N PS ARCHIVE

1968
ORTENGREN, R.

SURFACE CHARATERISTICS OF WINDROWS

by

Ralph William Ortengren, Jr,

**!l°3 '<K>|>poi5 SS2

*A 'N 'asnsoMs ■

M3QNia J13HS

DUDLEY KNOX LIBRARY

NAVAL POSTGRADUATE SCHOOL

MONTEREY CA 93943-5101

UNITED STATES
NAVAL POSTGRADUATE SCHOOL

THESIS

SURFACE CHARACTERISTICS OF WINDROWS

by

Ralph William Ortengren, Jr.

December 1968

Tkl& docwwvt ka6 been approved fan. pub tic n.z-
tza&z and 6alz; <LU diiVUbution i& untirruXzd,

MONTEREY. CALIF. »

SURFACE CHARACTERISTICS OF WINDROWS

by

Ralph William Ortengren, Jr.

Lieutenant, United States Navy

B. B.A. , University of Michigan, I960

Submitted in partial fulfillment of the
requirements for the degree of

MASTER OF SCIENCE IN OCEANOGRAPHY

from the

NAVAL POSTGRADUATE SCHOOL
December 1968

ABSTRACT

Aerial photographs were taken of windrow accumulations in
Monterey Bay on 1, 8, 15 and 22 October 1968. A Fairchild T-ll
aerial mapping camera was used, with photographs taken approxima-
tely every two minutes over 40 to 60 minute periods. Windrows
were marked with accumulations of computer cards, wind speed
measured by cup anemometer, and wind direction taken with the
aid of a MK. 6 Smoke Float. Sea surface temperature, depth of
the thermocline, and surface air temperature measurements were
taken concurrently.

An attempt was made to correlate windrow spacing and wind
speed, to find mean deflection of windrows relative to the wind, to
determine any relationship between row spacing and depth of the
thermocline, and to find the response time of windrow orientation
to a wind shift.

Windrow spacing was found to depend on other factors than
wind speed. Deflection angles varied between 20 left and 20° right,
with 0° being the most common angle. No correlation was found
between depth of the thermocline and row spacing. Response time
fell between two and four minutes.

TABLE OF CONTENTS

CHAPTER PAGE

1 7

Introduction 7

Motivation. 12

II. EQUIPMENT AND INVESTIGATIVE PROCEDURE . . 13

Equipment , 13

Sampling technique and dates 14

III. PRESENTATION AND DISCUSSION OF DATA .... 17

Daily results 18

Discussion of statistical significance . 24

Individual photographs . 27

IV. CONCLUSIONS AND RECOMMENDATIONS 45

Conclusions 45

Recommendations 47

BIBLIOGRAPHY 50

APPENDIX I Additional references 51

APPENDIX II Program for least-squares fit 53

INITIAL DISTRIBUTION 54

LIST OF FIGURES

FIGURE PAGE

1. Least-Squares Fit, First Degree, for Spacing

versus Wind Speed 29

2. Least-Squares Fit, Second Degree, for Spacing

versus Wind Speed 30

3. Least-Squares Fit, Third Degree, for Spacing

versus Wind Speed 31

4. Bathythermograph Traces for 1, 8, 15, 22 October . . 32

5. Spacing Histograms for 1, 8, 15 October 33

6. Deflection Histograms for 1, 8, 15 October 34

7. S/D Histograms for 1, 8, 15 October 35

8. Histogram of Total Spacing Data • • 36

9. Histogram of Total Deflection Data 37

10. Histogram of Total S/D Data 38

11. Photograph 81 . . 39

12. Photograph 87 40

13. Photograph 155 .... 41

14. Photograph 158 42

15. Photograph 1 59 ............ 43

16. Photograph 1519 44

ACKNOWLEDGMENTS

I wish to first thank my advisor, Professor Noel E. Boston, of
the Department of Oceanography, for his timely and numerous
critiques during the preparation of this thesis. The finished product
would not have been possible without his providing the proper sense
of direction and perspective over a period of nearly seven months.

Secondly, I am indebted to the many naval aviators attached to
the Naval Postgraduate School, who willingly flew in circles for an
hour at a time in order to take photographs of a blob of computer
cards floating in Monterey Bay.

I would like to express my special appreciation for the help
given me by Professor Ted Green of the Department of Oceanography.
His efforts were instrumental in initially defining the problem and
formulating the method of attack in order to obtain meaningful results.

Finally I must acknowledge the untiring typing efforts of my
fiance, Jean Kennedy. Through numerous drafts, she maintained
a sense of calm and order that enabled the author to complete this
thesis despite his innate frustrations.

CHAPTER I

I. INTRODUCTION

Lines of confluence, denoted by foam, seaweed, or other
floating material, are commonly observed on the surface of lakes
and oceans. These lines are commonly called "windrows" or "wind
slicks". They will be referred to as "windrows" hereafter in the
body of this thesis. Windrows are thought to be formed by the
horizontal convergence of surface waters, which then descend leaving
any less dense, floating matter behind. Winds appear to cause
windrows by setting up horizontal vortices in the upper waters of
lakes and oceans. The axes of these vortices are roughly parallel
to the wind, and adjacent vortices rotate in opposite directions.
The existence of this particular situation is often indicated by the
presence of lines of foam parallel to the direction of the wind, and
lying directly above the region of downflow between two adjacent
vortices.

Besides wind induced confluences, other flow conditions may
give rise to lines of convergence. Stommel (1951) observed that a
line often forms some distance from shore, where the surface water
descends and flows out again at some depth. This occurs along
shorelines where surface currents have an onshore component.
Ewing (1950) found that where conditions cause the formation of
internal waves on the thermocline, the internal waves are sufficient

to cause lines of confluence on the surface. These lines propagate
along with the internal waves. Such confluences may or may not
form in the presence of wind. They are usually easily distinguish-
able from windrows since they tend to parallel coasts, are often well
defined over several miles, are separated by fairly large distances,
and finally are independent of wind direction provided wind speed
is low. They are most easily seen in low to still wind conditions.
They are caused by processes originating within the body of the
ocean.

The mechanism of windrow formation is much different. The
triggering agency seems to be the external stress exerted by the
wind on the sea surface. The goal of this thesis is to present an
analysis of the surface characteristics of these wind-formed lines
of confluence.

