\ YSh wee (4520 me) Vice
,

A STIDY OF THE ERRORS
of the

BATHYTHERMOGRAPH

Contract No. NObsr 52348 June 1952

NATIONAL SCIENTIFIC LABORATORIES, INC.
Elechronies - Research - Development

f 2010 MASSACHUSETTS AVENUES iN. W.

——_ ‘{, WASHINGTON 6, D.’C.

rns Vata ey
ie) Tie oat

pets

MBL/WHOI

WON

Final Report

A STUDY OF THE ERRORS
of the
BATHYTHERMOGRAPH

by

A. L. Bralove
E. I. Williams, Jr.

ih

NATIONAL SCIENTIFIC LABORATORIES, Ince.

2010 Massachusetts Avenue, N. W.
WASHINGTON 6, D. C.

0 0301 O043ecb4 b

NAVY DEPARTMENT ° BUREAU OF SHIPS ° ELECTRONICS DIVISION
CONTRACT NObsr 52348

June 30, 1952
SERIAL NO. NE051247

TASK 7
PROBLEM 1-U
TYPE 1
(UNCLASSIFIED)

INDEX

Page
Tas, AXTSRSSIRT RAO EY ce See ae AA RUN ren) Sey Rue Sen SN LNAI SUL ae Fe ee sO a ee 1
BD LGN TIET RON DSW Ged ON aaa cet Mea Oa oe OE Sel SEE I a eee A 2
3.0 GENERAL DISCUSSION OF MATHEMATICAL ERRORS ~...W 3
Sale irunem Values sand. Probables Values: hes li ee oo ere Ve 3
SEMEN ORS peanut aba MAN Te, AVANT ae oe NA Laer ecwamiee te ERAN a Be Pu EAS Fd 3
Seo Nm D eterna te wHGR OMS |gmee ee cease reine eee ee teat suened ee ely Neue 1 ee 3
Sto 2 an de berIMiNa Cel EE OT Sieces eee eee een tate a eerie ater en gel BLN, Peer cee 3
Seo SATCEIG ental ashy 1e1.0 FS yes erase eae vs tonto it) eel ae Ree aa ee 3
SP 2e Aran ELMS ELUM ETIG PELE OLS ypeerene eee tan eet te sas take eel ete be Re aoaeen ey 4
3.2.2.3 Dependent and Independent Errors ._........----.---e 4
SP ENC GUT Cry ieeenrr ot sae brs Cale ie Se SLC A nS SEED RSE ORR AnN tee Ra LIM Slave Ueto ahs UROL 2 ole ete 4
Bl AMPEAGE Wellies earl AAQVETREYSe: TOTO oe cee cee eee cee eee e eee cece 4
SrAOe WAN GRA Cea Ves it ic nan tebe enn Lee Narn ERRN MN Ss AMR uve So MO Nugget ans Le 4
STAs DOVE GIOT Merete tates e an eae eke Sere Lh sage RRL Us ten) BO atte PUR ate 4
SAS VA Verage- Devi atlonc es we te Ree) ee ee le es 5
OAt2i2) Standard Weviatione =. lee es Meee lee ree Be eae 5
Sel OMe ISCO WANICYH twelk sett oot ALU Fecal ON veoh CR i ee ESS Ne 5
SrA AONE AN CG) eet fs sede obvi Ro nae eee ae AVN ere NE Re eal LO ae 5
See OMMMAV Gea eo LO TUN type ac et em teat ate eg uetrentie Trans we Pare STUUR SE ae. ches Nel Rot eee Lae ae 5
Seba OCC ea Cetra TT Ol Oyj pikes Ee TORE OP esky MA eg Dee Ye A ee i a 5
3.6 Overall Error as a Function of Individual Errors ............-...2----.ee---eceeeeeeeeeee- 5-6
4.0 APPLICATION OF ERROR TERMINOLOGY TO THE

TEAMED TaY@s TUE EUV @ (GR ACP ETS Wate ee ter St rib nO Pees SLY eA Mh SM Ee Yo ri
4.1 Determinate Errors of the Bathythermograph — 2222-20 7-8
4.2 Indeterminate Errors of the Bathythermograph —........0...02200-0--22-----2ee eee 8
4.3 Bathythermograph Deviations as a Function of Errors and Ocean Variability 8-9
4.4 Measurements of Instrument Error by Special Tests —....... 9
HO ITIVIRO IS, HINTED OI SINIL TONE ASOD) TONS ROMO OINTY 10
Bell, .7 RS ONeCOYG GMC PUD ties ee te cee, eee Sa WUE 2B fb Ve ence eee LUTE Ft | CRN es A ape 10
5.2 Accuracy of Registering Temperature Change with Change of Depth .......... 10-11
OS OME EL SUCHE SIS pein) <tr teak rn 10) BUMS RAY ere ee Sat De bh ope cory enone ie IR tht
Drovl Mechanically steresisy 2.0 se0 oo Ue 11
Dist a wMIReCSpoOnse: Py Steresis (26 fi 00s el 11-12
5.3.3 Hysteresis Due to Ocean Variability 20 --2220.-o0o2ceceeeeeeeecceeeeeeeeeeeeee 12
HBA Clears ino Ibner Cornhiniorns s220 ee 13
pe AReevesponser Characteristics weve cci etal ee ee ee 13
LAME BEVESD ONSC ULV SLEREGIS reer ae ne DAN ee ea et ane 13
Okan CMe eS PONSe Tate Un tl aly) ey thier eee oe eee Ce de ns Se 13
prAr ome eChani CalmelesitetlOmpe tu evum el tel eon aur) ec 13
rome aulny the ino strate GS ebSueten cre rn Se heen mune eee Acts a Lee 13-14
SLOmplins GEUIMIentaloainlin estima eee: sale Cauarcalns OLE MeN 8 We ee ee) ee es 14
SSCs Gemavenerevh oBL te OE Sah Ue ek ete aS ca ce ANT A 14
ROR MND EN OMT RO EXe TH ANus Tia ee altaya mee Rie ea By Acces uulnl Oech URS i) leet teal 14
5.6.3 Superimposition of Trace and Destruction of Trace _.....................- 14-15
DRO tam ellowsmivallUneneateat ae meen Mer musics ee ee ee Uae 0 eS 0) sl ae 15
DOr ose EOOSe MBAR Site eee WO ae eh Gh SP eee Ne Rs LR ye «ee 1165

6.0 ERRORS INTRODUCED BY CALIBRATION AND RECALIBRATION OF
EI DAUD Fah RM O GVAIRIMG (ieee bee al en 16
Glia G Cri ercalligee omens see cam Moe Usa cere aim algae wld Sik tae WA i Me 16
Online Re paliauhests pecs nies ee el WE Asana ere Gt 16-17

7.0

8.0

9.0

11.0
12.0

GalEZ Ere quency, ot C allio ree tric valet ee nomen Eee eee 17
6.1.3 Bathythermograph Test Log and Repair History —....................22 7
@Gil4! Wharirorarntigy Ieee Wes, SUNOS ares atu
G1 Cayomenigy oie Wesitilinse SWAMOMG coerce laces cere eceeetnnecceee oceania eemenneeeees 17
Gilt Ay SCORE STS occa ee Ae ices es ee 2 pe eect 17
Galey Rate OR URES ONS ey see ote Bet rh SST ene a ss Oa ee 17-18
Gaz) Reproduretiotiitiy, Mestists etext oko peace e. oes Ee Ov nls ee 18
GAO) “Smioothmess: Mest epee ee a ee 2 ee ec ene nel ee a oe 18
6.1.10 Standardization of Testing Procedure 18
ERRORS INTRODUCED BY OPERATING TECHNIQUES WW .. 19
7.1 Errors Produced by Handling the Bathythermograph _.......... 19
4.2| Errors’ Produced by Supplementary Equipment 2220 19
Yoss lovereorrs Ireoclyescl lony wae Ojorermentivayes Leva Oma a 20

ERRORS INTRODUCED BY PROCESSING OF BATHYTHERMOGRAPH DATA 21

Sule MemiperatunenC onkecti oral ec lini ciel sseeeeses eae een eens tuna een sense anu 21-22
8:27 Depth, Correction Gechnique 2 ee eee 22
So Idperores Iimyolhyerel ihm Crorereretemvayer ISCAS eee eee hee ee 22
Bavroll 1Proore SrerewrarGleyeyy Waa OVEN, eerie cee eens 22-23
8.352). Vela tlom ii Sets re. OO ee ie ES Se? Se a ad ae 23-24
8.3.3 Factors Affecting Zero Depth of the Bathythermograph 24
8.3.4 Reliability of Surface Bucket Samples 2 24-25
PROPOSED METHODS FOR DISCOVERING, EVALUATING, AND
MINIMIZING ERRORS IN THE EXISTING AND FUTURE BATHYTHERMO-
GURVAUPUT TO ATA ss a Sia oe ce Cre TRS ona he See eee ZR A er oe 26
9.1 Establishing the reliability of the Surface Temperature Correction Techniques 26
Oe Use ot Trend iGharts 2 2 ee en Sse eee 26-27
Gil Coronjorareisyorn coe Sieeyaclewrecl ISOS aces 27
9.1.3 Evaluation of Variability Data by the Coefficient of Correlation .... 27-28
9.2 Use of Two Bathythermographs Simultaneously — 0... 29
9.3. Other Techniques to Correct Bathythermograph Sets —...... 29
9.3.1 Correction of Bathythermograms through Heat Content Studies ... 29-30
93-2) (Curve Bitting Mecha Ges) ca ee Re 30

9.3.38 Correcting the Bathythermograms by the Permanent Thermocline ... 31

9.3.4 Use of Minimum Thermometer to Correct Bathythermographs _........ 31
Establishment of Average Bathythermograms — 23 31-32
Kistablishine Average Trends) oro ey oe GS AR AUN ge bd ls UC a Nga a eee a 32
REC OWPNUE NID ATIOINS 522) ee Uae Nels Nis ae Cl eerie Ue Selon PRUs Cee 33
UO@QJL — IVAW aSraMN@TACAV Oe) MTOR WUTMETONG a5 eae accent eneeeenoneonesoeeneeeeeesecenetonestecneeees ‘date ne 33
TIO) S24 Pha Gey GUN oe s4 610) cM Rn RL aU Dam ek CS re oe na ee a 33
1O@ Ojoswmaniomell Wise Oe BeinlonyaSreran@eyeajola ae nance ee accent cetee eee eee Perens 35
OFA Cone Chines aulayi ble rari gareeyy nese) Seer ee Un 33
10.5 Further Studies of the Errors of the Bathythermogrph —......... 34

HOF al Bathy thermo prapliy i Setsy xi van ee ele eGo te ee 34

OFSE AM AVA ELOTIy NE aie lait Me TINO S41;e (0 10S Css eee eee 34

NOMS ys Caltbr ation Pr Ors) se ee aw ened i eos 34-35

O54 Reproducillority. ioe 8 a See I eee ek ne a eee eee 35

MOSS. 5 Oyo Sratsl Orme] HOS sc cee et SID) Ese 35

10.5.6 Other Methods to Correct Bathythermograph’s Temperature —_...........- 36
(GIG OS SVAIRAV 62 OV ts RANTS ese oe cleo 2 ns kre yg ne RRS ge el Se a Ue 37-46
183] 6) B3) COX GH) Saye hc BS ieee lee ne Rens aoe Re Ag ea ete Dir coh A A 47-50

ii

SECTION 1.0

ABSTRACT

This report has been written with the following purposes in mind:
1. To identify and tentatively define the errors in BT data.

2. To place on record a preliminary study to indicate to persons not familiar with the
device the limitation of the bathythermograph as a scientific instrument.

The study of the errors of the bathythermograph is an investigation to determine the
possible sources of error which may exist in the bathythermograph data. The investigation
encompasses (1) intensive investigations of several anchor stations, i.e., data taken on a
ship while in a fixed position, and (2) a comprehensive conference program with members
of various Oceanographic Institutions to analyze their needs, methods, and processes.

In this report the errors are grouped together according to their error source, that is,
instrument limitation, calibration technique, operational technique and data processing.
These errors are classified wherever possible in accordance with standard mathematical
nomenclature and practices. Correlation methods are also discussed, and proposed tech-
niques for minimizing errors through correlation studies are suggested.

The result has led to a classification of possible sources of errors involved in obtaining
bathythermograph information, and several techniques and suggestions for handling the
bathythermograph data in order to recognize and minimize these errors. An extensive
bibliography in Bathythermography and a glossary of terms are also included.

SECTION 2.0

INTRODUCTION

The Bathythermograph has provided the Navy and the Oceanographic Institutions
with close to a half-million Bathythermograms. These Bathythermograms, taken over a
period of ten years, indicating temperature versus depth structures in the oceans of the
world, have been of immeasurable value in accumulating scientific data and thus further-
ing the general knowledge of oceanography.

The initial use of the Bathythermograph was in the qualitative studies of the tem-
perature depth relationships. However, the advent of World War II created a definite ap-
plication for the Bathythermograph as a military instrument. It was during this period that
a large number of these data were continuously collected. Following the war, the Bathy-
thermograph program was continued with the idea of collecting data which would be avail-
able for analysis at a later date.

Emphasis has gradually shifted from the qualitative contributions of the Bathyther-
mograph to a more precise and more exacting quantitative requirement. Quantitative ap-
plications imply that a certain accuracy of measurement is expected from the BT. It is
quite evident, therefore, that the extent to which the existing mass of data can be utilized
depends largely upon the reliability of the information contained therein. This reliability
is a function of many factors, such as the manufacture of the instrument itself, its calibra-
tion, state of maintenance, operation, and the methods of correcting, processing and cor-
relating the data.

The National Scientific Laboratories for the past year have been concerned with a compre-
hensive study to determine the possible sources of errors which may exist in the Bathythermo-
graph data. This study has included, (1) intensive investigations of several anchor stations,
i.e., data taken on a ship while in a fixed position; and (2) a comprehensive conference pro-
gram with members of various Oceanographic Institutions to analyze their needs, methods and
processes. The result has led to a classification of possible sources of errors involved in obtain-
ing Bathythermograph information and several techniques and suggestions for handling the
Bathythermograph data in order to recognize and minimize these errors.

In this report these errors are classified wherever possible in accordance with standard
mathematical nomenclature and practices. These errors are grouped together according to
error source, i.e., instrument limitation, calibration technique, operation technique and data
processing. Correlation methods are also discussed and proposed techniques for minimizing
error through correlation studies are suggested.

