The Unsinkable

“The Unsinkable” and Additional Thoughts on Watertight Subdivision in Small Yachts

In response to Eric Sorensen’s Parting Shot “The Unsinkable” in Professional BoatBuilder No. 140, Professional Engineer Christopher D. Barry brought to our attention a paper he’d written in 2005 on the subject of Sorensen’s text, namely, the possibility of designing watertight subdivisions in modest sailing and motor yachts. While reprinting the entire paper would have overburdened the magazine, thanks to the generosity of the Society of Naval Architects and Marine Engineers (SNAME), we are able to expand the discussion on the topic by reproducing it here at, after Barry’s letter to the editor.

(For a downloadable PDF of the paper, visit SNAME. Click on Buy Now and follow the prompts for a free download.)

First the Letter:

To the Editor:

I concur with Eric Sorensen’s Parting Shot “The Unsinkable” (Professional BoatBuilder No. 140), suggesting that flooding survival could be more common in recreational boats. It is actually fairly simple to have watertight subdivision in most recreational craft that are large enough not to be required to have foam flotation. In my short paper titled “Watertight

Subdivision in Sailing Yachts” for the SNAME Small Craft Sailing Panel Newsletter, Vol. 1, June 2005, I look at a couple of examples and propose an alternative method (to traditional floodable length curves) of calculating flooded flotation limits.

One tentative conclusion from this paper and a few others is that, without a significant adverse impact on the internal arrangements, as few as two bulkheads can enable a boat to survive with one main compartment flooded. It is also worth noting that ISAF [International Sailing Federation] Offshore Special Regulations for Category 0 Monohulls (transoceanic), Section 3.13, requires a watertight “crash” bulkhead within 15% of the boat length aft of the bow and two additional watertight transverse bulkheads (or permanently installed closed-cell-foam buoyancy effectively filling the forward 30% LOA of the hull), and strongly recommends that after flooding any one major subdivision, the yacht should be capable of not just floating but providing shelter and sustenance for two weeks. (The regulations are on the ISAF website

Though these studies involved sailing yachts rather than powerboats, the conclusion is probably even more applicable to powerboats, because they generally have higher freeboards and wider beams than sailing yachts, and don’t have ballast. This makes subdivision more effective. Also, most military and many commercial small craft that are basically similar to recreational boats already meet a one-compartment flotation-survival standard.

It may be worth considering watertight subdivision for larger recreational boats, especially those at increased risk of flooding due to environmental exposure or the likelihood of voyaging far away from safe refuge or rescue.

Christopher D. Barry, P.E.
Chair, Small Craft T&R Committee
Society of Naval Architects and Marine Engineers
Baltimore, Maryland

Here is Christopher Barry’s paper:


Originally presented in the SNAME Small Craft Sailing Panel Newsletter, Volume 1, June 2005. Reprinted with the permission of the Society of Naval Architects and Marine Engineers (SNAME).  Material originally appearing in SNAME publications cannot be reprinted without written permission from the  Society, 601 Pavonia Ave., Jersey City, NJ 07306

Watertight Subdivision in Sailing Yachts

Christopher D. Barry, P.E.

Watertight subdivision is a common technique for increasing safety in commercial and military ships, but is virtually unknown for yachts, especially sailingSeveral reasons for this are commonly given, the two most important of which are that watertight bulkheads would be cumbersome and divide the interior excessively, and that the calculations are difficult.  However, the proportions of most sailing yachts are such that watertight subdivision is relatively easily accomplished with few bulkheads, and a simple method for determining the weight and center of gravity limits for a given bulkhead scheme is presented.  Thus, watertight subdivision is shown to be feasible and should be considered for sailing yachts where the risk of loss by sinking is significant.
How many psychiatrists does it take to change a light bulb?  Only one, but the bulb has to want to change.

Thanks to a recent movie, most people are aware of watertight subdivision.  The Titanic had watertight bulkheads arranged so that even if any four adjacent compartments flooded from the sea, she would still remain afloat, upright, and stable — she was thus a “four compartment ship”.  Unfortunately, the notorious iceberg breached six compartments, dooming her.