Langmuir (1938) first noticed seaweed arranged in parallel
lines at sea in 1927. These lines seemed to be oriented parallel to
the wind. Through experiments on Lake George, Langmuir found
several significant features. The streaks were caused by the wind
on the water setting up longitudinal surface currents in the direction
of the wind. The wind produced a series of alternating right and
left helical vortices in the water having horizontal axes parallel to
the wind. On the water's surface, water converges toward the
streaks, and dives under them. Between the streaks, there are
rising currents, which flow out laterally toward the streaks upon

reaching the surface, Langmuir' s results led him to believe that
the helical vortices were bounded on the bottom by the depth of the
thermocline. In addition, he postulated that there might be some
relationship between the depth of the thermocline and the spacing of
the windrows.

At the present time the exact mechanism which generates
Langmuir circulation is not known. Three possible mechanisms are
presented in the literature, . . . shear flow instability, surface films,
and thermal convection. Faller (1963) demonstrated that cellular
rolls were formed in laminar flow due to instabilities of the Ekman
boundary layer. By analogy he has suggested that shear flow insta-
bility is a possible generating mechanism for Langmuir circulation.
As supporting evidence, Faller and Woodcock (1964) found that
windrows are not exactly parallel to the wind, but lie at some small
angle to the right of it.

Welander (1963) proposed a theory for Langmuir circulation,
in which the motions are generated by the shearing effect of the wind
upon an organic surface film of varying thickness. Stommel (1951)
also suggested that the generating mechanism for Langmuir circula-
tion was confined to a relatively thin surface layer because of the
rapid response of windrow alignment to changing wind direction.

Csanady (1964) thought that the presence of windrows was
invariably associated with a rapid vertical diffusion, On this basis,
he assumed that Langmuir vortices were the principal causes of

the vertical mixing of epuimnion waters. During the summer of
1965, Csanady had boat crews perform routine observations of wind-
rows on Lake Huron. His general conclusions confirmed the findings
of Faller and Woodcock, i. e. , that windrows are not correlated with
upward heat flux. He proposed that a convective, surface-cooling
mechanism is at least a reasonable cause of Langmuir vortices. A
doctoral candidate at the University of Wisconsin, P. D. Uttormark,
is presently working on a thermal convection model of Langmuir
circulation, but his results are not available at this time.

Recently McLeish (1968) has suggested that no other factor than
wind induced water turbulence is necessary for the formation of wind-
row patterns. He further stated that the windrow patterns, although
showing a dominance of lines parallel to the wind, are really con-
tinually changing networks with irregular line widths, spacings and
directions. Although his paper is generously provided with excellent
photographs, there is a dearth of quantitative results to support his
conclusions.

Very little work has been done since Langmuir concerning the
surface characteristics of windrows. Csanady (1965) reported on
246 observations of windrows, but noted only the basic occurrence
of windrows versus upward heat flux. He did notice that only once
were windrows observed below a wind speed of 2. 1 meters/second.
Csanady noted that windrows were much more noticeable and easier
to identify in the presence of surface evaporation. Faller (1964)

10

made 25 observations of windrows, 12 from a small boat and 13 by-
aerial photography. He marked the windrows with several thousand
sheets of 8-1/2 x 11 inch paper. With this small data base, he found
an average angle of deflection of the windrows to be 12.9 degrees to
the right of the wind, with a standard deviation of 6. 6 degrees.
Faller and Woodcock (1964) used a data base of only 14 samples to
postulate the dependence of windrow spacing upon wind speed as:
spacing in meters equals 4, 8 times the wind speed in meters/second.
It was during this project that Faller and Woodcock attempted to find
a correlation between the occurrence of windrows and upward heat
flux. They found no such correlation. There was in addition only
a small range of physical conditions present during the sampling,
and seasonal and latitudinal variations were not tested. Sullivan
(1964) made 27 sampling runs on Lake Huron using a small boat. He
obtained spacing measurements by dragging a marked polypropylene
rope behind the boat and driving through the windrows perpendicularly.
He found the windrows to be roughly parallel to the wind and inde-
pendent of wave direction. He observed no windrows below a wind
speed of 3, 5 meters/second. He also noted the lack of occurrence
of windrows on several occasions when the wind speed was signif-
icantly greater than 3, 5 meters/second. In addition Sullivan measured
14 samples of windrow spacing versus depth of the thermocline. The
ratios found varied between 1. 0 and 1. 72, indicating some relation-
ship may exist, as Langmuir had forecasted.

11

II. MOTIVATION

Prior to this thesis, investigations of the surface characteristics
of windrows have suffered from two important inadequacies. Fore-
most has been the lack of a sufficiently large number of observations.
As stated earlier, the empirical results of Faller, Woodcock, and
Sullivan were based on samples ranging between 14 and 27 total
observations. Secondly, and perhaps equally important, has been
the lack of adequate measuring equipment. Spacing measurements
were taken by "seaman's eye". Faller used aerial photography for
1 3 of his samples, but the camera was a hand-held movie camera
which was not described.

The intent of this thesis is to present a study of windrows based
on an adequate number of observations to be statistically significant,
and with measurement of length, spacing and direction inherently
more accurate than any previous work. The windrow characteristics
investigated include the following:

(1) Dependency of windrow spacing on wind speed.

(2) Angle of deflection of windrow orientation from the direction
of the wind.

(3) Response time of windrow orientation to a wind shift.

(4) Relationship(s) between windrow spacing and the depth of
the thermocline.

12

CHAPTER II
EQUIPMENT AND INVESTIGATIVE PROCEDURE

I. EQUIPMENT

The core of the data acquisition system was a Fairchild T-ll
Aerial Mapping Camera. This is the primary camera used by the
U. S. Air Force and U. S. Navy for accurate mapping, and has an
accuracy of linear measurement to one part within 100, 000. Coupled
with this is the feature of extremely good resolution of small-scale
objects. The camera was mounted in a U. S. Navy S-2 Grumman
aircraft attached to the Naval Auxiliary Landing Field, Monterey.
The camera uses a lens of fixed focal length and is mounted such that
if the aircraft is in a level flight attitude, the lens is pointed directly
down. As the shutter is tripped, the aircraft's altitude (by pressure
altimeter) and the time of the shutter trip (by instrument panel clock)
are printed on the margin of the negative. The camera operator
has only to note the aircraft heading at the time of shutter release.
Automatic film advance allows a picture to be taken every two sec-
onds. The responsibility of the pilot is to maintain as level an
attitude as possible during the actual photography, After the photo
runs, the camera's magazine is unloaded and the film airmailed to
the Mirmar Naval Air Station for development into negatives. Due
to the unofficial status of the project, contact prints of the negatives
were not produced.

