Proceedings ofthe United StatesNational MuseumSMITHSONIAN INSTITUTION ? WASHINGTON, D.C.
Volume 123 1967 Number 3609An Application of Automatic Data ProcessingTo the Study of Seabirds, INumerical Coding '
By Warren B. King, George E. Watson, and Patrick J. Gould
Seabird distribution, abundance, and movements in the open oceanare difficult to analyze because the limited data available have beencollected by various methods, the observers have varied in reliability,and the information is scattered widely. Furthermore, publication ofindividual sightings, except those of rare vagrants, has been at bestspotty with the result that abundance information is difficult to de-rive from the data. Records of at-sea bird observations are main-tained in several institutions in different countries, but most of theseare sortable only by hand, an exercise that is extremely time-consum-ing and invites clerical errors when the data are considerable. As aresult, few analyses have been published that involve complicatedcorrelations of extensive distributional and environmental data, andeven fewer seabird distribution maps show detailed documentationof seasonal occurrence and abundance.
1 Contribution no. 22 from the Pacific Ocean Biological Survey Program.A preliminary version of this paper was presented at the 14th InternationalOrnithological Congress, Oxford, England, July 25, 1966.2 King and Gould: Department of Vertebrate Zoology; Watson: Division ofBirds.
2 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 123Smithsonian Institution scientists recorded data on seabirds ob-served during 15 cruises aboard the Bureau of Commercial Fisheriesresearch vessel Tovmsend Cromwell while the ship was engaged inoceanographic research as a preliminary part of the Trade Wind ZoneOceanography Program. These data have been utilized in a pilotstudy to determine the feasibility of recording, storing, and analyzingseabird records in a computerized automatic data processing system.The system described herein is in part descended from a simpler onedeveloped by the Bureau to store and analyze its own bird observa-tions. Coding used in the various fields of data in this pilot projectis presented in this paper and details of the codes are given in thetables. Computer programming for the system is discussed byCreighton (in prep.). A full report on the results of the analysis ofbird observation data from these cruises will be published by King,but the preliminary results suggest that ADP may be applicable toa cooperative international system of seabird observation, datastorage, and analysis.The Bureau of Commercial Fisheries, Hawaiian Island area, hasbeen a vital partner in the development of this pilot study. John C.Marr, area director and Gunter Seckel, coordinator of the TradeWind Zone Oceanography Program, and their staff have made itpossible for Smithsonian scientists to observe seabirds aboard theTownsend Cromwell and have given us access to BCF oceanographicdata, punch cards, and computer programs. Reginald Creighton'sclose cooperation as programming analyst has turned the projectedaims of this system into a working reality. The original impetusto undertake the computer analysis of the seabird observation datacame from Philip S. Humphrey, Principal Investigator, Pacific OceanBiological Survey Program. Nicholas J. Suszynski, Jr., Director ofthe Information Systems Division at the Smithsonian, providedvaluable advice. Mrs. Anne Keenan Poulson patiently preparedthe figures for this publication. We are deeply grateful to all of them.Observations.?From March 1964 to July 1965 the SmithsonianInstitution conducted an at-sea study to determine the distribution,movements, and abundance of seabirds in a 500,000-square-mile areaof the central Pacific Ocean south and east of the main HawaiianIslands between 148? and 158? west longitude and 10? and 26?30' northlatitude (fig. 1). Each of the 15 replicate cruises covered a fixed trackof 4460 miles and lasted approximately 20 days. There was no cruisein August 1964.Two experienced observers alternated 2-hour watches each day fromsunrise to sunset. They identified birds at sea as accurately andspecifically as possible and recorded all data on standardized reportingforms. Some identifications were confirmed through collection of
no. 3609 DATA PROCESSING SEABIRDS KING, WATSON, GOULD 3
specimens. The time of observation was recorded to the nearestminute. Whenever possible, an actual count was made but largeflocks were only estimated within accuracy limits. The direction ofmovement, if any, was recorded. Notes were made on behavior,plumage, molt, age, sex, and any other data that appeared significant.Descriptions were recorded in cases of uncertain identification. Stand-ard marine meteorological observations were taken every six hours.Bureau of Commercial Fisheries personnel took bathythermographsand recorded weather conditions every one-half degree, hydrographicstations every one and one-half degrees, and surface plankton towsevery evening. Ship's position was reckoned hourly and early morn-
180? 160?W 150?W 140?W
? ".***,,
4 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 123
meteorologic, upon the distribution and movements of seabirds inthe area. And finally, ADP permitted more rapid statistical handlingof the data in order to ascertain (1) whether apparent monthly-population and distribution changes were real or were merely due toinadequate sampling, and (2) whether or not apparent correlationsbetween environmental factors and bird distribution were statisticallysignificant.The large number of different parameters to be analyzed requiredthe use of three decks of 80-column punch cards to retain a maximumamount of data in the system. The three decks were called "sightingdeck," code number 99, "environment deck," 03, and "day deck," 98(fig. 2). Of the three decks, the sighting deck was the largest, con-
Figure 2.?Three decks of 80-column punch cards. Datain columns 3-15 interrelate all three decks; environmentand sighting decks are further interrelated by data incolumns 16-28.