Nonetheless, watertight subdivision has saved many ships, both commercial and military.  Small Navy and Coast Guard boats are required to be at least “one-compartment ships” — flooding any one compartment will not sink them – small passenger vessels are required to have watertight subdivision depending on construction and the number of persons embarked and commercial fishing boats over 79 feet must meet a one-compartment standard.  Most cargo vessels also have watertight subdivision, though current international rules use a somewhat more complicated method of describing the level of subdivision.  According to Marco Polo, even many Chinese sampans of that period had watertight bulkheads to prevent sinking.

However, watertight subdivision is almost unknown in yachts.  The reasons generally given are that the calculations are too complicated; that the bulkheads would divide the interior up too much and that watertight doors are too heavy, expensive and complicated.  Finally, relatively few yachts sink, so most builders and designers don’t think the marginal increase in safety is worth it.

Each of these points has some merit, but the potential for greatly increased safety at relatively little cost is worth at least an examination of the possibilities.

Why Not?
First, the calculations are not taught in most yacht design courses and are tedious if done entirely by hand. However, they can be readily automated with most electronic spreadsheets, and professional naval architecture computer programs (such as GHS or SHCP) include modules for all the necessary calculations, so that performing the required analyses takes only a day or so for an experienced naval architect.   It is also worth noting that yachts have relatively few loading conditions and combinations of possible damage, so many fewer cases have to be examined than for a commercial ship.

Second, the notion that watertight subdivision requires many closely spaced bulkheads is based on cargo ships carrying heavy loads.  A fully loaded cargo ship is mostly below the water, having relatively little freeboard, (distance from the deck down to the water) compared to its draft.

Floodable Length Curve for Day Boat

Figure 1 — Floodable Length Curve for Day Boat

Yachts, on the other hand, are mostly empty space.  A forty-foot sailboat might have five feet of freeboard with less than two feet of draft to the bottom of the hull.  This gives the yacht tremendous reserve buoyancy compared to its weight — the total enclosed volume below the watertight deck in a yacht might be six or seven times its displacement, so only two bulkheads are typically required.

Watertight doors on ships are pretty impressive with their multiple dogs and heavy steel construction.  But these doors are designed to a standard that allows them to be installed anywhere aboard.  This means that they must resist more than thirty feet of water head, so they have to hold a load on the order of a ton per square foot.  They are also, incidentally, required to resist a fire for sixty minutes without leaking any smoke.  A watertight door in a yacht might not have to even be watertight all the way around, since the top would generally be above the deck edge, and any fire in a small yacht that burned for an hour would create more problems than smoke passing through a door.

Finally, it is true that relatively few yachts sink due to being holed in collisions (especially with icebergs), but the number that are flooded due to failures of through-hulls or shaft tubes is much larger.  During one single month in 1985, I salvaged two sailboats in a single marina that sunk through their toilet fittings (I guess this can be called going down by the head).

Arrangement for Aluminum Cruising Yacht

Figure 2 — Arrangement for Aluminum Cruising Yacht

In the recent Hobart race, and of course, the 1979 Fastnet, boats were lost or flooded to the point of being abandoned by seas entering through the hatches and windows.  In fact, water entering from above the deck causes some 20% of recreational boat sinkings.  Finally, in recent racing yacht disasters a number of boats that appeared to be sinking in fact did not.  The crews were lost when they entered a liferaft.  There have also been numerous sinkings of smaller yachts due to knockdowns that filled the cockpit and main hatch, most fortunately without loss of life.  If these boats had had watertight subdivision, the crews would have had the confidence to stay with the boat, and might even have been able to pump out the flooding water.

