-
extracted text (extracttext:extracted_text)
-
VALLEY CIRCULATION EXPERIMENT:
A CLASSIFICATION OF WIND FLOW IN THE
H.J. ANDREWS EXPERIMENTAL FOREST
by
Jerilyn R. Walley
A Thesis
Submitted in partial fulfillment
of the requirements for the degree
Master of Environmental Studies
The Evergreen State College
June 2013
©2013 by Jerilyn R. Walley. All rights reserved.
This Thesis for the Master of Environmental Studies Degree
by
Jerilyn R. Walley
has been approved for
The Evergreen State College
by
________________________
Judith Bayard Cushing, Ph.D.
Member of the Faculty
________________________
Date
ABSTRACT
Valley Circulation Experiment:
A classification of wind flow in the
H.J. Andrews Experimental Forest
Jerilyn R. Walley
Weak wind flow at night is a poorly understood phenomenon in the field
of meteorology. The Valley Circulation Experiment (VALCEX) aimed at creating
a wind climatology for the Lookout and McRae Valleys in H. J. Andrew (HJA).
Two Sound Detection and Ranging (SoDAR) systems and two sonic anemometers
were installed at two stations in adjoining valleys: Lookout Valley, near the HJA
Headquarters meteorological station, called Primet, and McRae Valley,
approximately 6 kilometers away. Among other variables, each station collected
wind speed and direction data in 10 m vertical increments starting at 15 m to 395
m above ground level, aggregated into 5-minute mean averages. Research
questions for the study include: 1) To what extent are wind speeds and directions
similar at the two stations? 2) Under what conditions do the two instrument
locations experience similar phenomena or connectivity? 3) Can a visual
classification of sodargrams based on 5 fundamental characteristics in this case
synoptic forcing, wind direction, valley jet, pulsing, and similar be used to
determine connectivity between the two stations?
Results indicate that based on the 5 criteria the wind flow in the valley was
considered connected between the two stations on 43 of the 96 nights classified.
The two stations, McRae and Primet, were classified as weak synoptic forcing on
79 of the 96 nights classified. The second criterion, wind direction was classified
almost equally at each station; 52 nights were classified as from the Northnortheast while 42 nights were classified as from the Southwest. The criterion of
valley jet present was determined visually on 49 of the nights..
Keywords: Micrometeorogy, Cold-air pooling, SoDAR, Visualization
Table of Contents
Figures......................................................................................................................v
Tables ......................................................................................................................v
Acknowledgements ................................................................................................ vi
1
Introduction ........................................................................................................1
1.1
The Atmospheric Boundary Layer ......................................................................... 3
1.2
Cold Air .................................................................................................................. 4
1.3
Valley Jet ................................................................................................................ 5
2
Materials and Methods .......................................................................................5
2.1
Instrumentation ....................................................................................................... 5
2.2
Speed and Directional Computations – Scalar v. Vector ....................................... 7
2.3
Site Characteristics ................................................................................................. 9
2.4
Instrument Placement ........................................................................................... 11
3
Analysis............................................................................................................13
3.1
VALCEX and HJA Climate Network Station Data Comparison ......................... 13
3.2
Sodargram Analysis of 94 days ............................................................................ 15
3.3
Cases based on classification criteria ................................................................... 23
3.4
Mean wind climatology: March 13 through June 13 ............................................ 25
3.5
Case 1 – Weak SF NNE Valley Jet No Pulse Similar .......................................... 27
3.6
Case 2 – Weak SF SW No Valley Jet Pulse Dissimilar ....................................... 28
3.7
Case 3 – Weak SF NNE Valley Jet Pulse Similar ................................................ 29
3.8
Case 4 – Weak SF NNE No Valley Jet Pulse Dissimilar (9 nights)..................... 30
3.9
Case 5 – Weak SF, SW, No Valley Jet, Pulse, Similar (6 Nights) ....................... 31
3.10
Case 6 – Strong SF, SW, No Valley Jet, Pulse, Dissimilar (6 Nights)................. 32
3.11
Case 7 – Weak SF, SW, Valley Jet, Pulse, Similar (5 Nights) ............................. 33
3.12
Case 8 – Weak SF, SW, No Valley Jet, No Pulse, Dissimilar (5 Nights) ............ 34
3.13
Case 9 – Weak SF, NNE, Valley Jet, Pulse, Dissimilar (5 Nights) ...................... 35
3.14
Case 10 – Weak SF, SW, Valley Jet, No Pulse, Dissimilar (4 Nights) ................ 36
3.15
Case 11 – Weak SF, NNE, No Valley Jet, No Pulse, Dissimilar (4 Nights) ........ 37
4
Discussion ........................................................................................................38
5
Conclusions ......................................................................................................41
5.1
VALCEX and HJA Climate Network Station Data Comparison ......................... 41
5.2
Valley Connectivity .............................................................................................. 42
5.3
Cases based on 5 Classification Criteria .............................................................. 43
6
Future Work & Recommendations ..................................................................43
iv
Figures
Figure 1: SoDAR installed at Primet ...................................................................... 6
Figure 2: Cartesian coordinate system X is oriented into the mean wind direction 7
Figure 3: H. J. Andrews Map, Courtesy of H.J. Andrews LTER ......................... 10
Figure 4: Topographic map of H.J. Andrews showing SoDAR locations ............ 10
Figure 5: Primet location showing surrounding elevations .................................. 11
Figure 6: McRae location showing surrounding elevations ................................. 12
Figure 7: Distribution of wind speeds of the sonic anemometer and propeller
anemometer at Primet .................................................................................... 14
Figure 8: Comparison of wind speed between the sonic and the cup propeller ... 14
Figure 9: Speed sodargram example ..................................................................... 16
Figure 10: Directional sodargram example........................................................... 16
Figure 11: Diagram of decision-making process. ................................................. 17
Figure 12: Strong Synoptic Forcing classification example ................................. 19
Figure 13: Weak Synoptic Forcing classification example .................................. 19
Figure 14: NNE wind direction classification ...................................................... 20
Figure 15: SW wind direction classification ......................................................... 20
Figure 16: Valley jet classification example ......................................................... 21
Figure 17: Pulse wind direction classification example ....................................... 22
Figure 18: Pulse wind speed classification example............................................. 22
Figure 19: Speed and direction averages by height for study period .................... 26
Figure 20: Speed and direction averages by height for Case 1 ............................. 27
Figure 21: Speed and direction averages by height for Case 2 ............................. 28
Figure 22: Speed and direction averages by height for Case 3 ............................. 29
Figure 23: Speed and direction averages by height for Case 4 ............................. 30
Figure 24: Speed and direction averages by height for Case 5 ............................. 31
Figure 25: Speed and direction averages by height for Case 6 ............................. 32
Figure 26: Speed and direction averages by height for Case 7 ............................. 33
Figure 27: Speed and direction averages by height for Case 8 ............................. 34
Figure 28: Speed and direction averages by height for Case 9 ............................. 35
Figure 29: Speed and direction averages by height for Case 10 ........................... 36
Figure 30: Speed and direction averages by height for Case 11 ........................... 37
Tables
Table 1: Classification Criteria ............................................................................. 18
Table 2: Number of Nights by Case ..................................................................... 24
Table 3: Number of nights meeting each criterion ............................................... 39
Table 4: Number of nights meeting each criterion ............................................... 39
Table 5: Connectivity by Case .............................................................................. 41
v
Acknowledgements
It is with immense gratitude that I acknowledge all those who provided me
the possibility to complete this thesis. Special appreciation goes to my reader, Dr.