The National Scientific Laboratories wishes to extend to the Oceanographic Fellowship its
sincere gratitude for the generous and helpful assistance which many of its members so gra-
ciously extended to us.

SECTION 3.0

GENERAL DISCUSSION OF MATHEMATICAL ERRORS

In order to understand more completely the various classifications into which the Bathy-
thermograph errors have been placed, it is advantageous to start with a general discussion of
mathematical errors.

3.1.

3.2.

3.2.1.

3.2.2.

3.2.2.1.

True Values and Probable Values

In determining an instantaneous temperature somewhere in the oceans one must
assume that there is a true value which can be called an absolute value without error.
However, the true value of most quantities can never be known because of the un-
avoidable errors in measurement and calculation, no matter how precise the instru-
ments employed are.

In lieu of obtaining a true value one usually obtains the most probable value and
along with the probable value a probable error. One can say that the probable value
will not differ from the true value by more than the probable error.

Errors

The word “error” is a most common term used to describe a multitude of dif-
ferent conditions. One type of error has already been discussed, the probable error.
In addition errors can be classified as determinate errors and indeterminate errors.

The term “‘true error” is often used, which mathematically defines the difference
between a true absolute value and any single measured value. However, the true
value is never known in the ocean, and therefore one needs to concern himself only
with the probable error.

Determinate Error

Any error that is discovered and allowed for in magnitude and sign in the form
of a correction for its effect is a determinate error. For example, comparison of an
ordinary thermometer with a standard may reveal errors in the graduation of the
former and result in a standard calibration. Every temperature measured with this
thermometer would then be subject to an error of definite magnitude and sign which
could be determined by reference to the standard calibration.

Indeterminate Errors

All errors that either cannot or are not properly allowed for in magnitude and
sign are known as indeterminate errors. It is obvious that the correction for deter-
minate errors will themselves be subject to errors and thus constitute one class of in-
determinate errors.

Accidential Errors

A particularly important class of the indeterminate errors is that of accidental
errors. Such errors are due to the combined effect of a large number of undeter-
mined causes. Experience has shown that these deviations are inevitable in all meas-
urements and that they result from small unavoidable errors of observations due to
more or less fortuitous variations in the sensitivity of our measuring instruments
and of our senses of perception.

3.22.2.

3.2.2.3.

3.3.

3.4.

3.4.1.

3.4.2.

Instrument Errors

In the instrument field one commonly encounters the ‘“‘instrument error”. This
can be split into two parts: determinate errors and indeterminate errors.

Where the error is determinate, one sometimes refers to this as a “set”. Where
the error is indeterminate, it is referred to as instrument precision, or instrument
reproducibility. The word precision denotes the extent to which a result is free from
accidental errors. It is important to note that a result may be extremely precise and
at the same time highly inaccurate, the precise result being in error by the magni-
tude of the determinate error.

Dependent and Independent Errors

Errors can furthermore be classified into dependent errors and independent er-
rors. Where a quantity Y is a function of a quantity X, an error in X will produce
a corresponding error in Y. This error in Y is known as the dependent error. Where
the measurement of Y involves an accidental error which is not a function of X, it
would be classed as an independent error. The total observed indeterminate error in
X would be a function of dependent and independent errors observed.

Where the term “error” is used without classification or enlightenment, it is
assumed that it means the indeterminate error, which in turn is a sum of the depend-
ent and independent errors.

Accuracy

The accuracy of a value is a measure of how close to the true value or probable
value the indicated measurement is likely to be. It in reality is the probable error,
and takes into account both the determinate and indeterminate components of error.

Average Values and Average Errors

Average values and average errors are distinguished from probable values and
probable errors by the fact that the former are arbitrarily defined by mathematics,
and do not necessarily have any physical significance unless the assumptions which
are made are valid.

Average Value

In considering a series of measurements of the same conditions, the average
value or arithmetic average of the measurements is mathematically defined as the
sum of the values of the individual measurements divided by the number of meas-
urements taken. This is also sometimes called the “mean value’’. It should be noted
that the average value is not necessarily the most probable value; it is simply the
arithmetic average of the measurements which were taken.

Deviation

The net difference between any one reading of the series and the average value
is called a deviation. Sometimes it is called a deviation. Sometimes it is called a de-
viation from the mean; sometimes it is called a variation; and it is sometimes loosely
termed as “‘error’’. In its most mathematical sense it is called the residual. However,
all these alternate terms should be avoided and it should simply be known as a de-
viation from the average or, more simply, deviation.

3.4.2.1.

3.4.2.2.

3.4.3.

3.4.4,

3.4.5.

3.5.

3.6.

Average Deviation

The average deviation is defined mathematically as the sum of the absolute
values of each individual deviation divided by the total number of deviations ob-
served. The absolute value of course is the value without regard to its sign.

Standard Deviation

The standard deviation is mathematically defined as the square root of the “sum
of the squares of each individual deviation divided by the number of deviations ob-
served”. The mathematical significance of the standard deviation can be found in
most mathematical text books and will not be taken up here. It suffices to say that
the standard deviation is considered the most probable average for the deviation
where normal or Gaussian distribution of the readings is evident.

Discrepancy

Where it is desired to compare the simultaneous readings of two different in-
struments, as opposed to a series of measurements by a single instrument, the dif-
ference in reading between the two instruments can be called the discrepancy. This
is an arbitrary term which will be used here to avoid ambiguity.

An average discrepancy will define the average of a group of discrepancies be-
tween two instruments measuring the same value. Its mathematical treatment is
similar to the average deviation.

Variance

When it is desired to describe the change over a period of time of an average
error or average deviation, it is convenient to call this change the variance.

Variability

In considering a series of measurements of conditions which do not change ap-
preciably with time, it is possible that conditions will vary slightly in the immediate
vicinity where the measurements are taken. The deviations in measurements will,
therefore, be a function not only of'the instrument variations, but also of variations
in the conditions being examined. This latter variation is called variability and can
be described in terms of average or standard deviations, where such deviations are a
true measure of the change of conditions. The term variability should not be con-
fused with the term variance used in the previous section.

Other Terminology

In an effort to clarify and perhaps to standardize the nomenclature used to
describe the errors of the Bathythermograph a glossary of pertinent terms is at-
tached as an Appendix to this report. It is hoped that this glossary will not only
serve to clarify the terminology used in this report, but will also act as a basis
for a more extensive and amended glossary in the field of Oceanography and Bath-
ythermography.

Overall error as a function of individual errors

Where the overall error of an instrument is a result of the contribution of
many individual types of errors as, for instance, reading error, correction errors,
instrument precision and dependent error, the overall error cannot be less than
the maximum of any one individual contribution, nor can it be greater than the

sum of all error contributions. The probability that the error will approach either
this maximum or minimum is very remote. It is much more realistic therefore to
establish a probable value based on each individual contribution.

The overall probable error is related to each individual contribution by the fol-
lowing equation:

ih 2 VCE ee aR? (1)

This asumes that the individual errors and their magnitudes are subject to com-
plete randomness and would tend to follow a normal distribution curve. Equation 1
does not take into account determinate errors since it is assumed that if an error can
be established with regard to its magnitude and sign a proper correction can be made
and such an error eliminated. However, in cases where the determinate is not
accounted for, Equation 1 should be modified to read as follows:

AB = ABoe, + ABiyoet (1a)
The establishment of such a relationship between the individual contributions

and the overall error can be found in any standard mathematical textbook and will
not be derived here.

SECTION 4.0

APPLICATION OF ERROR TERMINOLOGY
TO THE BATHY THERMOGRAPH

The previous sections have described the mathematical definitions of errors and form the
basis upon which the errors of the Bathythermograph can be analyzed and classified.

Consider first the true temperature value of any one point in the ocean at any one instant
of time. Since temperature is an arbitrary reference level of thermal energy, it is subject to
the errors of the reference instruments. However, a temperature can be defined within any ac-
curacy desired, say for instance, + 0.01°F.

Thus at time 6 and depth D the temperature T can be assumed to have a definite value
within the accuracy desired. At time 0 plus AO and at the same depth D a second temperature
T. can also be observed to the accuracy desired.

The question arises as to the exact relationship between T, and T». Ina liquid medium such
as the ocean the relationship between T, and T» can reflect the following four basic changes:
(1) inaccuracies in the temperature sensing element (since temperature is a relative condi-
tion, the accuracy of the sensing element cannot be eliminated, even theoretically) ; (2) the
change in the heat content of the mass of water, under measurement; (3) translation of the
mass of water during the interval AO, this change in mass reflecting a new temperature condi-
tion, or (4) where D is subject to accuracy of measurement, this change can reflect a change in
the location of the instrument making the measurement during time AO.

It is particularly important to note that any variations observed when a series of tem-
perature measurements are made can be caused by any one of the four factors mentioned
above. Thus when it is desired to study and make determinations concerning the change of the
temperature in the ocean or the mass displacement occurring in the ocean, one must consider
and recognize the variations which can arise from variations in the accuracy of the depth and
temperature sensing elements. Conversely when concerned with establishing accuracy of the
temperature and depth sensing elements one must be cognizant of the relative mass and tem-
perature changes which are occurring in the ocean at the same time.

The actual changes in the temperature of the ocean and mass displacement during time
theta (according to the above terminology) is ocean variability. The changes occurring as a
result of the depth measuring element and the temperature sensing element, according to our
mathematical definitions, are instrument errors, or more generally indeterminate errors.

Unfortunately, when dealing with a series of temperature measurements in the ocean, it
is impossible to distinguish between the two types of changes; hence, ocean variability is
always subject to the limitation of definition by the order of magnitude of the instrument
errors. Conversely a study of instrument errors is limited by any ocean variabilities which may
exist at the time the measurements are taken.

Probable values of the temperature of the ocean under various conditions can be
established through series of measurements which take into account ocean variability and in-
strument erorrs. The accuracy of such a probable value is limited by the instrument accuracy
and by the ocean variability which is occurring at that time.

4.1. Determinate Errors of the Bathythermograph

The location of the pen arm of the Bathythermograph with respect to the glass
slide is an arbitrary one, since there are no marks on the glass slide which indicate
actual values of the temperatures or depth. It is necessary to translate the tempera-

4.2.

4.3.

ture depth trace in a horizontal or vertical direction, or both, in order to have the
temperature depth trace coincide with the actual values of depth and temperature
of the ocean. This necessary translation is called depth set or temperature set. These
depth and temperature sets and their corresponding correction factors (known as
DCS and TCS) are, according to the terminology described above, determinate
errors. These errors are determinate because they have definite magnitude and di-
rection of sign and can be duly accounted for.

Indeterminate Errors of the Bathythermograph

The set of the Bathythermograph can change from time to time because of
various factors which will be discussed later. This variation in set is then the inde-
terminate error contained in the determinate error.

In addition there are many other indeterminate errors to which the Bathyther-
mograph is subject. (1) The Bathythermograph has an inherent reproducibility.
(2) It may be subject to a hysteresis error. (3) It may have an error due to the
rate of response of the instrument. (4) It may have an error due to operational
conditions. (5) It may also have an error imposed upon it when corrections of the
slides for the set are made. (6) When the slides are read they are subjected to read-
ing errors. (7) Errors in temperatures may result from errors in depth. (8) Errors
in initial calibration can produce calibration errors; and finally, (9) the error of
the instrument can be a function of the depth, producing a variance in the error.

Bathythermograph deviations as a Function of Errors and Ocean Variability

It may be assumed, therefore, that in a series of temperature measurements in
a fixed position in the ocean the average or standard deviation observed will be a
function of the ocean variability and the errors involved in the instruments used to
make the measurement. The relationship between the standard deviation and its
various components again is given by Equation 1 which relates overall error and the
contributions which make up the error.

ING = WV/ (RR) 2 (NE) soe (NR) SV E(t) (1)

Concerning ourselves with standard deviations which are the correct deviations
to employ where random error is involved, Equation 1 is rewritten as follows:

2 2
Oro TAL -\/ 29] ocean * 2lOlinsorument (2)

Where sigma equals standard deviation

Segregating the standard deviation into the standard deviation of temperature
and standard deviation of depth, the following equation can be written:

(Op) ronan = V 2G) jonny # (Gn) memes (3)

The standard temperature deviations of the ocean and the instrument can be
further broken down into their dependent and independent functions as follows:

(07) ocean ay U(Cp) imate 2 2 (Gq) ame (4)
(Cpoman = V 3GPime * Ow) pee (5)

4.4.

The dependent variability of the ocean is a function of the depth movement and
the slope of the temperature depth trace at that point. Thus the influence of internal
waves would produce a change in temperature at a given depth or a temperature
variability which is actually a function of the water mass movement. (It should be
noted that internal waves do not consist of random motion, and therefore (0 t)
DEP. will not represent a true average condition. Instead, an average based on
harmonic wave motion should be employed.) Equation 5 also indicates a dependent
error in the temperature sensing element due to the error in the depth sensing
element. This error can be related to the depth error as follows:

(op)
(Or) pep = a aad (6)

Combining (5) and (6):
Where S is the slope of the temperature-depth trace in ft/degree.

(Cp) ~str = V 2p) tnoee + = 2 (0p) twoep (7)

Similarly the standard deviation in depth is related to the temperature error
and the independent error as follows:

On) sane =V Gn) aeonp 4 S 2G) cane (8)

The significance of the dependent error of temperature due to errors in depth
can be best described as follows: Let us suppose we have an instrument capable of
measuring to within .01°F and we wish to determine the temperature at a depth of
100 feet. Let us further suppose that at a depth of 100 feet the temperature varies
from 70° at 95 feet to 60° at 105 feet, or 1° per foot. Let us further suppose that we
are not able to locate the exact depth of the temperature sensing element to better
than + 5 feet.

While the temperature recorder may be accurate to within .01°F, the location
will not be known except to within the nearest five feet ; hence the element may have
been anywhere from 95 to 105 feet. The temperature variation through this area is
ten degrees; hence, the absolute value of the temperature at 100 feet cannot be de-
termined with any greater accuracy than + 5°. This relatively poor accuracy is a
result of the dependent error in temperature due to the error in depth.

Measurements of Instrument Error by Special Tests

Since it is mathematically impossible to segregate the variability of the ocean
and the instrument errors by observing a set of measurements, it is desirous when-
ever possible to determine instrument error independent of such measurements. This
can often be done in special test such as will be discussed later. Whenever such
errors can be independently determined they may be assumed to be contributing in
accordance with Equation 1, and the independent value obtained may be subtracted
from the equation to determine by difference the desired ocean variability com-
ponents.

