtered (but not gap‐filled) data, complete months only
Hog Island Wind Distribution
The probability distribution function of the Hog Island met tower 60 meter (west) wind speed data
indicates a shape curve dominated by low‐to‐moderate wind speedsFigure 13, but interestingly, with
fewer calm winds (0 to 0.5 m/s) than measured at Pyramid (see Figure 24).
Figure 24: Hog Island wind speed probability distribution histogram
Variable
Ch1_Speed_
60m_E
Ch2_Speed_
60m_W
Ch3_Speed_
50m_E
Ch4_Speed_
50m_W
Ch5_Speed_
40m_E
Ch6_Speed_
40m_W
Measurement height (m) 59.7 59.3 50.3 49.8 40.9 40.4
Mean wind speed (m/s) 6.58 6.11 5.93 6.96 5.41 5.38
Max wind speed (m/s) 26.3 26.6 26.1 26.0 25.8 25.6
Max gust wind speed (m/s) 35.4 40.0 35.4 40.7 35.3 34.6
MoMM power density (W/m²) 442 351 318 463 278 284
Frequency of calms (%) 30.9 33.4 34.3 26.5 40.6 41.4
Data Recovery Rate (%) 48.0 95.8 95.9 32.2 95.9 78.1
City of Unalaska Wind Power Phase III Report Page | 32
Hog Island Wind Shear and Roughness
Hog Island met tower site wind shear is moderate by wind industry standards with a mean calculate
power law exponents of (west‐facing anemometers only) of 0.191 from all wind direction sectors. But,
with reference to Figure 26, wind shear is extremely high with northwest to northeast winds. This
reflects the topography of the met tower site area where a high hill lies to the north. This is an
unavoidable constraint of Hog Island. The high terrain cannot be developed due to conflict with the
Unalaska Airport Runway 13 instrument approach area, and the developable southwestern portion of
the island is lower elevation and partially shadowed by higher terrain to the north.
Figure 25: Hog Island vertical wind shear profile (calculated 0.191 power law exponent)
Figure 26: Hog Island vertical wind shear rose (0.40 power law exponent, outer ring)
Hog Island Turbulence
Turbulence at the Hog Island met tower site is moderate with a mean turbulence intensity (TI) of 0.13 at
15 m/s (refer to Appendix A for an explanation of turbulence calculation).
City of Unalaska Wind Power Phase III Report Page | 33
Hog Island Wind Direction
The measured prevailing wind directions at Hog Island are north and south‐southeast to southwest, with
southeast and southwest winds strongest (refer to Figure 27). This is largely consistent with the Pyramid
met tower wind rose although without the southerly winds channeling through upper and lower
Pyramid Valley.
Figure 27: Hog Island wind energy rose, 60‐meter west anemometer and 60‐meter wind vane
Hog Island and Pyramid Comparison
One objective of Unalaska’s wind study was simultaneous collection of wind data from two or more
primary sites. As noted, primary sites were only lower Pyramid Valley and Hog Island, both equipped
City of Unalaska Wind Power Phase III Report Page | 34
with 60‐meter met towers. Although 20 months of data have been collected at Pyramid, the delayed
installation of the Hog Island met tower resulted in only 10 months of data overlap. Although mid‐June
to mid‐August are missing from the comparison, this is a low wind time in Unalaska and does not
appreciably impact the analysis.
With reference to Figure 28, for comparable anemometers (60‐meter west‐facing) the mean wind
speeds measured at Pyramid are consistently higher, or at least equivalent, to those measured at Hog
Island. All other considerations aside, this is the definitive comparative assessment of the two sites. To
be the preferred site, Hog Island must be considerably windier than Pyramid, but clearly that was not
observed.
Figure 28: Hog Island vs. Pyramid wind speed comparison, 60 m W anemometers, overlap period only
Icy Creek Reservoir (upper Pyramid Valley)
Upper Pyramid Valley, for the purposes of this analysis, comprises the area between Icy Creek Reservoir
(ICR) and Icy Lake at the top of the valley. Although of secondary interest given the wind power
development advantages of the lower valley, upper valley was thought potentially promising should the
lower valley wind resource prove less robust than desired and/or wind power development in the lower
valley prove not feasible for other reasons.
Given the lower likelihood of wind power development in the upper valley compared to lower valley, a
34‐meter met tower was installed at a well‐exposed location immediately west of Icy Creek Reservoir.
Besides providing wind data to lend insight into the upper valley wind resource, data from the Icy Creek
Reservoir met tower was desired to serve as a reference point for a wind flow model using Pyramid met
tower as the model’s data set.
City of Unalaska Wind Power Phase III Report Page | 35
Figure 29: Icy Creek Reservoir 34‐meter met tower (D. Vaught photo)
Icy Creek Reservoir Site and Met Tower Information
The Icy Creek Reservoir met tower was installed in mid‐October 2018 at the same time as the 60‐meter
Pyramid and 10‐meter Bunker Hill met towers.11 The tower was decommissioned and removed from the
site by Department of Public Works personnel in October 2019 following failure of an outer guy wire
that resulted in an unrepairable “crack‐over” of the top of the tower.
Table 14: Icy Creek Reservoir met tower summary information
Data dates 10/16/2018 to 10/28/2019
Datalogger information NRG Symphonie PRO, 16 channel, site no. 3551
Site coordinates 53.82946 North, 166.55130 West (WGS 84 datum)
Site elevation 168 meters (551 ft.)
Wind speed, mean annual, 34 m 5.46 m/s (12.2 mph)
Wind power density, mean annual, 34 m 318 W/m
2
Wind power class 3 (fair), of 7 defined classifications (possibly Class 4 with
adjustment; see Pyramid met tower discussion)
Maximum 10‐min. avg wind speed 28.9 m/s
Maximum 2‐sec. gust wind speed 40.7 m/s (91.0 mph)
Wind shear power law exponent 0.0717 (very low; 0.14 considered nominal)
Calm wind frequency (winds < 4 m/s) Approx. 44%
Extreme wind probability (50‐year period) Not calculated
Turbulence intensity, 34 m 0.122 (moderately high)
IEC 61400‐1 3rd ed. classification Not determined
11 Met tower installation accomplished by V3 Energy LLC with contracted assistance from Bering Straits
Development Company and Solstice Alaska Consulting.