taining 18,000 cards for the 13,080 sightings. The environment deckwas smaller, containing 2800 cards. Each environment card wasassociated with all the sightings from the time of one set of environ-mental data observations to the time of the succeeding set of observa-tions. Similarly, each of the 305 day cards was associated with allsightings and all environment cards on that day. Thus, each obser-vation had a sighting card or cards, an environment card, and a daycard. But it shared the environment and day cards with othersightings.The three decks of cards were tied together by common dataconsisting of ship's name, cruise number, and date. The environ-ment and sighting cards were further linked by hour and position.
no. 3609 DATA PROCESSING SEABIRDS?KING, WATSON, GOULDThe 80-column punch card system works on a numerical code intowhich all verbal data recorded by the observer at sea must be trans-lated before entry. The Cromwell observations, oceanography,weather, and ship's position were checked for accuracy by a singleeditor and entered as numbers on intermediate coded sheets of differentkey colors. Data were punched on the cards directly from theintermediate coded sheets.Sighting deck.?The sighting card includes all data that pertain
.to each individual sighting. We define a sighting as the observationof a bird or group of birds acting as a unit. Data for the followingfields were entered in the indicated columns on the green intermediatesheets (fig. 3) for the sighting cards (card code 99 in columns 1-2)
:
Ship name 3-5Cruise number 6-9Date 10-15Local time 16-19Position 20-28Species identification 29-34Association code 35Species identification reliability 36Number 37-41Number reliability 42Direction of bird movement 43Behavior 44Color marking 45Method of marking 46Age 47Sex 48
Color-plumage phase 49Molt 50Whether or not a specimen was col-lected 51Food association 52Special weather conditions 53Special information 54Duration of sighting (minutes) 55-57Ship's speed (knots) 58-59Ship's direction 60-61Distance to nearest land (nauticalmiles) 62-65Distance to nearest breeding area(nautical miles) 66-69[Columns 70-80 were not utilized onthe sighting cards for this program.]Each ship used for observations was assigned a code number; theCromwell was 006. Cruises were coded serially for each ship. Thedate was expressed as month, day, and year with only the final twodigits being coded for year. Local and all other absolute times wereexpressed to the nearest minute in the 24-hour system.Position of the sighting and all other localities were expressed ascoordinates of longitude and latitude. In order to economize oncolumns, the world was divided into octants along the equator andeach 90? meridian. The octants (column 20) were numbered from0-3 in the Northern Hemisphere and from 5-8 in the Southern Hemi-sphere beginning at 0?-89?59'W and proceeding west. In this way itwas not necessary to express the 100's figure and east and west forlongitude nor north and south for latitude.The numerical code for seabird species is based on a revised check-list of the seabirds of the world. It is designed to permit data re-trieval hierarchically by class, order, family, genus, species, and/orsubspecies. The code requires only six digits because of the relatively
PROCEEDINGS OF THE NATIONAL MUSEUM VOL. 123
no. 3609 DATA PROCESSING SEABIRDS KING, WATSON, GOULD 7low number of forms involved, yet it includes all known taxa of sea-birds and has sufficient leeway for most synonymy and the inclusionof land-based accidentals, migrants, and marine mammals and reptiles.The first digit represents class; the second, order; the third, family;the fourth, genus; and the fifth and sixth, species or field-recognizable,nonoverlapping subspecies. The code for central Pacific Ocean birdsused in this pilot program is given in table 1. Codes for seabirds(presently being revised) and marine mammals of the world areavailable from the Smithsonian Institution. Birds that are identifiableonly to class, order, family, genus, and/or species may be entered andretrieved at the same level of reliable identification.When the code translation of a single sighting involves more thanone item of information in any category, additional sighting cards areneeded for complete and unique expression. Thus, a complex sightinginvolving four species of birds of three color phases and of three agecategories requires 10 cards. These 10 cards are tied together by theiridentical times and by a special code key known as the association key.A "0" in the association column (35) indicates a simple sightingexpressible by one card. A "1" in the column indicates that the cardhas information pertaining to the same sighting as all other cards withthe same time and a "1" in the association column. A sighting of amixed flock of sooty terns and brown noddies feeding together wouldbe represented by two cards each bearing a "1" association codenumber. When two or more cards of a sighting have identical datain the ship, cruise, date, time, position, species, and number fieldsbut differ in other fields such as age or behavior, then the numbers 2,3, 4, or 5 are used in the association column to preserve uniqueness.This is necessary because the computer automatically rejects, asduplicates, cards with identical data in columns 1-42. If a sighting-involves two sooty terns, one adult and one immature, the first cardbears the association number 1 and the second, number 2. A sightingof three sooty terns, one feeding, one sitting on the water, and a thirdsearching, would have three cards with association code numbers 1,2,and 3 respectively. If, on the other hand, two or more differentsightings have identical data in the ship, cruise, date, time, position,species, and number fields but the birds were not associated in thesame flock, we employ the numbers 6, 7, 8, 9, or in the code to indicateuniqueness but nonassociation.Note that "0" in the association code has a double function. It maybe used merely to denote a simple, one-card sighting or one of fiveconcurrent nonassociated sightings. When more than five nonasso-ciated sightings occurred simultaneously, it was necessary to adjustthe time of the supernumerary sightings by one minute to preserveuniqueness.