SECTION 5.0

ERRORS INHERENT IN THE INSTRUMENT

The Bathythermograph is an instrument primarily designed to measure change of tem-
perature with change of depth. The word change is emphasized to point out the fact that the
values of temperature and depth registered by the Bathythermograph are arbitrarily estab-
lished by superimposed grids and correction factors.

5.1.

5.2.

Reproducibility

Given a Bathythermograph trace which represents a temperature versus depth
structure, the ability of the instrument to reproduce that trace (in shape but not
necessarily in the same coordinate location) is termed reproducibility. The rated re-
producibility is usually assumed to be + 0.1°F. and + 2 feet in depth. The
instrument’s best demonstration of reproducibility is obtained by the superim-
position of the up and down trace on a slide from a single cast where the vari-
ability of the ocean is low during the cast.

Reproducibility can also be described as instrument precision. The reproduci-
bility or instrument precision is limited by the accidental errors which accumulate
in the mechanics of the instrument and by the limitations and the accuracy of
manufacture.

The reproducibility of the instrument is also limited by other sources of error
which may creep into the picture, such as mechanical hysteresis, mechanical hesi-
tations, and others. These will be discussed in separate sections. The value for
reproducibility given above assumes that such other errors as may mar the repro-
ducibility are not present. This statement is true in a carefully handled and cali-
brated instrument.

Accuracy of Registering Temperature Change with Change of Depth

While the reproducibility of an instrument may be good, the accuracy can well
be poor. One will recall that accuracy is a measure of the instrument’s ability to
reproduce the actual or true conditions. For instance, should the change in the
water be 10° and the instrument show a change of 15° repeatedly, the instrument
would be highly reproducible but poor in accuracy.

Actually the accuracy of the Bathythermograph is governed by the assumptions
which are applied when the grid by which these temperature changes are meas-
ured is made up. Remember that we are concerning ourselves only with temperature
changes and not actual values. However, the grid must be employed to register tem-
perature changes as well as actual values. The limitations on the accuracy which are

imposed by the assumptions involved in making up the grid can best be demon-
strated as follows: If it is assumed that an isothermal line is a straight line on the
temperature depth slide and the Bathythermograph registers this line with a slight
curve, the accuracy will be distorted by the amount of the curve. The accuracy
could be made as good as the reproducibility if the grid is made with a slight curve.
However, present techniques assume straight lines for isothermal conditions.

Likewise, in depth characteristics it is assumed that a movement of so many
millimeters on the scale in shallow water will represent the same changes in deep
water. Any failure of the instrument to meet this requirement will produce in-
accurate change measurements. Discussions of the grid makeup will be deferred
until the section on Calibration. It is well to point out here, however, that the in-

10

5.3.

5.3.1.

5.3.2.

strument is normally considered to have the same accuracy in measuring changes
of temperature with depth as quoted for reproducibility, that is + 0.1°F or + 2
feet in depth.

Hysteresis

The term hysteresis is borrowed from its more exact meaning (that is, a meas-
ure of the irreversibility of a closed work cycle) and is used here to describe the
failure of the up and down trace to superimpose upon each other in an ocean con-
dition of supposedly low variability. This failure to reproduce is manifested in these
particular cases by a space between the two curves, either in their entirety or in
any portion thereof, which is similar to the hysteresis curves obtained in irrevers-
ible work cycles.

Hysteresis can evolve from several sources listed as follows: (1) mechanical
(2) response (38) ocean variability (4) change in lag conditions.

Mechanical Hysteresis

Should the pen arm be loose in its mount, it is quite possible that a play in the
pen arm of several degrees could be present. This will produce up and down traces
which are not superimposed but displaced with respect to each other by the amount
of play in the pen arm. Under special ocean conditions, notably where there may
be positive as well as negative gradients at different depths, the distortion as
well as displacement can exist. Likewise, a loose glass slide or loose slide holder
can produce mechanical hysteresis.

Such a hysteresis is an indeterminate error unless one of the traces is repro-
ducible from slide to slide. One has then only to correct for the order of the magni-
tude of displacement between the up and down trace.

Slides have been noted where mechanical hysteresis of this nature has ap-
proached 3°F.

Response Hysteresis

The failure of the instrument to respond rapidly to sudden changes in tempera-
ture conditions (more accurately termed large temperature gradients) will produce
hysteresis in that portion of the curve represented by displacement between the up
and down trace.

This is most easily demonstrated by example. Consider a true state of condi-
tions in the ocean represented by a step-wise Bathythermograph; that is, an isother-
mal section followed by a horizontal section representing say a drop of 10° followed
by another isothermal section. Hysteresis will result because of the failure of the
instrument to follow this rapid gradient change. The down trace will be below the
horizontal portion of this step while the uptrace will be above the horizontal por-
tion of the step.

The response characteristics of the machine is rated such that: the 63% of the
temperature range can be traversed in the period of 0.27 seconds. This rate of re-
sponse should be adequate for almost all conditions which would be encountered in
the ocean. However, one must remember that the ability of the instrument to follow
the changes in gradient of the ocean is a function of the rate of lowering and raising
the instrument. This possible source of error is discussed in a later section under Op-
eration.

11

5.3.3.

Rate response curves can be calculated according to the following set of equa-
tions:

Let T = indicated temperature recorded by the bathythermograph
at any time @.

Let T. = initial temperature of bathythermograph.
Let T: = actual temperature of surroundings of bathythermograph.

Let tT = a constant.

Then: et -+ (T, - 7) (1)

For T:; = constant, the solution to (1) is as follows:

T-T
To- T,

Tau may be experimentally determined as follows: Assume the BT is moved
suddenly from T. to T: (which is experimentally done by plunging the BT suddenly
into a cold bath for instance).

The relative response conditions at time 9 = Tt, would be as follows:

a" (2a)
a
Botts
Rearranging :

T, -T 4

se Oe een

et ae 0.632 (2b)

Where T: = indicated temperature at time 6=T

Since T. - T: is the total change that the BT would undergo, and T. - T, is the
change actually gone through at time 8 =T, one has only to register the time at
which the BT records a temperature approximately 63% of the way between the
initial temperature and the bath temperature. Time value is numerically equivalent
to T, the response constant of the BT.

Normally the response time is reported for such instruments, and represents
the time to travel 63% of the desired scale limits.

There have been many questions which have arisen with regard to the response
characteristics of the instrument. These questions all point to a definite need for
more experimental programs to be undertaken with regard to rate response. Rate
response knowledge is important in order to differentiate between this type of hys-
teresis and the hysteresis resulting from ocean variabilities.

Hysteresis Due to Ocean Variability

The up and down traces may not coincide in portions of the BT trace because
of a change in the ocean conditions between the time the downtrace was made and
the uptrace was made. This notably occurs in the presence of internal waves. This is
not an error of the instrument, but is a true measure of ocean variability. However,
as previously pointed out, ocean variability cannot be distinguished from instrument
error in any one given set of data except by independent means of checking.

12

5.3.4.

5.4.

5.4.1.

5.4.2.

5.4.3.

5.5.

Change in Lag Conditions

In some rare cases hysteresis can result from a change in lag conditions. This is
brought about for example by allowing the Bathythermograph to plow into the
mud. When the Bathythermograph comes up again it is covered with mud, thus in-
sulating the thermal element. The response to temperature changes is thereby de-
layed and a curve displacement results. Such a displacement may result also from
sudden shock which will change the location of the pen arm. This type of displace-
ment is usually a permanent one, however.

Response Characteristics

Response Hysteresis

Response hysteresis has already been discussed in the above sections.

Response at Initial Depth

There is good cause to believe that the Bathythermograph may not respond with
respect to its pressure sensitive element in the initial depths.This is borne out by
experiments which compare the Bathythermograph trace to traces made by ther-
mister beads.

The lack of this response of the pressure bellows in the initial depths followed
by a sudden response of this element to the conditions would result in a straight or
isothermal line from zero feet to the point at which the bellows element responded.
The temperature registered for the zero depth value therefore would be that actually
found at the depth at which the Bathythermograph responded, say 10 feet. This is
vitally important when considering the desirability of checking the absolute value
of the zero depth reading of the instrument with an independent measurement such
as the bucket temperature. If the Bathythermograph is not registering an actual
zero depth value but rather a line representing the 10 foot value, such comparisons
can be in great error. This is further discussed under temperature correction ap-
plications.

Mechanical Hesitation

In just the same manner as the Bathythermograph is believed to hesitate in the
initial depths before responding, so it is quite possible for the Bathythermograph to
hesitate all along its path producing a series of minor wiggles in the Bathythermo-
graph trace. These minor wiggles, if they are due to mechanical hesitation, are
consequently a source of error.

There are no experimental data to support the presence and, if so, the magnitude
of these mechanical hesitations. It can only be stated at this point therefore that
such mechanical hesitations can exist as a possible source of error and should be
considered whenever wiggles in a trace become an important factor in analysis of
the data.

These mechanical hesitations could easily look like true ocean variability, or
thermal microstructure. Some data suggest that the Bathythermograph does not
reproduce thermal microstructure while other data indicates very fine wiggles equi-
valent to thermal microstructure. This indicates that a resolution of these errors be
made by further research.

Bathythermograph Sets

All of the conditions discussed above are representative of possible sources of
errors which will reflect in the reproducibility and accuracy of the instrument in its

13

5.6.

5.6.1.

5.6.2.

5.6.3.

ability to measure temperature change with change in depth. So far nothing has
been mentioned about the recorded values as related to the actual temperature values
of the ocean.

The process of establishing reference values for the Bathythermograph trace is
simply one of transposing the slide containing the grid calibrations until the mark
which was made by the Bathythermograph pen arm at the zero depth level corre-
sponds to the zero depth level on the grid and the temperature value corresponds to
the actual temperature of the ocean, possibly recorded by a separate instrument.

When the instrument is released by the manufacturer, or after it has been
checked, the positioning of the grid is made at that time so that the Bathythermo-
graph slides supposedly read not only the changes, but also the actual values rep-
resented there.

Any departure of the trace thereafter with respect to its absolute position rela-
tive to the grid will produce a false reading. This false reading will be in error by a
determinate amount, i.e., the magnitude and direction of the error can be estab-
lished. These determinate errors which manifest themselves in terms of displacement
of the trace from the actual point at which it should be located are called depth and
temperature sets.

It should be well noted that the Bathythermograph was not designed to correct or
maintain its position relative to the grid, but only to reproduce changes. The position
relative to the grid is a man-made one and is an arbitrary one. This attempt to posi-
tion the grid to read actual values and the correction techniques involved is a source
of indeterminate errors. These errors will be taken up in a later section.

Instrument Failures

General

The greatest source of errors occur naturally when the instrument components
fail or become damaged. The most common sources of failure are listed below.

Bending of Pen Arm

The pen arm position relative to the grid is established at the point of calibra-
tion or manufacture. Should the pen arm be bent the result would be a displacement
of the trace and the resultant error in the set. This condition has been quite preva-
lent as a result of two things — (1) if the Bathythermograph is left in the sun the
heating element carries the pen arm beyond its normal limits and thus bends it, and
(2) if the Bathythermograph is retrieved from the water too fast, the pressure
element overrides and jams against the pen arm, thus bending it, in some cases.

Corrective measures for both of these conditions have been suggested in the
literature. It merely suffices here to repeat that such corrective measures should be
employed so as to minimize these possible sources of error.

Superimposition of Trace and Destruction of Trace

At present the up and down trace are intended to superimpose on each other.
While this indicates the reproducibility of the instrument, it is often time disadvan-
tageous, whenever it is desirable to make accurate calculations with regard to tem-
perature structure. It is often impossible to determine which is the up and which is
the down trace and thus to follow it through continuously. It would be decidedly ad-
vantageous to design for scientific use an instrument which displaces the uptrace

14

5.6.4.

5.6.5.

from the downtrace by a standard amount, say 5°, thus producing two independent
traces which can be distinguished at all times, and which nevertheless still can be
superimposed for comparsion by any number of simple techniques.

A further disadvantage of the superimposition is that often times the entire top
portion of the trace is destroyed by the pen arm after the Bathythermograph has
surfaced and skips along the waves. The jarring and jolting of the pen arm causes
the destruction. A pen lifting attachment which removes the pen from the trace in
the last 50 feet of travel has been described previously in the literature. This is an
advisable procedure in that it will preserve the downtrace of the pattern.

Bellows Failure

The 900 foot Bathythermographs have been demonstrating a high percentage
of bellows failures through fatigue. Failures have been occurring after one hundred
to two hundred lowerings; there is even skepticism as to the accuracy of the depth
sensing elements even before complete failure. One example was cited where a cruise
vyhich was carrying eight 900 foot Bathythermographs failed to bring back one op-
erable instrument, all having failed because of bellows fatigue. The remedy seems to
be in the type of bellows used and the mechanism for strengthening and guiding the
bellows. Corrective measures are being taken by the Bureau of Ships.

Loose Parts

A great many Bathythermograph traces are rendered invalid because of loose
parts such as the pen arm, producing hysteresis and distortion. The best remedy for
this is more frequent checking of the Bathythermographs during operation by
qualified personnel.

15

6.1.

Galeie

SECTION 6.0

ERRORS INTRODUCED BY CALIBRATION AND
RECALIBRATION OF THE BATHYTHERMOGRAPH

General

The calibration tests to date are the only tests that a Bathythermograph re-
ceives which can in any way check its accuracy and reliability. These calibration
tests are in general too infrequent and not of sufficient extent. Calibration tests
under idealized conditions and employing secondary temperature and depth meas-
uring instruments can provide some reliable indications of the possible sources of
errors of the Bathythermograph, these being determined experimentally independ-
ent of ocean variability.

At the same time it should be noted that calibration techniques are also a pos-
sible source for the introduction of errors. These errors would originate primarily
from failure to determine whether or not the Bathythermograph has linear charac-
teristics. In particular there is evidence to show that the pressure elements are not
linear with respect to depth measurements. These conditions can be determined ex-
perimentally in the calibration procedures, however.

The present techniques for calibration and recalibration procedures are satisfac-
tory as far as they go, but do not cover sufficient grounds. Briefly the problems may
may be summarized as follows:

a. Notests are run before Bathythermograph is repaired.

b. Tests are not conducted often enough on each Bathythermograph unit.

ce. Test log and history of repair of Bathythermograph are not kept at all or is
incomplete.

d. There is no uniformity between test stations.

The capacity of the testing stations is too small for the amount of Bathy-
thermographs which must be calibrated.

f. The tests are not complete enough.

(1) Do not account for hysteresis.
(2) No rate of response tests.

(3) No stability tests (run through a second time to see if same as first
time) (initial fatigue of new parts).