City of Unalaska Wind Power Phase III Report Page | 36
Figure 30: Icy Creek Reservoir met tower location, view north, Google Earth image
Prior to installation of the met tower, a Federal Aviation Administration (FAA) obstruction evaluation
was requested. FAA issued Aeronautical Study No. (ASN) 2018‐WTW‐5349‐OE in July 2018 with a
determination of no hazard to air navigation. Obstruction lighting was not required although FAA
requested alternating bands of aviation orange and white paint on the met tower and orange high‐
visibility marker balls be attached near the top of the outer guy wires to improve visibility of the tower
for aviators. Both requirements were accomplished.
The Icy Creek Reservoir met tower was equipped with two anemometers at 34 meters and one
anemometer at 20 meters; one wind vane each 33 meters; and temperature and relative humidity
sensors at the tower base (refer to Table 15). Refer to Appendix D for detailed sensor technical
information.
Table 15: Icy Creek Reservoir met tower sensors
Pyramid met tower
Icy Creek Reservoir
met tower Upper
Pyramid
Valley
Lower
Pyramid
Valley
City of Unalaska Wind Power Phase III Report Page | 37
Icy Creek Reservoir Data Quality Control
As with data collected from the Pyramid and Hog Island met towers, ICR met tower data was manually
filtered to remove compromised records. This included startup sequencing, isolated periods of power
supply problems, icing events, tower shading, and poorly functioning sensors. Figure 31 and Table 16
demonstrate mixed results regarding data recovery at Icy Creek. There was minor data loss due to icing
in December 218, January 2019 and May 2019, but significant 34m ESE (channel 1) anemometer failure
beginning in July 2019 and the strange and unexplained behavior of the 34m WSW (channel 2)
anemometer during June and July 2019 where the data was obviously faulty for most of June, recovered
for two brief periods, was faulty again during most of July, and then recovered and appeared to respond
normally until the tower was decommissioned in October.
Figure 31: Icy Creek Reservoir met tower data recovery graphic (tower shading filtering not employed)
Table 16: Icy Creek Reservoir met tower data recovery rate table (tower shading filtering not employed)
Data Channel Height DRR (%)
Ch1_Speed_34m_ESE 34 m 68.6
Ch2_Speed_34m_WSW 34 m 84.1
Ch3_Speed_20m_ESE 20.5 m 98.1
Ch13_Direction_33m_W 33 m 98.7
Icing Data
Considering the cool, wet climate of the Aleutian Islands, significant data loss due to icing was expected,
especially at the higher elevation of ICR compared to lower Pyramid Valley. This concern proved
unfounded however as icing loss was a very minimal 0.9 percent over the one‐year data measurement
period.
Icy Creek Reservoir Wind Speed and Data Synthesis
Given the data recovery problems with both 34‐meter level anemometers, data reconstruction or gap‐
filling was employed, like with the Pyramid met tower data, to yield a more accurate dataset for analysis
than raw or filtered data alone would yield.
With reference to reconstructed or gap‐filled data, mean wind speeds at the 60‐meter level were
measured at approximately 5.44 m/s (12.2 mph) with a mean wind power density of 318 Watts/m2 (see
Table 17). This classifies lower Pyramid Valley as a Class 3 (description: fair) wind resource.
City of Unalaska Wind Power Phase III Report Page | 38
Table 17: Icy Creek Reservoir wind speeds with reconstructed (gap‐filled) data
Icy Creek Reservoir Wind Direction
The measured prevailing winds at Icy Creek Reservoir are northwesterly and southeasterly, which
strongly reflects the confining nature – due to enclosure by high mountains to the east and west – of
upper Pyramid valley where the met tower was located.
Figure 32: Icy Creek Reservoir wind energy rose
Icy Creek Reservoir and Pyramid Comparison
As noted earlier, one purpose of the Icy Creek Reservoir was to explore the wind potential of upper
Pyramid valley to determine possible suitability as a wind turbine site vis‐à‐vis the lower valley. It was
recognized that the upper valley is more constrained than the lower valley, which could prove
unfavorable in comparison. With a full year of overlapped data (October 2018 to October 2019), the
expectation of a superior wind resource in the lower valley proved true.
With reference to measured wind shear at the Pyramid met tower (see Figure 14), a virtual 34 meter
anemometer was created to enable direct comparison with the Icy Creek Reservoir wind speed data.
Figure 33 shows the comparative monthly mean wind speeds, with Icy Creek clearly lower for all months
City of Unalaska Wind Power Phase III Report Page | 39
except June 2018 and January 2019 when measured mean wind speeds were equal. The result for the
Icy Creek Reservoir site is classification one wind power class lower than at Pyramid (lower Pyramid
valley).
Figure 33: Icy Creek Reservoir vs. Pyramid wind speed comparison, overlap period only
Although detailed month‐by‐month wind speed and wind direction data could provide additional
insight, comparing the wind frequency roses (overlap period only, Figure 34) measured at the two sites
clearly indicates that Pyramid benefits from southwesterly winds blowing along Captain’s Bay while Icy
Creek Reservoir does not. The reason for this is the high mountainous terrain that borders the entire
eastern boundary of upper Pyramid valley.
Figure 34: Icy Creek Reservoir vs. Pyramid wind direction comparison
Bunker Hill (aka Little South America)
Bunker Hill (also known locally as Little South America because of the shape of the original island of
which Bunker Hill comprises the summit feature) was identified in the Phase II report as a suitable
location to install measure the wind resource – primarily wind directions – to validate meso‐scale wind
modeling of Cold Bay upper air data. There were two candidate sites – Bunker Hill and Ballyhoo
(Amaknak Island) – for this purpose. In some respects, Ballyhoo may have been preferable to Bunker Hill
City of Unalaska Wind Power Phase III Report Page | 40
as it is twice the elevation and hence better exposed, but the location of Bunker Hill between the main
prospective met tower sites – Lower Pyramid Valley and Hog Island – made it the more suitable choice.