8 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 123Codes for species identification and number reliability, direction ofbird movement, behavior, marking, and other sighting fields are givenin table 2. Reliability was determined and recorded by the observerin his at-sea notes. Some of the at-sea logs recorded doubtful identifi-cations of dark shearwaters as "sooty-slender-bill" with a notationthat the underwings appeared light. Such sightings were coded as
"sooty shearwaters," 122519, in the species identification field (29-34),as "either this or the next most similar form," 2, in the species relia-bility field (36) and as [consult] "identification" [data in at-sea log],1, in the special information field (54). The next similar species to thesooty shearwater in the central Pacific is the slender-billed shear-water, but in the southern Atlantic Ocean it might be the great-wingedpetrel, Pterodroma macroptera.Special means of marking birds to show points of origin or otherstatus were employed in studies of seabirds in the central PacificOcean. Birds marked by "painting" with body dye or by attachingleg bands and colored plastic streamers were expressed in code inthe system. Colors were used for identification of birds at a distancewhen they could not be captured. In seabirds, color or plumagephases may indicate geographic origin or reproductive condition.Light and dark color phases in petrels, boobies, and jaegers mayvary in relative proportions in different geographic populations.A pinkish or yellowish tinge in white tropicbird, gull, or tern plumagemay indicate reproductive activity.The special weather conditions field was used to refer to squalls,storms, or calms that were so local they could not be expressed inthe environment deck. The special information column was used asa warning signal to alert the scientist to consult the raw data for aparticular field in the observer's at-sea log. A routine sighting as abird crossed the ship's bow was considered arbitrarily to have aduration of one minute. Birds following in the wake were countedperiodically, and an attempt was made to determine duration of theirassociation with the ship. Direction of ship movement and windand wave directions on the environment card were recorded as thefirst two digits of the compass heading rounded off to the nearest10 degrees (e.g., 165?= 17, 24?=02). A stationary ship or calm airor sea was coded 00.Part of a page of an at-sea observation log and its numericallycoded translation onto the intermediate coded sighting sheet is shownready for punching into cards in figure 4.Environment deck.?The format of the environment card wasslightly modified from cards already in use by the Bureau of Com-mercial Fisheries in the Hawaii area. Complete environmental dataof this sort are recorded on standard forms and punched into cards
no. 3609 DATA PROCESSING SEABIRDS?KING, WATSON, GOULD 9on file at the BCF in Honolulu. The utilization of this existingcard format in our system saved duplication of effort and man hoursof work by permitting standardization of data collection and storage.The codes used by the BCF and by us in this program are basedon number codes developed by the Hydrographic Office in Publication607, "Instruction Manual for Oceanographic Observations," 1955,reprinted 1959, for use by deck officers of vessels. On the environ-mental punch card deck (code 03 in columns 1-2), the fields for ship's
OBSERVERS:
-;'?'
.^ ySMITHSONIAN INSTITUTIONDIVISION OF BIRDSAT SEA DAILY LOG - E Date /#4U/MPg.# /TIME SPECIES
Figure 4.?Bird observation data in at-sea log (above) and part of the numerical translationonto the intermediate coded sheet for sighting deck.
name 3-5, cruise number 6-9, date 10-15, time 16-19, and position20-28 provide linkage with a particular group of sighting cards.The following additional fields of data are found on the environmentcard and are coded on red intermediate sheets (fig. 3) in a manneranalogous to that used for the sighting deck.Bathythermograph file number for Surface water temperature 35-37reference 29-31 Surface salinity 38-40Thermocline depth 32-34 Water temperature at 10 meters 41-43
10 PROCEEDINGS OF THE NATIONAL MUSEUMWind direction 44-45Wind speed 46-47Barometer 48-49Weather conditions 50-51State of sea 52Visibility 53Surface oxygen 54-56Surface phosphate 57-59
Moon age 60Cloud cover 61Air temperature 62-64Relative humidity 65-66Wave direction 67-68[Columns 69-80 were not utilized onthe environment cards.]
