Introduction
The world has witnessed in the last 10 years the synthesis of the art of ocean flying, from a New York to Paris solo flight of international acclaim in 1927 to a “routine transfer of men and equipment” in mass flight from San Diego to Honolulu in 1936. We have but seen the beginning of an era of air transportation. Today man plans Atlantic schedules, schemes to join distant ports by almost nonstop day-and-night flying, seeks shortest routes from continent to continent by way of the North Pole. These ambitious projects may stem their claims to feasibility, in part at least, from the successful operation of a 2-year-old commercial air line across the Pacific from California to China. Besides the geographical trail it has blazed in its 26-ton Martin Clippers over almost 8,000 miles of the deep blue with but 5 stepping stones (at Honolulu, Midway, Wake, Guam, and Manila), Pan American Airways may rightfully claim to have blazed a trail for commercial operators in the art of air navigation as well. It has coordinated the three methods of position finding available today, namely, dead reckoning, radio direction hiding, and nautical astronomy, and it has demonstrated their applicability to regular and frequent long-range operations. The proficiency of the system of air navigation developed in the sturdy Clipper craft on the San Francisco-Hongkong route has been brought home by the weekly schedules maintained, the regularity of which has occasioned the transfer of news of arrivals and departures from the front page headlines of the daily newspapers to the shipping news section.
Great interest has been evinced by naval personnel in the flights of the Clippers and in the approach to the problem of aerial navigation practiced by the crews aboard them. It is with the idea of possible clarification on some of the phases of this problem that this article has been written. It is a running account of a scheduled operation between San Francisco and Honolulu with mention of some of the “wrinkles” of air navigation utilized in flight, which possibly set it a little apart from the rigid rules of surface navigation. In this connection, for instance, it is probable that surface navigators will object to the looseness of phraseology in aerial navigation. They may recoil from the suggestion of crossing a radio bearing 200 miles out with a line of position and calling the result a fix. They may object, too, to our term “fix” which actually means, with only one navigator in a crew of six aboard, a running fix with times of sights separated by 5 minutes and a ground speed varying anywhere from 90 to 140 knots. Our accuracy may sometimes be subject to question by the meticulous academician. But lights are usually visible over 60 miles, even on cloudy nights, and the air navigator is not unduly alarmed although wounded in vanity perhaps, as he approaches landfall, if someone ventures the opinion that his star fix may be 10 or 15 miles off.
There are many controversial points to consider in a detailed analysis of the task of navigating an airplane and no attempt is made here to cover the entire problem. Below is merely one navigator’s account of a flight and the methods by which he attempts to cope with the myriad of details which demand attention en route. The accompanying chart (Fig. 1) shows the flight path and the fixes found on a flight which took place on February 25 and 26, 1937. The leather maps (Figs. 2 and 3) are reproductions of the area covered by the flight, taken from the actual maps drawn up by the meteorologist from the 1200 GCT Weather Bureau reports of these two days. They show the sequence of weather over the 24-hour period. The 0000 February 26 map has been omitted because it carries only a transition between the two 1200 situations.
Flight Number 91
2259 Feb. 25 GCT Left Dock Alameda
2310 Took of San Francisco Bay
2318 Took Departure Fort Point. Set course 215° psc. 18° E. Var., 1° E. Dev. Wind
(Oakland upper air) 295° force 14 knots, drift angle 8° left at 8,000 ft. Desired
track 226°(T)
All time notations on these flights are carried in Greenwich Civil Time. The ship carries one chronometer and two Longines second-setting watches, the watches being attached to the octants. The chronometer is, of course, checked daily by radio time signal. Two Pioneer octants and a sextant are carried on board. At a thousand feet or less, the sextant is preferred, as a precision instrument, when visibility and horizon permit. The dip-correction error introduced by a difference in true altitude and altimeter reading is negligible at this level, being on the order of 1.5 minutes per hundred feet. Clouds generally obscure the horizon on the rare occasions when this 1,000-foot altitude is maintained. Other considerations govern to make the usual flight level above 8,000 feet pressure altitude and most reliance is placed, therefore, on the octant and its artificial horizon. Before a scheduled flight the crew assembles in the meteorology office for a discussion of the weather map, drawn up every 12 hours from Weather Bureau Ships and Land Stations Reports.