A short, 10‐meter met tower was chosen for Bunker Hill as the location, although presumably with a
comparable or stronger wind resource than lower Pyramid Valley, was not considered suitable for wind
turbines. The summit area of Bunker Hill is small and the existing road access expensive to improve.
More importantly, due to significant war usage with many historical features, nearly the entire island
and especially the Bunker Hill summit area is administered by the National Park Service as part of the
Aleutian World War II National Historic Area.
Figure 35: Bunker Hill 10‐meter met tower (K. Arduser photo)
Bunker Hill Site and Met Tower Information
The Bunker Hill met tower was installed in mid‐October 2018 at the same time as the 60‐meter Pyramid
and 34‐meter Icy Creek Reservoir met towers.12
Table 18: Bunker Hill met tower summary information
Data dates 10/18/2018 to 6/16/2020
Datalogger information NRG Symphonie PRO, 16 channel, site no. 3547
Site coordinates 53.87568 North, 166.55820 West (WGS 84 datum)
Site elevation 110 meters (361 ft.)
Wind speed, mean annual, 10 m 6.14 m/s (13.7 mph)
Wind power density, mean annual, 10 m 400 W/m
2
Wind power class 4 (good) to 5 (excellent), of 7 defined classifications
Maximum 10‐min. avg wind speed 30.9 m/s
Maximum 2‐sec. gust wind speed 43.6 m/s (97.5 mph)
Wind shear power law exponent Not calculated
12 Met tower installation accomplished by V3 Energy LLC with contracted assistance from Bering Straits
Development Company and Solstice Alaska Consulting.
City of Unalaska Wind Power Phase III Report Page | 41
Calm wind frequency (winds < 4 m/s) Approx. 35%
Extreme wind probability (50‐year period) Not calculated
Turbulence intensity, 34 m 0.147 (high)
IEC 61400‐1 3rd ed. classification Not determined
Figure 36: Bunker Hill met tower location, view north, Google Earth image
Prior to installation of the met tower, a Federal Aviation Administration (FAA) obstruction evaluation
was requested. FAA issued Aeronautical Study No. (ASN) 2018‐WTW‐5351‐OE in September 2018 with a
determination of no hazard to air navigation. Obstruction lighting was required in addition to alternating
bands of aviation orange and white paint on the met tower and orange high‐visibility marker balls near
the top of the outer guy wires to improve visibility. Obstruction lighting was accomplished with an LED
light from Unimar, Inc. and a 24 Volt battery power system with a 1 kW wind turbine supplied by
Renewable Energy Systems of Alaska.
The met tower was purchased as a NOW configuration from NRG Systems, Inc. As such, it had a
standard suite of instrumentation for a 10‐meter met tower, including two anemometers, one wind
vane, and one temperature sensor, plus a pyranometer (solar irradiance sensor) that was added to the
standard suite. In February 2019, a relative humidity sensor was added to aid the detection of icing
events (refer to Table 19). Refer to Appendix E for detailed sensor technical information.
City of Unalaska Wind Power Phase III Report Page | 42
Table 19: Bunker Hill met tower sensors
Bunker Hill Data Quality Control
As with data collected from the other met towers, Bunker Hill met tower data was manually and
automatically filtered to remove compromised records. This included startup sequencing, isolated
periods of power supply problems, icing events, and poorly functioning sensors. Figure 37 and Table 20
demonstrate several problems including a faulty boom arm on the channel 1, 10‐meter anemometer in
June 2019 that was not corrected until August 2019. Following, the direction sensor failed in October
2019 and was replaced in November 2019. The datalogger itself experienced unexplained and strange
data loss from mid‐March to mid‐April 2020, which seemingly resolved on its own. A check of datalogger
events was not revealing. On a positive note, data loss due to icing was extremely minimal.
Figure 37: Bunker Hill met tower data recovery graphic
Table 20: Bunker Hill met tower data recovery data (tower shading filtering not employed)
Bunker Hill Wind Speed and Data Synthesis
The Bunker Hill met tower was installed not with the intention of evaluating the wind resource at this
location for wind power, but rather to lend insight into wind pattern differences between Pyramid
Valley and Hog Island. As such, gap‐filling reconstruction of filtered anemometer data was not
employed, which explains the excessive wind speed variation measured by the tower anemometers (see
Table 21). Although a mean wind speed of 6.14 m/s at only 10 meters above ground level may seem
extraordinary compared to the same mean wind speed measured at 40 meters on the Pyramid met
tower, this can be highly misleading. Although wind shear on Bunker Hill was not measured (a minimum
Data Channel Height DRR (%)
Ch1_Speed_10m_NE 10 m 83.6
Ch2_Speed_10m_WSW 10 m 96.4
Ch13_Direction_9m_SE 9 m 89.8
Ch16_Temperature_3m_N 3 m 96.5
Ch17_RH_1m_E 1 m 76.9
Ch22_Solar_2m_S 2 m 96.5
City of Unalaska Wind Power Phase III Report Page | 43
of two levels of anemometers would be required), experience has demonstrated with the wind shear on
exposed high hills is very nearly zero to even negative. With this, the measured wind speed at 10 meters
on Bunker Hill is almost certainly representative of the wind speed much higher above ground level.
Table 21: Bunker Hill wind speeds with filtered data
Bunker Hill Wind Direction
The primary purpose of the Bunker Hill met tower was to compare to mesoscale13 winds from the Cold
Bay upper air data to validate the selection of sites for installation of met towers (refer to pages 13
through 20 in the Phase II report). With that, Figure 38 presents the measured wind rose on Bunker Hill
and Figure 39 the Cold Bay upper air data wind rose. As apparent, they do not match well, likely due to
the complex terrain and the channeling of low elevation winds through the complex topography of
terrain near Unalaska. Interestingly though, the Cold Bay wind rose better matches the Icy Creek
Reservoir wind rose (see Figure 32) and to a lesser extent the Pyramid wind rose (see Figure 17).