(4) No smoothness tests are run (change of temperature with change of
depth under idealized conditions to test smoothness of response).

Pre-Repair Tests

There are at present no tests conducted (except in certain special cases) to
determine how much in error an instrument is before the instrument is cleaned, re-
paired, and put into shape. Such a pre-repair test would give valuable information
which would aid in evaluating the reliability of the data taken by the instrument,
and the corrective factors to be applied when indicated, particularly if a BT log
was kept.

It is true that at present the majority of instruments sent in for recalibration
are completely inoperative to begin with. This condition lies partly with the length of
time an instrument is in use before it is sent back for recalibration.

16

6.1.2.

6.1.4.

6.1.7.

It is also true that many instruments may be cleaned up and repaired to a minor
degree by the field personnel before being returned to the testing stations, thus
rendering pre-repair tests ineffectual. But these conditions can be ironed out.

Frequency of Calibration

As mentioned above, most instruments are recalibrated when they are complete-
ly inoperative. There should be a properly scheduled system of recalibration of in-
struments without their having to become inoperative. Recalibration at frequent
periods serves to produce more reliable data and more reliable corrective factors
for data taken. A three months recalibration period has been generally suggest-
ed as a goal to shoot at. Of course, the more frequent, the better, commensurate
with reasonable cost, etc.

Bathythermograph Test Log and Repair History

In many cases a history of repair is kept of a given Bathythermograph. This
has been confined to special instruments used by scientific personnel.

A complete, up-to-date, test log and repair history for a bathythermograph
unit would (1) aid immensely in the evaluation of the data obtained, (2) produce
a statistical record of Bathythermograph performance, showing its main manu-
facturing weaknesses, and the major causes for damage by the personnel operating
them, and (3) establish the optimum time for scheduled recalibration before be-
coming inoperative.

Uniformity Between Testing Stations

Standardization of testing procedures requires that the various testing sta-
tions employ the same techniques for testing, and if at all possible, the same type
of testing equipment. In addition, the independent absolute measuring devices
should be cross calibrated between stations to insure standard absolute values.

At present there is some standardization of testing stations attempted through
the proper Navy manuals for testing procedures, but in general the actual
practices followed are quite varied. It seems quite in order that uniformity be-
tween stations be a part of a general program outlining standardization for
all test procedures.

Capacity of Testing Stations

The older installations hold one or two Bathythermographs for simultaneous
test cycles. The newer units on the West Coast can hold up to six units. In view
of the large number of Bathythermographs that need frequent recalibration, the
testing stations should be designed with the idea of being able to test several
units at one time. This is particularly important if the testing cycle is increased in
scope to account for the aforementioned features.

Hysteresis

It is important that a test pattern to inspect for hysteresis be made in the
calibration procedure. A concurrently acting temperature versus depth trace which
was then reversed would show any tendency to hysteresis in the instrument.

Rate of Response

No test at present computes the rate of response of the Bathythermograph to
changes in temperature and depth. Rate of response is particularly important in

17

6.1.8.

Opi),

6.1.10.

waters of high thermal gradients. Certain recent anchor stations indicate a need
to investigate the rate of response further. Incorporation of rate of response into
the testing procedures seems advisable as part of the procedure of checking all
phases of operation of the Bathythermograph.

Reproducibility Tests

The simple procedure of running a Bathythermograph through two simul-
taneous cycles would give adequate determination of the ability of the Bathy-
thermograph to reproduce results, and an indication of its stability. It is a fairly
well established fact that when a Bathythermograph is repaired, especially when a
bellows is replaced, the first flexing under water pressure it receives produces an
initial fatigue set. It is almost mandatory that a new or freshly repaired instrument
be flexed to overcome this initial fatigue set.

Smoothness Tests

Major and minor wiggles have long been observed in the traces of operated
Bathythermographs. There is much debate as to whether these wiggles are pres-
ent in the ocean or the instrument.

The same tests that would be conducted for hysteresis would also show the
smoothness of response, and would resolve the origin of the wiggles as to in-
strument or ocean. Operationally, the smoothness test would show if the instru-
ment had any inclination to hang or catch along its temperature depth route. This
is especially important in the very top layers, where small amounts of friction and
inertia will prevent the instrument from responding.

Standardization of Testing Procedure

It appears well in order that the whole picture of calibration be thoroughly
studied with a view to recommending the best and most desired testing proce-
dures. Adoption of such recommendations would serve to standardize the entire
testing program, giving more uniform and more reliable results.

18

SECTION 7.0

ERRORS INTRODUCED BY OPERATING TECHNIQUES

Errors introduced by the operating technique take into consideration two distinct features:
(1) Errors produced by the technique itself, and (2) Errors produced by the personnel oper-
ating the Bathythermograph.

7.1. Errors Produced by Handling the Bathythermograph

The major faults of the Bathythermograph from an operational standpoint
may be summarized as follows:

a. Can be operated with slide not in all the way.
b. Can be broken by leaving in the sun.
ec. Can destroy the trace in the surface layer when recovered from the cast.

While it would be considered advantageous if the instrument could be designed
not to operate unless proper seating was first accomplished, it is still the fault
of the personnel not to seat the slide properly.

Again a fault of the personnel, the instrument’s pen arm can be seriously
bent if left in the sun exposed to too high a temperature. Some manner of dis-
engaging the pen arm would eliminate this very common error.

Destruction of the upper part of the trace is often caused by the skipping and
bouncing of the instrument upon recovery. This has been eliminated in newer
instruments by a pen lifting device which disengages the pen on the “up” trace
of the cast in the last fifty feet approaching the surface. Sometimes, however, this
disengaging device, actuated by the protecting sleeve, renders the Bathythermo-
graph inoperative unless care is taken to have the sleeve in the proper position,
this proper position being left to the attention of the operating personnel. The
result has been blank slides.

7.2. Errors Produced by Supplementary Equipment

The principal supplementary equipment is the winch used to raise and lower
the Bathythermograph. One of the major operational difficulties is bringing the
Bathythermograph aboard, especially in rough weather. A one speed winch causes
the Bathythermograph to swing violently in an arc as it frees the water, and some-
times results in jamming the Bathythermograph against the davit blocks. An ad-
justable speed winch would facilitate handling and boarding of the Bathyther-
mograph.

As previously pointed out, there is evidence to support the fact that the speed
of raising and lowering the Bathythermograph can effect the trace. A number of
Bathythermographs lowered experimentally by hand showed no hysteresis whereas
other Bathythermographs lowered under the same conditions, but on the winch,
showed considerable response hysteresis. It is quite possible that the speed of raising
and lowering the Bathythermograph can introduce a source of error under certain
conditions.

In view of the high degree of damage the Bathythermographs suffer, it seems
advisable to study a little more thoroughly the design and operation of the winch and
davit system used to lower and raise the Bathythermograph. Extra expense in im-
proving the supplementary equipment might be more than compensated for by the
savings in the Bathythermographs and improvement in the Bathythermograph data.

19

7.3.

Errors Produced by the Operating Personnel

The chief comment which can be made concerning the operation of the Bathy-
thermograph is that the ultimate accuracy of the data taken by the instrument de-
pends as much upon the reliability of the personnel as it does upon the instrument
itself. It has been stated that portions of the data received from the tactical per-
sonnel cannot be regarded as reliable.

Most of the personnel professionally engaged in oceanographic work appreciate
and understand the limitations of the Bathythermograph instrument, i.e., its in-
ability to picture microstructure, its peculiarities with respect to hysteresis, its
dependent errors of temperature and depth, and its variations from the absolute
measurement of temperature and depth. These limitations are not generally under-
stood, however, by the majority of Navy personnel who operate the Bathythermo-
graph; nor is the care and use of the instrument understood and appreciated by
those engaged in gathering a great percentage of the Bathythermograph data.

It is strongly recommended that an abstract be prepared on the capabilities,
limitations, peculiarities, and characteristics of the Bathythermograph and directed
in its preparation towards the data procurement personnel. It is felt that if all levels
fully understand the importance of the data and the proper treatment of the instru-
ment in the collection of this data, more of the data collected would be reliable.

The errors manifest themselves in terms of (1) the lack of care and handling
of the instrument causing jarring and failure, (2) lack of maintaining proper and
complete records to the extent of falsifying or not taking readings, and (3) errors
involved in reading instruments and reading the Bathythermograph slides.

A considerable amount of data could be interpreted and analyzed if the op-
erating personnel were careful in maintaining complete notes on conditions that
existed at the time of the Bathythermograph cast. For instance, such things as
double traces, sudden changes in sets, abnormal BT traces, blotches, etc., if accom-
panied by the operating experience at that time would provide invaluable informa-
tion. This will require the taking personnel to examine each BT card when it is taken
to determine if such irregularities existed, and to report any abnormal conditions
which may have existed simultaneously. Failure to do this results in lack of relia-
bility of the data taken.

20

SECTION 8.0

ERRORS INTRODUCED BY PROCESSING
OF THE BT DATA

The method of processing BT data primarily involves superimposing the calibrated grid
of the Bathythermograph over the raw slide and thereby establishing their relative positions,
giving not only the correct temperature changes but also the correct temperature values.

Assigning correct temperature values involves locating the grid with relation to the slide
through the medium of temperature and depth set corrections. In the field and under the as-
sumption that no changes have occurred since calibration, the grid can be matched to the raw
slide by matching the edges of the grid with the edges of the slide. Whenever a set develops
in the Bathythermograph, the grid is then moved from its original coincident position by the
set correction to account for this error.

It is necessary to establish correct temperature values for each Bathythermograph in
cases where it is desired to compare a series of Bathythermograms taken over a period of
time. In this case one is not only interested in the relative changes which occur on any one
Bathythermogram, but also the relative changes which occur between Bathythermograms.
These relative changes which occur between Bathythermograms must not include the changes
in the temperature and depth correction factors of the Bathythermograph.

The change in sets of the Bathythermograph is a greater source of error than any
variance in its precision or reproducibility. Since sets are determinate errors, the magnitude
and direction can be calculated and allowed for. It is this process of establishing the determi-
nate error and making proper corrections which warrants careful analysis.

8.1. Temperature Correction Technique

The present practice employed is to correct Bathythermograph temperatures by
a supplemental independent surface temperature measurement. This is accomplished
by making simultaneous bucket temperature casts or by reading the water injec-
tion thermometer.

The technique for applying the temperature corrections varies considerably, the
technique being dependent primarily upon the reliability of the independent surface
temperatures taken. These in turn depend upon the experience of the personnel and
the care they exercise in taking the surface temperature measurements.

Where exceptional care is taken in obtaining bucket temperatures, individual
corrections are applied to each Bathythermogram. This technique is limited to special
scientific studies where the precision desired is justified by the experience of the
personnel taking the data.

The majority of corrections are not applied individually to each corresponding
Bathythermogram. Instead, the discrepancy between each surface temperature and
the corresponding Bathythermogram are accumulated for one specific Bathythermo-
graph over a period of time and averaged. This average discrepancy is used as a
correction factor, and is applied to each Bathythermogram involved in the averages
when corrected photographic prints are made.

21

8.2.

8.3.

8.3.1.

The reason for taking averages is twofold: first, it facilitates the photographic
process when only a single correction setting has to be made. Secondly, the inde-
pendent surface temperature measurements are themselves subject to considerable
inaccuracies. These inaccuracies are ironed out, or so it is believed, when a series of
measurements are averaged.

The length of time over which these averages are taken varies considerably.
These periods end whenever there seems to be some sharp break in the discrepan-
cies (difference between surface value of the Bathythermograph and the independ-
ent surface measurements) noted for a particular instrument.

Sometimes no correction factor is applied. This occurs when considerable doubt
is cast upon the validity of the independent surface thermometer readings (bucket
temperature or injection temperatures). One check for inconsistencies is accom-
plished by comparison with a chart which shows average surface temperatures for
various times of the year for various ocean areas. Another check is gotten by look-
ing for unwarranted variations in the reported surface temperature values while
the Bathythermogram shows a steady repetition.

Depth Correction Technique

Each Bathythermogram is subject to errors along its depth scale which are
noted by comparing the zero depth grid line to the horizontal trace usually found
on all Bathythermograph slides at the true zero depth level, caused by the rapid
change in temperature from air to water.

Individual corrections are seldom, if ever, applied. Again an average discrep-
ancy is recorded for a series of Bathythermograms, and an accumulated average of
varying periods taken. Interruptions of these averages are effected whenever a sharp
change in the values is noted. A single setting facilitates the photographic process,
as was the case in temperature corrections.

Errors Involved in Correction Techniques

The methods described above to establish temperature and depth correction
factors involve four basic weaknesses which are possible sources of error. These are
summarized as follows:

1. Correction of Bathythermograph to poor secondary information.

2. No allowance for variations in set unless there is a sharp discontinuity.
3. Lack of reliability of the zero depth value of the Bathythermograph.
4

Applying the corrections at the most unstable point of the ocean, the sur-
face.

Poor Secondary information

The ability to correct the Bathythermograph to read correct temperatures is
limited by the extent to which the independent measurements are reliable. Here is
where the human element plays a very important part. The independent tempera-
ture measurements, whether they are made by a bucket thermometer or injection
thermometer, must be correctly read and correctly recorded. As has been previously
pointed out under Section 7.3, the operating personnel often record erroneous data
either accidentally or otherwise. This is especially true where injection thermometers
are employed for the independent measurement. The thermometers are not read cor-
rectly and in many cases are not read at all, but the last reading is merely repeatedly
recorded.

22

8.3.2.

Another major contribution to incorrect data is the unreliability of the inde-
pendent measuring instruments themselves. Injection thermometers notably cannot
be read any better than + 0.5°F and yet are expected to correct the Bathythermo-
graph to within 0.1°F. More often than not the injection thermometers are stuck,
or are off calibration. There is at present no program requiring these injection ther-
mometers to be calibrated. A small expenditure to provide a much more accurate in-
jection thermometer would increase the reliability of tremendous quantities of
Bathythermograph data which are taken at great cost.

Similarly the bucket thermometer which is normally provided in the kit in
which the Bathythermograph comes is entirely inadequate in its readability. Here
again a better thermometer should be provided, one which has a readability to the
accuracy desired of + 0.1°F.

Another source of error is a poorly designed bucket with which surface samples
are taken. There are considerable data to show that the temperature can drop sev-
eral degrees between the time the sample is taken and the time it is read, due to
high evaporation losses. In other cases the temperature can rise from the radiation
of the sun causing heating of the bucket. There are buckets which have been de-
signed to provide for good insulation and low evaporation losses, so that the tem-
perature of the water when read will be the same as the temperature when the water
was actually sampled.

