John -
Seems that the basic goal is to balance both the 1) peak and 2) annual loading on your system more, by finding more heat sinks in both of those timeframes. I think DHW is a great suggestion, as is any measure that lowers your cooling load (economizer, data center efficiency measures, ice storage, higher heat pump efficiency). Adding heating would be the opposite approach - is there any building or process nearby that could benefit from some free heat (multifamily is perfect...)? Beyond that, yes you could consider any type of hybrid geothermal... There are of course many options for supplemental heat rejection, and snow melt is just one of them. If you stick with a hybrid with snow melt as the supplemental device, the thesis by Ramamoorthy (http://www.hvac.okstate.edu/research/Documents/Ramamoorthy_Thesis.pdf ) covers design and control of that particular configuration. And also, if you have access, there was a great presentation by Ed Lohrenz of Geo-Xergy at this summer's ASHRAE Montreal conference that discussed a hybrid with snow melt, I think in Alberta.

Of course, you want to take care in not making the system too complex - more than one supplemental heat sink added on and the system may get tough to control. Controls of basic cooling-dominated hybrid approaches are covered in a couple recent papers: http://www.hvac.okstate.edu/research/Documents/Theses%20and%20reports/Xu_Thesis_2007.pdf and www.ecw.org/hybrid . These don't deal with snow melt systems though, so they won't answer the question about placement of the control sensor in a snow melt system. Seems to me that what Nick alluded to, putting the sensor in the asphalt or concrete itself, would make the most sense.

Good luck!

Scott Hackel, P.E., LEED AP