Judy Cushing, whose remarks and encouragement throughout this project helped
me to complete my thesis, especially during the writing process.
I would like to thank Dr. Christoph Thomas, Oregon State University
Biomicrometeorology Group, for introducing me to the field of
Micrometeorology, and for so willingly sharing his precious time and equipment.
Dr. Thomas conveyed a spirit of adventure in regard to this research and the
daunting task of learning MATLAB.
I would like to thank my friends and family who have supported me
throughout this project by encouraging me to persevere. I will be forever grateful
for your support: Al Walley, Ardith Walley, Craig Walley, Donna Peeples, MJ
DeHart, Eryn Farkas and Donna Waters.
VALCEX was supported by Oregon State University
Biomicrometeorology Group’s Advanced Resolution Canopy Flow Observation
(ARCFLO) campaign (NSF Career Award 0955444). Evergreen’s Visualizing
Terrestrial and Aquatic Systems (VISTAS) (NSF-DBI-1062566) supported my
own time on the project.
vi
1
Introduction
The Valley Circulation Experiment (VALCEX) aims at improving the
understanding of larger valley-scale airflows in the H.J. Andrews (HJA) Long
Term Ecological Research (LTER) Forest. VALCEX goals include determining
the vertical structure of airflow dynamics of cold-air drainage in the HJA between
Lookout and McRae Valleys. The study analyzed image-based data with a
commonly used micrometeorological visualization, the sodargram. This process
used visualization to enable insights based on cognitive and perceptual principles
by applying human judgment to reach conclusions.
Air exchange between forests and the lower atmosphere plays an
important role in transport of heat, moisture, and other trace gasses between the
ground and the atmosphere, directly impacting human life and the environment
(Thomas et al., 2012). Information about air exchange helps to correctly predict
the transport of pollutants and contaminants for various atmospheric conditions
and to more precisely estimate carbon sequestration and evapotranspiration rates
for tall vegetation. Further, how topography influences weak-wind transport is
poorly understood, in spite of the fact that approximately 25% of the earth’s
surface is mountainous terrain.
Data was collected over an 8-month period using a pair of acoustic
ground-based remote sounders installed in two locations within HJA. A SoDAR
unit and sonic anemometer were installed at HJA Headquarters in Lookout Valley
(dubbed Primet) and McRae Valley (dubbed McRae). Data were analyzed in two
1
stages. First, data collected from November 30 to December 13, 2011, from the
SoDAR and sonic at Primet were compared to a wind cup anemometer that has
been collecting data for the past 20 years. Next, data from Primet and McRae
were analyzed to determine connectivity within the valley under different
mesoscale conditions. The second analysis was completed on 94 days of data,
from March 13 through June 23, 2012. This study differs from many
micrometeorological studies in its use of visual analytics to classify phenomena
(Whiteman et al., 2001; Clements, 2002; Mahrt, 2008; Whiteman and Zhong,
2008; Smith et al., 2009; Dorninger et al., 2011).
Rationale for placing the first installation, Primet, near Headquarters was
to compare existing meteorological instrumentation with the SoDAR and sonic,
and because of existing infrastructure (power and networking capabilities).
Placement of the second installation at McRae Valley was to determine
connectivity between Primet and McRae. The McRae installation is north of the
confluence of McRae and Lookout Creeks. Wind in the H. J. Andrews was
known to flow in two dominant directions, either from the North-northeast, down
McRae valley, or from the Southwest, through Lookout Creek valley. The
hypothesis was that when wind is from the Southwest, the two locations will
experience different phenomena, i.e., be disconnected, and when wind is from the
North-northeast, the two stations will show similar speeds and directions, i.e., be
connected.
2
1.1
The Atmospheric Boundary Layer
The Atmospheric Boundary Layer (ABL) is commonly defined as the
layer of the atmosphere where the earth’s surface influences wind dynamics. The
sun heats the earth’s surface during the day, which results in an unstably stratified
ABL. This in turn results in increasing convectively driven turbulence throughout
the ABL. The surface layer is the lowest portion of the ABL and is where the
surface fluxes are assumed to be constant with height. The largest area of the
ABL is a mixed layer, just above the surface layer, and is generally neutrally
stratified. At the top of the ABL is an inversion layer, which separates the ABL
from the free troposphere. At night, when net radiation is negative, an inversion
layer caps the mixed layer, creating a shallow Stable Boundary Layer (SBL). The
SBL grows after sunset due to the earth’s surface being colder than the air above.
The SBL has generally reduced wind speeds, and wind directions can be subject
to sudden shifts of up to 180°, termed “meandering” (Mahrt, 2008). Weak-wind
meandering has been studied less than cross-wind fluctuations yet are more
directly related to practical problems, such as vertical dispersion of contaminants.
Despite the prevalence of weak-wind conditions, diurnal weak-wind
transport is one of the least understood phenomena of micrometeorology (Smoot,
2012). Net radiant energy is the difference between incoming and outgoing
components of radiant energy. During the day, heat is transferred from the sun to
the earth, creating a surplus of energy at the surface, here defined as positive net
radiation; during the night, the net radiant energy is negative. Air in touch with
3
the ground surfaces gets cooled through conduction and can sink on sloped
surfaces because of increased density.
1.2
Cold Air
A cold-air pool is a topographically confined, stagnant layer of air that is
colder than the air above (Whiteman et al., 2001). Whiteman et al (2001)
characterized cold pools either as diurnal, forming during the evening or night and
decaying following sunrise the next day, or as persistent, lasting longer than a
normal night-time temperature inversion. In mountainous terrain on a calm clear
night, air in contact with the ground becomes cooled from radiative energy loss,
and being denser than the warmer air above, sinks to the valley floor (Lundquist
et al., 2008). This air can remain stagnant, trapped by the surrounding higher
terrain, resulting in long periods of poor air quality and fog, depending on
pollution sources and the amount of moisture in the air. With such very weakwind conditions, wind direction may be quite variable (Mahrt, 2008).
Cold pools begin to form in depressions, valleys, and basins in the early
evening when radiant energy is negative. Cold pools can capture moisture,
carbon, pollen and pollutants. Pools exchange energy through temperature
differences, creating micro currents and turbulence. Cold air drainage requires
large-scale wind to be weak, as strong winds will dispel the cold-air pool.
According to Mahrt (2010), well-developed cold air drainage has been studied
extensively from observations. Whiteman (1990) provides a detailed review of
prior studies on cold pooling.
4
1.3
Valley Jet
A feature of airflow in mountainous terrain is that wind speed increases
with height above ground level (agl) more rapidly than it does over level ground.
