Knowledge / Navigation and collision

Navigation in restricted visibility: detection, assessment and safe options

Understand restricted-visibility rule context, radar evidence, time margins and uncertainty through a bounded educational scenario.

On this page

Restricted visibility changes the evidence available to a bridge team and the rule context for vessels that are not in sight of one another. A well-defined target symbol does not restore visual information that is absent. The learning task is to connect detection, uncertainty, available response time and the applicable rules without treating a display as a complete account of the surroundings. This article is a public educational explanation, not a fog-navigation checklist, a maneuver recommendation or a substitute for qualified watchkeeping and current official publications.

Identify the visibility context accurately

Restricted visibility is not limited to fog. The COLREG definition includes conditions such as mist, falling snow and heavy rainstorms. Separately, whether vessels are in sight concerns visual observation, not whether an electronic system displays an identity. These are different questions and should be recorded separately in a training case. A visibility estimate based on one nearby object may not describe conditions across the entire encounter area.

The MCA’s MGN 369 Amendment 1 explains this framework and the continued relevance of rules applying in all visibility conditions. It is UK official guidance. The actual international provisions, applicable local rules and vessel procedures remain controlling; a teaching article does not replace them.

Keep Rule 19 distinct from in-sight encounter rules

International Rule 19 applies to vessels not in sight of one another when navigating in or near restricted visibility. The familiar in-sight crossing labels cannot simply be assigned from radar geometry alone. Read Rule 19 together with the general and all-visibility provisions, rather than isolating one sentence about a radar target. The USCG compilation marks International and US Inland text separately, which must not be silently combined.

For study, compare two fictional scenes with identical plotted tracks. In one, the vessels are visually in sight; in the other, they are not and are in restricted visibility. The geometric calculation may be unchanged, while the legal reasoning differs. This is why a simulator must preserve visibility and vessel-status assumptions as explicit inputs. A label such as “crossing” can describe geometry without establishing the complete rule-based responsibility.

Detection and reliable tracking are separate achievements

A radar echo first indicates that returned energy has been detected. Establishing a reliable track requires additional observations and processing. Target association, clutter, shadow sectors, own-ship inputs and a recent maneuver can affect interpretation. A newly acquired vector should not be treated as though it had the same history as a stable, well-observed track. Its displayed precision is not a record of its actual uncertainty.

The MCA’s electronic-aids guidance discusses limitations and effective use of radar and plotting aids. It does not justify a universal acquisition-time allowance for every device and condition. The engineering lesson is to keep detection time, track-establishment time and decision time distinct. A system can detect something while still having insufficient evidence for a confident motion estimate.

An original time-margin exercise

Imagine a classroom scene in which separation is 2.4 nautical miles and the assumed constant range-closing rate is 18 knots. The time to zero separation under that one-dimensional assumption is 2.4/18 hour, or eight minutes. This is not TCPA for a general two-dimensional encounter and is not a prediction of unavoidable collision. It deliberately isolates the relationship between distance, closure and elapsed time.

Suppose two minutes pass while the fictional team establishes and discusses the available information. At the assumed closure, 0.6 nautical mile has been consumed, leaving 1.8 nautical miles and six minutes in the same simplified model. No part of that six minutes is automatically available for a successful maneuver: execution, vessel response and verification take time too. The example recommends no speed, distance or waiting period. It shows why information delay belongs in the assessment.

Safe speed cannot be recovered from one number

A speed considered in isolation omits visibility, traffic, maneuvering characteristics, background conditions, hazards and radar limitations. A vessel may need time to detect, understand and respond to objects that have not yet produced a confident track. A reassuring prediction for the current target set does not establish that the set is complete. The absence of a plotted target is not equivalent to a verified empty waterway.

The USCG navigation-rules FAQ explains that there is no universal numerical safe-speed or passing-distance answer. In a learning exercise, ask which assumptions support the selected speed and which observation would invalidate them. The exercise should also identify maneuvering data and procedures that would be needed for a real assessment, rather than inventing a general stopping-distance rule from the classroom arithmetic.