13

Surface wind speeds were obtained with a Cassela cup anemometer.
The anemometer uses the number of revolutions of the cup assembly
per minute to give an accurate average wind speed over the previous
minute. Each instrument is provided with its own calibration sheet
of rpm versus wind speed.

Thermal structure was obtained by means of a standard U. S.
Navy shallow water bathythermograph. Bucket thermometer readings
were taken concurrently to obtain some estimate of the sea surface
temperature. An electronic psychrometer was used in conjunction
with a standard dry bulb thermometer.

With the exception of the camera, all instruments were mounted
on the Naval Postgraduate School's 63 foot oceanographic research
boat. The marking of windrows was accomplished by dropping
3-1/2 x 7-3/8 inch computer cards from the aircraft and from the
boat. Approximately 3, 000 cards were used during each sampling
period. Individual cards could be distinguished in the photographs
made by the T-ll camera from a height of 300 feet.

II. SAMPLING TECHNIQUE AND DATES

A complete sampling period consisted of twenty or more
photographs of a given windrow accumulation. The time interval
between photographs was roughly two minutes.

The boat would position itself in a likely area for windrow
occurrence in Monterey Bay, approximately two miles northwest of
Monterey Marina. Voice communications would be established with

14

the aircraft, and shipboard instrumentation rigged. Upon arrival
overhead, the aircraft would fly directly over the boat on a heading
upwind and release the load of computer cards from the bomb bay
at an altitude of approximately ZOO feet. The cards would immediate-
ly separate one from another and fall in a completely random pattern.
Moving around the periphery of this pattern, the boat would plant a
standard U. S. Navy MK. 6 Smoke Float slightly upwind of the com-
puter card pattern. The boat would then stand off downwind of the
pattern, commence taking wind speed measurements and make a
BT cast. Upon sighting the smoke from the float, the aircraft
would commence flying a racetrack pattern at an altitude of 300 feet,
so as to pass over the smoke float and computer card pattern roughly
every two minutes, The photographs would be taken directly over-
head and the aircraft's heading and exact altitude noted. Any
necessary changes to the flight pattern or any other segment of the
sampling run could be communicated to all parties by means of
frequency modulated transceivers located in the aircraft and in the
boat. On the boat, the anemometer was firmly attached near the
bow at a height of six feet four inches above the waterline. All
wind speed readings were obtained with the boat lying to with all
protrubances of ship structure being downwind of the anemometer.
Wind direction was obtained from the photographs of the smoke float
and its plume. Upon completion of approximately twenty photo runs,

15

a second bathythermograph reading was taken, and the day's
investigation concluded.

Sampling runs were conducted in Monterey Bay on 1, 8, 15
and 22 October 1968. Further runs were precluded due to inclement
weather and aircraft availability.

16

CHAPTER III
PRESENTATION AND DISCUSSION OF DATA

Data derived from the four sampling runs is presented and dis-
cussed on the following pages. Initially each day's results are
shown in an individual table. Following each table is a descriptive
analysis of the prevailing meteorological and bathythermometric
conditions during the sampling process. Following the tables is a
discussion of symbols used for the definitions of parameters. The
chapter concludes with graphs and discussions concerning the cor-
relation of the various windrow characteristics. The actual photo-
graphic negatives are included with this report under separate
enclosure. Final analysis and conclusions are presented in the next
chapter.

17

1 October 1968

PHOTO NO. TIME WIND NO. ROWS SPACING DEFL. S/ D

12 1516 4.2 10 135.0 4.36

13 1518 3.3 12 82.5 2.66

14 1520 3.4 5 103.2 3.34

15 1522 3.7 4

16 1524 2.8 5 90.0 3°L 2.90

17 1526 3.0 5 121.9 9°R 3.92

18 1528 3.1 11 140.7 7°R 4.53

19 1530 3.0 8 12CR

110 1532 2.7 6 100.0 4°R 3.22

111 1534 2.6 3 123.8 3°R 3.99

112 1536 2.6 3 8°R

113 1538 2.8 6 145.0 5°R 4.68

114 1540 2.5 5 110.5 4°R 3.57

115 1542 2.4 —

116 1544 2.6 4

117 1546 2.2 2

118 1548 2.2 3

119 1550 2.5 5

120 1552 2.1 3

121 1554 2.1 1

122 1556 1.9 4

Thermocline remained fixed at 31 feet during observational period. T \ T

s-^ w

during entire period. Data taken in 250 feet of water; visibility unlimited;
clear sky; 5 foot swell from southwest.

4°R

12 °R

9°R

5°L

2CR

8 October 11)68

PHOTO NO.

TIME

WIND

NO. ROWS

SPACING

81

1450

4.8

82

1452

4.1

83

1454

4. i

84

1456

4.8

5

85

1458

5.1

6

106.9

86

1500

4.3

4

37

1502

4. 2

9

88

1504

4.5

3

433.0

89

1506

4 6

6

472 0

810

1508

4.1

6

811

1510

4.2

11

150.0

812

1512

4.0

9

178.1

813

1514

3.5

2

814

1516

2.8

8

150.0

815

1518

2.9

9

178.1

816

1520

3.1

7

178.1

817

1522

2.4

8

818

1524

3.3

9

142.6

819

1526

4.1

9

820

1528

4.0

6

821

1530

3.7

8

DEFL. S/D

4.11

0°
0°
8°R

0°

5°R

16.70

0° 18.10

4°L 5.77
6.81

6°L

0° 5.78

0° 6.85

6.85

20°L 5.49

0°

L2°L

13:L

Thermocline fixed at 26 feet during observation period. TXT

s/ w

d uring same period. Data taken in 270 feet of water: visibility unlimited;
clear sky; calm sea (no detectable swell).

19

15 October 1968

PHOTO NO.

TIME

WIND

NO. ROWS

SPACING

DEFL.

