Well, let's see what we can do...
The first stop is NS6T's Azimuth map page here: http://ns6t.net/azimuth/azimuth.html
It doesn't recognize Kelso, but entering "Fayetteville, TN" is probably
close enough for this purpose. A distance of 4000km is a good starting
That gives you a sense of what directions will be useful for radiation.
Due West (270 degrees) puts you straight through Arizona and New Mexico,
and just South of the bottom of California. Also includes Hawaii (about 275
to 280 degrees.) Meanwhile 310 degrees puts you through Montana and
Vancouver, BC, while 330 degrees gives better coverage of
the Dakotas, Saskatchewan, Alberta, and much of Alaska. So a beamwidth
of 40 to 60 degrees should cover most of what you want: any sharper and
you'll either cover a smaller portion of the area or need to rotate it the pattern.
Distances range from 1000km (Kansas, Oklahoma, Iowa), 2000km (Wyoming
and the 4-corner states) and 3000km (California, Oregon, Washington.) That
gives you an idea of what vertical radiation angles are going to be best. Most
of Alaska is in the 4000 - 6000km range, and Hawaii is 7000km.
Approximate vertical angles for various distances:
1000 miles requires 12 to 22 degrees (about the edge of Colorado).
2000 miles requires 0 to 8 degrees (about the West Coast)
Longer distances typically require multi-hop propagation or other modes,
but generally radiation angles below 20 degrees. The vertical radiation
pattern from a horizontal antenna up 1 wavelength is maximum at about
15 degrees, with a null at 30 degrees, so you're at a good height.
Now, a wire running WSW to ENE puts it about at 68 / 248 degrees. So
the requires shift from the direction of the wire is about 20 to 80 degrees.
The radiation pattern of a wire antenna 1 wavelength or more splits into
lobes and nulls, and, as you increase the length in wavelengths (by operating
at a higher frequency) the maximum will tend closer to the direction of the
antenna. This is a very promising development: an end-fed full wave wire
has a maximum about 40 degrees from the direction of the wire, so an
antenna with two full wave sections fed in the center (or an 80m dipole
operated on 20m) will have a maximum through Colorado, Utah, Nevada,
and northern California. The same antenna on 10m would have a maximum
at about 25 degrees to the wire, pointing at Los Angeles and Hawaii.
On 40m such an antenna would point broadside, giving coverage of the
Dakotas and Canada.
Note that the lobes aren't pencil-thin beams - they have a beamwidth of
several tens of degrees, so you'll work stations well on either side of the
maximum direction. But it does give you an idea of just where it is
While you nearly have room for a 160m doublet, you may find that the
directional patterns are aimed too far to the South (except on 40m, where
it would be similar to the 80m dipole used on 20m described above.) The
simple 80m doublet probably gives a better azimuth pattern on most bands
relative to your areas of interest. But if you want to be creative and try
something new you could put up two such antennas end-to-end (or use
shorter doublets to fit the space.) With two of them using identical lengths
of feedline you can switch the phase at the feedpoint to make them in
phase or out of phase, which will shift the lobes and nulls in the pattern.
I'd need to run some simulations to confirm the patterns, but that would
give you more flexibility using the same amount of wire, with just the
addition of more ladder line and a reversing switch.