The meteorologist’s forecast is the basis for the decision as to the route to be followed from Alameda to Honolulu. The Pacific, for our operations, is divided into zones, the first being from Alameda to Longitude 125° W., the succeeding zones in 10-degree increments, with zone number five from 155° W. to Honolulu. Flight time analyses, using the winds forecasted, are computed for the great circle route and two alternative routes, one from Alameda to a point 300 miles south of the midpoint of the great circle course, at Lat- 27°-30'N., Long. 135° W., the other to a point 300 north, at Lat. 35° N., Long- 1450 W. The distance added to the great circle distance by either of these detours is about 100 miles, a scant price to pay for the elimination of bad weather or head winds. Frequently the distribution of areas of high and low barometric pressures makes the time on one of these alternatives shorter than on the great circle route. This was the condition over the Pacific which obtained on the 1200 February 25 map and the decision to fly the southern route was made. The center of high pressure was located at about 29° N., 140 W., with north to east winds predicted for the southern side. The succeeding map, drawn several hours after our departure, indicated that the high pressure area had moved southeasterly faster than had been anticipated.
Flights of the magnitude of the San Francisco-Honolulu hop demand, in the interests of conservation and economy of fuel, that the ships be flown in “best miles per gallon” condition. The Engineering Department has, for this reason, prepared tables by which the ship is flown. Flights are normally made at 8,000 feet pressure altitude, the optimum altitude for the engines. The first half hour is spent in climb at a predetermined air speed. When altitude is attained the 950 h.p. Pratt and Whitney engines are shifted into normal cruising, horsepower reduced initially to 55 per cent and the automatic pilot engaged. The flights are then flown at a constant air speed in four time periods, each period consuming one-fourth of the fuel load, with reduction in horsepower being made at the beginning of each period to maintain the predetermined speed. The navigator can then reasonably expect constant air speed and headings flown within one degree of those called for, this latter being subject to precession within the automatic pilot and to the attention of the individuals on watch. The automatic pilot, needless to say, is indispensable to the navigator; its accuracy leaves little to be desired.
0025 Bearing from Alameda 226° (T)
0110 Gonio on KGO Broadcasting Station 45° (T)
0125 Bearing from Alameda 226.2° (T)
Although the Communications Department constantly reiterates that direction-finding bearings over long ranges are not yet to be considered out of the experimental stage, the results already obtainable over distances up to 400 and 500 miles are very gratifying. Reliance on their accuracy is, of course, tempered by consideration of such phenomena as twilight effect and shifting from meteorological causes. The Adcock DF system is used at virtually all the ground stations. The planes are equipped with goniometers, the accuracy of which is phenomenal within 150 miles. The radio operators aboard have frequently demonstrated the practicability of homing with them to small islands like Wake and Midway. The aerial navigator finds it difficult at first to reconcile the gonio bearings with his dead reckoning but repeated demonstrations of their accuracy soon convinces him of their worth. Ships at sea are contacted by radio, whenever necessary and available, for weather data and position determination. U. S. Navy vessels and the Dollar and Matson liners give the most aid to the flying boats since they are equipped with the frequencies on which the radio operator can take gonio bearings. Flying the great circle track is a fairly easy task when two or three ships, well spaced, provide such ground checks. For example, on one occasion flying eastbound above the lower cloud layers, the plane homed on a westbound Dollar liner until the gonio indicated the ship abeam, when a Coston water flare was dropped. The liner reported sighting the flare 2 miles off the starboard quarter and gave its current position. There have been flights made, however, with a scarcity of ships on the Pacific, as during the recent marine strike; when out there they serve as an additional and welcome check for the navigator.