S/D

151

1428

5.1

10

19°R

152

1430

5.0

6

153

1432

5.1

7

143.9

0°

4. 11

154

1434

4.0

5

155

1436

6.0

11

168.8

16°R

4.80

156

1438

5.7

3

157

1440

5.3

6

420.0

3°R

12.00

158

1442

6.4

5

360.0

0°

10.30

159
1510

1444
1446

5.8
5.7

6
6

300.0
420.0

Wind

Shift

6°L

8.58
12.00

1511

1448

5.2

4

281.0

8.03

1512

1450

6.4

7

0°

1513

1452

4.8

5

1514

1454

4.2

5

281.0

18°R

8.03

1515

1456

5.7

6

345.0

12°R

9.86

1516

1458

5.3

6

3°R

1517

1500

4.4

5

317.0

4°R

9.06

1518

1502

4.9

5

440.0

0°

12.58

1519

1504

6.0

7

450.0

0°

12.88

1520

1506

6.3

4

441.0

12.61

1521

1508

6.4

5

480.0

11°R

13.71

1522

1510

6.4

5

231.5

0°

6.62

1523

1512

6.3

6

0°

Thermocline remained at 36 feet during entire period. T "^ T

s " w

during same period. Data taken in 240 feet of water. Visibility unlimited;
clear sky; approximately one foot swell from southwest .

20

22 October 1968

No useable data is available for this date. Observations were
attempted between 1330 and 1530 local time, but no windrows formed.
During this period the wind varied between full calm and 1. 1 meters/
second. The average and median speed was 0. 4 meters/second. At
no time did the computer cards begin to separate from their initial
grouping. Between 1330 and 1530, the depth of the thermocline re-
mained fixed at 31 feet, and there was absence of swell. Evaporative
cooling was again present during the two hours, with T ^^ T

w

throughout.

Remarks on Symbols

PHOTO - Number refers to date and sequence. PHOTO 11 is

first photograph taken on 1 October; PHOTO 1520
is 20th photograph taken on 15 October.

TIME - Local time on a 24 hour basis relative to midnight

as time 0 .

WIND - Wind speed in meters/second during the two minutes

immediately preceding the photograph.

NO. ROWS - The number of windrows clearly distinguishable on
the photograph.

SPACING - The predominant or average spacing in feet between

windrows.

DEFL,

S/D -

The angle between windrow orientation and the wind
direction. Wind direction (relative) taken from
plume of smoke float.

The ratio of SPACING over depth of the thermocline.
Thermocline depth used as depth where temp. 1°
lower than surface temp, as taken by bucket ther-
mometer.

21

T - Surface water temperature

s
T - Wet bulb temperature

Spacing measurements were determined by use of the standard optical

equation:

IMAGE _ GROUND COVERED

FOCAL LENGTH ALTITUDE

The image and focal length of the T-ll camera are fixed at nine and
six inches respectively. With the altitude of the aircraft at the time
of shutter-trip known, the ground covered in the photograph can be
found, and converted to feet of horizontal distance per inch of photo-
graph. Only those photographs were used for spacing analysis where-
in the windrows appeared reasonably parallel and straight.

Where no deflection values appear in the tables, the smoke
plume was not visible in the photograph. Often the windrows extended
beyond the smoke plume, and the photograph was taken of that section
of the windrows rather than being centered on the smoke float.
Clearly S/D values can be presented only if a spacing measurement
is available.

1 October 1968

Ten observations of windrow spacing yeilded an average of
115. 3 feet with an average wind speed of 6. 8 m/sec. The average
angle of windrow deflection was 5. 0° to the right of the wind, with a
standard deviation of 5. 0°. During the observation period, wind
speed decreased fairly steadily from 4. 2 m/sec. to 1. 9 m/sec.

22

During the last 15 minutes of observation, the rows became more
irregularly spaced and seemed to converge upon one another. No
spacing measurements were taken during this segment in order to
avoid biasing the statistics. Deflection measurements of this 15
minute segment were made relative to the dominant windrow(s) in
the particular photograph. The windrows nearest the smoke plume
were weighted more heavily in deflection calculation. If windrows
immediately adjacent to the smoke plume indicated a given deflection
angle, while rows further away showed evidence of some other
angle(s), the rows adjacent to the plume were taken as correct.
Evidence was found that individual row orientation is closely related
to the wind eddies immediately above the row. Thus the windrows
at some distance from the smoke plume may be reacting to wind
action different than that indicated by the smoke float. No measure-
able wind shift occurred during the photographic period, thus no
calculation of windrow response to a wind shift could be made.

8 October 1968

Only nine photographs were used for spacing analysis, with
an average of 220. 9 feet for an average wind speed of 3. 8 m/sec.
Mean windrow deflection was 3° to the left of the wind, based on
14 observations, with a standard deviation of 7. 2°. Fully seven of
the 14 measurements showed no angle of deflection, however.
Once again, no wind shift occurred.

23

15 October 1968

15 measurements of spacing were used to obtain an average
spacing of 338. 6 feet with an average wind speed of 5. 6 m/sec.
Using 16 values, the average deflection was found to be 5. 0 to the
right, with a standard deviation of 7. 6 . Once again the mode was 0 ,
occurring in seven of the 16 photographs. At 1444 a wind shift
occurred. Two minutes later, the wind had completed its shift and
had steadied. The angle of deflection at 1446 was 6°L. By 1450,
the angle of deflection had become 0 . In addition, row spacing
increased by over 100 feet during the shift, but was back to within
20 feet of its original value by 1450.

1 October 8 October 15 October

Summary of Statistics

Average wind speed (m/sec. )
Average spacing (ft. )
Average deflection

std. deviation

mode

median

I. DISCUSSION OF STATISTICAL SIGNIFICANCE

Graphs and figures are presented on the following pages. The
first three figures are best-fit curves for spacing versus wind speed.
Next are histograms of deflection angles, spacing, and S/D for each

24

6.8

3. 8

5. 6

115. 3

220. 9

338. 6

5. 0°R

3. 0°L

5. 0°R

5.0°

7.2°

7.6°

0°

0°

5. 0°R

0°

3°R

individual day's data. The last three figures are histograms of
deflection, spacing, and S/D using the total of all three days of data
as an input.

