Note: This blog post has become a more complete page on the subject of thin ice. Click here to go to the new page.
There are two aspects of deciding if the ice is barely thick enough to support a person on foot.
1) Is the ice under your feet strong enough?
2) How confident to you want to be that all the other ice that you are going to walk on will be strong enough?
Human Ice Tester (the author) lounging in an area broken up by testing. With 3 layers of long underwear inside the dry suit and a bit of exercise, it is comfortable in the water even with 20 deg air temperature.
Based on a few experiments on thin black ice over shallow water, 175 lbs will probably break through 1.1" (28 mm) but probably won't break through at 1.2". At 1.1", it looks like you should be able to get back on the ice in a prone position. If it is much thinner you can't! If you are on skis, a light iceboat or even skates you can get far out on ice that will be too thin to get back on.
The second question, about the acceptable level of risk of falling through, is a little harder to answer. Ice thickens quickly when it is thin so there are likely to be local differences in when it caught and how fast it grew, resulting in variable thickness. Thin ice often has very few cracks to give an indication of thickness. Thin ice is dark so there is not much contrast between ice that is thick enough and ice that is not. Also, white splash-out rims may not have developed around any holes. This makes recently skimmed over holes especially hard to spot. In general, ice thinner than two inches is more likely to have these issues and even much thicker ice often has areas of open water or thin ice.
Ice Testing: A test pole is the best way for a person on foot to test ice thickness for ice less than three inches thick (nordic skating poles, ice chisel, wood test pole). A pole is better than a drill as you can assess ice quickly and easily making it more likely to be used enough. Axes are popular in some areas. They are effective but are not as fast as a pole and they have a shorter handle. To properly measure ice thickness requires a tape or other scale. 1/8" accuracy matters. You can't get a good enough measurement by eye. You can measure at wet cracks, or recently refrozen ones opened by your test pole, fishing holes, a hole you drill or by knocking a piece of the plate out with your test pole. A single hole made with a test pole usually has a cone knocked out of the bottom of the ice sheet, making it hard to use for making an accurate measurement.
Cracking: A single crack that forms as you move over the ice may or may not mean it is too thin. Single cracks that follow you can occur in fairly thick ice that is under some stress. None the less, any cracking that appears to be related to you is cause to verify the ice thickness. If you are starting to see or hear cross cracks (zipper cracks) or to see double cracks going in your direction, you are probably pushing the envelope. This is a great time to have on a dry suit, a life jacket and a rope between you and a buddy on thick ice nearby (or, even better, to be doing something else while you wait for the ice to get thicker).
Shallow Water: The notion of staying over shallow water is reasonable but you may get just as wet if you break through. If you are on something that spreads your weight enough to get you far out on thin ice before you break through, you could still be in big trouble even though it is not over your head.
Reasonable Thickness: On ponds we know well, with full equipment (claws, test poles, ropes, clothing, floatation, etc), we will skate, with some trepidation, on subfreezing 2+ inch black ice and sail DN's on 3". So far we have had relatively few problems at that thickness. Over the years, we have had plenty of swims in the thinner areas that can often be found somewhere. On bigger ice or with less experience or less equipment or less risk tolerance, adding an inch or two or more to these numbers is prudent.
For a range of activities the Minnesota DNR recommendations for good black ice at temperatures below freezing are reasonably conservative. The strength of thawed ice is significantly less than when it is below freezing. Without a significant warm wind, the thawing process takes place mostly internally at the crystal boundaries so the thickness may not change much as it thaws and gets weaker. Ice less than a few inches thick thaws quickly: a couple hours might be enough. Even early in the thaw process it can lose 50% or more of its strength. In thick, highly thawed, small grain ice we have seen situations where it has lost 99% of its original strength. Snow ice gets particularly weak in thaw conditions as it turns to mush.
Click here for a Swedish video of part of the standard training procedure for skaters: finding out what it is really like to fall through thin ice. There are a bunch of these videos. They give a decent idea of what it is like to fall in at low speed, how getting back on the ice goes, etc.
Bob