In hindsight, installation of the Bunker Hill met tower was perhaps not strictly necessary as the options
for readily developable wind power sites in Unalaska were few, limited to lower Pyramid Valley and Hog
Island, and to a lesser extent upper Pyramid Valley, the Ptarmigan Road area of Iliuliak Valley, and on
the periphery of possibility, Ballyhoo. Further, the measured wind roses of lower Pyramid valley (Figure
17), Hog Island (Figure 27) and Icy Creek Reservoir (Figure 32) are understandable and make sense with
respect to their particular terrain exposure, and much less to the upper air wind resource from Cold Bay,
which lies far to the east.
13 Pertaining to meteorological phenomena, such as wind circulation and cloud patterns, that are about 1‐to‐100
km in horizontal extent (www.dictionary.com).
City of Unalaska Wind Power Phase III Report Page | 44
Figure 38: Bunker Hill wind energy rose, 10‐meter NE anemometer
Figure 39: Cold Bay upper air (4500 ft. level) wind rose (from Phase II report)
City of Unalaska Wind Power Phase III Report Page | 45
Solar Irradiance
Bunker Hill was the only met tower of the four equipped with a pyranometer (solar irradiance sensor) to
better understand Unalaska’s solar power resource. Although not the focus of this report, solar power
may be of interest to City of Unalaska and residents. Figure 40 and Figure 41 lend insight into the
potential.
Figure 40: Bunker Hill solar irradiance boxplot, units of Watts/meter2
Figure 41: Bunker Hill solar irradiance Dmap, units of Watts/meter2 on right‐hand scale
Other Wind Power Site Options
During the Wind Power Development and Integration Assessment Project, Phase II site selection
process, several site options besides upper and lower Pyramid Valley and Hog Island were considered
(see pages 22 through 31 of the Phase II report). Most were rejected due to proximity to the airport,
distance from existing power infrastructure, and other reasons. Two sites though – Ballyhoo (east
summer area of Amaknak Island) and Ptarmigan Road (mid‐elevation eastern flanks of Iliuliak Valley) –
stand out as possible alternatives to lower Pyramid Valley and have high modeled wind speeds. Ballyhoo
City of Unalaska Wind Power Phase III Report Page | 46
and Ptarmigan Road were considered for monitoring with met towers and ultimately rejected during the
Phase II planning process in favor of focusing on Pyramid Valley and Hog Island.
Ballyhoo (east summit area of Amaknak Island)
AWS Windnavigator software predicts exceptionally strong winds on Ballyhoo (referring here to the
formerly developed portion of Amaknak Island). At first glance this appears desirable, but the
Windnavigator model predicted excessively strong winds for wind power development. Additionally,
Ballyhoo is within the Aleutian World War II National Historic Area administered by the U.S. National
Park Service, there is no power distribution to the top of the mountain, and perhaps more significantly,
the access road is very steep with tight switchbacks. Challenges aside, Ballyhoo presents significant wind
power potential that may warrant re‐consideration.
Ptarmigan Road (eastern flank of Iliuliak Valley)
This site area is past the turnout of Upper Ptarmigan Road after it turns north and away from Ski Bowl
Road. AWS Windnavigator software predicts an excellent wind resource in this area due to its high
elevation (690 ft. vs. 305 ft. at Pyramid met tower). Ptarmigan Road consists of two sites, one near the
end of the access road and the other downhill and beyond it.
Access to the site area is reasonably easy on a well‐maintained road. Drawbacks however include lack of
high voltage service in Iliuliuk Valley that would be expensive to upgrade per Department of Public
Utilities personnel, location within the instrument approach area to Runway 31 (although this approach
is not used and the restriction perhaps could be successfully challenged), and nearness to housing
development with the potential for noise and shadow flicker14 complaints.
Wind Flow Modeling
WAsP (Wind Atlas Analysis and Application Program) is a Danish PC‐based software program designed to
model the wind resource across a landscape and estimate energy production for individual wind
turbines and/or for multi‐turbine wind turbine farms. This software was used to model the wind
resource in lower Pyramid Valley.
WAsP Software Methodology
WAsP modeling begins with import of a digital elevation map (DEM) of the subject site and surrounding
area and conversion of coordinates to Universal Transverse Mercator (UTM). UTM is a geographic
coordinate system that uses a two‐dimensional Cartesian coordinate system to identify locations on the
surface of Earth. UTM coordinates reference the meridian of its zone (60 longitudinal zones are further
subdivided by 20 latitude bands) for the easting coordinate and distance from the equator for the
northing coordinate, with meter units. Elevations of the DEMs are converted to meters (if necessary) for
import into WAsP software.
A met tower reference point is added to the digital elevation map, wind turbine locations identified (if
desired), and a specific wind turbine (if desired) selected to perform calculations. WAsP models the
orographic (terrain) effects on the wind, plus surface roughness variability and obstacles if added, and
14 Shadow flicker is the effect of the sun shining through the rotating blades of a wind turbine, casting a moving
shadow that will be perceived as a “flicker” due to the movement of the blades. Computer models can predict
when, where, and degree of problem (hours per day, month, and year) of the flicker.
City of Unalaska Wind Power Phase III Report Page | 47
calculates wind flow increase or decrease at each node of the DEM grid. The mathematical model,
although robust, has several limitations, including an assumption that the prospective turbine site(s) and
met tower reference wind regimes are similar, prevailing weather conditions are stable over time, and
the surrounding terrain is sufficiently smooth to ensure attached, laminar air flow. Note that WAsP
software is capable of modeling turbulent wind flow resulting from sharp terrain features such as
mountain ridges, canyons, shear bluffs, etc., but only by accessing the supercomputer resources at the
Technical University of Denmark in Roskilde, Denmark.
Lower Pyramid Valley Wind Model
Previously noted, lower Pyramid valley is demonstrated to have a superior wind resource than upper
valley, likely due to exposure of the lower valley to southwesterly winds flowing along Captain’s Bay.
With that, a WAsP model was created only for lower Pyramid valley. This was accomplished with the
long‐term average‐adjusted Pyramid met tower dataset with a virtual 46‐meter level anemometer and
referencing the 50‐meter level wind vane. A 15‐meter resource grid was generated with 110 columns
and 170 rows to create 18,700 calculation nodes where values such as mean wind speed, wind power
density, wind turbine energy production (if turbines are included), Weibull characteristics, etc. can be
displayed on a Google Earth overlay.