A close examination of the data tables at the beginning of this
chapter indicates the significance of correctly interpreting any sta-
tistic developed from the data therein. Only 30 values of spacing
versus wind speed have been used to generate the best-fit curves.
These represent only 46. 1% of the total number of observations.
Since over 50% of the samples could not be used for determination of
spacing/wind speed relationships, any statistics or empirical rela-
tionships derived from the remaining observations must necessarily
be at best highly suspect. The reasons for using only 30 observations
have been explained earlier. Since nearly 50% of the total sample
represented windrows that were not reasonably parallel or straight,
it would seem hazardous to acknowledge any relationship of spacing
to wind speed derived from the remaining portion of the sample.

In the same vein, a study of the frequency of occurrence of
various deflection angles reveals the danger of applying any high
degree of confidence to the statistics that can be manufactured. Al-
though the deflection angles varied within the small range of 20°
left to 20 right, the angle of 0 predominated.

In the investigation of Faller, Woodcock, Sullivan, and Csanady,
only Csanady refused to publish fairly sweeping empirical relation-
ships derived from field observations. Yet Csanady was the only

25

one of the mentioned group who could reasonably lay claim to a
sufficiently large sample (246). Faller (1964) proposed a linear
relationship of spacing versus wind speed based on a very few ob-
servations. Faller used only 25 observations to derive a particular
mean windrow deflection angle. Sullivan had available only 27
observations. It is the contention of this writer that samples of
less than several hundred observations over a considerable range of
meterological and oceanic conditions cannot and should not be used
to propose any relationships between wind speed versus spacing and
deflection angle other than the following: (1) windrows are generally
oriented in the direction of the prevailing wind over the sea surface,
and (2) the spacing between windrows in highly variable and seems
to be dependent upon other features than wind speed alone. Best-
fit curves up to fourth degree yielded almost identical degrees of
fit: poor. Again it is emphasized that even these curves are based
on the best of the samples taken. Surely those samples showing
spacing to be random or convergent are more "statistically signif-
icant" than that minority which showed only some correlation between
spacing and wind speed.

A particularly interesting aspect concerning deflection can be
noticed by close examination of the photographs used for this report.
Although various angles of deflection were noted in the majority of
cases, windrows near or on the plume of the smoke float tended to
parallel the plume. This tends to strengthen the proposition that

26

windrows orient themselves parallel to the wind in the absence of
other factors. In several photographs, there can be seen extremely
good correlation between the meanders of the smoke plume and the
track of adjacent windrows. This feature will be mentioned in the
next section concerning significant individual photographs.

II. INDIVIDUAL PHOTOGRAPHS

Photograph 81 is an excellent example of the forceful nature
of Langmuir circulation. The card pattern in this picture had been
seeded by boat, and the photograph taken 35 seconds after comple-
tion of card dumping. Already individual rows can be seen forming
in large numbers and parallel to each other. In subsequent photo-
graphs, the number of visible rows was significantly smaller. This
led to a possible conclusion that the various areas of convergence
may attract the computer cards, overcoming the attraction of weaker
areas. This in turn may lead to erroneous values of windrow
spacing.

Photograph 87 shows the difficulty involved in measuring both
deflection and spacing. Although the majority of visible windrows
show general parallelism, there can be seen significant crossings
of rows. In the area of row crossing, spacing measurements were
impossible. In such a situation, deflection angles are at best dif-
ficult to define.

The close correlation of windrow orientation and smoke plume
meander can be seen in the lower portion of photograph 155. The

27

picture was taken at a considerable distance downwind of the smoke
origin, yet one windrow directly underneath the plume follows the
path of the plume almost identically.

Photograph 158 also indicates the close response of neighboring
windrows to slight shifts in the wind. Where the plume appears to
break up in the middle of the photograph, the row immediately above
also appears to break up in the same interval and direction.

Photograph 159 is mentioned specifically, since it is used as
the definition of a quick wind shift. The wind had shifted abruptly
to the right only eight seconds prior to the time of shutter-trip.

Photograph 1519 is another good example of correlation be-
tween the smoke plume meanders and a neighboring windrow's
orientation. The windrow just below the smoke plume follows the
path of the plume quite well throughout its length.

28

500.0

450.0

400.0

350.0

300.0

-250.0

200.0

150.0

100.0

50.0

1.0

2.0 3.0 4.0 5.0 6.0

WIND SPEED (meters per second)

7.0

Fig. 1 Least -Squares Fit, First Degree, for Spacing versus Wind Speed

29

500.0

450.0

400.0

350.0

300.0

- 250.0

D

Z

j 200.0

n

150.0

100.0

50.0

I

— J

X X

X

<

I

!

I

X

X

x/

i

X

X

/ x

^ v

X /

X

i

X

v

X / s

<

i

i

i

3

x /

<

X

A.

X

X

/ X

X

/

i /

1.0 2.0 3.0 4.0 5.0 6.0

WIND SPEED (meters per second)

7.0

Fig. 2 Least— Squares Fit, Second Degree, for Spacing versus Wind Speed

30

500.0

450.0

400.0

350.0 —

300.0

a 250.0

<

Q.

<s> 200.0

150.0 -

0

50.0

\

>

X
f

\

X
X

X X

X

\

X

\

X >

/ x

\

X /

A.

X

X

v

X / 3

v

8

X
v

<

x

7\

X

X

X

X

1.0 2.0 3.0 4.0 5.0 6.0

WIND SPEED (meters per second)

7.0

Fig. 3 Least-Squares Fit, Third Degree, for Spacing versus Wind Speed

31

1 October

8 October

56.0° F

56.0° F

' 31 ft.

55.0° F

26 ft.

55.0° F

1 5 October

57.5° F

22 October

58.0° F

I 35 ft.

56.5° F

Fig. 4 Bathythermograph Traces for 1, 8, 15, 22 October

32

1 October

>
z

LU

a

5 -

ipillll

100

1

200

SPACING (ft.)

300

8 October

o

2
LU

D

a

LU
QC

5 -

L_L

100

200

I

300

SPACING (ft.)

J—U

400

500

15 October

85

LU

cc

J_L

J-L-Ll

500

100

200 300 400

SPACING (ft.)