Thermocline depth was recorded in meters. Water temperaturesat the surface and at 10 meters and air temperature were given to thenearest .1? C. Surface salinity was recorded in parts per thousand,omitting the initial 3 for any value. Wind and wave direction werecoded in the same manner as direction of ship's movement, column 43on the sighting card. The code for weather conditions is given intable 3 and those for state of sea, visibility, and cloud cover, in table 4.Wind speed was recorded in knots. Barometric pressure was recordedas the last two digits of a millibar reading, e.g., 1015 millibars wascoded as 15. Surface oxygen was recorded in milliliters per liter tothe nearest .01 and surface phosphate in microgram atoms per literto the nearest .01. Relative humidity was expressed in whole percent.Moon age was not recorded for this study but a BCF code is given intable 4.Day deck.?The day card includes information pertaining to allsighting and environment cards for one day and is concerned mainlywith positions, distances, and times. From this card, in combinationwith the sighting cards from the entire day, the computer calculatesdaily density figures in terms of birds per hour and birds per linearmile. Through the use of such density figures, we can make directcomparison between bird density in different areas of the ocean andfrom day to day, month to month, and even year to year within thesame area. It is important to be able to express density as a factorboth of time and distance because the two are not correlated neces-sarily on a cruise with periodic stops such as the BCF Cromwellcruises. The day cards, code 98 in columns 1-2, are linked to the othertwo decks through information on the ship's name 3-5, cruise number6-9, and date 10-15. In addition, the day card contains the followinginformation coded on the blue intermediate sheets (fig. 3)
:
Diurnal hours of observation 16-18Diurnal miles of observation 19-21Nocturnal hours of observation 22-24Nocturnal miles of observation 25-27Position at midday 28-36Position at sunrise 37-44Position at sunset 45-52
Local time of sunrise 53-56Local time of sunset 57-60Number of observers on each watch 61Quantitative plankton analysis 62-64[Columns 65-80 were not utilized onthe day cards.]
no. 3609 DATA PROCESSING SEABIRDS?KING, WATSON, GOULD 11Hours and miles of observations were recorded to the nearest .1hour and whole mile respectively. The ship's location at noon wasrecorded as midday position for a full day of observations but themidpoint of the diurnal track was recorded for a partial day of obser-vation. Surface plankton was collected in a 25-minute tow with a1 meter .308 millimeter terminal mesh net. Volume was measuredby BCF personnel and recorded in cubic centimeters per cubic meterof water strained. Computer AnalysisThe coded information was transferred from the cards onto magnetictape in the computer with the result that all data pertinent to eachsighting were stored and printed out as a unit. Thus, environmentaldata and the raw materials for calculating densities were included withthe print-out for any individual sighting.The system has been programmed in COBOL [Common OrientedBusiness Language] for a General Electric 425 digital computer having128,000 characters of core storage.The data flow sheet for the system thus includes the sighting thatis recorded in the observer's field notes, edited and coded on theintermediate sheets, punched into the cards, and transferred ontotape in the computer where it is available for print-out, plot-out, andanalysis (fig. 5).With this information fed into the system, various questions may beasked of the data. The simplest questions are descriptive: whatspecies occur in what numbers, where, and when? We can ask for adetailed print-out of all the data for all sightings arranged by species,by locality, or by time. Part of such a code print-out of red-footedbooby observations is shown in figure 6.Relatively simple correlations of species distribution with environ-mental conditions may be requested as tallies to provide raw materialsfor graphs or statistical analysis. Tallies of black-footed albatrossand sooty tern sightings at various surface water temperatures yielddata for plotting bar graphs (fig. 7) . The two species differ somewhatin water temperature preference.More detailed compound questions involving several conditionalclauses will give more restricted print-outs; thus, by asking for allrecords of sooty terns in aggregations greater than 25 seen feedingwithin the first two hours after sunrise during May in association withwater 25?-27? C and having a salinity greater than 34.5 parts per thou-sand we obtain a much more selective fisting. Indeed, only one sightingqualified (fig. 8). The instructions are in COBOL code language,which the computer understands. Asking this question of our 13,080
12 PROCEEDINGS OF THE NATIONAL MUSEUM
sightings took 20 minutes of computer time. Twenty-five questionsof this complexity may be asked at the same time.The most graphic application of automatic data processing to theobservation data is for plotting distribution maps. Seasonal move-ments and changes in abundance may be depicted directly by machineprint-outs in the form of annotated cruise tracks. Such machineprinted distribution maps of the two color phases of wedge-tailedshearwaters from May to July are shown in figure 9. Light phase birdsbreed in the Hawaiian Islands to the north; dark phase birds, below
Observer
Observation
At SeaLog Editor
ResearchScientist
Print-OutandTally
IntermediateCodedSheets
DATA PROCESSING SEABIRDS?KING, WATSON, GOULD 13
others are suggested only for possible incorporation into a larger scaleinternational program.On the sighting cards (p. 5), the fields "distance to nearest land"(columns 62-65) and "distance to nearest breeding colony" (66-69)will be deleted. These distances have proved laborious to calculatebut will be provided automatically by the computer when the neces-sary program has been written. It may be possible to incorporate afourth deck of cards into the system that will give the phenology of
640316
14 PROCEEDINGS OF THE NATIONAL MUSEUM
of a Secchi disc provides a good absolute measure for water clarityand one that is taken easily as a routine oceanographic measurement.At the same time, apparent water color may be recorded in the Forelscale. The BCF records and files both measurements on their punchcards. Photometric depth is another measure of water clarity madeon oceanographic cruises, but it demands much more complicatedinstrumentation and the figures vary with intensity of illumination.It may be useful to record the presence of flotsam, large masses of weed,ice, or other floating debris. Some observations suggest that marinelife and birds may be attracted locally to foreign matter in the water.