The navigators of the Clippers use H.O. 208, Dreisonstok, almost exclusively, although an Ageton is always included in ship’s equipment. With no desire to reinvoke the old arguments about the relative merits of the two systems, it may be permissible to extend the remarks of the Resume of Navigation Methods, copyrighted in 1934 by the Naval Institute, by the following addenda: the use of the dead-reckoning position necessitates more chart work prior to solution of the line than Dreisonstok; using the dead-reckoning position does not always eliminate the long intercept, especially after several hours of night flying; and the loss of accuracy of fix in some cases on account of integral degree of latitude and hour angle tabulations can be disregarded in commercial aerial work. Confidence in his sights is the only weapon with which the navigator can combat the doubt the long intercept may raise. One last word for Dreisonstok: interpolation becomes with practice a nimble operation.
H.O. 214, Precomputed Dreisonstok, has every advantage of simplicity. At the moment it is not yet available for general use. We have had some experience with the Hagner Position Finder, an instrument which solves the astronomical triangle mechanically. Its size, weight, and cost would seem to offset the advantages it may offer at the present time.
0026 Assumed wind found practically correct. Used until later determination was made.
For mechanical solutions of the wind-track-heading triangle and for the slide-rule computations involved in dead reckoning, the Mark VII computer is used extensively. Since the usual procedure at night is to alter course only at times of fixes, unless other factors necessitate a change, the wind determination at times of fixes is a simple matter on the computer. Applied to our problem the method is as follows: The wind found at time of first fix is used to determine the heading to be flown the next two hours. If the weather map and other observations show clearly a change expected within the period, a “best guess” is used for the wind. A second fix is made two hours later. The heading and track, air- and ground-speeds being known between the two fixes, these data are applied to the face of the Mark VII and the wind force and direction are read directly from the grid plate.
In our operations, the 2-hour interval for fix determination is found to be practicable. If sights are taken at shorter intervals, there is apt to be too much amplification of error in the wind determination. For instance, if the second fix were taken at, say, the end of 1 hour and were 10 miles in error, the entire error would be added to the actual wind. If taken at the end of 2 hours, the same error in position applies only 50 per cent error to the wind- Naturally, the longer the interval between fixes the less the hourly error component will be. However, the approach to actual wind in this way holds true only if the wind remains constant. In 3 hours the plane travels more than 300 miles and the winds are quite likely to change appreciably over this area. Furthermore, the 2-hour period lends itself nicely to the cycle of work: a 2-star fix computation takes about 25 minutes, from the beginning of the star shooting to the plotting; correcting and plotting radio bearings filling out hourly log and radio position, report entries, decoding and plotting weather information, and studying the weather map are some of the paper work for which the navigator is responsible. On average clear nights, he can routinize his work so as to give himself a short rest period in this time. It is the writer’s practice to take sights commencing on the half-hour. The task of working the sights takes him up to the hour and changes of course are made then, the track and ground speed made-good between fixes being extended to the hourly position.