Fig. 5 Spacing Histograms for 1, 8, 15 October

33

1 October

>-
o

z

UJ

a

UJ

5 ~

— r—
10L

JJ U

J_li

o i6r

DEFLECTION

8 October

z

LU

3

a

LU
DC

5 -

1J LL

20L

1
10L

DEFLECTION

J_i

10R

>
u

z

UJ

D

a

LU
CC

5 "

15 October

- r—
10L

II I II

10R

20R

DEFLECTION

Fig. 6 Deflection Histogram for 1, 8, 15 October

34

>-

is

LU

or

0-

1 October

II HI II HI

>-
o
z
in 5 -

a

LU
OC

LL

Jj

S/ D

-r—
10

15

LA

8 October

>
u

z

LU

85-

LU
DC

S/ D

,U I I I I 1 1 |_I_L

20

15 October

12

S / D

16

Fig. 7 S / D Histograms for 1, 8, 15 October

35

■8

in

o
o

■*■>

Q

§

a

<

CO

c
'o

CO

o

CN

A3N3nD3bd

36

15n

10

z

LU

a

LU
CC

5 "

J II , Nil

20L

10L 0

DEFLECTION

10R

111

20 R

Fig. 9 Histogram of Total Deflection Data

CD

_ <o

_ Tf

_o

to

Q
Q

CO

■«5

*•>
o

E
re

—

§

O

-CD

Lfi

A0N3nD3dd

38

Figure 11. Photograph 81

39

Figure 12. Photograph 87

40

Figure 13. Photograph 155

41

Figure 14. Photograph 158

42

Figure 15. Photograph 159

43

Figure 16. Photograph 1519

44

CHAPTER IV
CONCLUSIONS AND RECOMMENDATIONS

I. CONCLUSIONS

The primary conclusion derived from the data of Chapter III
is that a greater number of observations is required before any
significant statistics concerning spacing and deflection can be de-
rived. This in turn leads to the conclusion that any statistical
relationship concerning deflection angle or spacing must be viewed
in light of the number of observations taken and the physical condi-
tions present. Based on the photographs taken for this report,
windrow spacing must be considered variable and related to other
factors than wind speed alone.

Deflection angle measurements indicate that windrows generally
orient themselves roughly parallel to the wind. Although angles
between 20° left and 20° right of the wind were observed, the angle
of zero deflection was most common. A statement is needed here
concerning the derivation of a given angle from a particular photo-
graph. Faller (1964) used a photograph of mica strips and smoke
floats to derive deflection angles. Faller used the angle each wind-
row made with the smoke plume as an individual sample. In this
report, the photograph itself is considered to be a single sample,
and the rows seem used to determine a predominant or average angle
of deflection, This method would seem to take into account a given
set of physical circumstances during a given sample period.

45

Although a particular row can be traced from one photograph
to the next in several cases, there is sufficient photographic evidence
to lend credence to the findings of McLeish (1968). McLeish postu-
lated that the rows were constantly shifting in direction, spacing and
width within some range of values. This range depended upon the
turbulent conditions present within the surface waters.

Response of the windrow orientation to major wind shift appears
to occur within minutes. The single observation of an abrupt wind
shift indicated a response time for row orientation of between two
and four minutes. Minor deviations of wind direction, as taken
from the meanders of the smoke plume, are closely coupled with
identical deviations of neighboring windrows.

No specific mention has been made in this report of the
parameter S/D. No correlation could be found for spacing versus
depth of the thermocline, and this may be due to one of three reasons.
First, there may be no correlation. Secondly, the spacing evident
in the photographs may not be indicative of the total number of rows
present. Lines or paths of convergence weaker than their neigh-
bors may lose the marking cards to possession by the more powerful
areas of convergence. Third, there may be no well defined ther-
mocline. This was the case on 8 October, when only a very weak
negative temperature gradient was observed.

A summary of this writer's conclusions is as follows:
(1) Windrow spacing appears to be related strongly to factors other
than wind speed alone.

46

(2) Deflection of windrows to the right or left of the wind's direction
occurs, but these deflections are small and can be expected to be
within 20° of the wind.

(3) Windrows respond in their orientation to a wind shift within a
few minutes-

(4) The path of a windrow is most likely a good indication of the
wind field immediately above that windrow,

II. RECOMMENDATIONS

It is strongly recommended that the project of windrow in-
vestigation be continued at the Naval Postgraduate School. The data
derived in this report resulted in much less conclusive evidence than
was anticipated, and several corrections to the investigative tech-
nique are indicated.

Foremost is the need for better marking material. If there
exist areas of convergence more powerful than others, enough
material should be scattered in the water to indicate the presence
of both the major and minor convergences. The computer cards
used in this investigation show the areas of strong convergence.
It was noted, however, that the number of visible rows decreased
shortly after initial seeding of the cards. One way of overcoming
this problem would involve second seeding of material between
formed windrows. Another method involves the use of small confetti-
like particles of paper, or powder. The second seeding method
seems preferable, but the probability exists of disturbing the pattern

47

already formed. If reasonable care is taken, the recommended
method consists of using confetti or a metallic powder for an initial
planting, then following this at specific intervals with second and
perhaps third seedings of the same material between rows already
formed.

The second recommendation is to investigate the use of a heli-
copter incorporating the T-ll camera. Some of the most well-
defined and prevalent windrows can be found just offshore of the
Oceanography Beach Laboratory. This is an area impossible to be
photographed by fixed wing aircraft due to the proximity of the Mon-
terey County airport's traffic pattern. In such an area, convection
plays an important role. If thermal convection is present, only a
light wind seems required to produce windrows. In the absence of
convection, a significantly stronger wind would probably be needed.
McLeish postulated that turbulence alone could produce windrows,
but the photographs of this thesis contain too many observations that
contradict his conclusion, i. e. , observations of well-defined parallel
rows. In addition, several sequences of photographs showed wind-
rows retaining their identity over time. A beach- mounted movie
camera could be well used in conjunction with the T-ll to confirm
this fact. In conclusion a helicopter has the capability of hovering
for close-coupled time sequenced photographs of the windrows. This
should allow accurate observation of windrow response time to a
wind shift plus the variability with time of windrow width, direction,
and spacing.

48

In order to develop meaningful statistics, it is recommended
that sample photographs be taken over a longer period of time, and
at various times within a given 24 hour period. Infrared photographs
could be taken night or day, for example. The photographs of this
report were taken invariably in the afternoon due to aircraft and
boat availability. The number of observations should total in the
hundreds to allow for poor camera placement.

Since the wind field immediately above the water's surface is
turbulent, there is an eddying motion present. Thus at any instant
in time, the rows may not be parallel over all parts of a photograph.
Based on this conclusion, the only reliable deflection measurements
would be related to rows adjacent to the smoke plume, and these
are nearly always parallel. To confirm the coupling of windrow
orientation and wind eddies, it is recommended that multiple smoke
floats be used in future research. An array of some type would be
best, provided it does not obscure the cards or other marking
material.