Let’s look at real numbers.  In 1976, a friend and I designed a cold molded wood weekend racer for another friend in San Francisco.  The boat was 27′6″ long and displaced about 3,000 pounds.  That summer a small offshore racer had sunk off Santa Cruz with loss of two lives and another small yacht had sunk, fortunately without loss of life, in the Bay.  We decided that the boat should have positive flotation, but since it was wood, we were concerned that it would rot if spaces were foam filled.  I ran a floodable length calculation on the hull at the full load weight and came up with the curve shown in Figure 1.  The curve indicates the length of the vessel that can be flooded without sinking enough to immerse the margin line (a line set 3 inches below the deck edge) at every point along the hull.  Since the curve is based on the center of the flooded space, an isosceles triangle bounded by the two watertight bulkheads (bordered heavy dotted lines) indicates if a given length of flooding will sink the boat by whether or not the point is above the curve.  In the case shown, all three compartments result in triangles lower than the curve, so despite flooding any one of these spaces, the deck will be at least three inches above the water.  The curve assumes that 100% of the interior volume of the boat can fill with water, which is called 100% permeability.  This is a somewhat conservative calculation, as most regulations assume about 5% of the volume of berthing spaces is unavailable for flooding, resulting in 95% permeability, due to the volume occupied by furnishings and so on.  Nonetheless, the main compartment in this boat is 13 feet from bulkhead to bulkhead.  Most of space lost was in the extreme bow or under the cockpit and was useless as accommodation.  The forward bulkhead could even have been moved another foot forward, if desired.

(Note the unusual internal arrangement, incidentally.  The berths are head to head, and against the shell, and the joinery is in the middle at the forward and aft ends of the center aisle, with the aft joinery under the cockpit.  This arrangement was developed by Larry Payne, who also worked on the design, as a construction technique.  The boat was to be built as frames and joiner and other modules in a third floor apartment.  The modules would then be assembled and the deck, joinery and berths would form the strongback supporting the frames.  The hull would have then been strip-planked and covered with two layers of veneer.  Thus the owner would only need to rent a large space for a short period.)

The way the curve is high in the middle and dips down at both ends is typical of all ships.  Amidships flooding mainly causes sinkage, when the boat gets deeper in the water, but remains level, because the lost buoyancy is near the center of gravity (which is above the intact center of buoyancy).

Weight and Center Limit Plot

Figure 3 – Weight and Center Limit Plot

When an end floods, the lost buoyancy is away from the center of gravity, so there is a moment that also causes the end to trim down as well as sinkage.  Considering flooding in the center compartment, it’s easy to see that the lost buoyancy in the middle causes the whole boat to sink straight down.  The waterline rises and this adds buoyancy in the forward and aft intact compartments.  When these two compartments gain as much buoyancy as the middle compartment lost, the sinking stops.

Flooding one end, for example the forward end, not only causes downward sinking, but causes the forward end to trim down, often enough to lift up the intact compartment at the aft end, causing still more loss of buoyancy.  The intact middle compartment must not only make up the lost buoyancy, but it must produce enough bow up trimming moment to hold up the flooded end.  This trimming moment comes from the forward end of the middle compartment trimming deeper than the aft end.  This, plus possible loss in buoyancy in the aft end due to it rising out of the water, shifts the center of buoyancy forward to make up the lost buoyancy forward and thereby support the flooded end.   It is worth noting that raised forecastles and poop decks provide more volume in the ends and raise the deck edge in the ends as well, enhancing floodable length in the ends.  This is one reason why these features are often seen in small commercial vessels.

The slight upward rise of the extreme ends of the floodable length curve is caused by the fact that the ends of the boat have relatively little volume — note that the rise forward, where the bow comes to a point is larger than the rise aft with the larger transom.

The cold molded boat has relatively long ends and is light, both of which favor floodable length, but it also has very low freeboard and narrow beam, which do not.  Though very few floodable length curves have been published for sailing yachts, typical bulkhead placement of small Navy and Coast Guard boats, and small passenger vessels (which also have considerable above deck volume for their weight) suggest that a midships floodable length of about half the overall length is reasonable.  Also, at the 1975 Chesapeake Sailing Yacht Symposium, Charles Curtze, a retired Navy admiral, published a floodable length curve for his double-ended 43-foot cruiser Thule (Curtze, 1975).