In Figure 42, mean wind speed is displayed and presents an interesting perspective of wind power
options for the lower valley area. Clearly, the low elevation areas of the valley floor where the Icy Creek
Reservoir access road routes through is modeled as a low wind resource as expected. Importantly
though, as evident in the image, the met tower site is not necessarily the most ideal location for wind
power in the lower valley.15 The WAsP model demonstrates that higher winds exist on the higher
elevation terrain on either side of the valley, as one would expect, but also on the northern edge of the
plateau area – north of Veronica Lake – on which the met tower is located.
15 Note that the Pyramid met tower location was chosen because it is on City of Unalaska land, is accessible by ATV
and on foot, and is spacious enough to contain the large footprint of a guyed 60‐meter met tower. Monotube‐
supported wind turbines require much less ground area than a met tower.
City of Unalaska Wind Power Phase III Report Page | 48
Figure 42: WAsP software wind flow model of lower Pyramid valley; purple/blue color ~ 5.2 m/s, red color ~ 9.7 m/s
Wind Turbines
The project’s 2017 Request for Proposals referenced interest in 500 kW capacity wind turbines. It should
be noted though that there are no new‐manufacture 500 kW wind turbines on the world market. The
500 kW Vestas V39, of which there are two presently operational in Alaska (both in Sandpoint on the
Alaska Peninsula), meets the capacity criterion, but this turbine is no longer manufactured. The V39 may
be obtainable as a remanufactured unit from Halus Power Systems in San Leandro, California, but
subject to supply availability from wind farm re‐development projects in Denmark or elsewhere in
northern Europe where most Vestas turbines were installed. Halus Power Systems offers several
warranty options.
If considering only new‐manufacture wind turbines, models of approximately 1,000 to perhaps as high
as 2,000 kW capacity would be suitable for Unalaska. Turbines in this range are available from well‐
known and highly regarded manufacturers who provide excellent warranties and support. These are
exceptionally large machines though, with blade tip heights ranging from 75 to 120 meters (250 to 395
ft.) above ground level.
Recommended Wind Turbine
It is assumed that City of Unalaska would prefer a new manufacture wind turbine to obtain not only the
newest design features, but also the highest levels of warranty coverage and support. With that, the
recommended wind turbine for Unalaska is the Emergya Wind Technologies B.V. (EWT) DirectWind
series of wind turbines. EWT is based in Amersfoort, The Netherlands and manufactures turbines in
relatively wide use in Alaska. At 900 to 1,000 kW capacity, the EWT is an optimal capacity for Alaskan
City of Unalaska Wind Power Phase III Report Page | 49
hub community wind‐diesel systems. The EWT is pitch‐controlled, direct‐drive, variable speed, and
available with 52, 54, 58, and 61‐meter rotor diameters.16 Because the turbine is a direct‐drive design
with a permanent magnet, synchronous generator, there is no gearbox. This simplifies the mechanical
complexity of the machine as there are fewer rotating components than standard wind turbine designs
that employ an asynchronous (induction) generator and gear box.
EWT notes in their product literature that the turbine is optimized for weak grids and micro‐grids with
use of a back‐to‐back, full power converter to control real and reactive power output (enabling power
factor correction) and voltage. If desired to help balance wind energy production with diesel generator
output and load demand, the turbine can de‐tune or purposefully lower its power output by rotor blade
pitch control to 250 kW.17 Eight EWT wind turbines are presently operational in Alaska: two each in
Delta Junction, Kotzebue and Nome, and one each in Bethel and St. Mary’s (refer to Figure 43 for a
photo of the EWT wind turbines in Kotzebue).
The EWT turbine can be ordered with a cold climate package enabling continuous operation to
temperatures as low as ‐40°C. Although necessary for the eight EWT turbines presently operational in
Alaska, a cold climate package would not be necessary in Unalaska’s much more moderate maritime
environment. Of more importance for wind turbine operations in Unalaska is saltwater intrusion from
high onshore winds. The author of this report is not aware of water intrusion problems with EWT’s
existing turbine installations in Alaska, but Unalaska would be a unique installation for the turbine and
hence this issue should be discussed at length with the manufacturer. This recommendation though
applies to any wind turbine contemplated for use in Unalaska.
Figure 43: EWT DW52‐900 wind turbines in Kotzebue, Alaska
Utility‐scale Wind Turbines
Table 24 demonstrates that 6,000 kW of wind power capacity will be necessary to achieve
approximately 33% fuel savings, assuming a 35 percent turbine capacity factor. Should City of Unalaska
implement a project with 33% or higher wind penetration goal, a larger wind turbine model may be
16 EWT’s legacy IEC 61400‐1, 2nd edition‐certified models are 900 kW capacity; new 61400‐1, 3rd edition‐certified
models will be 1,000 kW capacity
17 See http://www.ewtdirectwind.com/ for additional information
City of Unalaska Wind Power Phase III Report Page | 50
desired to reduce a wind farm footprint compared to the 900‐1,000 kW capacity EWT models. At about
2 MW capacity, wind turbines are considered utility scale.
For a high penetration wind project, City of Unalaska may want to consider General Electric (GE) wind
turbines as their models are well represented in Alaska with eleven units presently on Fire Island in
Anchorage and six units in Kodiak. The Fire Island units are 1.5 MW capacity and the Kodiak units are a
mix of 1.5 and 1.6 MW capacity models. Time moves on however and the wind industry continues to
evolve with production of ever larger capacity models. At present, the smallest capacity GE wind turbine
likely suitable for Unalaska is the GE 1.85‐87, rated at 1.85 MW with an 87‐meter diameter rotor. Hub
height is 80 meters, resulting in a tip height of 124 meters (407 ft.) above ground level. GE manufactures
an even more robust version of this turbine, with a 1.85 MW capacity, 82.5‐meter rotor, and a lower 65‐
meter tower.18 More information can be found at https://www.ge.com/renewableenergy/wind‐
energy/turbines.