If in any case the water sampled for independent measurement is not the same
as the water through which the Bathythermograph was lowered, the independent
measurement will bear no accurate relationship to the Bathythermograph itself.
This is more completely discussed in Section 8.3.4.

Variation in Sets

The correction of the Bathythermograph involves an independent measure-
ment of the water by another instrument. It is assumed that these measurements
are made simultaneously. The difference between the reading of two instruments
taking simultaneous measurements has been termed discrepancy. This discrepancy
is a determinate condition with an indeterminate portion represented by the con-
tribution of the errors of the two measuring instruments.

The variance in a discrepancy can be theoretically caused by three different
conditions. (1) The Bathythermograph can contribute 100% of the variation due
to its variation in set. (2) The independent measuring instrument can contribute
100% of the variation, due either to its inaccuracy or to the fact that it is not really
measuring a simultaneous condition. (3) The variation and discrepancy can be
caused by the combined contribution of both instruments, as pointed out in (1)
and (2).

Where condition (1) is in evidence, then the application of an average discrep-
ancy as described in Section 8.1 above does not allow for the variation in set of the
Bathythermograph, therefore, it does not correct the Bathythermograph to the ex-
tent to which it could be corrected if each individual discrepancy were applied to
each Bathythermograph.

If condition (2) exists, then the varations in the secondary measuring instru-
ment can be smoothed out by employing average discrepancies such as described in
section 8.1. Where the secondary measurements are considered to be less accurate
than the Bathythermograph itself, consideration can even be given to applying no
correcting factor at all. Where condition (3) exists, it must be determined whether
these conditions lie closer to condition (1) or condition (2) and thence where the
individual or average correction should be applied.

23

8.3.3.

ie)
(SS)
ay

Actually all three of these conditions can and do exist. Where exceptionally
careful personnel are employed and exceptionally good secondary measuring in-
struments are used, individual correction should be applied, since this corresponds
to condition (1). Where Bathythermographs have been taken using injection ther-
mometers for corrections, it has been shown that these correspond more closely to
condition (2) and average corrections are in order or even no corrections at all.
All Bathythermograph data should be studied to determine whether they correspond
to condition (3), and if so which would be the better method to employ, individual
correction or average corrections. Failure to use the best method for correcting the
data can inject greater sources of error than not employing any correction to the
Bathythermograph slide.

Factors Affecting Zero Depth Reading of the Bathythermograph

It has been pointed out previously that the zero depth value of the Bathythermo-
graph may in actuality not be a zero depth value at all but is the value at whatever
depth the pressure element of the Bathythermograph responded. Should, for in-
stance, the Bathythermograph pressure element stick until the depth of five feet is
reached before it suddenly moves, the line it forms from 0 to 5 feet will register
the temperature at five feet. Furthermore, it is at the surface where considerably
high gradients as well as thermal microstructure exist. Failure of the Bathythermo-
graph to respond either in temperature or pressure in the initial depth of even one
foot will produce erroneous results when attempting to apply correction factor at
that point.

Reliability of Surface Bucket Samples

Not only does the extreme temperature gradient in the few inches of the surface
have an effect upon the actual value recorded by the Bathythermograph, it also has
a considerable effect upon the reliability of the sample which is taken by the bucket.
Thermal gradients exist not only vertically, but also horizontally. The surface bucket
scoops up a sample of the first few inches of the surface of the water. The temperature
that this represents is being compared to the zero depth reading of the Bathyther-
mograph which is in reality some average of the first few feet due to the inherent lag
of the instrument in the initial descent. Should there be any time lag between the
taking of the bucket sample and the lowering of the Bathythermograph, it is quite
possible that the bucket will scoop up a sample which is separated from the point
at which the Bathythermograph was lowered by an extreme horizontal gradient. In
other words, the bucket sample is a poor representation of the surface conditions of
the ocean since it samples such a small amount of the surface, and this amount is
subject to random surface variations which will not be reflected in the Bathyther-
mograph record. This is particularly true on the edges of currents.

Where injection temperatures are read, the injection intake is considerably be-
low the surface of the water. Thus the injection temperature does not represent a
surface reading, but a reading in the subsurface, the distance below the surface
being a function of the pitch and roll of the ship. The injection thermometer cannot
be regarded, therefore, as a reliable indication of the surface conditions at the point
where the Bathythermograph was lowered.

The attempt to compare Bathythermograph surface value with an independent
surface temperature measurement is, in theory at least, a poor technique because of
the fact the sampling is being done in the most unstable portion of the ocean. Com-
parisons would be much better made at some considerable depth where the ocean is
more stable and where conditions do not vary so radically in their spacial relation-

24

ships. While this in practice may be hard to achieve, nevertheless definite considera-
tion should be given to establishing comparison at some other point, rather than the
surface of the ocean.

Another approach to the problem would be to establish whether the Bathyther-
mograph reads an average value of the first few feet or an actual value. If the
former is true, then some means should be made for obtaining an average bucket
sample rather than a superficial surface sample. Improving the sampling conditions
might well be the answer to increased accuracy in establishing Bathythermograph
corrections.

25

SECTION 9.0

PROPOSED METHODS FOR DISCOVERING, EVALUATING,
AND MINIMIZING ERRORS IN EXISTING AND
FUTURE BATHYTHERMOGRAPH DATA

The recognition and classification of the various sources of error which are involved in the
collection and processing of Bathythermograph data provides definite indications as to meth-
ods of attack for evaluating and minimizing these errors when employing these data for
theoretical studies.

These techniques for minimizing errors must bear a direct relationship to the use to which
the Bathythermograph data is to be put. For instance, should the study only require a knowl-
edge of temperature differences between one depth and another, it is not necessary to be
concerned with the absolute values which the Bathythermograms represent. On the other hand,
should the studies be concerned with the variability of the ocean with respect to time, then all
variations in correction factors must be eliminated, or at least minimized as best one can.

Minimization of the error in reproducibility in the Bathythermograph lies for the
most part in experimental investigations. Such investigations will not be taken up here,
but suggestions will be found in Section 10.0, Recommendations. On the other hand, mini-
mization of errors in Bathythermograph sets lend themselves to mathematical studies. These
are taken up in detail below.

9.1. Establishing the Reliability of the Surface Temperature Correction Technique

9.1.1. Use of Trend Charts

Visual comparison of the variations and trends in the bucket or injection tem-
peratures as compared to the uncorrected zero depth values for the corresponding
Bathythermograms is one of the best methods for determining the reliability of
correcting the Bathythermograph sets by this independent measurement.

When sudden changes or discontinuities appear in the trend charts, three pos-
sibilities are evident:

(1) <A change has occurred in the set of the BT (noted by a change in the
zero-depth plot which is not followed by a corresponding change in the
bucket temperatures).

(2) A change has occurred in the injection thermometer, due to error of read-
ing, or failure of the instrument (noted by discontinuity in the bucket
temperature plot not observed with the Bathythermograph zero depth
values).

(3) A change has occurred in both curves, representing a decided change in
ocean conditions.

The degree of “randomness” can be observed in both curves, which is a good
indication as to the relative accuracies of both instruments. Where the random-
ness of the bucket temperatures is small compared to that of the zero depth value
of the Bathythermograph, then this is good indication that individual, rather than
average, correction factors be applied. If the converse is true, then average correc-
tion factors are better.

26

A notable exception to this rule occurs when the bucket or injection tempera-
ture shows no variation whatsoever over a period of time. This is not an indication of
a good instrument, but rather it is an indication that no readings were made, and
the value of the last reading taken was repeated over and over. Contrary to the rule
of low randomness, this indicates a very unreliable bucket or injection thermometer
reading, and the average values should be relied upon.

9.1.2. Comparison of Standard Deviations

When the average value is calculated for a series of zero depth values, and a
corresponding average is obtained for the injection or bucket thermometer values,
their standard deviations from this average can also be obtained.

With reliable instruments, both standard deviations should be approximately
equal, representing for the most part the variations in the ocean. (See Section 4.3,
Equation 3). If the two standard deviations are not similar, either one of two con-
ditions is prevalent: (1) the one with the larger standard deviation has less ac-
curacy and reliability; (2) the one with the smaller deviation represents a case of
repeated readings rather than measured ones. The latter case can be quickly de-
termined from the individual values.

9.1.3. Evaluation of Variability Data by the Coefficient of Correlation

The extent to which the data represent true ocean variability as opposed to
the random errors of the two measuring instruments can be obtained through the
coefficient of correlation. For any two independent measurements, the extent to
which one measurement is dependent upon the other, i.e., Y is a function of X,
is given by the correlation coefficient.

If the coefficient of correlation is high, then the variations of the readings of
both instruments are a measure of variations in the ocean itself. If the coefficient of
correlation is low, then the variations of both instruments are random variations and
no conclusions about ocean variability should be attempted.

The coefficient of correlation may be derived mathematically as follows:

ule Nim = XS Vip nn VE
r=7.2" @ 4
Wee) cee 1)
Where r = coefficient of correlation
N = number of readings cigs
xX = arithmetic average of the raw score, or i - i X,
xX: = each individual reading

y and y: are defined as ¥ and x: (specifically applied, x: and X are the
individual and average bucket temperatures, while y: and y are
the corresponding individual and average zero depth values of
the Bathythermogram.

standard deviation of x:

standard deviation of s:

ox

II

Oy

The coefficient of correlation of two sets of variates, expressed in their standard
deviation as units, is the arithmetic mean of the products of deviations of correspond-
ing sets of values from their respective means.

Since ox and os are constant;

a(x, i x) (vy, a y)
Ty Sie ae eet a Eee ets (2)
N 0,0,

fe

27

From the coefficient of correlation, the extent to which the readings of both in-
struments expressed in standard deviations measure ocean variability can be ob-
tained. This is derived as follows:

Sn 2 Oy (hae) 2 (3)

2
Sex = O.. (1—r \2 (4)
Where S; = standard error of estimate of y

Ss = standard error of estimate of x
If the correlation is extremely high, the standard error of estimate corresponds

to the random errors of the instrument:
From Section 4.3:

Sy = (on) more. (5)
Thus:
(On) rota, =V 2 (55) ocean + 2 (or) tusre. (6)

Oy =V (on) ocean + (Or) instr. (7)
Combining Equations (3), (5) and (7):
(Gq) 5
1) INSTR. = (o,)?

2
ae ocean * (Or) instr. (8)

Which reduces to

(O79) ocean J ie
(Op)instr. V 1-r* (9)

The coefficient of correlation does not indicate the relative reliability of the in-
struments (Bathythermograph and bucket or injection thermometer). The magni-
tude of the standard deviations must also be considered.

When the coefficient of correlation is high and the standard deviations are
equal, the resulting standard errors of estimate are low, and both instruments
are relaible to the accuracy of the standard error of estimate. The ocean varia-
bility can then be measured by the standard deviations.

When the coefficient of correlation is low and the standard deviations are also
low, there may be negligible ocean variability, assuming the conditions of 9.1.2 are
observed.

When the coefficient of correlation is low, and the standard deviations are high,
one or both of the instruments has a high random error. If the magnitudes of the
two standard deviations are dissimilar, probably the instrument with the higher
standard deviation is the culprit. In any case, no reliability can be placed on the data
representing ocean variability.

28

9.2.

9.3.

9.3.1.

Use of Two Bathythermographs Simultaneously

The use of two Bathythermographs simultaneously rather than one Bathyther-
mograph and bucket thermometer will give invaluable information in regard to ocean
variability versus random error.

Employment of the coefficient of correlation as outlined above, but using the
values recorded by the two Bathythermographs, (not only at the zero depth value
but at many depths) will provide a measure as to the extent of ocean variability
and random error of the instruments.

High correlation coefficients between the two Bathythermographs will indicate
that the variations observed are really in the ocean. Low correlation coefficients will
reveal that these variations are random errors of the instrument.

The two Bathythermographs should be on the same cable and as close together
as possible. Also they should be reliable instruments, and matched, to the extent
that the random errors should be about equal in magnitude.

Other Techniques to Correct Bathythermograph Sets

In lieu of employing an independent temperature measurement to correct the
Bathythermograph to read absolute temperature values, other techniques present
good possibilities. These are discussed in categories below.

Correction of Bathythermograms through Heat Content Studies

Theoretical studies can be undertaken assuming that no significant change in
the heat content of the water will exist over this time of study. This condition is
especially true where the ship which took the Bathythermograms was in a relatively
fixed position during the time the Bathythermograms were taken. Any apparent
large change in heat content represented by the temperature measurements of the
Bathythermograph could therefore be considered to be produced by an error in the
Bathythermogram values. Furthermore, this error can be considered a franslation
of the slide, or in other words a variation in the set of the Bathythermograph.

By integrating each Bathythermogram with respect to temperature an arbi-
trary heat content value can be established. This is in accordance with the following
equations:

Z —
A ; Jas Ap Cp (Tz, -Tz, dz (1)

where Z, and Z; are the depths from which the integration takes place.

Where A, 7, Cp (area, average density, and specific heat) are constant, as would
be the case in comparing Bathythermograph traces representing nearly constant
conditions, the equation reduces to the following:

Q vA
= = — 2
ABC, K Je Gr Ty ,)dZ (2)

where the integration takes place from the surface to the bottom of the trace.
The Bathythermogram trace represents temperature as a function of depth,

ie: T, = £(2) (3)

29

9.3.2.

Combining (2) and (3): Keon nel), (4)
Thus solve for K for each Bathythermogram.

Without regard as to which Bathythermogram represents the true heat content
of the ocean, each Bathythermogram in the series can be compared to some arbitrary
reference heat content. Having established a quantity which must be added or sub-
tracted from each individual Bathythermogram in order to make the heat content
equal, this quantity can be re-translated into terms of a AT. This AT is the correction
factor which should be applied to the Bathythermogram to cause the Bathythermo-
gram to register the same heat content.

This technique cannot be applied to ocean conditions where considerable internal
waves are present. Internal waves will represent an actual change in heat content
of the ocean at that point from BT reading to BT reading. This technique would
erroneously provide for a horizontal displacement of the Bathythermogram to over-
come a vertical displacement in the water. In cases where internal waves are present,
other techniques such as curve fitting techniques to be described below will have to
be employed.

Curve Fitting Techniques

Thus far the processes for comparing Bathythermograph data have involved
the use of their absolute grid values and various applied correction factors. A much
more accurate picture of variation in the ocean can be obtained if the determinate
errors of the instrument and the man-made determinate errors are first eliminated.
This can be accomplished if the individual Bathythermograms are compared, not
by their absolute grid values, but rather by superimposing one curve on top of an-
other and noting the variations between the curves. This is a method for correcting
the Bathythermograms; i.e., superimposing the non-varying portions of the curves
upon each other.

