The purpose of this paper is to present data and draw conclusions on the chances of collision in different areas of the waters around the UK. The subject of risk analysis has recently grown in importance, and it is often necessary for an operator to demonstrate that a proposed operation is acceptably safe before he will be given permission to begin. Risk estimates are generally of two basic types, backward and forward. Backward estimates are based upon the number of accidents that have occurred in a particular situation, and if there have been many there will be a fair degree of confidence that the risk has been accurately estimated. Forward estimates have to be made where there are so few actual accidents that no backward estimates are possible: they are based upon an analysis of the situation into a sequence of events, the probability of each being known by comparison with data obtained from other situations. Although these data may themselves be reliable, the final risk estimates will depend upon the comparability of the situations and also upon whether the sequence of events has been correctly identified.

In areas of high marine traffic density it has been necessary for IMCO to introduce traffic separation schemes to reduce the total number of encounters, and in particular the dangerous head-on encounters. It is now widely acknowledged that separation schemes can and do improve safety. However, as a side effect they unfortunately increase the number of overtaking encounters. For example the introduction of a separation scheme in the Dover Strait (with 1 per cent of ‘through rogues’) has increased threefold the proportion of potential overtaking encounters.

Of all forms of transport, ships are unique when it comes to avoiding one another. Compared with both land vehicles and aircraft they have remarkable freedom of action, so that collisions are avoided less by external regulation than by mariners exercising their individual skill. A framework for their actions is provided by the collision regulations which have evolved over hundreds of years, a process which seems unlikely to reach an end in the near future. Since it is reasonable that any changes should be informed by events, information on what mariners actually do to avoid one another can only contribute to the safety of ships at sea.

This paper is concerned with obtaining information on collision avoidance from radar records. It describes the data extraction and the selection of encounters between vessels, and presents some sample statistical results. Finally, probable applications of the information are discussed.

The presence of any structure in the sea is a potential hazard to shipping navigating in its vicinity. It is therefore of considerable interest both to mariners and to engineers placing structures such as oil platforms in the sea to try to assess the attendant risk of collision, with such a structure.

We have over a period of time attempted several methods of assessing these risks and it was felt important to link this with our work on marine traffic flows, as experience of the traffic patterns in the southern North Sea was a useful input to the study. Hence we considered an area in the southern North Sea, to the eastwards of the outer Thames Estuary as far as the Middlebank and between the parallels of 51° 08´N and 52° 15´N, since we wished to look at the offshore situation; the area is nowhere less than 3 miles from the coast.

The memory of the Amoco Cadiz disaster of 1978 is still fresh in the minds of most people. The loaded VLCC drifted on to rocks 3 miles off l'Abervrach and caused almost a quarter of a million tons of crude oil to engulf the Brittany coast. Almost immediately some far-reaching measures were put into effect by the French authorities through IMCO to lessen the likelihood of a similar incident occurring again. Perhaps the most important of these were the fundamental alterations which were made to the traffic separation schemes off Ushant and the Casquets to route the large oil carriers well away from the French coast.

When in 1969 the specific task of scheduling shipping movements in and out of a tide-restricted port had been given as an exercise to a group of harbourmasters attending a symposium at UWIST, Cardiff, the varying methods employed and the apparently large workload involved indicated that a formal study of the problem would produce a more rational solution. The early part of the study, by K. D. Jones with the help of his assistant B. J. Moody, was to develop suitable computer programs to automate a scheduling process such that an ideal pre-movement plan could be arrived at, using as criteria the earliest movement times for given ship-draught and available water-depth.

Data relating to a total of 74 tanker casualties in the Baltic area during 1971–5 have been analysed and compared with the earlier collected data on 268 tanker casualties. The discussion concerns local and seasonal distribution. The statistics also cover the nationality, age and size of ship involved. Emphasis is laid upon the need for a more efficient casualty reporting system and new methods of data collection are presented.

The Baltic, Gulf of Bothnia and Gulf of Finland form a limited sea-water basin, where the change of water is small and the evidence of man marked. One of the worst threats to the sea water is the oil leaked from ships. This oil gets into the water in ship casualties or by accidental or deliberate discharge of spill oil into the sea. The oil amounts leaking into the water are very large in ship accidents if the ship or ships in question are tankers in loaded conditions and tanks are damaged.