Other utility‐scale wind turbines are manufactured by very highly regarded Denmark‐based Vestas or
Germany‐based Siemens. Vestas for instance offers the V90‐2.0 MW model, which is optimized for
robust wind regimes. The V90 has been offered for many years and has an outstanding track record of
performance. One consideration though regarding Vestas and Siemens, or another utility‐scale wind
turbine manufacturer such as Enercon (Germany) is that, except for several remanufactured legacy
Vestas units, Vestas, Siemens and Enercon have no new turbines in Alaska and hence no existing support
network.
Estimated Wind Energy Production, Pyramid Met Tower Site
Wind turbine energy production is estimated with a representative (and tentatively recommended) EWT
wind turbine, specifically the new DirectWind (DW) 58‐meter rotor, 1,000 kW capacity model. Other
turbine models are available from EWT, but the DW58‐1000 appears at the time of this report to be
most suitable for Unalaska. A 46‐meter tower is suggested as the low wind shear measured at Pyramid
means only a marginal wind speed gain at higher heights. Higher turbine towers are naturally more
expensive to purchase and require much more robust foundations to resist overturning moments. But a
69‐meter tower is available for the DW58‐1000 model and could also be considered.
There are several approaches to estimate wind turbine annual energy production (AEP) beyond the
selection of a representative wind turbine model noted above. A basic approach is to assume that the
turbine is located at the site of the met tower (the Pyramid met tower). With that, one can calculate AEP
with reference to filtered, reconstructed (gap‐filled), and long‐term adjusted data. One must also
consider a loss factor. Most optimally a wind turbine is available 100 percent of the time and there are
no losses between energy production at the turbine and energy consumption by consumers. This
assumption is not realistic though as many factors conspire to produce losses. These include distribution
line loss, wake loss (for multi‐turbine arrays), rotor blade soiling and icing, downtime for maintenance
and repairs, purposeful curtailment, etc. For an initial evaluation, 10% aggregate losses are assumed,
but they are often higher for wind‐diesel power systems in rural Alaska.
18 Robust in this context means a wind turbine optimized for higher wind speeds, higher extreme wind events
and/or higher turbulence.
City of Unalaska Wind Power Phase III Report Page | 51
With the assumptions noted above, wind turbine energy production for a wind turbine located at the
Pyramid met tower site is presented in Table 22. Note that 10% energy production losses, reflecting a
combination of all loss factors, are assumed. The first column group represents mean annual
reconstructed (gap‐filled) data as measured from October 2018 to June 2020. The second column group
represents the measured data adjusted against the long‐term Dutch Harbor airport weather station data
set. As demonstrated in the table, approximately 5 percent higher AEP can be expected with the EWT58‐
1000 on a 69 meter tower versus a 46 meter tower. More substantively, when adjusted against airport
data, the AEP increase is approximately 17 percent higher than non‐adjusted data.
Table 22: Wind turbine annual energy production estimate, Pyramid met tower site, includes 10% AEP loss; NCF = net capacity
factor19
Estimated Wind Energy Production, Pyramid Valley Alternative Sites
As discussed previously, WAsP wind modeling software allows one to model and visualize wind flow
across the terrain surrounding a data reference site such as the Pyramid met tower. This enables one to
consider site options beyond the met tower location itself and, if multiple turbines are contemplated, to
consider the inter‐turbine effects of multiple wind turbines arranged in an array. Referring again to
Figure 42, locations of interest where the estimated wind speed is relatively high and presents
reasonably constructability potential can be modeled for wind turbine energy production.
Six locations are identified, as shown in Figure 44. WTG 1 (wind turbine generator 1) is the Pyramid met
tower site. WTG 2 is not a site of high interest by itself but is contained within the City of Unalaska
property ownership corridor and hence a possible site option for a two turbine (2,000 kW capacity) wind
power project with the other turbine at or near the met tower site, labeled here as WTG 1. WTG 3 is on
the north edge of the plateau area between the Pyramid Valley access road and Westward Seafoods.
WTG’s 4, 5, and 6 are higher terrain east of the access road. WTG 6 is likely too high of elevation to be
practical, but WTG 4 and WTG 5, plus the green‐color overlay surrounding them, should be considered
of interest as access would not be particularly difficult to construct and potentially highly cost effective
in order to access a superior wind resource.
19 Net Capacity Factor (NCF) or Capacity Factor (CF) is mean power (or energy production) divided by rated power
(or energy production at rated power), or actual energy production divided by maximum energy production
possible (100% power, 100% of the time). 100% capacity factor is never realistic, even for a fossil fuel or nuclear
power plant.
Hub Ht. Hub Ht.
Wind Spd Zero Rated Net AEP NCF Wind Spd Zero Rated Net AEP NCF
Turbine (m/s) Power Power (MWh/yr) (%) (m/s) Power Power (MWh/yr) (%)
EWT DW58‐1000 (46m) 6.27 15.3 4.3 2,247 25.7 6.91 14.3 6.6 2,651 30.3
EWT DW58‐1000 (69m) 6.51 15.1 5.2 2,382 27.2 7.17 14.1 7.6 2,781 31.8
Measured Met Tower Data Data Adjusted against Long‐term Average
% Of Time At Annual % Of Time At Annual
City of Unalaska Wind Power Phase III Report Page | 52
Figure 44: Lower Pyramid valley sites of potential interest for wind power, WTG = wind turbine generator
Table 23 indicates the relative benefit of alternate wind power sites in lower Pyramid valley. Note that
the WAsP software model was run with measured data, not the data set adjusted against the long‐term
airport weather station reference data. Also note that 10% losses were assumed to match the preceding
analysis. With that, one can compare the 2,185 MWh/y estimated AEP of WTG 1 in Table 23 with the
2,247 MWh/y estimated AEP of the EWT DW58‐100 (46 m) in Table 22. They should be the same but
differ by approximately 3 percent. This speaks to the nature of mathematical modeling where differing
methodologies yield slightly different results, even from the same data set. Given the significant natural
variability though of the wind resource, small percentage differences of modeling results are
inconsequential.