The basic characteristics of the Thule curve are similar and the midships floodable length is also on the order of half the boat’s length despite having much less buoyancy in the ends. Thule has a displacement/length ratio of nearly 400, whereas the weekender is more nearly 150.  These two boats represent the extremes in characteristics for watertight subdivision, so it is probably reasonable to assume that a typical sailing yacht can sustain flooding of a space about half its length.  Motor yachts are probably a bit better, because they don’t have lots of ballast, and generally have even more freeboard and beam for their weight.

Figure 2 shows a medium/light displacement aluminum hard chine cruising yacht forty-five feet long.  It has a modern aft cabin, aft cockpit arrangement; with what is normally the owner’s cabin partly tucked under the cockpit and bridge deck.  The watertight bulkheads are again shown as bordered dotted lines and the margin line is shown as a light dotted line. The bulkheads on either side of the head would have watertight doors.  The aft cabin could be directly accessed from the cockpit, or it could have a watertight door to the main cabin.  This boat is partly a two-compartment ship — if either of the head bulkheads is breached, it still floats.

Weight and LCG Limits
Figure 3 shows the limits of displacement and longitudinal center for this compartmentation scheme.  This figure is an alternative method of showing the subdivision performance and is more appropriate than floodable length curves for smaller vessels, since it shows displacement (weight) and center limits for a given bulkhead scheme, and is easier to calculate.  It is also much easier to explain.

The figure is generated by calculating the remaining buoyancy after flooding when the boat is placed at a series of waterlines just touching the margin line.  The horizontal axis is the longitudinal center of buoyancy, (LCG), the fore/aft location where the displacement apparently acts.  The vertical axis is the displacement required to sink the vessel at that LCG so it just touches the margin line in a given flooding condition.

Draw a waterline touching a point three inches below the deck edge and parallel to it at any point.  This is one post-flooding waterline.  If you have access to a hydrostatics program that provides for flooding spaces such as SHCP or GHS, then just input this water line with the flooded space designated and get the hydrostatic properties at this point.

If you do not have such a program, use the Bonjean’s curves:  Mark the waterline on the curves and pick off the area at each station and plot it with stations horizontal and area vertical.  Mark out the flooded portion and calculate the area and center of the remaining intact curve.  This is the displacement and LCG of the buoyant volume remaining to support the flooded boat, and you can plot it above the LCG as one point on a curve like Figure 3.  Mark out any other cases of flooding and repeat the area and center calculation.  When you have calculated all the cases of interest, strike another waterline and repeat the process, sort of “rocking” the margin line on the waterline.  Obviously, the remaining buoyancy is also the limit of displacement prior to damage, and hence the limits of weight and LCG to survive flooding.

Also obviously, this process would be pretty easy to implement in a spreadsheet.    One layout for a spreadsheet approach in Excel would be to have a summary sheet that determines the waterline at each station and does the longitudinal integration, and a separate sheet for each station that contains the Bonjeans.  I have found that most Bonjean’s curves can be fit to a cubic on draft, draft squared and draft cubed, and this in turn can be derived from the Excel “linest” function.  Points for the Bonjeans can be done in the spreadsheet by the same methods as would be done by hand, entering offsets and numerically integrating.

The other points on the plot show actual combinations of displacement and LCG corresponding to given draft and trim conditions.  The open circles are displacement and LCG at level trim at the load waterline (LWL) and at 3, 6, 9 and 12 inches above (resulting in about 40,000 pounds).  The closed circles are LCG and displacement with the stern trimmed down a foot relative to the bow at the LWL and at 3 and 6 inches above it.  The X’s show the same drafts, but with the bow trimmed down a foot.

For example, with a 22-foot LCG, about 58,000 pounds is required to sink the boat with the main cabin and head flooded.  The two critical flooding cases are shown as dark lines and are aft cabin flooding and the combination of fore cabin and head flooding (i.e., the forward head door was left open).  All of the other cases are less severe.  Any condition of initial LCG and displacement that is below both of these curves will result in surviving all the cases of flooding.