This in actuality is correcting the temperatures at many depths rather than
trying to correct the Bathythermograph by one single temperature measurement at
the surface. The absolute value of the superimposed curve is still undetermined, but
nevertheless the ocean variability can be shown by comparing the variations of these
superimposed curves.

The assumption involved in curve fitting techniques are that the areas which are
being investigated are relatively stable to begin with, and it is quite unlikely that the
ocean will change in displacement throughout the entire exploration of the BT unit.
It is considered then that the average deviation which is obtained by this type of
comparison comes closer to approximating the true ocean variation than comparing
BT’s by their absolute values.

Two techniques are available for fitting these curves. The simplest and perhaps
the most reliable is visual comparison. This is accomplished by superimposing
transparent Bathythermograph slides of a given series until the curves are matched
in their greater proportion. A grid can then be placed over top of these to determine
the magnitude of the variations between the two curves. Scale differences must
be considered.

If is is desired, a rigid mathematical approach can be applied to the curve fitting
techniques. Several mathematical techniques have been suggested (refer to literature
survey at the end of this report). It will require further study to determine the
practicality and desirability of applying mathematical techniques in lieu of visual
superimposition.

30

9.3.3.

9.3.4.

9.4.

Correcting the Bathythermograms by the Permanent Thermocline

Where variations in the ocean are expected to be considerable, curve fitting
techniques cannot be employed. However, there is indication that the bottom-most
portion of the Bathythermograph will remain fairly stable even in conditions of high
variability in the upper level. This bottom portion of the curve will be stable if it
reaches the top of the permanent thermocline. This occurs only with 900 foot Bathy-
thermographs.

It has been shown that the top of the permanent thermocline is fairly stable
all year round, having variations not greater than present limitations of correcting
Bathythermograph data. Where the top of the thermocline is reached by the bottom-
most part of the trace, the trace can be superimposed. One is not depending upon an
independent temperature measurement at this point, but depending upon the fact
that this point does not vary in time, at least to any significant extent. This is a much
better point to correct the Bathythermograph than the surface, which does vary
considerably.

The additional information which is received and the ability to reach into a
fairly stable portion of the ocean warrants the use exclusively of the 900 foot Bathy-
thermographs. The present state of development of the 900 foot Bathythermograph
is such that large failures in the bellows exist at this time. However, it is felt that
emphasis should be placed upon extended use of the 900 foot Bathythermograph.

Use of Minimum Thermometer to Correct Bathythermographs

In cases where the bottom of the trace does not reach the top of the permanent
thermocline and the variability of this part of the ocean is high, an independent
temperature measurement would have to be made at the bottom of the trace. This
is still an advantageous position to take an individual temperature measurement as
opposed to the surface; because whatever the variability at this point, it is less than
the top.

This suggests in lieu of the bucket or injection thermometer the use of an instru-
ment which would measure the temperature at the maximum depth to which the
Bathythermograph went. Such a minimum registering thermometer will have to
have a good response characteristic and would have to be designed to indicate the
temperature at that point, when brought out at the surface. The depth value would
not have to be recorded since it would be fixed to the Bathythermograph.

While not much consideration has been given to the practicality of such an in-
strument, nevertherless in theory it would be highly advantageous over the pres-
ent techniques of correcting Bathythermographs at the surface.

Establishment of Average Bathythermograms

In many studies it is desirable to establish a typical average Bathythermograph
profile which is representative of the conditions over which the average was taken.
The recording of temperatures at given specific depths from which an average tem-
perature at each depth is then taken and from which an average Bathythermograph
is drawn often leads to erroneous results. This is especially true where internal
waves exist or where large variations in Bathythermograph sets exist. Such an av-
erage will round off any sharp corners in the Bathythermograph profile and distort
the actual gradients which exist. This technique must be foregone in lieu of the more
accurate curve fitting techniques as described above.

Proper curve fitting techniques can be used not only to determine the variations
of the ocean, but also to obtain a typical average Bathythermograph. This type of

31

9.5.

averaging technique will maintain the basic profile structure without distortion,
and will provide a more accurate picture than the normal averaging of isothermal
charts or isodepth charts. This technique can either be a visual one or a mathematical
one.

Establishing Average Trends

There are many instances of study where long range trends are desired. These
long range trends can be obscured by the individual variations such as standing
waves, spurious data, etc., which are of short-term duration. In its place an average
trend is desired which minimizes or eliminates these short-term variations.

The technique employed to establish an average trend is to establish a series of
average Bathythermograms from which the average trends can be taken. The ques-
tion arises as to the length of time over which an average should be made. Since for
any average Bathythermogram which is desired the average or standard deviations
can be obtained, the duration over which the Bathythermographs are averaged should
be such that the deviations of the average Bathythermogram should not exceed the
minimum differential which is desired in establishing the trend. If at any time the
deviations of the average Bathythermogram are greater than the desired differen-
tial in trends, these trends have no mathematical significance, and should either be
ignored or reworked with a smaller average Bathythermogram.

It should be noted that the average Bathythermogram which is made will pro-
duce erroneous or distorted results unless obtained as described in Section 9.3.

32

SECTION 10.0

RECOMMENDATIONS

Based on this study of the errors of the Bathythermograph, the following recommendations
are made. These recommendations are broken down into the various categories below.

10.1.

10.2.

10.3.

10.4.

Bathythermograph Instrument

1. Eliminate the bending of the pen arm caused by excessive heating by the sun
and the bellows table override.

Displace the uptrace from the downtrace of the BT for further clarity.

3. Place greater emphasis on the 900 foot Bathythermograph and correct its high VA
incidence of bellows failure. Wherever possible, use the 900 foot Bathythermo-
graph exclusively, in preference to the shallower ones.

4. Check the Bathythermograph frequently for loose parts during its use.

5. Re-evaluate the Bathythermograph’s specifications in the light of the extensive
experience which has been obtained in its use since the specifications were
written.

Calibration

6. Use a pre-repair test on all Bathythermograms before recalibration.
7. Calibrate Bathythermographs quite frequently, at least every three months.
8. Keep a Bathythermograph log of repair history.
9. Obtain uniformity between testing stations.
10. Increase capacity of testing stations.

11. Test all Bathythermographs for the following things in addition to what is now
being done:

a. Hysteresis

b. Response Rate
ce. Reproducibility
d. Smoothness,of Response

Operational Use of Bathythermograph

12. Use an adjustable speed winch.

18. Instruct personnel as to the importance of the data and the care required in
taking it. This should be developed into an extensive training program.

14. Maintain complete notes on the conditions which prevail at the time each Bath-
ythermograph cast is made, including every irregularity observed.

Correcting Bathythermograph Traces
15. Use better injection thermomeiers.
16. Use better bucket thermometers.

17. Use reliable, well insulated buckets.

18. Study the possibility of using independent check point located at a consider-
able depth of the ocean rather than at the surface.

33

10.5.

10.5.1.

10.5.2.

10.5.3.

Further Studies of the Errors of the Bathythermograph

Bathythermograph Sets

1G),

20.

At present, average depth correction factors are used for an entire series of
Bathythermograms. It is felt that a study of anchor stations representing a
cross-sectional range from accurate data to poor data should be made in order
to compare the advantages of applying depth correction factors to each card in-
dividually versus the present practice of applying an overall average correction.
It is hoped that a formula can be evolved which will show when each practice
would be most advantageous.

The bucket temperatures used to correct Bathythermograms are often subject
to greater error than the Bathythermograms themselves. It is desired to study
these variations between bucket and surface values of Bathythermograms
through statistical analyses. It is hoped in this manner that factors can be
found which will indicate whether to use correction factors or not. Other
methods of investigating temperature correction factors would be curve fitting
through superimposition and least square fit.

Variation in Bathythermograph Sets

2

22s

23.

24.

25.

A study of change or variation in set with each lowering should be further in-
vestigated, being easily accomplished through analysis of data from the anchor
stations employed to accomplish Recommendations 19 and 20. An evaluation of
the extent and amount of this variation should be forthcoming from such in-
vestigations.

Variation from fatigue and extended use of the Bathythermograph is best
evaluated by the direct measure of the change which occurs in its calibration.
This can be accomplished in the calibration procedures by measuring the cali-
bration before the Bathythermograph is repaired. This can be compared to the
last calibration.

Abnormal handling is indicated by a quite sudden change in the set. Sudden
changes in set affect the averaging techniques employed in set corrections and
should be studies with respect to their lack of recognition in computing average
corrections. This should be studied in the anchor stations employed for Recom-
mendations 19, 20, 21 and 22.

There are studies which show the depth set varies with depth. Further inves-
tigation is in order to determine the change as a function of depth. This is best
accomplished by further experimental procedures involving comparisons to
independent measuring instruments. Comparison of calibration slides periodi-
cally can also show the presence of a depth error.

Since pressure effects the bourdon tube, there may be a variation in the tem-
perature set which is a function of depth. The study is best accomplished as
outlined in Recommendation 24.

Calibration Errors

26.

27.

Calibration sets can occur unless duplicate tests are run, especially when new
bellows are installed. Calibration sets are best evaluated experimentally
through the calibration procedure.

A study is in order to determine the extent of variation present due to the as-
sumption of linearity over the temperature scale and depth scale. At: present,

34

equidistant spacing is employed for both temperature and depth.
Comparison between the actual grid from a calibration tank and the approxi-
mated grid employed to make up the standard calibration would show these
errors best.

10.5.4. Reproducibility

28.

Reproducibility is best measured by the closeness with which the uptrace
matches the downtrace, ocean variability being accounted for. A study of a
series of Bathythermograms will reveal the extent of the following possible
characteristics contributing to reproducibility.

(a) Mechanical Hysteresis

Mechanical hysteresis is represented by parallel displaced lines in a
Bathythermogram study. Other effects of this mechanical lag might be
observed.

‘

(b) Response to Initial Depths

Response of the Bathythermograph to the first ten feet of water should
be studied further experimentally. This is particularly important to un-
derstand because of its bearing upon the studies of keying the surface
value of the Bathythermograph to a bucket temperature. It ties in with
BT-bucket correlation analyses.

(c) Rate of Response to Ocean Variations

Response rates should be experimentally ascertained. This is important
in relation to variability studies, internal wave studies, and will dictate
operational procedures in proper rates of lowering, etc.

(d) Response Hysteresis

It is felt that a study of Bathythermograms can incorporate a study of
all hysteresis effects. Response hysteresis should be able to be distin-
guished by the characteristic shape of the curves, and possibly some
quantitative information can be produced from a study of the hysteresis
curve shapes.

(e) Mechanical Hesitation

The wiggles on Bathythermograms can be due to mechanical hesitation
through sticky bellows, pen play, etc. This must be experimentally evalu-
ated through control tests.

10.5.5. Operational Errors

29.

30.

ol.

Studies have been made in order to evaluate personnel error. It is considered
feasible to incorporate into the before-mentioned anchor station studies a per-
sonnel factor, this being accomplished by comparing anchor stations of various
grades of accuracy, i.e., good, average, poor. Other studies along this line have
been fairly quantitative in their results.

With regard to speed of raising and lowering the Bathythermograph, direct
experimentation is necessary, along with the hysteresis analysis mentioned in
Recommendation 28.

Bathythermogram reading errors can be established also on an experimental
basis.

5)

10.5.6. Other Methods to Correct Bathythermograph’s Temperature

32.

33.

Other methods of keying Bathythermographs together are open to study and
should be investigated. For such purposes a range of anchor stations should
be studied: superior, typical, and even injection temperature stations.

The technique of lowering two Bathythermographs simultaneously and apply-
ing correlation studies to minimize instrument error should be investigated.

36

SECTION 11.0

GLOSSARY OF TERMS

The following is a partial list of terms employed in the oceanographic and mathematical
fields. These terms are submitted in the hope that they will not only serve to clarify the termi-
nology used in the report, but will also act as a basis for a more extensive and amended
glossary in the field of Oceanography.

Absolute Value:
Accidental Error:

Accuracy:

Advection:

Anchor Station:

Average Bathythermogram:

Average Deviation:
Average Difference:

Average Value:

Bathythermogram:

Bathythermogram, Average:

Bathythermograph:

A
See Value, Absolute.
See Error, Accidental.

The degree of concordance of a number representing the value
of a quantity and the number representing the true value of
the quantity ; it may be expressed in either absolute or relative
terms. This is a qualitative definition, since the true value of
many quantities can never be known.

Movement of water mass into or out of a fixed locality because
of ocean currents, upwelling, ete.

Position occupied by a hydrographic vessel which is in a fixed
locality. The more correct application of the term is made to
vessels specifically at anchor, although it is sometimes applied
to ships adrift.

See Bathythermogram, Average.
See Deviation, Average.
See Difference, Average.

See Value, Average.

B

Photographic reproduction of a Bathythermograph slide
which has the Bathythermograph grid superimposed for read-
ing the temperature and depth values, the necessary tempera-
ture and depth correction having been applied in superim-
posing the grid.

A Bathythermogram calculated by employing average tem-
perature values for a fixed locality over a period of time.
Average Bathythermograms may be obtained by a number
of averaging techniques, such as curve fitting, mathematical
regression lines, least square, etc.

Instrument which measures the change of temperature witn
the corresponding change in depth of the ocean. It employs
a pressure bellows element to register depth changes and
liquid filled temperature element to register change in tem-
perature. The data are given in the form of a curve of tem-
perature versus depth, or a smoked slide.

37

Bathythermograph Card:
Bathythermograph Curve:

Bathythermograph Data:

Bathythermograph Grid:

Bathythermograph Profile:
Bathythermograph Slide:

Bathythermograph Unit:
Bathythermography:
Blotches:

BT:
BT gm:

Bucket Temperature:

Characteristic Shape:

Component of Error:

Consistency:

Correction Factor:

DCS:

A Bathythermogram.

Trace made on a Bathythermograph smoked slide by the pen
arm. Represents change of temperature with change of depth.

Data in the form of temperature values, depth values, slope
of temperature-depth trace, and points of inflection which
are obtained from Bathythermograms.

Graph lines, the abscissas of which represent line of con-
stant depth, and the ordinates of which represent lines of
constant temperature, photographed on a photo-sensitive glass
slide and mounted in a viewer. Purpose of the grid is to
translate directly the Bathythermograph curve into quan-
titative values of temperature and depth.

See Profile, Bathythermograph.