Table 23: WAsP wind flow modeled EWT DW58‐1000 wind turbine AEP at six selected sites in lower Pyramid valley
WTG 1
WTG 2
WTG 3
WTG 4
WTG 5
WTG 6
City of Unalaska Wind Power Phase III Report Page | 53
Wind Power System Capacity
With respect to the wind power penetration levels discussed in Wind‐Diesel Concepts section of this
report (see Table 1 and Table 2) and considering Unalaska’s minimum, average, and maximum load
demands of 4.0 MW, 6.1 MW, and 12.5 MW respectively, Table 24 demonstrates the number of wind
turbines that will achieve AEA and the revised penetration categories with use of the representative
EWT DW58‐1000 wind turbine generating at net 30 percent capacity factor (this assumes that turbines
are at or near the met tower location with the wind resource modeled after reference to long‐term
airport reference data). Although there is not necessarily a 1:1 wind kW‐to‐diesel kW fuel savings when
replacing diesel‐generated power with wind‐generated power, the ratio is equal to or near unity in low
penetration situations or when storage batteries are employed. Modeling with HOMER20 software can
lend insight into this dynamic. For the purposes of this analysis, the wind kWh‐to‐diesel kWh
replacement will be considered as 1:1.
Table 24: Wind Power Penetration Categories at Increasing Wind Power Capacities
As observable in Table 24, to achieve highly meaningful annual fuel savings – more than approximately
25 percent total – medium to high wind penetration is necessary. This requires at least 5,000 kW of wind
power capacity and integration and control features in the powerhouse to augment the frequency
regulation capabilities of the diesel generators. This can include a flywheel, a battery energy storage
system (BESS) with an accompanying grid‐forming converter, ultra‐capacitors, or a combination of these
elements. Note however, that the control features that enable medium wind penetration (without
curtailment of the wind turbines as a means of control) also include some elements to enable high
penetration. With a full integration package, which could include a remote node secondary load and
controller (to augment a hydronic heat system), BESS, diesels‐off operation and significant fuel savings
may be possible.
20 Hybrid Optimization and Multiple Energy Resources (HOMER) software; see www.homerenergy.com for further
information.
City of Unalaska Wind Power Phase III Report Page | 54
Appendix A – Wind Resource Technical Information
The purpose of a met tower(s) is to collect one or more years of data pertinent and necessary to
develop a wind power project. At a minimum, this includes data to support an evaluation of wind speed,
wind direction, turbulence, extreme wind behavior, general environmental conditions, icing potential,
and in some circumstances, vertical wind flow.
For the Unalaska wind power development, the following data will be collected or calculated:
Wind Speed and Power
Wind speed is the most important wind characteristic measured at a potential development site. It can
vary widely across the landscape and generally increases with elevation above ground level. At sites with
significant topographic relief, vertical wind flow may be important, and measurement of both horizontal
and vertical wind speed allows one to calculate a wind flow vector.
The power in the wind is proportional to the cube of the wind speed (see Equation 2). This means that a
20 m/s wind has eight times more inherent power than a 10 m/s wind, although by theory21 and the
practicalities of engineering design not all wind power can be converted to wind turbine power. This will
be discussed in much greater detail in the project Phase III report following wind data collection.
Equation 2: Power and wind speed relationship
𝑃ൌଵ
ଶ 𝜌𝑉ଷ (units of power: Watts/m2)
Direction
Direction of the wind is important for turbine siting and performance. A wind resource that is unimodal
or bimodal (with prevailing directions 180 degrees apart) are the most optimal as this will minimize
wake interference between turbines. If a unimodal or 180 degrees bimodal wind resource, the turbines
can be spaced close together (e.g., as near as two turbine rotor blade diameter separation). On the
other hand, if the resource has a multimodal wind direction, wind turbines are likely to require a
minimum 4‐to‐5 blade rotor diameter spacing.
Temperature
Temperature is directly measured with a sensor mounted to the met tower. Equation 3 shows that the
power of the wind is directly proportional to the air density (ρ). Density and temperature are related by
a derivation of the ideal gas law where: where P is site air pressure and R is the specific gas constant
(287.04 Joules/kg, expressed in degrees Kelvin), as shown in Equation 3.
Equation 3: Density, pressure, and temperature relationship
𝜌ൌ𝑃𝑅𝑇ൗ (units of density: kg/m3)
21 Betz’s Law, published in 1919 by the German physicist Albert Betz, demonstrates that the maximum power that
can be extracted from the wind in open flow, independent of turbine design, is 16/27 (59.3%) of the kinetic energy
of the wind (energy is the product of power and time). See https://en.wikipedia.org/wiki/Betz%27s_law for a
detailed explanation.
City of Unalaska Wind Power Phase III Report Page | 55
Pressure
Pressure is a variable of Equation 3Equation 3. If pressure is not directly measured it can be estimated to
within 0.2% accuracy per reference to measured temperature and site elevation.
Wind Shear
Shear is the change in wind velocity with height. This is important as a large variation in wind speed with
height can result in excessive load differential between top and bottom of the rotor which results in
poor turbine performance and increased blade fatigue. The wind shear power law exponent, α, used to
extrapolate wind speed above the upper measurement height of the met tower, is a logarithmic
relationship of wind speeds measured at two or more anemometer heights on the tower. α is non‐
dimensional and typically varies from 0.1 to 0.5. If multi‐height wind speed is not available, α = 0.14 is
assumed because it adequately describes wind shear in terrain with low surface roughness, such as the
grassy tundra of Unalaska. The general form for wind shear is shown in Equation 4 where U is wind
speed, z1 is a selected height on the met tower, z2 is another selected height on the met tower and α is
the power law exponent.
Equation 4: Wind shear relationship
𝑈௭భ
𝑈௭మ
ൌ ൬𝑧ଵ
𝑧ଶ
൰
ఈ
Turbulence
Turbulence Intensity is characterized by the standard deviation of 2‐second wind speed data in a 10‐
minute sampling interval divided by the mean wind speed during that interval as shown in Equation 5
where I is turbulence intensity, σ is standard deviation, U is wind speed and I refers to a specific time
step. Refer to IEC Classification and Table 25 for a discussion of acceptable levels of turbulence intensity
for wind turbine operations.