<
?
.20
16-
5 .12
o .08
.04-
QOO
r10
-.6
-A
-.2
(0za.m ?h l> L.o(A
<22 e 22-23.9? 24?- 25.9? 26?>Figure 7.?Bar graphs of abundance of black-footed albatross (shaded) and sooty terns(white) in birds per mile seen at various surface water temperatures on Cromwell cruises.On the day sheets several changes should be made. All observa-tions for the pilot program were made in a single octant of the world,but other sets of observations may be made in several octants or acruise track may cross the equator or a 90? meridian and thus requiretwo octant codes for the same day. For this reason octant columns(37 and 46) should be added to sunrise position and sunset positionfields. As a result, the fields for sunrise and sunset time and numberof observers would be moved two columns to the right. It has notbeen feasible to make these octant additions in the Smithsonianprogram.
no. 3609 DATA PROCESSING SEABIRDS?KING, WATSON, GOULD 15Three additional fields have already been added to the day sheets:time of beginning of observations (65-68) , time of end of observations(69-72), and special parametric analysis (73-75). The times of begin-ning and end of observations coincided with the times of sunset andsunrise on the Cromwell cruises, but in other studies observations maybegin or end earlier or later.In the pilot study, density figures were computed on the basis of afull day's observations. In future studies, however, it may benecessary to determine abundance in areas that were covered in less
PROVIDE ALL RECORDS OF SOOTY TERNS IN AGGREGATIONS GREATERTHAN 25 SEEN FEEDING WITHIN THE FIRST TWO HOURS AFTER SUN-RISE DURING MAY IN ASSOCIATION WITH WATER 25 TO 27 DEGREESC AND HAVING A SALINITY GREATER THAN 34.50 PARTS PER THOU-SAND.
REQUEST PARAMETER CARDS, PACIFIC PROJECT, JUNE 8, 1966.IF IDENT EQUALS 146867 ,AND NUMB IS GREATER THAN 00025 ,AND HOURS -AFTER-SUNRISE IS GREATER THAN 00AND LESS THAN 03 ,AND BEHAVIOR EQUALS 3 ,AND WATER-TEMPERATURE IS GREATER THAN 249AND LESS THAN 271 ,AND SURFACE -SALINITY IS GREATER THAN 450AND MONTH EQUALS 05 ,GO TO SROl.
16 PROCEEDINGS OF THE NATIONAL MUSEUM
IOFNTi??5lC ?7 26 25 ?4 23 22 21 20 1$ IB l? 16 15 H 13 12 11
flO 00 00 00 01 1ft 01 00 00 07 02 05 07 04 04 02 01 02 03 02 00 00 04 0(1 01 02 Ofl 06 Ol^oY^lO 03 0400 02 00 0?
?01 00 0005 00 01
ni oi oo oi00 06 31 04 02 06 16(24j99] > 7 fl4 00 03 0000 OO 0?0?Light Phase Wedge -tailed Shearwater4 1
?(5) G3?(??) 17 O5 (30?S) 06 ob ofl ? 5 02Qy i2 l3 ?" ? 9 02 ? n 01^3^53)n5 00 06^2)01 On (5o)ntl07 fi" 00 (To)00 18 04no 02
???00?00.- " " ? 3 ? 10 ? 8 ?'? 15 "
26 25 24 23 22 21 20
00 00 00 00 o>) on on oo 01 oo on 01 oo on oo oo 02 o? 01 01 00 00 02 on no 02 04 0100 00 0.000?
?
?00 00 ni 1600 00 00 00 01 on 00 00 00 00 00 01 02 01 00 01 00 01 02 01 no 05 02 ni oo 14 00
00 01 Dark Phase Wedge -tailed Shearwater
01 03 02 00 00 00 02 02 04 02 03 00 00 00 00 01 OOUMm 03 00 01 01 '2 ozQljQni'yfecS00 00 00 06 10oo oo oo Qe)
oi oo(6?)oa00 00 OS02 00 00 0300 01 00 01 00 00 01 OO 00 01 OO OO 04 02 01 01 00 00 00 02 On 02 17 03 04 06^^(2^00 00 00 01 ??