49

BIBLIOGRAPHY

Csanady, G. T. , 1965
Windrow Studies

Baie Du Pore Report 1965, Great Lakes Institute Report
No. PR, pp. 60-82

Ewing, G. , 1950

Slicks, Surface Films, and Internal Waves

J. Mar. Res. , V. 9, pp. 161-181

Faller, A. J. , 1963

The Angle of Windrows in the Ocean

Tellus, V. 16(3), pp. 363-370

Faller, A. J. , and A. H. Woodcock, 1964
The Spacing of Windrows in the Ocean
J. Mar. Res. , V. 22, pp. 22-29

Langmuir, Irving, 1938

Surface Motion of Water Induced by Wind

Science, V. 87, pp. 119-123

McLeish, W. , 1968

On The Mechanisms of Wind Slick Generation

Deep Sea Research, V. 15, pp. 461-469

Sullivan, P. J. , 1964

Windrows; in Hydrodynamic Studies on Lake Hurron at Baie Du Dore

Great Lakes Institute, Univ. of Toronto, and Univ. of Waterloo, PR19

Uttormark, P. D. , 1968

A Thermal-Donvection Model of Langmuir Circulation

Unpublished Thesis for Doctor of Philosophy (Civil Engineering) at the

Univ. of Wisconsin

Wellander, P. , 1963

On Generation of Wind Streaks on the Sea by Action of the Surface Film

Tellus, V. 15(1), pp. 67-71

50

APPENDIX I
Additional References

Csanady, G. T. , 1963

Turbulent diffusion in Lake Huron

J. Fluid Mech. , V, 17, pp. 360-384

Csanady, G. T. . 1964

On Windrows

Univ. of Waterloo, Dept. of Mech. Engr. (unpublished report)

Csanady, G. T. and H. K. Ellenton, and R. E. Deane, 1962

Slicks on Lake Huron

Nature, V. 196, pp. 1305-1306

Faller, A. J. . 1963

An Experimental Study on Instability of The Laminar Ekman Boundary

Layer.
J. Fluid. Mech. , V. 15(4), pp. 363-370

Ichiye, T. , 1965

Upper Ocean Boundary Layer Flow Determined by Dye Diffusion

Physics of Fluids, Supplement, 1965, pp. 5270-5277

Kraus, E. B. , 1967

Organized Convection in The Ocean Surface Layer Resulting from

Slicks and Radiation Stress
Physics of Fluids, Supplement, 1967, pp. 5294-5297

Munk, W. H. , 1947

Effect of Earth's Rotation upon Convection Cells

Ann. N. Y. Acad. Sci. , V. 48, pp. 815-820

Palm, E. , I960

On Tendency toward Hexagonal Cells in Steady Convection

J. Fluid Mech. , V. 8, pp, 183-192

Pearson, J. R. A, , 1958

On Convection Cells Induced by Surface Tension

J. Fluid Mech. , V. 4, pp. 489-500

Rayleigh, Lord, 1916

On Convection Currents in a Horizontal Layer of Fluid, When the

Higher Temperature is on the Upperside
Phil. Mag. , V. 32, pp. 529-546

51

Stommel, H. , 1949

Trajectories of Small Bodies Sinking Slowly through Convection Cells

J. Mar. Res. , V. 8(1), pp. 24-29

Woodcock, A. H. , 1944

A Theory of Surface Water Motion Deduced from The Wind- Induced

Motion of The Physalia
J. Mar. Res. , V. 5, 196-205

52

Appendix ii

Program for least-squares fit

— D

•

K *

<*

—

IS)

0

to o

#

>-

_J

-4-

< 00

#

oo

<J

w

_/ ••

CC

<

*—

u

» X

at

-J

s

en

*-< *

«-'

*

O

»

» ♦-

♦

•-4

z

—

o </>

•»

pi

>

**

«-< ••

tf\

o

_J

* >

•^

^t

c

X

a *

00

•>

Q.

•>

•» CD

♦

LUK

«*

« •>

o

QtU-t

w »

O OD

•"^

•" »

»

Ouj

»-

•> • •>

r<^

rri

o

2U

CO

©«/>>-

■)i

»

«

lu

•>

»«J_J

*

«

o

e?

— .

©<ruj

Ct

or

»

LLi—.

«*

•<— o

or

or

o

—J

■V

ox ►

»-

•—

»

o

>

•o>-

«

*

o

>-z

•>

©z •»

*•

*»

•

GCm

—

l^>3J

>t

>t

c

o

*t

»_) »

■*— *

•^

^

t-<r

«^

OOLU

cc

ec

»

•—a

— •

00

*o.u

o

LUCO

o

to

LU <
_JLUQ.

+

+

•>
LU

LUX

«— ■ w

«■».

•»-a:oo

•>—

■"— ■

^

-J

CCC

<\J >

«*

co»~0 •>

<M

QC

cc

»-

<a

^ V

V

ji-zj

*

QC

cc

►— <

DO

— V

oc

O — LULL'

*

*

#

>-

oz

UJ •>

»

-J *o>

a:

a:

a

•—1

tot—

-1 —

*»

JJUJ •>

cc

a

cc

»

J*

1- c

c

oiuj<

•^

w

•**'

-J

»-

hJ(«

rr

LUCO X

#

«

*

LU

coot

h- »w^

•«—

<:luo

— .

■i— •

^

CC

<r

nUXO>-

tOJO*-"

f^l

m

ro

<

LUCO

*vixrn-j

H •> tO

<-»

M>

•^

_l

-j a:

— N. •>X«-

'LU

Zi-«oo •>

CD

OC

CO

•>

u:

KJ »- *UJO

»-« »HO

U-ir>

UJ>

1 O—U »

o«-iz •»

■f

4>

■f

> •■

X

a *o<~

a mjjc

am

I-O

•• eooao

>-•- »

Op-KMCX-

o:

CC

cc

~ ••

LU

I co<M>Offi

o-uo

«NJ cm (Net:

a:

a

cc

o>

cooo

i »w i

H->«- ••

*

#

♦

*

>-

LU>v

«t >->-

>v

< MtX

Cl7>C->

^>

— »

— »

K •■

KI

— -J>C -

OXLUiC

Kt-l-fsj

(\i

(M

f\!

i—X

3

ec ilu »m— •

«M

m

— »

LU XUJ *

w

«B>

*p>

•_•

U-X

az

#coa-»«-

'©

1-4

•— i

*-*

-) l-XCO

ooocd

OC

CD

.00

•>

Xi-

.j'v »o-jm

• LU *

«-»w

K- 3 ►OfO

ocu

»-o

o

<_J<IOLU«—

>*

>tu»o

.-.LU

H-ao — c

c

4

■f

+

♦ X

COO

uc

ujuujra>3:

»

_iU-«ru.o

«-'OH-0nZfr>LL_J *

o

— ».<«■»

o

LUO*

LU

cceco—

rH

U-if^Q-fnm«j<

t-<«»000

o

(NJC^^--.