Above a shallow surface shear layer, a low-level jet, a.k.a. valley jet, can form as
a vertical band of stronger winds in the lower part of the ABL (Arya, 2001). This
requires calm and widely non-turbulent conditions called stable stratification of
the ABL; however some intermittent turbulence is almost always present. Under
these conditions, a valley jet can occur at 10 to 300 m agl (Folken, 2008). Mayer
(2005) defines a low level jet as a thin stream of fast moving air with maximum
wind speeds of 10 ms-1 to 20 ms-1 usually located at 100 to 300 m agl. For this
study, I have expanded Mayer’s definition of a valley jet to include wind speed
occurrences above 3.5 ms1 at 100 to 250 m agl, with lower wind speeds above and
below.
2
2.1
Materials and Methods
Instrumentation
The VALCEX study utilized a paired data collection system consisting of
a Sound Detection and Ranging (SoDAR) array and a sonic anemometer for data
collection (Figure 1). SoDAR is a meteorological instrument and is an acoustic
profiler to measure the backscattering of sound pulses for observing wind speed,
wind direction, atmospheric turbulence and stability classes. SoDAR systems are
used to profile the lower atmosphere from 15 m to ~˃ 1,000 m agl. VALCEX
5
used a Meteorologische Messtechnik GmbH (METEK) mono-static phased array
acoustic profiler, the PCS.2000-24.
Figure 1: SoDAR installed at Primet
The SoDAR system emits an audible pulse at a defined frequency and
listens for the return signal. A small fraction of the energy that travels through the
atmosphere is backscattered and received by SoDAR antenna. The scattered
elements are small-scale air density variations due to small-scale turbulence in the
air column. The return signal intensity and the frequency are processed by the
METEK SoDAR to determine wind speed, wind direction, turbulent character of
the atmosphere and atmospheric reflectivity at multiple heights (Mayer, 2005;
Smoot, 2012). The two SoDARs were programmed to take measurements
averaged over 10 m vertical increments, or gates, starting at 15 m and continuing
to 395 m agl, depending on atmospheric conditions, at a frequency of 2200 Hz.
6
Measurements were taken every 8 seconds, then averaged to 5-minute increments
by the SoDAR. Smoot (2012) provides comprehensive information on the use of
this SoDAR system, SoDAR limitations, fixed echoes and data processing.
Sonic anemometers use ultrasonic sound waves to measure wind speed
and turbulence based on the time of flight of sonic pulses between pairs of
transducers. The VALCEX sonic anemometer sampled wind speed, wind
direction, turbulence and temperature, at the instrument height, at a temporal
resolution of 10 Hz, averaging those data to 5-minute and hourly increments.
This device collects wind speed and direction using three wind components, with
the x-axis aligned to the mean average direction of the wind, the y-axis horizontal
and the z-axis vertical with positive upwards (Figure 2).
Figure 2: Cartesian coordinate system
X is oriented into the mean wind direction
2.2
Speed and Directional Computations – Scalar v. Vector
Wind has both direction and speed. Both scalar and vector quantities are
used to describe wind. Scalar measurements refer to magnitude only, e.g., 27° C.;
vector quantities refer to the both magnitude and direction, e.g., 20 ms-1, East.
7
SoDAR wind speed and direction data are collected using vector averages, i.e.,
wind components x, y, and z, as shown in Figure 2, are combined to form a wind
vector at selected averaging intervals.
Averaging vector quantities differs from averaging scalar quantities since
both the vector’s direction and magnitude need to be accounted for in the
calculation. For example, suppose a constant wind from 360° at 5 ms-1 for 5
minutes changes to 5 ms-1 from 180° for 5 minutes; if vector and scalar averages
were calculated for quantities over this 10-minute period, the vector-averaged
speed would be zero, whereas the scalar-averaged speed would be 5 ms-1.
In a polar coordinate system, the horizontal wind vector is defined by its
mean scalar wind speed (U) and mean horizontal wind direction (ϕ) for each
height computed by the SoDAR software as temporal averages over 5 minutes.
Temporal variability (σ) over any period with N 5-minute data points is then
computed by a MATLAB script as:
(1) 𝜎! = 𝜎!! =
!
!!!
!
!!!
𝑈! − Ū !
and
(2) 𝜎! = ∆∅,
with the wind direction difference between subsequent intervals ΔΦ,
which is restricted between 0 and 180°, defined as
(3) ∆∅! =
∅ − ∅!!! 𝑖𝑓 ∆∅ ≤ 180°
∅ − ∅!!! − 360 𝑖𝑓 ∆∅ ≥ 180°
for
!
!
(4) 𝑝 𝑥 ≤ 𝑥 = ! and 𝑝 𝑥 ≥ 𝑥 = !
8
where p is the probability in the probability density function of the
elements of variable x.
2.3
Site Characteristics
HJA is situated in the North-central Oregon Cascades, near the town of
Blue River (Figure 3). HJA is a 15,815-acre drainage basin of Lookout and
McRae Creeks, which are tributaries of the Blue River, which flows into the
McKenzie River. Elevation within the HJA ranges from 410 to 1630 meters
above sea level with roughly 1220 m elevation difference between the lowest
point, near the headquarters’ complex at the southwestern edge, and the highest
point at Carpenter Mountain in the northwest corner (Figure 4). McRae Creek
flows southeast from its headwaters near Carpenter Mountain toward the Blue
River Reservoir. Roswell Ridge rises to an elevation of 1100 m between McRae
and Lookout Creeks that flow from the east, joining to become Lookout Creek at
44.233147 N, -122.206421 W, just below VALCEX McRae instrument location.
9
Figure 3: H. J. Andrews Map, Courtesy of H.J. Andrews LTER
Figure 4: Topographic map of H.J. Andrews showing SoDAR locations
10
2.4
Instrument Placement
One SoDAR array and one sonic anemometer, referred to as Primet, were
installed near the headquarters complex and its meteorological station and situated
at 44.211777 North, -122.255954 West, 443 meters (m) above sea level (asl). On
October 14, 2011, the SoDAR was installed on a compacted gravel pad with
access to electricity and an Internet connection. The sonic anemometer was
mounted on an instrumentation tower at 6.81 m above ground level (agl). As
shown in Figure 5, this location is at the bottom of Lookout Creek Valley, north
of the confluence of Lookout Creek and the Blue River Reservoir. North of
Primet is a ridge that runs northwest at an elevation of 600 m to 800 m. To the
south is a ridge running east-west which rises to 1,000 meters.
Figure 5: Primet location (44.211777 North, -122.255954 West)
showing surrounding elevations
11
The second SoDAR array was installed October 28, 2011, up McRae
Valley at 44. 24037 North, -122.19897 West, elevation 594 m asl. On November
8, the sonic anemometer was relocated to an instrument tower at 6.74 m agl. Due
to lack of line power at the McRae Valley site, data collection was limited to the
period between 18:00 and 8:00 Pacific Standard Time (PST). On March 12,
2012, a gasoline powered generator to charge batteries was installed that enabled
24-hour data collection. The McRae location is 161 m higher in elevation than
Primet. It is nestled in the McRae Creek Valley with a ridge to the south reaching
elevations just over 700 m. To the north is the Blue River Ridge that reaches
elevations of 1,300 m, turns west and rises to the summit of Carpenter Mountain
at 1,630 m, just off the map to the top right in Figure 6. Due west of McRae is
Roswell Ridge which climbs to 1,100 m before descending to Lookout Creek.