Figure 9.?Computer printed distribution maps of two color phases of wedge-tailed shear-waters May-July. Four-digit numbers in extreme left margin refer to Cromwell cruisesby month and year on which first observation in that line was made. Note that, onApril and July cruises, observations were made in May and August. North is to theleft, latitude is read from right to left, and longitude from top to bottom. The cruisetrack is outlined by the number print-out. Asterisks indicate no observations were madein that half-degree square, 00 indicates observations but no birds. Abundance maybe read directly from the printed numbers. Circles indicate 20 birds or more per half-degree square.
no. 3609 DATA PROCESSING SEABIRDS?KING, WATSON, GOULD 17day cards. Only those sightings that were made within the timeinterval on the fractional day card are flagged for special parametricanalysis. Because the data are flagged on the basis of times ofobservation, the computer does not require geographic informationto define the area; thus, bird densities may be computed for areasof any size or shape. These areas may be mutually exclusive or mayoverlap wholly or in part (fig. 10). The computer also will providedensity figures based on total observations made within a given areaduring all fractional days within a requested time span. Areas may
Figure 10.?Three overlapping areas along a day's cruise path with the coded representa-tions for various subareas. All sightings made within the time the ship was in the squareare flagged with a 1 in first column of the SPA field; those in the circle and the triangle,with l's in the second and third columns respectively.be selected for special parametric analysis either before the datahave been entered on punch cards or at any future time by fractioniz-ing the appropriate day cards.The revised intermediate coded sheet for the day card is shown infigure 3.Plankton measurements were made only once daily on the Cromwellcruises, but in other programs plankton abundance may be sampledmore frequently. The plankton field (62-64), therefore, should bemoved to the environment card where it would occupy columns 69-71.
18 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 123Primary productivity, a measurement of the richness of environ-ment and potential food supply, eventually may prove useful forcomparison with bird distribution. At present, however, becausethere is no universally agreed-upon measure of productivity, we haveused surface phosphate and plankton abundance as measures of themarine environment.The data stored on tape reels in the computer may be alteredthrough a revised program if these or any later modifications becomenecessary.With some modifications, the coding used in this pilot seabird ADPsystem is suitable for storage of seabird records voluntarily sent infrom all over the world to a central office or offices such as WorldData Centers A and B for Oceanography in Washington and Moscow.Information submitted on standardized forms would be available foruse by marine ornithologists and ecologists. The establishment of aworld center for the storage of seabird distribution data and inter-national agreement on uniform methods of observation, recording,and coding of data were two basic recommendations of a seabirdcommittee authorized by the 14th International OrnithologicalCongress, Oxford, July 1966.Varying reliability of observations and different techniques forgathering data, however, probably will necessitate uniform editingbefore the data are fed into the system. This suggests the need for acentral editor for all the data and a staff to prepare them for entry.As an indication of the extent of the work that may be needed toprepare data for computer storage, two persons worked full timeeight hours a day for 50 days transferring the raw at-sea data fromthe Cromwell cruises onto intermediate coded sheets before they couldbe punched into the system. Because of the far larger volume ofdata that may be collected by many more observers from all over theworld in an international program, it might be necessary for them tosubmit data to the proposed editor or central files already translatedonto intermediate coded sheets.The worldwide scope of such a seabird data storage project and thefact that many species of seabirds associate with commerciallyvaluable schooling fish make this an attractive candidate for supportunder the International Biological Program, the major theme ofwhich is productivity and man.
no. 3609 DATA PROCESSING SEABIRDS?KING, WATSON, GOULD 19TABLE I.?Codes for sighting deckSpecies Identification for Central Pacific Ocean: 29-34100000 Aves120000 Procellariiformes121000 Diomedeidae121100 Diomedea121110 D. nigripes121111 D. immutabilis121112 D. melanophris122000 Procellariidae122100 Fulmarus122101 F. glacialis122500 Puffinus122501 P. carneipes122504 P. creatopus122506 P. pacificus122518 P. bulled122519 P. griseus122520 P. tenuirostris122521 P. nativitatus122524 P. puffinus newelli122534 P. Iherminieri122600 Pterodroma122601 P. hasitata122607 P. h. phaeopygia122608 P. h. externa122609 P. h. cervicalis122615 P. solandri122620 P. rostrata122624 P. inexpectata122629 P. hypoleuca122630 P. h. hypoleuca122631 P. h. nigripennis122633 P. neglecta122636 P. alba122642 P. arminjoniana heraldica122645 P. coohi122646 P. cooki cooki122649 P. leucoptera122653 P. I. masafuerae
