I have given this some thought and I want to check my reasoning or see if there is an alternative explanation.  Was the water in the pond so still that as the first ice froze it acted as a nucleus for the next layer which froze underneath and in doing so lifted the first ice up out of the water.  Obviously, there would have to be no convection present to even out the freezing process.

I live in New England and park my car outside, so I routinely need to scrape ice off the glass. Some mornings the ice peels off smoothly in large sheets and needs little effort. On others it is bonded very tightly to the glass and I need to exert considerable force to remove even a thin sliver. I i have kept track of the temperature under both extremes, and there is no correlation: very cold and very mild (though still freezing) mornings can give either type of response. Can someone suggest what IS responsible for the different binding patterns?