Understand information quality without tuning by guesswork

Radar range scale and display settings affect what is visible and how a scene is interpreted. A setting useful for one purpose may obscure another feature. Clutter reduction is not a proof that every real echo remains visible. The appropriate use and checking of the installed equipment belongs to its manual, training and procedures. This article does not prescribe gain, clutter or alarm settings.

A useful after-action review retains the observation conditions and the evidence actually available at each moment. It distinguishes a target that was never detected from one that was detected but not tracked, a track that was misunderstood and an understood situation that did not receive an effective response. These different failures can lead to superficially similar outcomes but require different learning and corrective actions.

Sound, lights and non-electronic observation still matter

Reduced visual range does not remove sight and hearing from the lookout function. Lights, sound signals and other observations remain governed by the applicable rules and circumstances. Their interpretation has limitations too: the direction, identity or distance of a sound may be uncertain. An incomplete observation should be recorded as incomplete, rather than upgraded to a precise target report because the team needs a simple answer.

In a fictional debrief, separate “a sound was heard” from “the sound was confidently associated with the tracked vessel.” The second statement requires evidence beyond the first. The same distinction applies to a light appearing briefly in variable visibility. Combining observations is useful when their timing, direction and identity are examined; merely merging all information into one symbol can hide unresolved disagreement.

AIS and radio do not remove the rule obligations

AIS can provide identity and motion information, but its presence is not guaranteed for every object and its data can be incomplete or wrong. A radio conversation also consumes attention and can introduce ambiguity about the vessel or intended meaning. The MCA’s MGN 324 Amendment 2 warns about these limitations and the danger of using VHF discussion in place of compliance with COLREG.

An analytical record should distinguish message sent, correct recipient established, meaning understood and action observed. These are separate milestones. A confident voice and a familiar vessel name do not prove that both teams are discussing the same encounter. The lesson is not to avoid all communication; it is to preserve the difference between added information and demonstrated resolution of the navigational problem.

Review the whole scene after a change

Any own-ship or target change alters the evidence for future motion. A prediction made before the change is not automatically valid afterwards. In a multi-target scene, improving one geometric relationship can worsen another. Shorelines, channel limits and shallow water remain relevant even when the collision plot focuses on ships. A display organized around the selected target should not define the entire scope of the assessment.

For a training sequence, record the expected effect of an action and then compare it with subsequent observations. Keep prediction and realized track separate. A favorable outcome at the end of the exercise does not prove that the earlier reasoning was sound; chance or unmodeled action may have contributed. Conversely, a carefully reasoned assessment can be overtaken by new information and must remain open to revision.

Design a debrief that tests understanding

Ask learners to reconstruct the information timeline before seeing the complete replay. What was observed directly? What was inferred? When did the visibility context change? Which rule section applied to each relevant vessel relationship? Which measurements were stale or unconfirmed? This prevents hindsight from replacing the uncertainty that existed during the scenario.

Then review the physical and human response chain. Were assumptions about machinery readiness, steering response, bridge roles and communication explicit? Did the team verify outcomes rather than treating a command as completion? These questions support learning without assigning a generic maneuver or operational threshold. Any real emergency response must follow the applicable rules, the full situation and the vessel’s approved arrangements.

Limits and common errors

Common errors include treating a radar symbol as visual sighting, importing in-sight priorities without checking applicability, waiting for perfect information, assuming all targets carry AIS and accepting one favorable CPA as proof of safety. Another is describing a successful simulated encounter as evidence of certified navigation capability. A simulation only establishes what its model, inputs and tested conditions support.

A useful educational output is a traceable explanation of context, evidence, uncertainty and reassessment. It should name the limits of the example and the official provisions that need fuller study, including relevant signals and all conditions within Rule 19. This article deliberately does not extract a turn direction or speed command for a live vessel from an incomplete scenario.

Sources