Smoked glass slide on which the Bathythermograph trace
is drawn by the pen arm.

A Bathythermograph; a Bathythermograph instrument.
The art of taking, analyzing, and utilizing Bathythermographs.

Marks left on BT slide by sporadic movement of the pen arm,
usually caused by a sudden shock to the BT itself. Marks
resemble jagged lines or ink spots. ;

Abbreviation for Bathythermograph.
Abbreviation for Bathythermogram.

See Temperature, Bucket.

Cc

The term refers to the BT trace, and specifies the general
overall appearance of the curve. Since the BT trace often
represents a typical ocean condition, this ocean condition is
recognized by the “characteristic shape’ of the Bathy-
thermogram.

In formulating an average BTgm, it is most important to
maintain the characteristic shape of the individual BTgms.

That individual error which stems from a distinct source,
this error being one in a group of errors, each emanating
from a distinct source, which together produce the overall
error.

Regularity with which the data seem to fit the particular
conditions.

The numerical value which is added to or subtracted from
the location of the BT slide with respect to the BT grid to
correct for the determinate errors or sets of the Bathy-
thermograph.

Average difference between the 0 depth trace of the slide
and the surface line (0 feet) of the reading grid. If the
depth trace is above the zero foot line, the DCS is plus.

38

DSP:

TCS:

TSP:

Correlation:

Correlation Coefficient:

Cross Plotting:

Curve Fitting:

Cyclical Variations:

DCS:

Dependent Error:
Depth Displacement:
Depth Set:
Determinate Error:

Deviation:

If the depth trace is below the zero foot line, the DCS is
minus.

When the BTgm is photographed, this correction is applied,
thus giving a correct BTgm.

Depth correction of the BTgm or photographic print. In
case the error was not fully corrected in the making of the
BTgm, there is a correction factor to be applied to the final
print. The DSP is recorded in such a manner that the set
correction to be applied is obtained by adding the DSP to
the DCS. When no correction is necessary, the DSP will be
equal in magnitude but opposite in sign to the DCS.

Average difference between the surface temperature as read
from bucket or intake thermometer and the surface tempera-
ture as read from the slides. If the surface temperature
is greater than the temperature of the slide, the TCS is
plus. If it is less, the TCS is minus.

Temperature correction of the BTgm applied in the same
way as described under DSP.

The degree of relationship which may exist between two
variables.

The correlation coefficient of the two sets of variates ex-
pressed in their respective standard deviations as units, is
the arithmetic mean of the products of deviation of corre-
sponding values from their respective means.

Process of replotting a set of data by interchanging the var-
iables with the parameter. For instance, temperature versus
time with depth as the parameter can be replotted as depth
versus time with temperature as the parameter.

Process of relating two or more BTgms together by com-
paring how close the curves will match in characteristic
shape. This process may be visual and qualitative, or may
be precise and mathematical. The mathematical solutions
also include the process of fitting a smooth curve through
a set of discreet points. In the case of BTgms, the discreet
points are replaced by a whole curve.

See Variations, Cyclical.

D
See Correction Factor.
See Error, Dependent.
See Displacement, Depth.
See Set.
See Error, Determinate.

Arithmetic difference between the average value of any one
single value which contributed to the average value.

39

Deviation, Average:

Deviation, Standard:

Discrepancy:

Displacement:

Displacement, Temperature:

Displacement, Depth:

Distortion:

Distribution, Gaussian or

Normal:

Down Trace:
Drifting Station:
D.S.P:

Error:

The Arithmetic mean of the deviations of a series of meas-
urements of equal weight taken without regard to sign. The
sum of the absolute values of deviations divided by the total
number of deviations.

The square root of the following quantity: the sum of the
squares of the deviations of a series of measurements di-
vided by the total number of deviations.

The mathematical difference between two measurements of
the same conditions, these measurements being taken si-
multaneously by two different measuring instruments.

Translation of the origin of a set of coordinates; the dif-
ference, both in magnitude and direction, between the origins
of two or more coordinate systems. Specifically, the displace-
ment of a BTgm is the amount that the BT curve lies to
the right or left of, up or down from a reference curve, or
reference point. This displacement can be caused by sets in
the instrument, can represent correction factors, comparisons
between curves, etc.

Translation of the origin of a coordinate system of tempera-
ture versus depth, such displacement occurring in the di-
rection of temperature.

Translation of the origin of a coordinate system, temperature
versus depth, such displacement occurring in the direction of
depth.

Change in the characteristic shape of the BT trace in any
one portion not attributable to true ocean condition, but such
change being caused by erroneous measurements, faulty in-
strument, or poor averaging technique.

A measure of the occurrence of accidental errors in a series

of measurements of the same degree of precision. The mathe-
matical treatment is based on three axioms:

(1) Small errors are more frequent than large ones, and
the probability of occurrence of an error is a function of
its size.

(2) Positive and negative errors of the same size are about
equal in number.

(3) Accidental errors of large magnitude do not occur.

See Trace, Down.
See Anchor Station.
See Correction Factors.

E

A class of deviations which describe the absolute or prob-
able limits of reliability of a single measurement or a group
of measurements. The term is too general for quantitative
evaluation, since there are many kinds of errors. Generally
the term error is used when in actuality the term indetermi-
nate error should be used. Hence the more specific meaning
is understood in the use of the more general term.

40

Error, Absolute:

Error, Accidental:

Error, Dependent:

Error, Determinate:

Error, Independent:

Error, Indeterminate:

Error, Instrument:

Error, Mean:

Errer, Mean Square:

Error Probable:

Error, Reading:

Error, True:

Gaussian Distribution:

Geological:

Gradient:

See Error, True.

Those deviations, inevitable in all measurements, which re-
sult from small unavoidable errors of observation due to
more or less fortuitous variation in the sensitivity of our
measuring instruments and the keenness of our senses of
perception. Such errors are due to the combined effect of
a large number of undetermined causes.

An error produced in a dependent variable by the error of
the independent variable. Where Y is a function of X, an
error in X will produce an error in Y. Referring to BTgms,
where temperatures are a function of depth, an error in
depth will produce a corresponding error in temperature.

An error that is discovered and allowed for in magnitude
and sign in the form of a correction allowing for its effect.

That portion of the error of a dependent variable which in
itself is not a function of the error of the independent var-
iable; these are accidental errors inherent in the measure-
ment of that variable.

That error which cannot or is not properly allowed for in
magnitude and sign. All errors that are not determinate
errors.

Those errors which are specifically related to the measuring
instrument. These comprise both determinate errors (some-
times called sets) and indeterminate errors (called instrument
precision or reproducibility).

See Deviation, Average.
See Deviation, Standard.

An error of such magnitude that from the standpoint of
probability the true error of the observation is just as likely
to be greater as it is to be less than this magnitude.

More generally, it is a measure of the closeness with which
the probable value approaches the true value.

That error introduced into the measurements by the ac-
curacy with which the observer can read the measured value
of the instruments used to take the measurement.

The difference between the observed or calculated value and
the true value.

G
See Distribution, Gaussian or Normal.

Pertaining to the science which treats the history of the
earth and its life, or the materials of the earth itself.

In Oceanography, geological pertains to the study of the
ocean floor formations.

Rate of change of one dependent variable with respect to
the independent variable; slope; first derivative. As applied
to BTgm, it is the slope of the curve.

41

Gradient, Negative:

Gradient, Positive:

Gradient, Temperature:

Grid:

Gross Changes:

Hysteresis:

Independent Error:

Injection Temperature:

Instrument Deviation:
Instrument Error:

Instrument Precision:

Instrument Reproducibility:

Instrument Set:

Interpolation:

Isodepth Chart:
Isodepth Line:

Isodepth Profile:

Isotherm:

Isothermal Profile:

Keying Bathythermograms:

Dependent variable decreases with increasing independent
variable. As applied to the BTgm, it represents a decrease
in temperature with increase in depth.

Dependent Variable increases with increasing independent
variable. As applied to the BTgm, it represents an increase
in temperature with increase in depth.

Rate of change of temperature with change of depth of the
ocean.

See Bathythermograph Grid.

Changes in conditions of the ocean which are of an order of
magnitude greater than the thermal microstructural changes
which may occur.

H

As applied to Bathythermography only, hysteresis represents
the lack of the uptrace to follow the downtrace, resulting in
a typical hysteresis loop. Such loop resembles a new moon
or parallel curves that converge at both ends.

I
See Error, Independent.
See Temperature, Injection.
See Deviation, Instrument.
See Error, Instrument.
See Precision, Instrument.
See Reproducibility, Instrument.
See Set.

Method of predicting the location of a point or points on a
graph by assuming the data on either side of the interpolated
region closely follow a smooth curve.

See Profile, Isodepth.

Line connecting points of equal depth. Usually such a line
represents temperature versus time or distance at a con-
stant depth.

See Profile, Isodepth.

Line connecting points of equal temperature. Usually such
a line represents depth versus time or distance at constant
temperature.

See Profile, Isothermal.

K

Process of relating the reference temperatures of a series
of BTgms so as to compare the changes from BTgm to BTgm
without being subject to the errors involved in establishing
and maintaining these reference temperatures.

42

Lateral Variation:

Mass Movement:

Mean Square Error:

Microstructure, Thermal:

Normal Distribution:

Occupation:

Ocean Stability:
Ocean Variation:

Oceanography:

Parameters:

Plot, Simultaneous:

Precision, Instrument:

Pressure Sensing Element:

Probable Error:
Probable Value:
Profile:

Profile, Bathythermograph:

L

See Variation, Lateral.

M

Displacement of water by advection, ocean currents, or in-
ternal waves.

See Error, Mean Square.

Minute changes in the temperature of the ocean, in both lat-
eral and vertical directions. These minute changes do not
follow a predictable pattern, but are discontinuous and fluc-
tuating.

N

See Distribution, Normal.

O

(Refers to usage such as “occupation of Anchor Station.’’)
Maintenance of a fixed position in the ocean by a hydrographic
vessel.

See Stability, Ocean.
See Variation, Ocean.

Oceanography embraces all studies pertaining to the sea and
integrates the knowledge gained in the marine sciences that
deal with such subjects as the ocean boundaries and bottom
topography, the physics and marine biology.

P

A third variable in terms of which the two coordinates x
and y can be expressed, rather than in terms of each other.

Graph of two or more BTgms (single or average) on the same
coordinate system. These BTgms are not superimposed as
would be done by curve fitting techniques, but maintain their
specific reference values as individually recorded.

Measure of the ability of an instrument to reproduce a read-
ing or set of readings. Equivalent to Reproducibility. This
bears no relation to accuracy, since the instrument may pre-
cisely reproduce an erroneous value.

See Sensing Element, Pressure.
See Error, Probable.
See Value, Probable.

Graph of temperature, depth, and time, using any two as
variables with the remainder as a parameter. Also called
charts, and plots.

Plot of temperature versus depth, with a time parameter.
Normally called a bathythermograph slide, if uncorrected,
or a BTgm if corrected. Also called temperature profile.

43

Profile, Isodepth:

Profile, Isothermal:

Profile, Temperature:

Readings:

Reading, Average:

Readings, Zero Depth:

Reading Error:

Regression Line:

Reproducibility:

Residual:

Salinity:

Sensing Element:

Sensing Element, Pressure:

Sensing Element,
Temperature:

Set:

Set Variation:

Simultaneous Plot:

Plot of temperature versus time with depth as the parameter.
Also called isodepth charts, isodepth profiles; sometimes
called isobaric profiles.

Plot of depth versus time with temperature as the parameter.
Also called isothermal charts and plots.

Usually refers to BTgms. Incorrectly refers to isothermal
profile.

R
Discreet values or sets of values as read from the BT gms.
Arithmetic mean of several values as read from BTgms.

Temperature values as read from the BTgms corresponding
to the surface of the ocean or abscissa coordinate labeled zero
feet in depth.

See Error, Reading.

A straight line plotted among a scatter diagram of points
(plotted to determine the extent of correlation between two
variables) in such a way as to make the sum of the squares
of the vertical distances from the points to the line a minimum.

A measure of the closeness with which a measuring instrument
will repeat its response to the same set of conditions. A
measure of the closeness with which a series of repeated
values resemble each other. See also Precision.

The difference between the most probable value and any in-
dividual measurement of the series.

)

The total amount of solid matter in grams contained in one
kilogram of sea water when all of the carbonate has been
converted to oxide, the bromine and iodine replaced by chlo-
rine, and-all organic matter completely oxidized.

Physical equipment which responds by change in some one
of its characteristics to temperature, depth, or salinity.

Evacuated spring loaded compression bellows. Change of
depth compresses or expands the bellows, changing the po-
sition of the BT slide.

Copper tubing filled with liquid and connected to a Bourdon
tube. Change of temperature causes the liquid in the copper
tube to expand or contract, transmitting the developed pres-
sure to the Bourdon tube.

Mathematical Value, both with magnitude and sign, which
describes the amount by which the measured temperatures
or depths depart from the true, or most probable, or cali-
brated value.

See Variation in Set.

See Plot, Simultaneous.

44

Slope:
Stability, Ocean:

Standard Deviation:

Superimposition:

TCS:

Temperature, Bucket:

Temperature Displacement:
Temperature Gradient:

Temperature, Injection:

Temperature Profile:

Temperature Sensing
Element:

Temperature Set:

Thermal Microstructure:

Thermocline:

Trace:

Trace, Down:
Trace, Up:
Trend Chart:

True Error:

True Value:

TSP:

See Gradient.

Mathematical term describing the extent to which water is
in buoyant equilibrium at any point in the ocean. Refers
to Dynamic stability.

See Deviation, Standard.

Act of placing the trace from one BT over the trace of another
BT so as to coincide in all points possible by translating,
but not rotating, the traces the necessary amount.

mT
See Correction Factor.

Value of surface temperature of the ocean as read from a
regular measuring thermometer which is placed in a bucket
sample of surface water scooped up at the desired point of
measurement by a specially designed bucket. ~4

See Displacement, Temperature
See Gradient, Temperature.

Surface temperature of the ocean as read from the injection
thermometer which is located aboard ship at the point of
cold water intake. The location of the intake is not actually
at the surface, but may be a considerable distance below the
water level.

See Profile, Temperature.

See Sensing Element, Temperature.

See Set.

See Microstructure, Thermal.

In its strictest sense, a transition layer of water at mid
depths between the top surface layer which is isothermal
and mixed, and a very deep mass of much colder water in
which the temperature decreases very slightly and uniformly
with depth.

The term, more generally applied, refers to any layer in
which the temperature decreases markedly with depth.
Impression on the smoked BT slide etched by the tempera-
ture pen arm which represents a graph of temperature versus
depth.

