5m in height could make a difference in very still conditions, but I think the shelter of the trees is the main factor here.

Pete could very well be right here. Anything that interferes with the radiation of heat from the ground to the sky is a real killer for frost.

This said, let me point out that 5 m is more than an adequate difference in height for the boundary between frost and no frost, particularly when the situation is not particularly cold but very quiet, very clear, and with a valley or basin in which cold air can form a pool. The surface of that pool can be well under a metre thick. I used to live on a farm on a hillside and cycle in to university of a morning. I had been well aware that on frosty mornings the air was much colder at the bottom of the hill than up at the house, where the air was merely cool. The first time I cycled downhill into the valley in suitable weather in winter, I got a considerable shock halfway down. My face hit a layer of cold air so abruptly that I responded with the diving reflex as if I had jumped into cold water. For weeks on end this would happen day after day, and if you were alert you could walk into the pool of cold air and tell when it had hit your face. The difference was nothing like 5 m, nor yet one.

And how is it possible? Well, if the air forms a cold, undisturbed pool, then it makes sense that the surface might be well defined.

Partly this is a microclimate thing similar to katabatic winds.

Hot air rises, cold air falls.

Air is too transparent to cool itself by radiation. The main effect is that roads and vegetation cool by radiation into the night sky, and they chill a thin layer of air close to the ground by conduction. On level ground, this is seen as ground frost.

On a slope, this cool air slides away to find the lowest level, in the same direction that water would, and is replaced by air from above that has not yet been chilled. So the effect is actually that the air chill on a slope brings in a current of (slightly) warmer air.