Since a given LCG and displacement condition results in a specific combination of trim and draft, this plot could also be cast in terms of draft and trim or even draft at the forward end and draft at the aft end.  However, in most cases, just marking the bow, midships and stern with limiting draft marks can express these draft limits.  As long as none of the marks are submerged, the boat is safe.  The cold-molded wood boat would have had designated marks for maximum draft for safety and minimum draft to control weight for racing.

Watertight Doors
The aluminum boat would require watertight doors, but if the boat was arranged as a center cockpit, it might be possible to avoid any doors, provided it is acceptable to enter the cabins and engine space from the deck only.

Doors don’t have to be a big problem though.  First, a door on a yacht like these will only have to hold a few feet of water head at the most — about 2 psi or so at a sill five feet below the deck.  This could made of plywood or single skin fiberglass, or possibly even sheet molded plastic with some corrugations.  The relatively small pressure would not require heavy dogs — readily available quarter turn marine hatch catches are rated for a hundred pounds or so, and high gasket compression is not required at such low pressures either.  It may even be possible to arrange the door so flooding water pressure will help hold it closed.  In general, one flooding case will result in a relatively low waterline at the door and might even be below the sill.   In this case, the door could open toward the deep flooding case.

Since the tops of the doors will generally be above the deck, they only need to be splash resistant at the top.  A door could be dropped in, either as one piece or as separate slats like a conventional cockpit hatch.  The opening would get narrower towards the bottom, providing the wedging action for a gasket.  One scheme uses a wedge shaped door on a loose hinge.  To make the door tight, it is lifted slightly before closing, then pushed down and latched.   This also solves the shipboard problem of wiring, ductwork and piping.  Many of the penetrations can be above deck, in the raised portion of the deckhouse, so they don’t need to be absolutely watertight.  Yachts also simply don’t have the many piping and electrical systems of a commercial vessel.

Maintaining subdivision can also be made easier because there is a varying degree of flooding hazard depending on conditions.  Watertight doors may only need to be secured in heavy weather or at night.  Otherwise they could be left open or latched but not dogged, so long as they are ready to be closed in case of trouble.  Military craft set “damage control conditions”.  Each door and valve is designated by a letter that corresponds to one of three conditions, X, Y or Z, of increasing hazard.  Doors designated X are closed in most conditions, Y, in more severe conditions, and Z in the worst conditions.   In addition, each compartment is provided with a “Compartment Check Off List” (COCL) that describes each fiiting and designates its position in each condition.

It is worth remarking that a variant of this scheme is worthwhile even without watertight subdivision, (especially for yachts with multiple users, such as charter boats or club boats).

Designating and marking the engine cooling water valves and similar functional items might ensure that the boat is left tied up in the safest possible condition, but no one ever tries to run the engine dry or floods the sink.  (Valves involving sewage can be especially important.  Though they may not cause flooding, the results of accidents can be quite unpleasant.)

It is also worth noting that there is a small risk of capsizing as well as sinking after flooding.  The boat may lose enough buoyancy so that it is no longer stable enough to resist the overturning effect of wind and waves.  There are various standards for stability following flooding depending on the type of craft, but it would not be very useful if the boat were to float upright, then turn over, so this should be checked.

Fortunately, this is rarely a problem for ballasted sailing yachts and even motor yachts, provided there are no bulkheads running fore and aft that would cause flooding on one side only.  It is also worth noting that the same procedure as proposed for weight and center limits can provide the KM of the remaining buoyant volume.  This can then be used as a minimum to assess a limiting KG for the pre-flooding condition.

Watertight subdivision is feasible for most yacht designs.  It can be a fairly difficult retrofit, but is easy to design in from scratch.  Foam flotation is another alternative but is much more expensive and costs a great deal of otherwise useful volume.  Either alternative can greatly improve safety, not only by preventing sinking but also by giving the crew the confidence and time to stay with the boat.  It is worth considering.

The opinions expressed herein are those of the author and do not represent official policy of the Coast Guard.

Curtze, Charles A. “A Cruising Boat” Chesapeake Sailing Yacht Symposium, 1975; Marine Technology, Oct 1976; Society of Naval Architects and Marine Engineers, Jersey City, NJ

One Comment

Comments are closed.