Why the empty tank is the whole question
Run the default box — 8.5 by 5.5 by 5 feet outside, 8,500 lb of concrete, 1.5 ft of cover, groundwater a foot down — and the water covers the lid, so the tank displaces its whole 233.8 cubic feet. That is 14,586 lb of uplift. Against it stand the tank at 8,500 lb and the soil over the lid at 6,606 lb, which is 15,106 lb, and the whole thing is holding on by 520 lb. Fill the same tank with 1,000 gallons and you add 8,342 lb, and the margin goes from 520 to 8,861. The empty case has about six percent of the margin the full case does.
Nobody sizes the ballast for the full condition, and that is the point of running both. The day the hauler drives away is the day the tank is at its lightest, and in wet ground it is also, quite often, the day the water is highest.
Soil under water holds down at half weight
The cover soil in the default case is 110 lb per cubic foot above the water table and 62.6 below it — that is the saturated 125 minus the 62.4 of the water it is sitting in. Backfill that is under water is doing well under half the job you would credit it with, which is why adding cover to fix a flotation problem returns less than the multiplication suggests. A foot more soil on the default tank buys 2,927 lb rather than the 5,143 that 110 lb per cubic foot over 46.75 square feet would imply, because most of that extra foot ends up submerged.
The crossing point
The page solves for the water table depth at which uplift and hold-down are equal, and for the default tank that is 0.77 ft below grade — about nine inches. Above that the arithmetic reverses. Nine inches of standing water in a wet spring is not an exotic scenario on the kind of site where a tank gets set in the first place, and the number is worth carrying into the conversation with whoever is designing the system.
Change the shape to a horizontal cylinder and the arithmetic changes character, because a partly submerged cylinder displaces a circular segment rather than a slab. A poly tank at half depth has displaced exactly half its volume, but at a quarter of its diameter it has displaced only about 20 percent. That is why a light plastic tank in shallow groundwater behaves better than the weight difference alone suggests, right up until the water covers it and it becomes the worst case in the yard.
What the numbers cannot see
Side friction, any bond between the tank walls and the backfill, the risers and lids, and the soil outside the plan area of the tank all resist as well, and none of them are counted here. That makes the figure conservative, which is the right direction for a page that is not a design. Working against it is the excavation itself: a hole backfilled with free-draining stone in tight soil collects water faster than the ground around it drains, so the water table at the tank can be higher than the water table twenty feet away.
None of this is a verdict. Whether a tank needs an anti-flotation slab, straps, deadmen, or nothing at all is a decision that comes out of a site evaluation and a permitted design, and this page has no opinion on it. It multiplies volumes by densities.
Questions people ask
Can a septic tank really float out of the ground?
The failure is almost never a tank standing proud in the lawn. It is an inch or two of movement, enough to break the inlet and outlet seals or shear the pipes, and the symptom shows up at the house rather than in the yard. Concrete tanks move less often than plastic ones simply because they weigh more, but the balance is the same arithmetic for both and both get pumped out.
Does the soil over the lid count as hold-down?
The soil directly over the plan area of the tank does, at its own weight above the water table and at its weight minus 62.4 below it. This page counts only that prism. Soil outside the footprint would also have to be lifted by any real failure surface, and ignoring it makes the answer conservative — which is the right way for a page that is doing arithmetic rather than a design.
Why is the empty tank the governing case?
Because buoyancy is set by the outside of the tank and the hold-down includes whatever is inside it. A working tank carries thousands of pounds of liquid; a tank the day after a pump-out carries none. On the default figures the margin drops from 8,861 lb to 520 lb between the two conditions, so anything sized on the full tank is sized on the easy case.
What water table depth should I put in?
The seasonal high recorded in the site evaluation for your own ground, not what you saw in the hole on a dry day. This page takes whatever number you type and has no way to check it. If you do not have that figure, the person who did the evaluation does, and it is the input the whole answer turns on.
How much ballast do I need?
This page will not tell you, and you should be wary of one that does. Whether ballast is needed, what form it takes and what margin it is designed to carry are all part of a permitted design for your site. What the page gives you is the force balance on your own numbers, so the conversation with the designer starts from the same arithmetic.