Figure 6: McRae location (44. 24037 North, -122.19897 West)
showing surrounding elevations
12
3
3.1
Analysis
VALCEX and HJA Climate Network Station Data Comparison
Initial analysis focused on validating wind measurements at the HJA
existing meteorological climate station, which includes a propeller anemometer.
In contrast to the propeller anemometer, a sonic anemometer is thought to be
more accurate since it has no moving parts to measure wind speed and direction,
and thus has no starting threshold. The motivation was to conduct a feasibility
study to determine if VALCEX observations could be extended to the long-term
historic data record measured in the HJA. The period of November 30 to
December 13, 2011, was selected for comparison because it was dominated by
weak, down-valley winds that represent one common mode of the local
circulation. Both instruments were mounted at almost the same height (sonic:
~ 8 m agl and propeller: 10 m agl and in close proximity to each other. The
comparison between the hourly wind speeds and directions measured with the two
different instruments showed that the flows measured with the sonic anemometer
were consistently stronger than those observed with the propeller anemometer
(Figure 7). The wind speed over the period November 30 to December 13
measured with the sonic anemometer averaged 0.35 ms-1, while that measured by
the propeller anemometer was zero, significantly less than the sonic anemometer.
The significant scatter in the readings of the propeller anemometer was an
additional concern that arose from overspeeding of the sensor during strong gusts
and scalar averaging in week-wind flow characterized by meandering. Wind
directions, however, agreed reasonably well with much less scatter between the
13
two different sensors (Figure 8). We therefore concluded that the historic wind
direction data collected from the propeller anemometer is meaningful and can be
used to extend the wind climatology, but the wind speeds cannot be used to
investigate the strength of the flows.
Sonic & Propeller Wind Speed
1
Propeller 10.0 m agl
SNC 7.2 m agl
0.9
0.8
Wind Speed [m s−1]
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
11 27
12 04
12 11
12 18
Month Day
Figure 7: Distribution of wind speeds of the sonic anemometer
and propeller anemometer at Primet
Westerly and North−Northeasterly Flows
450
W
NNE
405
Propeller 10 m agl [°]
360
315
270
225
180
135
90
90
135
180
225
270
315
360
405
450
Sonic 7.2 m agl [°]
Figure 8: Comparison of wind speed between the sonic
and the cup propeller
14
3.2
Sodargram Analysis of 94 days
This study classified wind speed and direction phenomena for the 94
nights between March 13 and June 23, 2012, through a visual analysis of
sodargrams for each 12-hour period centered around midnight. A sodargram is a
visual representation of aggregated data for each 5-minute period, with time on
the x-axis and each of the 41 gate heights on the Y-axis. The variable of interest
is shown by a colored pixel within this 2D coordinate system using a user-defined
color legend. Errors and weak signals are visualized in gray. Sodargrams are
visualizations commonly used by micrometeorologists; sodargrams for this study
were produced using MATLAB. For this study, wind speeds are represented on a
scale from black to yellow (Figure 9). Wind direction was represented on a
circular scale, with black indicating North, green representing West, blue
representing South, and red representing East (Figure 10).
15
Figure 9: Speed sodargram example
Figure 10: Directional sodargram example
Note: Color bar indicating black = N on both ends
16
As cold-air pooling occurs just after the sun has set, when net radiation
becomes negative, the classification focused on nighttime phenomena, the period
between 18:00 and 06:00. Four sodargrams: Primet Speed and Direction; and
McRae Speed and Direction, for each 12-hour period, centered on midnight, were
classified visually based on five phenomena (Table 1):
1.
2.
3.
4.
5.
Synoptic forcing (Strong or Weak)
Wind direction (NNE or SW)
Valley jet (Presence/Absense)
Pulsing (Presence/Absense)
Similar phenomena at both locations (Yes/No)
Each criterion was evaluated independently. The first four phenomena
were based on the visual review of Primet and the fifth criterion was based on
phenomena at both stations ((Figure 11).
(Figure 11: Diagram of decision-making process.
17
Table 1: Classification Criteria
Criteria:
Synoptic
Forcing
Direction
Classification
Strong = 1 Weak = 0
Elevation
50 - 300
Indicator:
Speeds > 5 ms-1
NNE = 1
SW = 0
0 – 50
Valley Jet
Present = 1
Absence = 0
100 – 200
Pulse
Present = 1
Absent = 0
50 - 100
Similar
Similar = 1
Dissimilar =
0
NA
Black & Red (NNE) or
Green & Blue (SW)
Speeds > 3 ms-1 with?
lower speeds above and
below
Pulse of speed and
direction
Yes/No
The category of strong or weak synoptic forcing was determined by
presence or absence of wind speeds at or above 5 ms-1 for more than 6 hours
during a 12-hour period. When wind speeds at or above 5 ms-1 were present, the
period was classified as strong (Figure 12). When wind speed remained below
5 ms-1 for the more than 6 hours of the 12-hour period it was classified as weak
(Figure 13).
18
Figure 12: Strong Synoptic Forcing classification example
Figure 13: Weak Synoptic Forcing classification example
19
At low wind speeds, wind direction changes frequently and drastically, as
can be seen in Figure 10 on page 16. To determine directional classification for a
12-hour period, heights between 0 and 50 m agl were evaluated to determine the
primary flow direction (Figure 14 and Figure 15).
Figure 14: NNE wind direction classification
Figure 15: SW wind direction classification
20
The third classification criterion is the presence or absence of a valley jet,
a band of higher wind speed at approximately 100 – 200 m agl. This phenomenon
does not fit the classic definition of low-level jet as the speeds are not higher than
10 ms-1. However, as this area is dominated by weak winds, a valley jet was
defined as winds at heights of 100 – 250 m agl with slows speeds above and
below. The valley jet can be seen as a band of speeds between 2.5 and 4 ms-1,
nested between lighter wind speeds, as shown in the circled area on Figure 16.
Figure 16: Valley jet classification example (circled)
The fourth phenomenon evaluated was the presence or absence of pulsing,
defined as wind directions alternating between 50 and 100 m agl. A pulse
typically starts at lower elevations, around 40 m agl, then rises to approximately
100 m agl, and can be seen as a pattern of directional shifts, Figure 17 and Figure
18.
21
Figure 17: Pulse wind direction classification example (circled)
Figure 18: Pulse wind speed classification example (circled)
22
The final classification criterion was similar or dissimilar. This criterion
indicates whether the two locations (McRae, Primet) are similar or dissimilar for
wind speed and direction. This criterion evaluates the connectivity within the
valley. When evaluating this criterion, all four sodargrams were visually
evaluated for similar phenomena for the Synoptic Forcing and Direction criteria.