20 PROCEEDINGS OF THE NATIONAL MUSEUM122700 Bulweria122701 B. bulwerii123000 Hydrobatidae123100 Oceanites123101 0. oceanicus123400 Oceanodroma123404 0. pastro123409 0. leucorhoa123420 0. tristrami123423 0. furcata123500 Miscellaneous genera123502 N esofregetta albigularis130000 Pelecaniformes131000 Phaethontidae131100 Phaethon131106 P. rubricauda131112 P. lepturus133000 Sulidae133200 Sula133206 S. dactylatra133213 5. sula133217 S. leucogaster136000 Fregatidae136100 Fregata136107 F. minor136114 F. ariel140000 Charadriiformes141000 Charadriidae141200 Pluvialis141209 P. dominica141400 Arenaria141401 A. interpres142000 Scolopacidae142100 Numenius142101 N. tahitiensis142200 Heteroscelus142201 H. incanum142700 Crocethia143000 Phalaropodidae143100 Phalaropus143101 P. fulicarius
no. 3609 DATA PROCESSING SEABIRDS?KING, WATSON, GOULD 21144000 Stercorariidae144100 Stercorarius144101 S. pomarinus144103 5. longicaudus144200 Catharacta144201 C. skua145000 Larinae145100 Larus145101 L. artri cilia145145 L. delawarensis145184 L. glaucescens145185 L. hyperboreus146000 Sterninae146100 Anous146101 A. stolidus146110 A. tenuirostris146200 Procelstema146201 P. cerulea146300 Gygis146301 G. alba146800 Sterna146801 S. bergi146829 5. hirundo146835 S. paradisaea146857 S. sumatrana146866 5. lunata146867 S. fuscataTABLE 2.?Codes for sighting deckSpecies identification reliability: 36 Number reliability: 42Sighting as a whole reliable Actual count1 Sighting as a whole questionable 1 ? 5 per cent2 Either this or next most similar 2 ? 10 per centform 3 ? 15 per cent3 This or any similar form 4 ? 20 percent4 Genus reliable, species 5 ? 25 percentunreliable 6 ? 30 or 33 per cent5 Genus unreliable, species 7 ? 40 per centunreliable 8 ? 50 per cent6 Sighting as whole unreliable 9 Present, but no count possible
22 PROCEEDINGS OF THE NATIONAL MUSEUMDirection of bird movement: 43No data1 No apparent direction2 North3 Northeast4 East5 Southeast6 South7 Southwest8 West9 Northwest
Behavior: 44No information1 Traveling2 Sitting3 Feeding4 Searching5 Following ship6 Breeding (display)7 Parasitism8 Dispersing9 Other
Color marking: 45
no. 3609 DATA PROCESSING SEABIRDS?KING, WATSON, GOULDSpecimen: 51 Special weather conditions:No data1 Birds in rain2 Bird at edge of rain3 Very high winds4 Calm5 OtherSpecial information: 54No data1 Identification2 Species association3 Behavior4 Abundance5 Location6 Direction of movement7 Food association8 Other9 Parasitism
2353
Not collected1 Collected
Food association: 52
24 PROCEEDINGS OF THE NATIONAL MUSEUM04 Visibility reduced by smoke,e.g., veldt or forest fires,industrial smoke, or vol-canic ashes.05 Dry haze.06 Widespread dust in suspen-sion in the air, not raisedby wind at or near the shipat the time of observation.07 Dust or sand raised by windat or near the ship at thetime of observation, but nowell-developed dustdevil (s), and no duststormor sandstorm seen.08 Well-developed dust devil(s)seen at or near the shipwithin last hour, but noduststorm or sandstorm.09 Duststorm or sandstorm withinsight of the ship or at theship during the last hour.10 Light fog (visibility 1,100yards or more).
Haze, dust, sand, orsmoke.
11 Patches of.12 More or less continuous.13 Lightning visible, no thunderheard.14 Precipitation within sight butnot reaching surface at theship.15 Precipitation within sight,reaching surface, but dis-tant (i.e., estimated to bemore than 3 miles from theship).16 Precipitation within sight,reaching surface, near tobut not at the ship.
Shallow fog at the ship notdeeper than about 33feet.
DATA PROCESSING SEABIRDS?KING, WATSON, GOULD 25
Within sight during thelast hour.
y Not falling as showers.
17 Thunder heard, but no pre-cipitation at the ship.18 Squall(s)19 Funnel cloud(s) (tornado orwaterspout).20-29 Precipitation, fog or thunderstorm at the ship during the pre-ceding hour but NOT at the time of observation.20 Drizzle (not freezing).21 Rain (not freezing).22 Snow.23 Rain and Snow.24 Freezing drizzle or freezingrain.25 Shower(s) of rain.26 Shower(s) of snow, or of rainand snow.27 Shower(s) of hail or of hailand rain.28 Fog.29 Thunderstorm (with or withoutprecipitation).30-39 Duststorm, sandstorm or drifting snow.30 Slight or moderate duststorm Has decreased duringor sandstorm.31 Slight or moderate duststormor sandstorm.32 Slight or moderate duststormor sandstorm.33 Severe duststorm orsandstorm.34 Severe duststorm orsandstorm.35 Severe duststorm orsandstorm.36 Slight or moderate driftingsnow.37 Heavy drifting snow.38 Slight or moderate driftingsnow.39 Heavy drifting snow.
preceding hour.No appreciable change dur-ing preceding hour.Has increased duringpreceding hour.Has decreased duringpreceding hour.No appreciable change dur-ing preceding hour.Has increased duringpreceding hour.Generally low.
Generally high.
26 PROCEEDINGS OF THE NATIONAL MUSEUM40-49 Fog at the time of observation.40 Fog at a distance at the timeof observation, but not atthe ship during the lasthour, the fog extending to alevel above that of theobserver.41 Fog in patches.42 Fog, sky discernible.43 Fog, sky not discernible.44 Fog, sky discernible.45 Fog, sky not discernible.46 Fog, sky discernible.47 Fog, sky not discernible.