0026 Sun line, 6 sights averaged. Crossed with DF. to establish fix at Lat. 36°-13’N., Long. 124°-15’W.
Drift is measured in our planes by the use of drift bombs by day, and Coston water flares by night. The bombs are glass flasks filled with one pound of aluminum powder, which, upon striking the water, make a spot 10 to 20 feet in diameter. This target is clearly visible from 10,000 feet, even in moderately rough water. The bombs, incidentally, could be used to advantage as targets for dive-bombing practice in the fleet or at reserve bases. A Pioneer drift sight, a refinement of the Mark II Navy Drift Sight, is used in a bracket outside the lounge after-window. This method gives the drift with fair accuracy. That the observed drift angle will always be smaller than the actual drift in observations of this type may be illustrated by an example. Given the following data: Heading north, air speed 120 knots, wind from west, force 30 knots. A bomb is dropped from altitude 8,000 feet. At instant of striking water, assuming vertical homogeneity in wind, the bomb is directly behind and at an angle from the vertical of about 30 degrees. Thus the bomb will be
tan 30° X 8,000 = 4,620 feet
or roughly, 1 mile behind the plane. The diagram (Fig. 4) has been drawn to scale. This introduces an error in which the apparent drift angle a approaches the true T as the plane draws away from the bomb. The bomb remains visible for about 5 minutes and it is the practice, when drift is appreciable, to take the fading as late as possible. By reference to the diagram it is apparent that drift is greatest when the wind is abeam of the heading of the ship. With wind from any other point the wind effect on amount of drift decreases and its effect on ground speed increases. For a wind of this magnitude and direction the correction is 1.3 degrees. An average correction of 1 degree added to the observed will suffice in cases of beam winds.
The best determination of wind is by the wind-star method. The procedure is to fly on three headings, the original and two 60° on either side of the original for equal periods of time. The drift is observed on each heading and plotted on the computer to find the wind. This utilizes an equilateral triangle, putting the plane hack on course at the end of the third heading, and simplifies the hour’s dead reckoning because that hour’s run is shorter by the number of minutes on any one heading of the wind-star problem.
0241 Fix by Sirius and Venus advanced: Lat. 33°-21’ N., Long. 127°-46’ W. Wind 327°-21 knots. DA
11° left
0300 a/c 219° psc, 17° E. Var., 1° E. Dev. Desired track 226° (T)
The number of shots taken of each celestial body is a matter of discretion. In smooth air 3 shots will suffice to give a good sight. In rough air as many as 10 may be necessary for a good average. Similarly, our discretion governs whether to take 2 or 3 stars. Unlike the surface navigator who might take 3 or 4 stars, using the third and fourth as anchors to windward in the limited time the horizon is visible, the air navigator works usually with a host of celestial bodies at his beck and call throughout the night. On the rare occasions when a star sight appears “sour” he has but to go back into the cockpit with his octant for another try. Generally, the plane will be above the lower cloud level throughout the flight and only overrunning alto-stratus or cirrus clouds or frontal conditions may give him some brief hours of darkness with no stars to shoot. On these occasions he may elect to sit in the cockpit, octant in hand, to wait for a brief opening. If the weather indicates a frontal passage within a short time, he may postpone his sights, reassured by the knowledge that a clear sky awaits him a brief time hence. There are also times, naturally, when a star fix is impossible for hours on end. It is for these trips that the navigator has little appetite. His dead reckoning and radio bearings bear the navigating burden. The weather .map again plays an important part here, as he estimates wind, applying his knowledge of weather analysis and forecasting to the meteorological data available. At all times during the flight a division meteorologist is available in Alameda to radio advice and information concerning developments in the weather situation.
Three-star fixes are, without doubt, more reliable than 2-star fixes. The customary procedure, however, is to rely on 2-star fixes in good weather. For practice and to acquire confidence in his work, the inexperienced navigator is encouraged to utilize a third star. The resulting triangle can usually be covered by the rubbered end of a lead pencil and the navigator soon learns to rely on his first two stars, taking a third only when in rough weather or when he feels the need of the moral support a corroborating line can give.
0435 Fix by Polaris and Venus: Lat. 30°-45’ N., Long. 130°-29’ W. Wind 320°-27 knots, DA 14° left.
0500 a/c 253° psc, 15° E. Var., 1° E. Dev. Desired track 234° (T)
Venus at 0435 had an altitude of 19°-12’, and had about reached its lower limit for reliability. In the March, 1937, issue of the Naval Institute Proceedings (page 396), mention is made of the use of Venus for noon longitudes. This planet, when a sufficient distance from the sun, is always used by the navigators, who like to report “fixes” every two hours during the daylight flights. The moon is used similarly and in special circumstances we have been able to get three lines for a fix by these three solar system members.