3.3
Cases based on classification criteria
The 5 classification criteria create 32 potential combinations. The three
most commonly observed combinations were:
•
•
•
Case
1:
Weak
SF,
NNE
flows,
Valley
Jet,
No
Pulse,
Similar
(17
nights);
Case
2:
Weak
SF,
SW
flows,
No
Valley
Jet,
Pulse
Dissimilar
(13
nights)
Case
3:
Weak
SF,
NNE,
Valley
Jet,
Pulse,
Similar
(12
nights).
Table 2: Number of Nights by Case) shows the 32 possible cases and the
number of nights meeting criteria for each case. The cases with 4 or more nights,
which comprised 86, or 91% (86 of 94), of the nights classified were analyzed in
detail. The 13 cases with only 0 to 3 nights meeting the criteria were not analyzed.
These 13 cases comprised only 8 of the 94 nights classified (9%).
For each case analysis, below, it is noted whether wind speed and
direction are similar or dissimilar at each location, and whether (and how) those
measurements vary with height in m agl. For the nights meeting each case,
measurements for wind speed and direction were ensemble-averaged. Graphs for
the ensemble-averaged data are included with error bars showing standard
deviation, as in Figure 19. It is also noted whether the ensemble averaged speed
and direction agreed with the original assessment of similar or dissimilar for each
individual night.
23
Table 2: Number of Nights by Case
Case
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Synoptic
Forcing
Flow
Direction
Valley
Jet
Pulse
Similar
Number
of Nights
Weak
Weak
Weak
NNE
SW
NNE
Yes
No
Yes
No
Yes
Yes
Yes
No
Yes
17
13
Page
Number
27
28
12
29
Weak
Weak
Strong
Weak
NNE
SW
SW
SW
No
No
No
No
Yes
Yes
Yes
No
No
Yes
No
No
9
6
6
30
31
32
5
34
Weak
SW
Yes
Yes
Yes
Weak
Weak
Weak
NNE
SW
NNE
Yes
Yes
No
Yes
No
No
No
No
No
5
5
4
33
35
36
Weak
Strong
Strong
Strong
NNE
NNE
NNE
SW
Yes
Yes
No
No
No
No
Yes
No
No
Yes
Yes
No
4
3
3
2
1
37
N/A
N/A
N/A
N/A
Strong
Strong
Strong
Strong
Strong
SW
SW
SW
SW
NNE
No
No
Yes
Yes
No
No
Yes
No
No
No
Yes
Yes
No
Yes
No
1
1
1
1
N/A
N/A
N/A
N/A
Strong
Weak
Weak
Weak
Weak
Weak
Strong
Strong
NNE
SW
SW
SW
NNE
NNE
SW
SW
No
No
Yes
Yes
No
No
Yes
No
Yes
No
No
Yes
No
Yes
Yes
Yes
No
Yes
Yes
No
Yes
Yes
No
Yes
1
1
0
0
0
0
0
0
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
Strong
Strong
Strong
Strong
NNE
NNE
NNE
NNE
No
Yes
Yes
Yes
No
No
Yes
Yes
Yes
No
No
Yes
0
0
0
0
0
N/A
N/A
N/A
N/A
N/A
24
3.4
Mean wind climatology over study period: March 13 through June 13
Wind speed and direction data for all 94 nights between the hours of 18:00
and 06:00 were averaged by gate, i.e., 10 m height increments (Figure 19).
Directional averages between Primet and McRae are dissimilar, starting at
Southerly below 100 m agl, diverging above 100 m agl. Directional averages at
Primet are 180° at lower heights and shift with height to 90° between 100 and 200
m agl; above 200 m agl directional averages change to 0°. At McRae directional
averages at heights between 50 m agl and 100 m agl vary between 180° and 270°;
between 100 and 300 m agl, directional averages are 270°; above 300 m agl,
directional averages vary between 180 – 0°.
Wind speed averages are similar below 200 m agl at both stations. McRae
wind speed averages for the entire period are generally very weak, below 1 ms-1 at
gates below 195 m agl. Above 205 m agl speed averages increase and are
generally greater at the McRae station than at Primet; error bars show variation
increasing with elevation. Primet wind speed averages are very weak at all
heights. Near zero wind speed variation, shown by error bars, is the result of
opposing flows with similar magnitude, so the resultant vector speed is near zero.
25
94 Nights
400
Height [m]
300
200
100
0
−90
0
90
180 270 360 450
Direction [°]
U 94 Nights
400
Height [m]
300
200
100
Primet
McRae
0
−3 −2 −1 0 1 2 3 4 5 6 7 8
−1
Speed [m s ]
Figure 19: Speed and direction averages by height for study period (94 nights)
26
3.5
Case 1 – Weak SF NNE Valley Jet No Pulse Similar
Seventeen nights met the classification criteria for Weak SF, NNE flow,
Valley Jet, Pulse, Similar. The 17 nights were ensemble-averaged (Figure 20).
Wind directional averages for both stations were between 0° and 45° at gates
below 195 m agl. Above 205 m, directional averages shift to between 300° and
260°. Both McRae and Primet show similar wind speeds between 0 and 2 ms-1 at
all heights. McRae has greater average speed variability than Primet between 205
and 305 m agl. The valley jet is weak, with average wind speeds approximately
0.5 m faster between 95 and 165 m agl; this phenomenon can be seen in the
sodargram as increased speeds create a slight ‘nose’ between 75 and 105 m agl.
Weak NNE VJ NoPulse Similar (17 Nights)
400
Height [m]
300
200
100
0
−90
0
90
180 270 360 450
Direction [°]
U Weak NNE VJ NoPulse Similar (17 Nights)
400
Height [m]
300
200
100
Primet
McRae
0
−3 −2 −1 0 1 2 3 4 5 6 7 8
Speed [m s−1]
Figure 20: Speed and direction averages by height for Case 1
27
3.6
Case 2 – Weak SF SW No Valley Jet Pulse Dissimilar
Thirteen nights met the classification criteria for Weak SF, SW flow, No
Valley Jet, Pulse Dissimilar. Ensemble-averages for the 13 nights were plotted.
(Figure 21). Wind directional averages for these nights are different, with McRae
directional averages varying at the lower heights, turning to 280° above 275 m
agl. Primet directional averages vary at all heights, but tend towards 275° at
higher elevations. The two locations show dissimilar wind speed averages at
heights above 130 m agl, with speeds above 1 ms-1 and average wind speeds
increasing at McRae, while remaining below 2 ms-1 at Primet. Wind speed
averages increase in variability as height increases, with McRae having larger
standard deviation at all elevations.