Trace made during the descent of the BT.

Trace made during the ascent of the BT.

Plot of temperature or depth (single or average value) versus
time, primarily used to indicate the trends or changes in these
conditions with time.

See Error, True.

See Value, True.

See Correction, Factor.

Uncorrected Value:

Uptrace:

Value, Absolute:

Value, Actual:

Value, Average:

Value, Probable:

Value, Reference:

Value, True:

Value, Uncorrected:

Value, Zero Depth:

Variance:

Variations:

Variation, Cyclical:
Variation, Lateral:

Variation in Set:

Variation, Ocean:

Variability, Ocean:

Zero Depth Value:
Zero Depth Reading:

U
See Value, Uncorrected.

See Trace, Up.

Vv

Math Def.: Value without regard as to sign.
General: True value; correct value to the limits of high ac-
curacy when the true value is impossible to determine.

Representative of the thermal state in reference coordinates
corresponding to the standard temperature scale as opposed
to arbitrary reference coordinates.

Arithmetic means of a series of measurements, all of which
have equal weight or are properly weighed and handled
accordingly.

Average value which must closely resemble the true value.
It can be stated that the Probable Value is the value which
does not exceed the true value by the Probable Error.

Magnitude as related to a particular reference origin. When
standard origins are used, the reference and actual value
are identical.

Value without error. Value by definition or standard.

Values of BTgms as read from a glass slide with grid placed
in last calibrated position without application of set cor-
rection factors.

See Reading, Zero Depth.

Measure of the change with time or direction of an average
or standard deviation.

Small random changes in measurements due either to true
changes in the ocean parameters or in the measuring instru-
ments themselves.

Harmonic variation of temperature or displacement due to
internal waves.

Change in conditions in a horizontal traverse.

Change in magnitude of set which cannot be accounted or
allowed for. Indeterminate portion of the determinate error
of set.

See Variability, Ocean.

Indeterminate variations or changes in the ocean with time
or direction which cannot be accounted for in magnitude
and sign.

Z
See Reading, Zero Depth.
See Reading, Zero Depth.

46

SECTION 12.0

BIBLIOGRAPHY

The bibliography contained below covers only those reports relating to the following

subjects:

(1) Construction of BT

(2) Operation of BT

(3) Maintenance of BT

(4) Improvements in design of BT

(5) Accuracy of BT

(6) Comparison of BT with other instruments

(7) Methods

of treating data from BT

The bibliography contains references to some classified reports, but the titles themselves

are not classified.

Adams, Richard M.

Anderson, E. R., Anderson,

L. J., and Marciano, J. J.

Anderson, Ernest R., and
Burke, Arthur T.

Bralove, A. L., and
Williams, E. I.

Bristol Company,
Waterbury, Conn.

Burt, Wayne V.

Kalbfell, D. C.

Dept. of National Defense,
Halifax, N.S.

Dietz, Robert S.

Fleming, R. H.

“Atmospheric Influence on the Thermal Structure of the
Oceans,” Progress Rpt., Texas A & M R. F., October 18,
1951. Errors in the taking of bucket temperature and other
errors.

“Review of Evaporation Theory and Development of Instru-
mentation,” U.S.N.E.L., Lake Meade Water Loss Investigation,
Rpt. No. 159, Feb. 1, 1950. Describes equipment necessary
to obtain accurate temperature measurements.

“Notes on the Development of a Thermister Temperature
Profile Recorder (TPR)” Journal of Marine Research, Vol. X,
No. 2 (1951). For use with temperature profiles in upper 30
meters of inclosed bodies of water; compares TPR with BT.

“Tnterim Development Reports I through IX for the Analysis
of BT Data,” May through December, 1951. Study of data
taken from several anchor stations of the San Pablo and
Rehoboth.

“Instructions for Care & Use of Surface-vessel BT Types
CTB40080A & CTB40120A & Winch Assembly CTB 10318.”
Nav. Ships 900, 231-B.

“Letter to the Hydrographer, U.S. N.H.O.,” Washington 25,
D.C., March 1949. An examination of the differences between
injection temperatures and temperatures read from BT traces.

“Oceanographic Temperature Measurement Equipment,”
U.C.D.W.R. Rpt. (March 28, 1942).

“Design and Construction of BT Calibrator.’ Report PHX69,
December 23, 1949. Describes calibration equipment.

“Some Oceanographic Measurements on Operation HIGH-
JUMP,” N.E.L. Rpt. No. 55 (7 July 1948).

“Microstructure Instrumentation,” U.C.D.W.R. Rpt., Janu-
ary 12, 1943.

47

Fuglister, Frederick C.

Iselin, C. O’D.

Jacobson, A. W.

Sverdrup, H. U., Johnson,
Martin W. and Fleming,

Richard H.
La Fond, Eugene C.

La Fond, Eugene C.

Leipper, Dale F.

Leipper, Dale F.

Leipper, Dale F. and
Burt, Wayne V.

McNown, John S.

Miller, R. C., and
Spilhaus, A. F.

Mosby, Hakon

Murray, H. G., and
Shipway, G. D.

N.D.R.C.

N.D.R.C.

“Report on Weather Station in North Atlantic.” W.H.O.I.
Rpt. (unpublished). No date. Comparison of BT surface
Temperature reading with injection and bucket temperatures
is made along with an analysis of the heat content.

“The Application of Oceanography to Sub-Surface Warfare.”
Summary Technical Rpt. of Div. 6, N.D.R.C., Vol. 6A, 1946.
Describes use of surface vessel BT and Submarine BT.

“An Instrument for Recording Continuously the Salinity,
Temperature and Depth of Sea Water.’”’ The Int. Hydrographic
Rev., X XVII, No. 1, pp. 107-122 (1948).

“The Oceans,” Prentice-Hall (1942). Basic text book in
Oceanography.

“Factors Affecting Temperature Gradients in the Sea.” S.I.0O.
Publication, September 25, 1945. Describes relations of var-
tables in the ocean.

“Processing of Oceanographic Data,’’ H. O. Publication No.
614, 1951. Describes Techniques for handling oceanographic
data.

“Determining Local Sea Temperature Changes from BT
Data.” §S.I.0. Unpublished Rpt., August 8, 1949. Describes
mathematical approach to obtaining correct average bathy-
thermograms.

“Averaging Methods and the Distribution of Variables” (un-
published). Describes mathematical approaches for obtaining
correct average Bathythermograms.

“Annual Report, Bathythermograph Processing Unit,” 8.1.0.
Rpt., July 1947 and July 1948. Describes techniques and ac-
curacies of processing BT data.

“Adjustment of BT for Errors in Surface Temperature.”
U.C.D.W.R. No. M103, September 28, 1943. Describes change
in BT to include normal correction for BT.

“The Sea Sampler.” W.H.O.I. contribution No. 440 (1948).
Describes use of BT for sea sampling.

“The Thermo Sound, an Oceanographic Temperature Re-
corder.” Naturvitenskakelig rekke Nr. 1, April 1943. De-
scribes another type of instrument for obtaining temperature
versus depth.

“Service Test Report: Preproduction Model of Surface Ves-
sel BT,” May 38, 1951. Recommendations for improvements
of BT for Military use.

Physics of the Sound in the Sea, Parts I thru IV, Summary
Technical Rpts., Div. 6, Vol. 8. Contains relationship be-
tween temperature depth structure and various meteorological
conditions.

Principles of Underwater Sound, N.D.R.C. Summary Tech-
nical Rpts., Div. 6, Vol. 7. Contains discussions of Bathy-
thermograms.

48

Pritchard, Donald W.

Pritchard, D. W.

Raytheon Mfg. Co.

Revelle, Roger

Scripps Institution of
Oceanography

Scripps Institution of
Oceanography

Scripps Institution of
Oceanography

Seaboard Electric Co.

Seary, P. A. and
Gordon, W. E.

Spilhaus, A. F.

Shipway, G. D. and
Murray, H. G.

Stommel, H.

Stommel, H.

Ufford, C. W.

Urick, R. J., and
Searfoss, C. W.

U. S. N. BuShips

“Determination of Vertical & Horizontal Translation.” S.I.O.
Rpt. (unpublished). Theoretical technique for comparing
BT’s after removing horizontal and vertical translations of
BT traces.

“Cruise VIII, January 10 to 23, 1951,” Data Rpt. No. 6.
References are made as to the accuracy of the temperature
data.

“Instruction Book for Hoist E-2/S BT,’ Nav Ships 91175,
March 2, 1949. Instruction Book on Mechanical Equipment.

“Development & Testing of Oceanographic Instruments,”
8.1.0. Progress Rpt., Jan. 1 - June 30, 1949.

Quarterly Progress Rpt. No. 13, July - September 1949,
excerpt: “Accuracy of BT Temperatures & Depth under
Field Conditions.” Gives estimated accuracies for field
conditions.

“Procedure for Handling BT Data,” S.1I.0. Rpt., July 1,
1946. Describes S.I.O. technique for processing data.

“Analysis of BT’s of Operation HIGH JUMP.” Oceanographic
Rpt. No. 10, October 1, 1947.

“Bathythermograph System Buoyancy Recorder Set AN/BSH-
1, BT AN/BSH-2.” Progress Rpts., 1949 and 1950. De-
scribes experimental temperature measuring equipment.

“Temperature Recording with Thermisters.” U. of Texas
Elec. Eng. Research Lab., Rpt. No. 16, June 1, 1948.

“Bathythermograph” by Spilhaus. Journal of Marine Re-
search, Vol. 1, No. 2, p. 95 (1988).

“Evaluation of Preproduction Model Surface Vessel BT
OC-1/s, 2/s & 3/s.” U.S.N.E.L. Rpt. 252, June 1, 1951.

Test and evaluation of BT with regard to meeting specifi-
cations and recommendations.

“Memo on Mechanical Sorting of BT Data,’ W.H.O.I. Rpt.
(unpublished). Stommel suggests placing BT data on stand-
ard graph chart and punching cards for IBM. Basic pro-
posal only.

“An attachment for Standard BT for Making Multiple Traces.”
W.H.O.I. Rpt., September 21, 1945.

“An Electronic BT.” U.C.D.W.R. No. M439, July 10, 1946. De-
scribes a sensitive instrument using a thermocouple to meas-
ure temperature and a rheostat operated by a sylphon to
measure depth. Compares trace with BT trace.

“The Microstructure of the Ocean Near Key West, Fla.”
Parts 1 & 2. N.R.L. Rpt. No. S-3392. December 7, 1948.
A description of the thermopile, its location on the sub-
marine and a qualitative presentation of the records obtained.
Describes Averaging Techniques.

“Tnstructions for Installation, Care & Use of Sub. BT, Type
CTB40131,” Nav. Ships 943-E2.

49

U. S. N. BuShips Tech.

U.S. N. E. L.

U.S. N. E. L.

(Ufa Sp Nie 186 Ibs

W. H. O. I.

W. H. O. I.

W. H. O. I.
W. H. O. I.

W. H. O. I.

“Military Specifications BT, Surface Vessel,” MIL B15237
(Ships), March 15, 1950.

“Oceanographic Cruise to Bering & Chukcki Seas Summer
1949.” Rpt. No. 148, October 19, 1949. Data analysis is shown
here.

Quarterly Progress Rpt. “New Developments in Oceanog-
raphic Equipment at NEL” (1948). Describes latest de-
velopments in temperature measuring equipment.

“Oceanographic Measurments from U.S.S. Nereus on a Cruise
to the Bering & Chukcki Seas.” N.E.L. Rpt. No. 91, Prob-
lem 2A5, February 25, 1949.

“Instruction Manual for BT Observers.” W.H.O.I. Rpt., May
1942.

“Memo on Microstructure Observation and Field Methods.”
W.H.O.I. memo (not published). Suggests modifications for
BT.

“Lecture Notes on the Use of the Submarine BT.” July 1945.

“Tests on Submarine Signal Co. BT Type 781C” For Sig.
Corps Lab., Fort Monmouth, N. J. October 30, 1941. De-
scribes tests of Bathythermograph.

“BT Lifting Attachment and Expanded Scale BT,” Code 327
Files, U.S.N. BuShips, June 4, 1945.

50

DISTRIBUTION LIST
Navy Contract No. NObsr 52348

. Office of Naval Research 11. Marine Physical Laboratory

10.

Code 416
Washington 25, D. C.

. Office of Naval Research

Code 466
Washington 25, D. C.

. Director, U. S. Navy Electronics

Laboratory
New London, Connecticut
Attn: Mr. H. E. Nash

. U. S. Navy Hydrographic Office

Division of Oceanography
Washington 25, D. C.

. U. S. Naval Research Laboratory

Washignton 25, D. C.
Attn: R. J. Urick

. Research Analysis Group

Providence 6, Rhode Island

Via: Office of Naval Research
Boston Branch Office
150 Causeway Street
Boston 14, Massachusetts

. University of Washington

Department of Oceanography

Seattle, Washington

Via: Office of Naval Research
Washington 25, D. C.

. Chesapeake Bay Institute

RFD #2

Annapolis, Maryland

Via: Office of Naval Research
Washington 25, D. C.

Texas A & M College
Oceanography Department
College Station, Texas
Via: Commanding Officer
Office of Naval Research
Branch Office
Chicago, Illinois

51

12.

14.

15.

16.

17.

18.

Scripps Institution of Oceanography

San Diego 25, California

Via: U.S. Navy Electronics Laboratory
San Diego 25, California

Scripps Institution of Oceanography
San Diego 25, California

Laboratory é Attn: Bathythermograph Section
San Diego, California Via: U.S. Navy Electronics Laboratory
eeu Code 550) San Diego 25, California
. U. S. Navy Underwater Sound
13. Director

Woods Hole Oceanographic Institution

Woods Hole, Massachusetts

Via: Inspector of Naval Material
Boston, Massachusetts

Columbia University

Department of Geology

New York 27, New York

Via: Inspector of Naval Material
Brooklyn, New York

University of Texas

Austin, Texas

Via: Inspector of Naval Material
Houston, Texas

U. S. Coast Guard Headquarters
Test & Development Branch
Washington 25, D. C.

Department of the Navy
Bureau of Ships

Code 816

Washington 25, D. C.

Via: Inspector of Naval Material
Baltimore, Maryland

Department of the Navy

Bureau of Ships

Code 847

Washington 25, D. C.

Via: Inspector of Naval Material
Baltimore, Maryland

ayy:

t ‘ ii A q
iyi Pra) : { rt
sak saat Sane Care aie ae

ark
ie ea,

NP a's,