Certain measures now before IMCO require all vessels of particular classes to carry automatic radar plotting devices with a specified level of performance. Tankers over 10 000 tons trading to the United States will be required to carry such equipment from 1982 and the IMCO requirements will apply progressively to different categories of vessel over 10 000 tons from 1984–89.

The Technical Committee met in open session in London on 16 January 1980 to discuss the implication of these measures and some of the contributions to the discussion are printed below.

It is usually accepted that the visual estimation of miss distances at sea is difficult and that it is advisable to employ plotting methods with visual observations (or radar echoes) to assess the positions and movements of neighbouring vessels. Nevertheless, direct visual assessment is likely to remain a useful adjunct to more sophisticated methods, at least under conditions of good visibility, particularly as the use of radar does not absolve the mariner from keeping a good look-out at all times. It seemed, therefore, useful to examine some of the factors involved in the visual estimation of miss distance, in the hope that such an analysis might suggest ways in which the direct assessment of close-quarters situations could be improved.

The steering and sailing rules are changing with the development of available sources of information. This latest development is of special interest, as radar navigation has prompted a return to the principles of the ‘Larboard Helm Rule’ -if the compass bearing of an approaching vessel does not appreciably change, each of the vessels involved shall alter course so as to place herself to the right of the line of compass bearing o t the approaching vessel’.

This rule offered simple sailing instructions: each party to an encounter was required to take action. There was no stand-on vessel and no right of way. Our modern steering and sailing rules are based on the right of way of the vessel on the starboard side and this gives rise to difficulties unknown in the days of the larboard helm rule:

(i) There are two vessels on collision course, one large and slow-acting, the other small and manoeuvreable; the smaller vessel can easily take avoiding action. As, however, she is approaching from starboard she is the stand-on vessel. The large vessel, though restricted in her ability to manoeuvre is thus obliged to take avoiding action.

(ii) The officer on the bridge of the stand-on vessel recognizes risk of collision. The navigator on the bridge of the burdened vessel is distracted from his duty of keeping a proper look-out. Our existing rules require avoiding action on the part of the vessel that has not recognized the risk of collision.

A recent oceanographic cruise of the Spanish research vessel, B. O. Cornide de Saavedra, provided a suitable platform from which to survey ship traffic through the strait of Gibraltar covering several days. This was long enough for a representative sample to be taken, 472 ships. It was observed that only 65 per cent of the ships passing through the strait observed the traffic lanes. The current structure of the Alboran Sea may offer economic advantage to a different traffic convention. The present mean traffic is 5·8 ships per hour which represents a density greater than that of the Strait of Dover. This suggests that traffic regulation, in the interests of preventing loss of life and oil spills, may be necessary in the near future, particularly when the Suez Canal has been opened to larger tankers.

Sailing Directions nowadays implies ‘Pilots’, with which all mariners are familiar. However, the term in a broader sense embraces a wide range of printed and descriptive material of help to the seaman. In this paper, which was presented at a meeting of the Solent Branch of the Institute in December 1979, Dr Cotter traces the origins of the publications – pilot books, light lists, tide tables, and indeed charts – which are now available to the navigator. The paper has been abridged.

Men sailed the seas and made extended voyages, without benefit of instruments now considered vitally necessary, thousands of years before the invention of chart and compass, the pre-eminent instruments of navigation. To the question: ‘How did an ancient sea-trader shape his course in the confident expectation of reaching his destination safely?’ there is no complete answer. But we can be fairly certain that early seafarers learnt the techniques of their craft through observation and experience supplemented by verbal instruction and advice from their more-experienced contemporaries.

In 1873 Marcq St Hilaire introduced the concept of the intercept method of constructing a position line, with the use of an assumed position chosen to facilitate calculation or tabulation. This same basic principle is here used to facilitate direct calculation without the use of logarithms or calculator; a simple table of a natural trigonometrical function is required and, preferably, some aid to multiplication. Apparently the method has not been proposed before, and it is ironic that it should be considered when the availability of calculators has essentially removed any requirement for it! It is put forward mainly for theoretical interest, but there may be a place for it for emergency use when (for some reason) neither conventional tables nor a calculator is available.