Weak SW NoVJ Pulse Dissimilar (13 Nights)
400
Primet
McRae
Height [m]
300
200
100
0
−90
0
90
180 270 360 450
Direction [°]
U Weak SW NoVJ Pulse Dissimilar (13 Nights)
400
Height [m]
300
200
100
0
−3 −2 −1 0 1 2 3 4 5 6 7 8
Speed [m s−1]
Figure 21: Speed and direction averages by height for Case 2
28
3.7
Case 3 – Weak SF NNE Valley Jet Pulse Similar
Twelve nights met the classification criteria for Weak SF, NNE flow,
Valley Jet, Pulse, Similar, and were also ensemble-averaged (Figure 22). Wind
directional averages for McRae and Primet were between 0° and 45° below 295 m
agl, shifting to North at gates above 305 m agl. McRae directional averages
change from South to North and back to South as height increases. Variability of
wind direction averages is small in this case. Both McRae and Primet experience
average wind speeds at elevations below 305 m agl as weak, and less than
1.5 ms-1 at all gate heights.
Weak NNE VJ Pulse Similar (12 Nights)
400
Height [m]
300
200
100
0
−90
0
90
180 270 360 450
Direction [°]
U Weak NNE VJ Pulse Similar (12 Nights)
400
Height [m]
300
200
100
Primet
McRae
0
−3 −2 −1 0 1 2 3 4 5 6 7 8
Speed [m s−1]
Figure 22: Speed and direction averages by height for Case 3
29
3.8
Case 4 – Weak SF NNE No Valley Jet Pulse Dissimilar (9 nights)
Nine nights met the classification criteria for Weak SF, NNE flows, No
Valley Jet, Pulse, Dissimilar (Figure 23). Wind directional averages were
different at the two stations, with McRae shifting extensively at all gate heights
while Primet varied at heights below 95 m agl, then averaged Southerly at heights
above 115 m agl. Average wind speeds for both stations are below 1 ms-1 below
200 m agl. Above 205 m agl, speeds increase steadily to an average of 3 ms-1.
Weak NNE NoVJ Pulse Dissimilar (9 Nights)
400
Height [m]
300
200
100
0
−90
0
90
180 270 360 450
Direction [°]
U Weak NNE NoVJ Pulse Dissimilar (9 Nights)
400
Height [m]
300
200
100
Primet
McRae
0
−3 −2 −1 0 1 2 3 4 5 6 7 8
Speed [m s−1]
Figure 23: Speed and direction averages by height for Case 4
30
3.9
Case 5 – Weak SF, SW, No Valley Jet, Pulse, Similar (6 Nights)
Six nights met the classification Weak SF, SW, No Valley Jet, Pulse,
Similar (Figure 24). Wind directional averages are approximately 90° different at
the two locations. McRae averages South moving to North then back to South as
heights increase. Average speeds for both McRae and Primet were relatively
similar; however, a weak valley jet was seen at McRae, but not at Primet.
Weak SW NoVJ Pulse Similar (6 Nights)
400
Height [m]
300
200
100
0
−90
0
90
180 270 360 450
Direction [°]
U Weak SW NoVJ Pulse Similar (6 Nights)
400
Primet
McRae
Height [m]
300
200
100
0
−3 −2 −1 0 1 2 3 4 5 6 7 8
Speed [m s−1]
Figure 24: Speed and direction averages by height for Case 5
31
3.10 Case 6 – Strong SF, SW, No Valley Jet, Pulse, Dissimilar (6 Nights)
Six nights met the classification Strong SF, SW, No Valley Jet, Pulse,
Dissimilar (Figure 25). Wind directional averages at McRae are East at the lowest
gates, but shift to North at 45 m agl. As height increases, directional averages
move from North to East and back to North. McRae varies from North to West
and back again, while Primet directions averaged South as height increases. Wind
speed averages for McRae were below 1 ms-1 at lower gate heights with the
exception of the first measurement, 15 m agl. Average speeds increase above
235 m agl to greater than 1 ms-1. Wind speed averages at Primet increase with
gate height.
Strong SW NoVJ Pulse Dissimilar (6 Nights)
400
Height [m]
300
200
100
0
−90
0
90
180 270 360 450
Direction [°]
U Strong SW NoVJ Pulse Dissimilar (6 Nights)
400
Height [m]
300
200
100
Primet
McRae
0
−3 −2 −1 0 1 2 3 4 5 6 7 8
Speed [m s−1]
Figure 25: Speed and direction averages by height for Case 6
32
3.11 Case 7 – Weak SF, SW, Valley Jet, Pulse, Similar (5 Nights)
Five nights met the classification Weak SF, SW, Valley Jet, Pulse, Similar
(Figure 26). Directional averages for the two stations were approximately 90°
different below 45 meters agl, but were very similar at heights above 55 m agl.
Wind speeds at the two locations were relatively similar. A valley jet can be seen
at both locations at 105 m agl.
Weak SW VJ Pulse Similar (5 Night)
400
Height [m]
300
200
100
0
−90
0
90
180 270 360 450
Direction [°]
U Weak SW VJ Pulse Similar (5 Night)
400
Primet
McRae
Height [m]
300
200
100
0
−3 −2 −1 0 1 2 3 4 5 6 7 8
Speed [m s−1]
Figure 26: Speed and direction averages by height for Case 7
33
3.12 Case 8 – Weak SF, SW, No Valley Jet, No Pulse, Dissimilar (5 Nights)
Five nights met the classification Weak SF, SW, No Valley Jet, Pulse,
Disimilar (Figure 27). McRae directional averages vary at lower heights,
becoming southerly at above 100 m agl and speeds increase steadily above 205 m
agl. Primet speed and directional averages were consistent at all heights. Standard
deviation indicates variability in speed particularly at lower heights at McRae.
Weak SW NoVJ NoPulse Dissimilar (5 Nights)
400
Height [m]
300
200
100
0
−90
0
90
180 270 360 450
Direction [°]
U Weak SW NoVJ NoPulse Dissimilar (5 Nights)
400
Height [m]
300
200
100
Primet
McRae
0
−3 −2 −1 0 1 2 3 4 5 6 7 8
Speed [m s−1]
Figure 27: Speed and direction averages by height for Case 8
34
3.13 Case 9 – Weak SF, NNE, Valley Jet, Pulse, Dissimilar (5 Nights)
Five nights met the classification Weak SF, SW, Valley Jet, Pulse,
Disimilar (Figure 28). Directional averages at McRae vary at low heights and
becomes similar to Primet above 55 m agl. Average speed for both locations was
similar and both stations experience a valley jet at approximately 155 m agl.
Weak NNE VJ Pulse Dissimilar (5 Nights)
400
Height [m]
300
200
100
0
−90
0
90
180 270 360 450
Direction [°]
U Weak NNE VJ Pulse Dissimilar (5 Nights)
400
Height [m]
300
200
100
Primet
McRae
0
−3 −2 −1 0 1 2 3 4 5 6 7 8
Speed [m s−1]
Figure 28: Speed and direction averages by height for Case 9
35
3.14 Case 10 – Weak SF, SW, Valley Jet, No Pulse, Dissimilar (4 Nights)
Four nights met the classification Weak SF, SW, Valley Jet, No Pulse,
Dissimilar (Figure 29). McRae directional averages were East at the first two
heights, and then changed to North until 245 m agl, while Primet directional
averages vary at low heights. Below 245 m agl, average speeds for both locations
were below 1 ms-1 except the first McRae gate, which was 2 ms-1. McRae showed
a valley jet between 155 and 205 m agl, which was not seen at Primet.