— .

— .

*m

CO*'

iu

<hX22

II

>,_llO_4 »LU0l

* X 0.0

»

• • •»—

m

»H

•-*♦

X

<a

KJWXDC-

>

<m*

w»H

>U)

*wLL<ZOtr

•-hi

KOOOwC- c«-c

;^«- LU

LL«LU

ai/<

M MHO^

»

► It

II nir> oaOwv

ii a u-u-.LL'a — 'cc«-

*C »■

-CCCC3

Oct 3

o

cjcc^^ocoro

«KNhw

«M*

CMK 0*--J^O —

II • • • II

II

II

ii a z

QZZ

cc

•-* * *2

2 ►«-«

<1

<

N*»M

<0 K •

•

-iss-u-c-c* •-

c «-"a

ccz

j-ij_jiuiLCHC2 cxm

ww

OXZj<JOO<C<

'-iiir^t-'^

i-»-»-'l-»- II H

Z_iKO

CU»-

a<«rxm

<(X<TQ-

x>-<o£i~.j»-_j ii

II

to> w««-^w

ta>

V

*- z

HJZK

Z

X LULULU»— ^-'<—

OLUOU-CC

x>ujo»-0<xccoXluulll>o>-o>-c>-cco

K<OLU

to

•-'eCeCeC003

DaiLau.Dx>au.»-oo.uoaDXMMt-.vc>-c>-c>-QCU

i-uua

•—CC

♦

O X

c

IO

■-I

(M

(T>

IT. -J O>0

O

r-

cvjccc

— 1 r-

KU.

rvJO. u»-4

t\J

(M

CM i-4

o. ^«

ooo

u u

uu

53

INITIAL DISTRIBUTION LIST

No. Copies

1. Defense Documentation Center 20
Cameron Station

Alexandria, Virginia 22314

2. Library 2
Naval Postgraduate School

Monterey, California 93940

3. Oceanographer of the Navy 1
The Madison Building

732 N. Washington Street
Alexandria, Virginia 22314

4. National Oceanographic Data Center 1
Washington, D. C. 20390

5. Office of Naval Research 1
Special Projects (Code 418)

Department of the Navy
Washington, D. C. 20360

6. Office of Naval Research 1
Geophysics Branch (Code 416)

Department of the Navy
Washington, D. C 20360

7. Department of Oceanography 3
Code 58

Naval Postgraduate School
Monterey, California 93940

8. Department of Meteorology 3
Code 51

Naval Postgraduate School
Monterey, California 93940

9. Professor N. E. Boston 10
Department of Oceanography

Naval Postgraduate School
Monterey, California 93940

10. Lieutenant R. W. Ortengren, Jr., USN 2

USS EVERSOLE (DD-789)
FPO San Francisco, California 96601

54

N CLASSIFIED

M.unt\ Classification

DOCUMENT CONTROL DATA -R&D

-,, , ,rifi eliisafification of title, hm/i . I ..'■■ rr.it I .,i„l mi/.-xmri .uinotnti.m mu.-.t be entered nlwi the „verull report is ch,s^ilieil)

Naval Postgraduate School
Monterey, California 93940

a. RtPOR ! SECURITY CLASSIFICATION

U A1 CLASSIFIED

2b. GROUP

NFI'OR ' TITLE

Surface Characteristics of Windrows

[ , H'l IVL NOTES fTVpe o/reporl .ind.mrlij'iiii (iafes)

The si S

tulfORiSi hirst name, middle initial, last name)

Ralph William Ortengren, Jr

Rtl-ORT PATE

December 196i

C.1NHKT OR GRANT NO

b. PROJEC T NO

7a. TOTAL NO. OF PAGES

55

7b. NO. OF REFS

9a. ORIGINATOR'S REPORT NUMBER(S)

9b. OTHER REPORT NO (SI (Any other numbers that may be assigned
this report)

C DISTRI*. TION STATEMENT

Distribution of this document is unlimited.

1. SUPPLEMENTARY NOTES

12. SPONSO RING Ml LI T AR Y ACTIVITY

Naval Postgraduate School
Monterey, California 93940

ABSTRACT

Aerial photographs were taken of windrow accumulations in Monterey Bay
on 1, 8, 15 and 22 October 1968. A Fairchild T-ll aerial mapping camera
was used, with photographs taken approximately every two minutes over 40 to
b0 minute periods. Windrows were marked with accumulations of computer
cards, wind speed measured by cup anemometer, and wind direction taken with
the aid of a MK. 6 Smoke Float. Sea surface temperature, depth of the ther-
mocline, and surface air temperature measurements were taken concurrently.

An attempt was made to correlate windrow spacing and wind speed, to find
mean deflection of windrows relative to the wind, to determine any relationship
between row spacing and depth of the thermocline, and to find the response time
of windrow orientation to a wind shift.

Windrow spacing was found to depend on other factors than wind speed.
Deflection angles varied between 20° left and 20° right, with 0° being the most
common angle. No correlation was found between depth of the thermocline and
row spacing. Response time fell between two and four minutes.

DD

FORM I47O

I NOV 65 I "T / SJ

S/N 01 01 -807-681 1

(PAGE 1 )

UNCLASSIFIED

Security Classification

A-3M08

55

UNCLASSIFIED

Security Classification

KEY WO ROS

LINK A

Windrows

Langmuir circulation

Wind streaks

DD,F°0RvM651473 (back)

UNCLASSIFIED

S/N 0 101-807-682 1

Security Classification

56

t^<*~*&^£*221

3 2768 001 97382 9

DUDLEY KNOX LIBRARY