With antenna simulation software we can estimate the possible performance of antennas down to fractions of a dB - but only under the assumption of perfectly flat terrain. While still looking for such a QTH I got interested in the influence of the real world terrain on our antennas. It seems to be a terrain with not too much guru wisdom but nevertheless a lot of dB to gain or loose.

One of the most helpful tools is the HFTA-software by N6BV, supplemented with the ARRL antenna handbook, analyzing the effects of antenna height, terrain contour and ground properties for a given band. It helps to visualize the basic effects of slopes which can roughly be described as: The steeper a slope the more (virtual) height it adds to an antenna - thus resulting in more low angle radiation.

More information about this software and the ways to feed the topografic data can be found in the manual (.pdf) on the ARRL-site or to be googled depending on your map software.

The graph indicates the calculated gain in dBi at takeoff-angles between 1 and 34 degrees showing strong lobes as well as nulls for some take-off agles. On this graph we have a dipole at 25 ft/8m. The green line is for flat terrain in a specific direction. The blue line is the southeast direction on one of my favourite hills along a smooth slope and the red line down a rather steep slope: more than 10 dB advantage for the hilltop location against flat terrain at the important very low angle.

But: the height advantage can´t be generalized because from some height on the lower lobe decreases and additional lobes with higher angle appear. So here we go with the difference between the steep hillslope (red) and a 90 ft (30m) high cliff (blue).

As instructive the HFTA-graphs look there is one real world thing to consider: its calculations base on straight two-dimensional lines. But the radiated wave and the surface of slopes is three-dimensional being plain, concave or convex. In his antenna-book Les Moxon, G6XN, describes interesting experiences: "...very good results can be expected, particularly if the reflecting area is bowl-shaped so that it acts like a concave mirror to focus the signal in the desired direction. Some relatively poor results (including total failures) have been attributed to convex ground which disperses the wave..." (HF Antennas for all Locations, 2nd ed., p.167).

HFTA can also calculate the effect of the distance between antenna and edge of the slope. BUT: it has one restriction because it can only calculate antennas with horizontal polarization. For verticals there is very little information available about the influence of terrain.

Mostly we have to depend on experiences and guru wisdom - of course resulting in different opinions. Some of the experiences are summarized in this .rtf-file. It seems that verticals don´t like steep hills and seem to deteriorate on cliffs. Still open is the question from what slope angle on verticals lose their important ground under the feet. Moxon (quoted below) argues that from slope angles steeper than 3-5 degrees the horizontal antennas win. ON4UN shows a graph in the fourth edition of his book stating still a 2 dB advantage for a vertical at a 8 degree downslope over a vertical in a plain at 10 degree elevation angle - and an even bigger advantage up to 7 dB at 5 degree takeoff angle (Low-Band DXing, pg. 9-5).  

He shows also a graph indicating that a 80m-dipole at 30m/100ft takes more advantage of an 8-degree-slope than a vertical at the same slope (page 5-6).

Besides the software ON4UN uses there is "Terrain Analyzer" (TA) developed by K6STI capable of calculating terrain influence on antenna behaviour. But this program is rarely available since K6STI is no longer active in ham software. At least I got the help by Peter, DJ2ZS (SK), who tried some calculations with TA. There is one important remark by K6STI who stated in the manual that the absolute gain figures are not "valid" for long verticals close to the ground. The software uses a single point source as a simulation of the antenna. It should work with antennas being small compared to their height, i.e. a 20m-quad on a regular tower. But that shouldn´t refrain from comparing the sheer terrain-influence with one given "single-point"-antenna. 

Here are figures for a 7 MHz-vertical close to the ground. On flat terrain it shows about 1 dB more than the same vertical calculated with an EZNEC-software. This may be due to the single-point-restriction. But now to the figures:

Gray = flat terrain
Green = smooth slope (five degr.)
Blue = initial steep slope with 13 degrees (8 degr. average)
Red = Cliff (100ft/30m)

More to follow, especially comparisons for one terrain with differing frequencies and terrains with "optimized" straight slopes and defined angles.

Those having access to scientific librairies may look for an article mentioned by K6STI. It should be about aircraft-measured comparisons between verticals and dipoles in hilly terrain:

Modeling and measurement of HF antenna skywave radiation patterns in
irregular terrain
Breakall, J. K.; Young, J. S.; Hagn, G. H.; Adler, R. W.; Faust,
D. L.; Werner, D. H.
in: IEEE Transactions on Antennas and Propagation (ISSN 0018-926X),
vol. 42, no. 7, p. 936-945
Publication Date: 07/1994

If you want to read more about the basics of reflections in the foreground of an antenna and the influence of antenna-height and terrain I recommend an online-article by Palle, OZ1RH.

Any comments, corrections, experiences and additions are welcome and will be added here from time to time.