Weak SW VJ NoPulse Dissimilar (4 Night)
400
Primet
McRae
Height [m]
300
200
100
0
−90
0
90
180 270 360 450
Direction [°]
U Weak SW VJ NoPulse Dissimilar (4 Night)
400
Height [m]
300
200
100
0
−3 −2 −1 0 1 2 3 4 5 6 7 8
Speed [m s−1]
Figure 29: Speed and direction averages by height for Case 10
36
3.15 Case 11 – Weak SF, NNE, No Valley Jet, No Pulse, Dissimilar (4 Nights)
Four nights met the classification Weak SF, NNE, No Valley Jet, No
Pulse, Dissimilar (Figure 30). This case shows variation in both wind speed and
directional averages at all heights. Even though only 4 nights met this
classification, this is an interesting case because of the very strong flows only at
McRae.
Weak NNE NoVJ NoPulse Dissimilar (4 Nights)
400
Primet
McRae
Height [m]
300
200
100
0
−90
0
90
180 270 360 450
Direction [°]
U Weak NNE NoVJ NoPulse Dissimilar (4 Nights)
400
Height [m]
300
200
100
0
−3 −2 −1 0 1 2 3 4 5 6 7 8
Speed [m s−1]
Figure 30: Speed and direction averages by height for Case 11
37
4
Discussion
As the primary goal of this work was to characterize wind flow within the
McRae Valley, we conclude the thesis with a discussion of connectivity and ask
whether the methods used (1) appear to be good predictors of connectivity and (2)
whether cases could be combined in order to facilitate the characterization of
airflow.
Connectivity between McRae and Primet is said to occur when wind flows
down the McRae Creek Valley to the Blue River Reservoir. The two SoDAR
locations were determined to be connected when both McRae and Primet
experienced similar mean wind speed and directional profiles. This occurred
most often when the dominant direction was from North-northeast, when wind
from the North is able to travel the length of the McRae Creek uninterrupted into
Lookout Creek. During periods of weak synoptic direction and dominant wind
direction from the Southwest, wind flow within McRae was disconnected at the
confluence of the two creeks, McRae and Lookout.
During the study period, the valley was dominated by weak synoptic
forcing. Wind direction flows under two dominant regimes, downvalley from the
North-northeast, or upvalley from the Southwest. The valley jet was present 49 of
the 94 nights. Pulses were present 56 nights, generally when the dominant
direction was from the North-northeast. There were 45 nights with similar
phenomena at the stations (Table 3).
38
Table 3: Number of nights meeting each criterion
Criterion
Synoptic Forcing
Wind Direction
Valley Jet
Pulsing
Similar
Classification
Weak
SW
Absent
Absent
Similar
# Nights
79
42
45
38
49
Classification
Strong
NNE
Present
Present
Dissimilar
# Nights
15
52
49
56
45
When combined with the category of Synoptic Forcing, the study period
had 38 nights of strong synoptic forcing with flow coming from the Southwest, 14
nights of strong synoptic forcing and flows coming from the North-northeast, 11
nights of weak synoptic forcing with flows coming from the Southwest, and 31
nights of weak synoptic forcing with flows from the North-northeast (Table 4).
Table 4: Number of nights meeting each criterion
Criterion
Strong Synoptic
Forcing
Weak
Classification
# Nights
SW
SW
38
11
Classification
NNE
# Nights
14
NNE
31
In the 11 cases evaluated, weak synoptic forcing dominated with only one
case showing strong synoptic forcing (Table 5); this demonstrates that the two
SoDARs generally experience wind speeds slower than 2 ms-1, so speed is not a
good indicator of connectivity. Weak synoptic forcing with wind flow direction
from the Southwest occurred in 6 of the 11 cases; however, these cases had fewer
nights, 33 of the 94 or 35%. Five cases were classified as weak synoptic forcing
with wind flow from the North-northeast, which represents 47 nights or 5% of the
nights.
39
As described above, connectivity was inferred from the classification of
the five criteria in 5 of the 11 cases. Of the five criteria, connectivity is most
closely related to dominant wind direction, with connection occurring generally
when wind direction is from the North-northeast. The large variation in direction
that was noted for several cases appears to be due in part to the weak wind
regime. There also seems to be a relationship between connectivity of the valleys
and the presence of the valley jet, which occurred in every case of valley
connectivity and appeared in only one case of non-connectivity. It is also
interesting to note that in 3 of the 5 cases where the sodargrams were classified as
similar, the valleys appear connected.
The results are noteworthy when the last two criteria, pulsing and similar,
are removed (Table 5). In cases 1, 3 and 9, the sodargrams were all classified as
weak synoptic forcing with winds from the North-northeast with a valley jet
(highlighted in yellow). When these three cases are combined, 35, or 36%, of the
nights in the study period were classified the same in the three remaining criteria.
What is more interesting is that connectivity within the valley can be inferred in
each of these cases. There are two cases of weak synoptic forcing with winds
from the North-northeast that were classified as valley jet absent, cases 4 and 11
(highlighted in blue). The two SoDAR locations appeared to be disconnected on
these 13 nights, or 14% of the study period.
Of the 5 cases of wind from the Southwest, there were 3 cases of weak
synoptic forcing with no valley jet, cases 1, 5 and 7 (highlighted as pink). These
three cases, which comprised 19 nights, or 20% of the study period, were
40
evaluated as disconnected. The two cases of weak synoptic forcing with wind
flow from the Southwest with a valley jet, highlighted in blue, comprised 13
nights, or 14% of the study period. For both these cases, the valley was evaluated
as disconnected.
Table 5: Connectivity by Case
Case
Number
Synoptic Flow
Forcing Direction
Valley
Jet
Pulse
Similar
Number
of Nights
Valleys
Connected
1
2
3
4
5
6
7
8
9
10
11
Weak
Weak
Weak
Weak
Weak
Strong
Weak
Weak
Weak
Weak
Weak
Yes
No
Yes
No
No
No
No
Yes
Yes
Yes
No
No
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
No
No
Yes
No
Yes
No
Yes
No
No
Yes
No
Yes
No
17
13
12
9
6
6
5
5
5
4
4
Yes
No
Yes
No
No
No
No
Yes
Yes
Yes
No
5
Conclusions
5.1
NNE
SW
NNE
NNE
SW
SW
SW
SW
NNE
SW
NNE
VALCEX and HJA Climate Network Station Data Comparison
With regard to the first objective of this study, the comparison between the
SoDAR and the wind cup installed at Primet for the past 20 years, we conclude
that the historic wind direction data collected from the propeller anemometer is
meaningful and can be used to extend the wind climatology, but the wind speeds
cannot be used to investigate the strength of the flows.