The frequency with which we shoot the celestial objects offers some additional possibilities for checking track, speed, and position. With the celestial object abeam, we may check our track very nicely. This is frequently done at night with the stars, and at least once during the day with the sun, the approximate time of sight being determined by the Blue and Red Azimuth Tables. For best results, of course, the drift angle should be known within at least 1 or 2 degrees in order that the line of position obtained may be parallel to the track. A running fix of reasonable accuracy may be obtained with the sun alone at that time of the year when the declination comes close to the latitudes traveled in by the plane, and the course is nearly east or west. The sun’s travel is fast enough to assure a healthy change in azimuth within an hour’s time and allows navigator to take three lines of position to cross for his running fix. True, the altitudes are in the vicinity of 75 to 85 degrees, which may be considered quite a bit too high, but the sights seem steady enough to give accurate results. The lines, when plotted, have rarely failed to agree with each other and the dead-reckoning calculations.
In winter, of course, the azimuth change is so slow that the time interval necessary is much longer. In this case, we take a sight and plot a line about 1 hour ahead for at a time sufficiently in advance of calculated meridian passage to assure a reasonable azimuth difference). With the LHA of this sight we apply Todd’s Method find the interval to LAN. The slide rule enthusiast can work this in six operations using the formula:
Int. to Lan = (LHA X 60) / 900 + (C X G.S. / sin(90-MLat))
The third line of position is found an equal length of time after meridian passage, in order to obtain a good ground speed check. Halfway between the two lines of position on the track is the “noon fix.” The course must be nearly east or west and the meridian passage sight must be a good one, otherwise the lines of position will give an erroneous ground speed and fix. An interesting note: when checking our compass deviation by sun and Red Azimuth Tables, at sunrise, for instance, we find that we must delay shooting the sun some minutes after sunrise because our altitude anticipates surface condition by some time. Of course, H.O. 208 can still be used if the sun is beyond the scope of H.O.71.
0637 Fix by Sirius and Capella: Lat. 28°-57’ N., Long. 134°-03’ W. Wind 300°-10 knots. DA 4° left.
0700 a/c 241° psc, 15° E. Var., 1° E. Dev. Desired track 253° (T)
This fix indicated a backing wind and supported the theory that the High had moved too far to the southeastward for us to obtain advantageous winds. The subsequent 1200 February 26 map (Fig. 3) shows the movement of the high-pressure area. The course was laid for Makapuu Point. Fixes were established on schedule. In laying the course from fix to destination, we determine the Mercator course and then apply backwards the correction for radio bearings to be found inside the cover of the Radio Aids to Navigation. This gives the great circle initial course and serves until the next fix indicates a needed change.
Beginning of morning twilight is computed in the usual manner and the lightening horizon finds the navigator taking his final set of sights for the night.
1445 Fix by Antares, Arcturus, Spica: Lat. 24°-02’N., Long. 147°-44’ W. Wind 317°-11 knots. DA 5°
left
1450 Gonio bearing from U.S.S. Louisville, in Lat. 25°-03’ N., Long. 148°-14’ W. - 345° (T)
1530 a/c 246° psc, 12° E. Var., 1° E. Dev. Desired track 254° (T)
Having finished his plotting he finds the final track for Makapuu Point, studies his weather map for a guess at the average wind, and figures a course and probable arrival time. Except for half-hourly position reports, he can do little work for the next few hours. The ship is usually about 4 or 5 hours from its destination, and flying over cloud tops, making ground observations impossible. The sun will take several hours to climb into position for a longitude check. Radio bearings, meanwhile, are suffering from twilight effect. The navigator’s work is practically finished for the night. Only drift observations and DF bearings later in the morning will alter the heading in.
2000 Sighted Makapuu Point dead ahead
2017 Landed Pearl City, T.H.
2020 At bowboat, secured