there are some studies that have been linked elsewhere on this site that talk about how sand of a certain size (16grit) can be a barrier to termites because it is too big for the termites to easily haul away, and not big enough to support tunnel structures spontaneously on its own. This technique is supposedly used in surrounding foundations, or other areas where wood is in contact with soil. I forget the suggested thickness, but basically since you would want these areas to be as dry as possible anyway to prevent rotting, you would have to engineer a dry space via hill slow, ditches, french drains, or what have you. then surround your buried wood with this 16 grit sand , and the termites, according to this study, wouldn't be able to get through. i will try to find the link that i saw earlier and repost it on this thread.
so my thought was that if you mixed some diatomaceous earth with this sand barrier, it would be an added safeguard. so if the termites do find a weak spot where they could haul away sand or get it to form tunnels, they would also find DE in that spot, further limiting their expansion. And with regard to the wetness issue, as i mentioned before, you are going to want these areas to be as dry as possible anyway, so hopefully there will be limited moisture to mess with them. also since the sand would naturally drain well, it ought to dry out quickly.
i dont know if becoming wet then re-drying affects the effectiveness of DE but it seems to me since the stuff is formed by little ocean creatures in the first place, that even if water while present limits their effectiveness, once the water is gone (dried up) it shouldnt have permanently altered their structure. (of course the possibility exists that due to their microscopic structure, and water's incredible attractive strength at such scale that once wet it may take a very long time for the individual diatoms to become dry, even if the surrounding environment is. This is only speculation, i have no expertise in this matter, just trying to think of all angles)
So anyway, I have not tried this, nor have I tested the sand barrier method alone, simply read about it. However it just seemed to me that it might be a good fit. All just theoretical, any refinement would be welcome. and if anyone is in a position to test this theory, i would love to hear how it worked out, because alas i am not currently able to test it myself.