For short range air navigation the most general method of determining the distance between two points, and the course to be steered from one to the other, is by means of a grid system of coordinates based upon some map projection. Problems must arise, however, in routing aircraft between Traffic Control centres in adjacent regions which employ different grid reference systems, even though the airway markers are identified by their latitude and longitude. A difficulty which may result in steering-angle errors arises when these markers, which are generally VOR stations, are referenced to the local magnetic field rather than the geographic pole.

Animal navigation is magic. Nobody really knows how they do it. We can only suggest the senses they must be using, but how their brains convert this information into action is a mystery. In their heads are computers infinitely smaller, lighter and more flexible than the fantastic silicon chips that clumsy humans employ. Let us first consider what animals have to measure and, to make sure we cover the whole field, we shall look first at what giant tankers, nuclear submarines, Concordes and Apollo space-craft require. [Wing Commander Anderson's article was presented at the Young Navigators’ Meeting in London on 18 December 1979.]

In a recent paper the author reported a distribution theory which rationalizes some fairly well established phenomena in navigational errors. Adequacy of the implied model and other existing models was investigated in the light of empirical evidence. A summary of the findings has been available in a previous issue of this Journal.

This paper provides further insights into several theoretical models, establishes formal procedures for the estimation of parameters crucial to the evaluation and design of navigational systems and investigates the potentials of a broader system of distribution models. In addition, some simplified results aiming at practical use for navigators and air traffic control authorities are provided. A real-world application of the suggested methods and theoretical results is also demonstrated.

The major purpose of an air traffic control system is to ensure the separation of two or more aircraft flying in the same airspace, with an efficiency that can be expressed in terms of capacity and cost. As air traffic grows in numbers it becomes necessary to reduce the workload of the controllers by relieving them of many monitoring tasks, and eventually some decision-making tasks, through computerized automation. In this context many developments tend to build up an efficient conflict-alert subsystem.

The problem of conflict-alert in the air needs strategic tools, to make collision unlikely or even impossible, and tactical tools to detect impending collisions. The latter detect potentially hazardous aircraft encounters and alert the controller in time to warn the pilots (if necessary) and should obviously provide this capability with a minimal number of false alarms and no increase in workload.

This paper, which was presented at an Ordinary Meeting of the Institute in London on 21 November 1979, with S. Ratcliffe in the Chair, traces the evolution of airborne doppler navigation systems over the past two decades and discusses the treatment of errors introduced when flying over water. Both authors are with Decca Radar Ltd.

The object of this paper is to describe briefly the major developments which have taken place in doppler navigation systems during the past 15 years or so, to indicate the current state of the art and to examine in some detail the behaviour of doppler systems when flying over water. The year 1960 is taken as a starting point since by that time commercial doppler systems were established. A typical general purpose system of that generation (Fig. 1) consisted of the following units:

1. An antenna, probably mechanically stabilized.

2. A transmitter/receiver using either pulse or FMCW modulation.

3. A tracking system.

4. An analog computer.

5. A display of ground-speed and drift-angle.

6. A display of present position.

When airline fuel is calculated before flight, ‘contingency fuel’ has been included either separately or in combination in various forms over the years. The primary calculation has always been the fuel required in expected conditions to fly to the destination and thence to a nominated alternate. The next is the provision, in the words of the Air Navigation Order, of a ‘safe margin of fuel’ for contingencies. For aircraft operated under an Air Operator's Certificate, CAP 360 specifies ‘an allowance for contingencies such as errors in forecast winds and temperatures, navigational errors and ATC restrictions on altitude and route’. Also, ‘45 minutes holding reserve’ is stated to be a normally overriding requirement. The wording of CAP 360 is, perhaps intentionally, not precise but is broadly in line with Annex 6 of the ICAO Chicago Convention which requires:

(i) a reserve ‘to provide for contingencies’ plus, for turbojet aircraft,

(ii) 30 minutes' low level holding, plus

(iii) an amount ‘to provide for the increased consumption on the occurrence of any of the potential contingencies specified…’