Has become thinner duringpreceding hour.No appreciable change dur-ing preceding hour.Has begun or has becomethicker during precedinghour.48 Fog, depositing rime, skydiscernible.49 Fog, depositing rime, skynot discernible.50-99 Precipitation at the ship at the time of observation.50-59 Drizzle at time of observation.50 Drizzle, not freezing,intermittent.51 Drizzle, not freezing,continuous.52 Drizzle, not freezing,intermittent.53 Drizzle, not freezing,continuous.54 Drizzle, not freezing,intermittent.55 Drizzle, not freezing,continuous.56 Drizzle, freezing, slight.57 Drizzle, freezing, moderateor thick.58 Drizzle and rain, slight.59 Drizzle and rain, moderateor heavy.
Slight at time ofobservation.
Moderate at time ofobservation.
Thick at time ofobservation.
no. 3609 DATA PROCESSING SEABIRDS?KING, WATSON, GOULD 27
Slight at time ofobservation.
Moderate at time ofobservation.
Heavy at time ofobservation.
60-69 Rain at time of observation.60 Rain, not freezing,intermittent.61 Rain, not freezing,continuous.62 Rain, not freezing,intermittent.63 Rain, not freezing,continuous.64 Rain, not freezing,intermittent.65 Rain, not freezing,continuous.66 Rain, freezing, slight.67 Rain, freezing, moderate orheavy.68 Rain or drizzle and snow,slight.69 Rain or drizzle and snow,moderate or heavy.70-79 Solid precipitation not in showers at time of observation.70 Intermittent fall of snowflakes.71 Continuous fall of snowflakes.72 Intermittent fall of snowflakes.73 Continuous fall of snowflakes.74 Intermittent fall of snowflakes.75 Continuous fall of snowflakes.76 Ice needles (with or withoutfog.77 Granular snow (with orwithout fog).78 Isolated starlike snowcrystals (with or withoutfog).79 Ice pellets.
Slight at time ofobservation.
Moderate at time ofobservation.
Heavy at time ofobservation.
28 PROCEEDINGS OF THE NATIONAL MUSEUM80-99 Showery precipitation, or precipitation with current or recentthunderstorm.80 Rain shower(s), slight.81 Rain shower(s), moderate orheavy.82 Rain shower(s), violent.83 ShoweKs) of rain and snowmixed, slight.84 Shower(s) of rain and snowmixed, moderate or heavy.85 Snow shower(s), slight.86 Snow shower(s), moderate orheavy.87 ShoweKs) of soft or small hailwith or without rain, or rainand snow, slight.88 Shower(s) of soft or small hailwith or without rain, or rainand snow mixed, moderateor heavy.89 Shower(s) of hail with or with-out rain, or rain and snowmixed, not associated withthunder, slight.90 Shower(s) of hail, with orwithout rain, or rain andsnow mixed, not associatedwith thunder, moderate orheavy.91 Slight rain at time ofobservation.92 Moderate or heavy rain at timeof observation.93 Slight snow, or rain and snowmixed, or hail* at time ofobservation.94 Moderate or heavy snow, orrain and snow mixed, orhail* at time of observation, i
Thunderstorm during pre-ceding hour but not attime of observation.
"Hail, small hail, soft hail.
=600 DATA PROCESSING SEABIRDS KING, WATSON, GOULD95 Thunderstorm, slight or mod-erate, without hail* but withrain and/or snow at time ofobservation.96 Thunderstorm, slight or mod-erate, with hail* at time ofobservation.97 Thunderstorm, heavy, withouthail* but with rain and/orsnow at time of observation.98 Thunderstorm combined withduststorm or sandstorm attime of observation.99 Thunderstorm, heavy, withhail* at time of observation.
29
*Hail, small hail, soft hail. J
Thunderstorm at time ofobservation.
TABLE 4.?Codes for environmental deckState of sea: 52Flat calm1 Less than 1 foot2 1 to 3 feet3 3 to 5 feet4 5 to 8 feet5 8 to 12 feet6 12 to 20 feet7 20 to 40 feet8 40 feet and over9 Very rough, confused sea
Visibility: 53Dense fog. 50 yards1 Thick fog 200 yards2 Fog 400 yards3 Moderate fog 1000 yards4 Thin fog or mist 1 mile5 Visibility poor 2 miles6 Visibility moderate 5 miles7 Visibility good 10 miles8 Visibility very good ......... 30 miles9 Visibility excellent .... .over 30 miles(Use range-finder readings of known land)
Moon age: 60No dataNew1/41/23/4FullNot up
Cloud cover: 61No clouds1 Less than 1/10, or 1/102 2/10 and 3/103 4/104 5/105 6/106 7/10 and 8/107 9/10 and 9/10 plus8 10/109 Sky obscured
U.S. GOVERNMENT PRINTING OFFICE : 1967?0-24 8-334