10M band sounds perfect for this.
Why do you think 10m would work given the path involved?
Assuming you are trying to talk to Santa Fe from Edgewood the path is actually fairly open without
major blocking hills (as opposed to the path to Albuquerque with mountains that rise another 3000'
in the way.) In fact, I'm surprised you can't hit the repeaters on Sandia Peak and/or those on the
hills outside of Santa Fe.
10m has very limited ground wave coverage - a few miles at best. The direct wave might cover
the path if the antenna is high enough and/or enough antenna gain, but in that respect it is no
different than any of the other VHF bands (except that gain antennas are more convenient on
6m and/or 2m.) Height is important to clear the local obstructions: getting an antenna up to
30' or so will probably make a significant improvement on any of the bands. The combination of
that and a beam with a boom length of 6' to 12' or so may be sufficient to cover the path, at
least on SSB.
One way to verify this would be to take a quick drive to the top of Thunder Mountain, a quarter
mile or so to your West. See if you can hit any repeaters with a good mobile antenna from up
on Snowflake Trail that you can't hit from your house. One reason I'm recommending repeaters
here is that they will be elevated at the far end, which makes the path much easier.
FM is a strong signal mode - it doesn't work well for weak/marginal signals. SSB, CW and some
of the digital modes are much better when signals are weak.
One thing to remember with NVIS propagation is that conditions change throughout the day.
At this point in the sunspot cycle it may be that 80m is better at night and 40m during the day,
though here in Oregon at the bottom of the recent cycles we have needed to use 80m during
the day and 160m at night to maintain coverage. You can run forecasts for the expected
coverage areas for each band using the propagation tools on the Australian Ionospheric Propagation
Service here: http://www.ips.gov.au/HF_Systems/7/1
. The "LAMP" charts ("Local Area Mobile
Prediction") are the most convenient ones I've found for forecasting NVIS coverage out to
several hundred miles. (And the tools have an option that has the ham bands already built in.)