A ground-mounted vertical is not simply a radiating pole connected to earth. It is a monopole whose return path and near field are engineered by the ground system. When that system is weak, RF current spreads through lossy soil close to the feedpoint and a material share of transmitter power becomes heat. When it is strong, the return current is carried largely by conductors and the vertical approaches the performance expected from its geometry. The practical conclusion is straightforward: for a ground-mounted vertical, improving the radial field often yields more radiated power than changing the radiator itself.
This study compares five useful conditions: poor, moderate, and good terrestrial installations; a salt-water site; and an elevated-radial vertical. It separates three effects that are often conflated: near-field dissipation in earth, far-field ground-wave propagation, and ordinary conductor/matching losses. The numerical ranges are engineering scenarios, not promises for a specific site. Soil is layered, moisture-dependent, and frequency dependent; an installation should be modeled or measured before a close performance claim is made.
Executive findings
- The radial field is a low-loss RF return network, not a safety ground. A ground rod may be important for lightning and electrical bonding, but it does not substitute for a radial system at RF.
- Loss is concentrated near the base because surface-current density is greatest there. Adding conductors in the inner part of the field is especially valuable; uniform azimuthal coverage is usually more important than exact wire length once practical length is reached.
- Four short or sparse buried radials can leave ground loss comparable to, or greater than, a quarter-wave monopole radiation resistance. ARRL reported about a 3 dB radiated-power increase in its illustrated case when forty 0.4-wavelength radials replaced a sparse system.
- Salt water is exceptionally favorable, but shoreline geometry and the return path still matter. One ARRL field account found only a 0.1 dB signal change after adding sixty quarter-wave radials at a particular seawater site with a substantial existing system.
- An elevated system is a different design class: a small number of resonant, balanced elevated radials can largely avoid soil-return loss, but it must be installed and tuned as part of the antenna.
1. What the ground system does
An ideal quarter-wave monopole above a perfectly conducting infinite plane has a radiation resistance near 36 to 37 ohms. The image in that conducting plane completes the other half of the equivalent dipole. Real earth is neither an infinite perfect conductor nor a lossless mirror. The base current must return through a combination of radial wires, soil displacement and conduction current, nearby metal, and the feedline unless common-mode current is suppressed. The real part of the feedpoint impedance includes radiation resistance plus loss resistance.
For a first-order efficiency estimate, η = Rr / (Rr + Rg + Rc + Rm), where Rr is radiation resistance; Rg is ground-system loss; Rc is conductor loss; and Rm represents loading or matching loss. If a full-size quarter-wave radiator is used and other loss is small, the ground-only estimate ηg = 36 / (36 + Rg) is useful. A 36-ohm ground loss wastes roughly half of accepted power in the ground; 18 ohms wastes one-third; 4 ohms wastes one-tenth. The corresponding reduction relative to zero ground loss is 3.0 dB, 1.76 dB, and 0.46 dB, respectively.
1.1 Radials versus a connection to earth
A radial field presents low impedance over area; a rod presents a small, localized interface to earth. At RF, the useful question is not whether the base is “grounded,” but whether return current has a broad, low-loss path. A radial system also controls the near electric field at the soil surface and reduces the chance that the coax shield becomes the accidental return conductor. The safety and lightning system should be bonded appropriately, but it should be designed in addition to the RF return network.
2. Why soil and radial geometry matter
The ground surrounding a monopole carries outward radial current and supports a strong near field. In resistive soil, power density is dissipated as I² times effective resistance. Loss is not uniformly distributed over the radial field: it is heaviest close to the antenna, where current density is highest. This explains three established engineering practices: use many conductors, distribute them evenly in azimuth, and do not neglect the first fraction of a wavelength around the base.
