Much of this work paralells Bobs ( W7PUA )paper to be published at some "unspecified" time this summer or fall (2003)in QEX. I'm trying to avoid specific details of his work, since there is some issue with publication rights... etc.
This relates my experience from my station, and what Ernie ( W7LHL ) and I have done on the shorter path this winter 2002 -2003. These observations have all been done with two foot dishes and a couple of watts.
I doubt we have unique conditions here. More likely no one has mounted a similar effort to research this.
The DSP-10 we use for the IF has a function to alternate between transmit and receive automatically in any mode. We use this function to run for hours, or days while elevated at the expected common volume. This goes on until signals build, at which time we try to peak the elevation for best return. The only variable is the weather moving through the common volume.
Both stations are very similar using an SSB transverter (the one in 4 or 5 boxes) and the Qualcomm synth at 2556. I also use a 1 Watt Q'comm PA, and Ernie has a homebrew 3 Watt PA. Ernies station is mounted inside a window of his home. He is limited to 17 degrees elevation, and has around 10 db loss through the glass. My station is all on a tripod on the roof with a Yaesu 5400 type AZ-EL. Feed line is 6 inches of 141 to the stock TVRO feed which was hogged out a little to match a WR-75 transition. We both use a Shera design GPS controller with an HP 10811 OCXO.
Using the Shera GPS controller board for a 10 MHZ reference allows full utilization of long term averaging of the received signal.This level of equipment is not required, but makes possible useful observations in a normal lifetime. This is not a phenomenom that occurs very often, but we are learning what weather indicators are coincident. If you are impatient, you might want to just drive up a mountain.
The HP/GPS reference is nice, but these HP oscillators would be adequate for this type of work without the GPS discipline, if you are sure of your frequency. Stability is important, but even if only one end of the path has stabilized equip't that would be better than none. If things are pretty stable with both ends, then give it a try. When signals are enhanced, it's just like any other contact. The thing is, you have to be ready when the conditions are there. That's what makes a home station attractive.
If you look at the WV2 or WV4 maps, and see dark green or, even better, light blue in the middle of the path, then it is very likely that there is some propagation available. SO far this has proven to be 100% accurate. If it's raining, then you might as well work rainscatter. The antenna headings need to be directed at each other. This has not worked well if the headings are not direct. Signals are quite strong, Q5 on CW. We have not tried SSB, but it should be easy on SSB as well. We tried JT44 and found it worked quite well. Easier though is to stick with CW.
Another indicator is having a ring around the moon. This is a rather temporal indicator for the Pacific Northwest given the amount of cloud cover we have.
If all you have is light green over the path, then it might be difficult to tell if anything is there. It might, or might not. It has not played well for us with the light green. A DSP-10 isn't required.You can use something like "Spectran" to spot signals below the noise. There is also the advantage of spotting the frequency before siganals are high enough to hear.
I wouldn't spend a lot of time with this unless there is at least a dark green showing on the water vapor chart over the midpoint.
It might be possible to see beacons, depends a lot on the beacon. My guess would be probably not.
It seems unlikely that there is any "knife edge" effects in our observations, since there don't appear to be any features which might cause it. However, I can't say there aren't. It just seems "knife edge" would be a constant feature. When there is no signal, there is NOTHING....
My hope is that someone might have enough interest in this to give it a try.
Mike KD7TS