Equation 5: Turbulence intensity
𝐼ൌ𝜎 𝑈ൗ (dimensionless)
Extreme Wind
The extreme wind speed, or Vref, is the highest ten‐minute average wind speed in a 50‐year return
period. Because much less than 50 years of wind data is collected for a wind power project, Vref must be
calculated as a probability. This is accomplished with a Gumbel distribution, the explanation of which is
quite complex. The interested reader is directed to https://en.wikipedia.org/wiki/Gumbel_distribution
for a detailed explanation. Refer to IEC Classification section below and Table 25 for discussion of
acceptable levels of extreme wind for wind turbine operations.
IEC Classification
Standard 61400‐1 of the International Electrotechnical Commission (IEC), a Switzerland‐based technical
standards organization, forms the basis of evaluation of turbulence intensity and extreme wind. The 3rd
edition was released in 2005, is the most recent and differs in some respects from the 2nd edition which
was released in 1999. Per IEC 61400‐1, 2nd and 3rd editions, wind turbines are classified by extreme wind
probability, or Vref, and turbulence intensity, hence the importance of calculating this information from
site wind data collected by the met towers.
City of Unalaska Wind Power Phase III Report Page | 56
Classification is a matrix of extreme wind (Classes I, II, and III) and turbulence intensity (Categories A, B
and C) as shown in Table 25. Class IA represents the most extreme wind behavior for wind turbines and
Class IIIC the most well‐behaved wind behavior.
Note that IEC classification is different from wind power classification (class 1‐poor to class 7‐superb)!
Although generally a wind power class 1 site will be IEC Class IIIC and a wind power class 7 site will be
IEC Class IA, this rule of thumb does not always hold true. Most desirable is a high wind power
classification with a low IEC classification.
Table 25: IEC 61400‐1, 3rd edition basic parameters for wind turbine classification
Vref Class
I II III S
(‐)/m/s 50.0 42.5 37.5 Values
specified
by the
designer Iref Category A 0.16
B 0.14
C 0.12
Icing
Icing can affect the performance of a wind turbine. Information gathered from wind speed and direction
data includes frequency of icing occurrence, which can be translated to frequency and degree of impact
to wind turbine performance. This is not an exact science, but reference to Scandinavian wind industry
methodology and experience is useful.
City of Unalaska Wind Power Phase III Report Page | 57
Appendix B – Pyramid Valley detailed met tower information
Table 26: Pyramid met tower complete sensor installation information
City of Unalaska Wind Power Phase III Report Page | 58
Table 27: Pyramid met tower monthly anemometer data (gap‐filled data set)
MonthMean (m/s)10‐min max (m/s)Max Gust (m/s)Mean (m/s)10‐min max (m/s)Max Gust (m/s)Mean (m/s)10‐min max (m/s)Max Gust (m/s)Mean (m/s)10‐min max (m/s)Max Gust (m/s)Mean (m/s)10‐min max (m/s)Max Gust (m/s)Mean (m/s)10‐min max (m/s)Max Gust (m/s)Jan 7.15 29.8 40.3 7.15 29.8 40.7 7.01 28.9 39.1 7.03 28.6 39.0 6.92 28.5 37.06.95 28.6 38.3Feb 8.25 28.2 39.5 8.27 27.9 40.7 8.10 27.6 41.4 8.07 26.8 40.6 7.90 27.3 40.87.96 26.8 39.0Mar 7.35 26.7 37.2 7.35 26.2 36.9 7.25 26.0 36.0 7.26 25.3 35.3 7.12 25.3 36.27.15 24.9 34.5Apr 5.58 22.1 27.2 5.57 22.1 27.0 5.47 21.9 26.9 5.50 22.1 26.9 5.40 21.9 27.05.41 21.5 25.4May 4.90 25.9 34.9 4.89 26.1 35.4 4.82 25.3 33.8 4.83 25.5 33.7 4.74 25.1 32.44.80 24.9 32.2Jun 4.50 18.4 23.4 4.51 18.5 25.5 4.41 18.0 24.6 4.52 18.2 24.6 4.35 17.9 24.04.45 17.8 23.9Jul 5.84 17.4 22.6 5.79 17.4 22.5 5.76 17.0 21.6 5.71 17.1 21.6 5.65 16.8 21.75.75 16.8 21.6Aug 5.67 21.9 28.7 5.66 21.9 28.5 5.65 21.3 28.4 5.70 21.3 28.4 5.60 20.8 28.65.61 20.8 29.2Sep 7.29 22.3 29.5 7.25 22.2 28.5 7.21 21.8 30.7 7.14 21.6 29.2 7.00 21.3 28.67.04 21.2 28.4Oct 6.45 28.3 40.3 6.42 28.9 40.7 6.35 27.9 39.1 6.32 27.4 39.0 6.20 27.2 39.36.21 27.0 39.0Nov 6.08 18.2 34.1 6.01 18.0 32.4 5.92 17.6 34.5 5.89 17.6 33.7 5.79 17.5 37.75.84 17.4 31.4Dec 7.40 28.7 41.0 7.39 28.6 41.5 7.26 28.1 42.1 7.25 27.3 41.3 7.13 27.8 40.87.18 27.8 40.5Annual 6.36 29.8 41.0 6.35 29.8 41.5 6.26 28.9 42.1 6.26 28.6 41.3 6.14 28.5 40.8 6.19 28.6 40.5maximum maximum maximum maximum maximum maximum60m E 60m W 50m E 50m W 40m E 40m W
City of Unalaska Wind Power Phase III Report Page | 59
Appendix C – Hog Island detailed met tower information
Table 28: Hog Island met tower complete sensor installation information
City of Unalaska Wind Power Phase III Report Page | 60
Appendix D – Icy Creek Reservoir detailed met tower information
Table 29: ICR met tower complete sensor installation information
City of Unalaska Wind Power Phase III Report Page | 61
Appendix E – Bunker Hill detailed met tower information
Table 30: Bunker Hill met tower complete sensor installation information