41
5.2
Valley Connectivity
With regard to the second objective of this study, which analyzed the
connectivity of the flow within the valley, the valley was considered connected on
88 of the 96 nights classified. The two stations, McRae and Primet experience the
criteria synoptic forcing as weak on 79 of the 96 study nights. The second
criterion wind direction was classified almost equally at each station; 52 nights
were classified as from the North-northeast while 42 nights were classified as
from the Southwest. The criterion of valley jet present was determined visually
on 49 of the nights. Three of the 5 classification criteria – synoptic forcing,
direction and valley jet – were most effective in determining connectivity within
the greater McRay - Lookout Creek Valley.
Wind speeds at both McRae & Primet are generally very weak, below
1 ms-1 at heights below 195 m agl. Above 205 m agl speeds increase and are
generally greater at the McRae station than at Primet; error bars for each gate
show variation increasing with elevation. Directional averages show a disconnect
between Primet and McRae, with averages at Primet being South at lower
elevations and changing to East between 95 and 205 m agl and North-northeast
above 205 m agl. At McRae, on the other hand, flows above 45 m agl and below
105 m agl vary between South and West, moving to from the West between 115
and 305 m agl; above 305 m agl, flows again vary between South, Southeast and
North. Near zero wind speed variation, shown by error bars, is the result of
opposing flows with similar magnitude, so the resultant vector speed is near zero.
42
5.3
Cases based on 5 Classification Criteria
Of the 5 criteria used to classify each sodargram of a night of the study
period two – flow direction and valley jet – were most effective in determining
connectivity within the McRae Valley, given that weak synoptic forcing
dominated the study period. The classification of Synoptic Forcing would be
effective if the valley were not dominated by weak winds.
The problem was essentially a classification problem. Could the 96 daily
sodargram analyses by 3, 4 or 5 criteria be input to a machine-learning tool, to
produce algorithms that would correctly analyze sodargrams of other
topographically similar locales? Or is a classification based on a visual review of
the raw data in the form of a sodargram effective? While the study may have
been more robust had the researcher used multiple visualization techniques, it is
apparent that 3 of the 5 criteria were effective as determining connectivity.
6
Future Work & Recommendations
This study may have been more robust had the classification criteria been
limited to two or three phenomena: synoptic forcing and dominant wind direction,
with possible inclusion of the valley jet. Limiting the number of phenomena
classified would generate fewer cases, which would create a stronger relationship
between phenomena and valley connectivity. Further, a smaller number of
phenomena evaluated might have allowed for multiple evaluations of the
sodargrams, which in turn would have allowed for a comparison of each
43
evaluator’s set of evaluations. Such an analysis might strengthen the validity of
the visual classification process.
Wind direction variability typically accompanies weak winds. The
classification of wind direction was limited by using only two dominate
directions, NNE and SW. Using four directional categories – NW, NE, SW, and
SE – would allow for classifying sodargrams with wind directions not easily
grouped into NNE or SW. It would be interesting to evaluate how many of the 94
nights of the study period were not easily classified into NNE or SW.
The next step in this study would include analysis of multiple visualization
techniques. Perhaps comparing the visual analysis of sodargrams to windroses, or
other visualization tools. A study of that type could be extremely interesting in
determining which methods of visualization most effectively display and
communicate this data-rich information. Previous work for the Visualizing
Terrestrial and Aquatic Systems (VISTAS) project resulted in reviewing hundreds
of visualizations from four hydrology journals(Cushing et al., 2012). During that
review, visualizations and modeling software used by hydrologists were noted
with interest. It would be extremely helpful and telling to input the Primet and
McRae daily and aggregated data into tools that could portray the physical
phenomena (wind direction and speed) in three dimensions superimposed onto the
topographic surface, and animate these data through time. It would be interesting
to explore how visualization techniques used in other disciplines, such as
hydrology, that also study fluids in motion could be used to visualize (then
44
characterize and qualitatively or quantitatively analyze) micrometeorological
phenomena.
45
7
Bibliography
Arya, P. S. (2001). Introduction to micrometeorology. London, Academic Press.
Clements, C. (2002). "Cold-Air-Pool Structure and Evolution in a Moutain Basin:
Peter Sinks, Utah." Journal of Applied Meteorology and Climatology 42.
Cushing, J., E. Hayduk, J. Walley, L. Zeeman, N. Stevenson-Molnar, K. Winters,
D. Lach, M. Bailley, B. J. Bond, J. Bolte, C. K. Thomas and S. Stafford
(2012). How Ecologists ‘Visualize’ Research in Publications: VISTAS
Project Ecology Journal Survey. Poster presented at Ecological Society of
America Annual Meeting. Portland, Or., Ecological Society of America.
Dorninger, M., C. D. Whiteman, B. Bica, S. Eisenbach, B. Pospichal and R.
Steinacker (2011). "Meteorological Events Affecting Cold-Air Pools in a
Small Basin." Journal of Applied Meteorology and Climatology 50(11):
2223-2234.
Folken, T. (2008). Micrometeorology. Verlag Berlin Heidelberg, Springer.
Lundquist, J. D., N. Pepin and C. Rochford (2008). "Automated algorithm for
mapping regions of cold-air pooling in complex terrain." Journal of
Geophysical Research 113(D22).
Mahrt, L. (2008). "Mesoscale wind direction shifts in the stable boundary layer."
Tellus Series A-Dynamic Meteorology and Oceanography 60A: 700-705.
Mahrt, L., S. Richardson, N. Seaman and D. Stauffer (2010). "Non-stationary
drainage flows and motions in the cold pool." Tellus A: no-no.
Mayer, J.-C. (2005). Characterization of the atmospheric boundary layer in a
complex terrain using SODAR-RASS. Master of Science in Geoecology,
University of Bayreuth.
Smith, S. A., A. R. Brown, S. B. Vosper, P. A. Murkin and A. T. Veal (2009).
"Observations and Simulations of Cold Air Pooling in Valleys."
Boundary-Layer Meteorology 134(1): 85-108.
Smoot, A. R. (2012). Linking dynamics of the near-surface flow to deeper
boundary layer forcing in the nocturnal boundary layer. Masters of
Science, Oregon State University.
Thomas, C. K., A. M. Kennedy, J. S. Selker, A. Moretti, M. H. Schroth, A. R.
Smoot, N. B. Tufillaro and M. J. Zeeman (2012). "High-Resolution FibreOptic Temperature Sensing: A New Tool to Study the Two-Dimensional
Structure of Atmospheric Surface-Layer Flow." Boundary-Layer
Meteorology 142(2): 177-192.
Whiteman, C. D. (1990). "Observations of thermally developed wind systems in
mountainous terrain." Atmospheric processes over complex terrain,
Meteor. Monogr 23(45): 5-42.
46
Whiteman, C. D. and S. Zhong (2008). "Downslope Flows on a Low-Angle Slope
and Their Interactions with Valley Inversions. Part I: Observations."
Journal of Applied Meteorology and Climatology 47(7): 2023-2038.
Whiteman, C. D., S. Zhong, W. J. Shaw, J. M. Hubbe and Z. Bian (2001). "Cold
Pools in the Columbia Basin." Weather and Forcasting 16.
47