Soil complex permittivity and conductivity change with texture, moisture, salinity, temperature, frequency, and subsurface layering. “Good” and “poor” are therefore design labels, not fixed material categories. Wet soils commonly behave more favorably than dry, rocky, or sandy soils, but seasonal variation can be large.
| Term | Effect on a ground-mounted vertical | Practical implication |
|---|---|---|
| Conductivity | Lower conductivity increases resistive dissipation and weakens the ground image. | More radial coverage is valuable; do not infer performance from SWR alone. |
| Permittivity | Changes field distribution, input reactance, and the real-ground reflection. | Use site-appropriate values in modeling; retune after installation. |
| Radial count | Reduces gaps in the conducting surface and shares return current. | Add wires evenly; diminishing returns arrive gradually rather than at a magic count. |
| Radial length | Extends low-loss coverage outward from the high-current base region. | Quarter-wave is a strong conventional target; shorter radials can still help if numerous. |
| Height above soil | Controls coupling of the counterpoise to lossy earth. | Elevated radials must be resonant, symmetric, and kept clear of conductive objects. |
2.1 The radial field need not be a perfect circle
Radials need not be exactly equal in length or arranged as a perfect star to be useful. Obstacles, property boundaries, paths, and shoreline constraints are normal. However, strongly asymmetric coverage can make the current distribution and pattern asymmetric, especially where the missing sector contains the strongest base-region currents. When a full field is impossible, favor many conductors and coverage close to the base in every available direction; record the geometry so modeling and later measurements can account for it.
3. Five installation conditions compared
The comparison fixes the radiator class: a full-size quarter-wave vertical near its design frequency, with an RF choke at the feedline and conductor/matching losses excluded from the ground-loss column. These values are transparent resistance-to-efficiency scenarios, using 36 ohms as the radiation-resistance reference; they are most useful for understanding scale.
| Condition | Representative RF return system | Illustrative Rg | Ground-only efficiency | Ground power lost |
|---|---|---|---|---|
| Poor ground | Dry/rocky or low-conductivity soil; four short or sparse buried radials. | 25–60 Ω | 59%–38% | 41%–63% · 2.3–4.2 dB |
| Moderate ground | Average soil; about 16 reasonably distributed radials near 0.25 λ. | 10–25 Ω | 78%–59% | 22%–41% · 1.1–2.3 dB |
| Good ground | Moist/conductive soil; 32 or more radials near 0.25 λ or longer. | 3–10 Ω | 92%–78% | 8%–22% · 0.35–1.1 dB |
| Salt water | Beach/over-water site with 30+ substantial radials or suitable counterpoise. | 1–3 Ω | 97%–92% | 3%–8% · 0.12–0.35 dB |
| Elevated radials | Four symmetric tuned radials, raised enough to reduce soil coupling. | 1–5 Ω | 97%–88% | 3%–12% · 0.12–0.57 dB |
The energy column follows Pground / Paccepted = Rg / (36 + Rg), and loss dB = 10 log10[(36 + Rg) / 36]. It should be read as loss caused by the assumed RF return only. A 50% ground-power-loss result does not mean half of transmitter output necessarily becomes soil heat in a complete station; mismatch, feedline, loading-coil, and common-mode effects may change accepted power and add separate losses. Conversely, a low SWR can coexist with high loss because a lossy ground resistance can make a poor antenna appear easier to match.
3.1 Energy flow and takeoff-angle illustrations
The revised figures use one visual convention. Blue arrows represent the relative importance of return current and dissipation in the ground/radial path; red rays show elevation direction. Each panel has a deliberately distinct, labeled illustrative dominant takeoff angle so the comparison is visually clear. These are teaching scenarios, not polar plots or universal numerical predictions: actual elevation behavior depends on the complete antenna geometry, frequency, radial or counterpoise system, terrain, and nearby conductors.
Ground quality can change the power available to every part of the pattern, and an especially lossy near-field return can make low-angle signals disappointingly weak. It does not by itself set a universal takeoff angle.
3.2 Poor ground
Poor ground usually means a combination of unfavorable soil and insufficient low-loss coverage, not merely a low-conductivity number. Dry sand, rock, thin soil over bedrock, or only a few short radials can force a substantial fraction of return current through lossy earth. The result is reduced radiation efficiency and often a changing feedpoint impedance with rain and seasons. If the radiator is electrically short, the problem worsens because radiation resistance falls while ground resistance does not necessarily fall in proportion.
First remedies are geometric: increase radial count, improve azimuthal coverage, add wire in the base region, and suppress feedline common mode. Deep burial, very heavy wire, and a ground rod are usually lower priorities for RF performance at HF than adding distributed wire.
3.3 Moderate and good terrestrial ground
With a moderate field, radials increasingly carry the high-current return component and reduce the soil volume that participates in loss. There is no universal “enough” number because the cost-benefit curve depends on frequency, radial length, soil, antenna height, and required field strength. The broadcast-engineering benchmark is instructive rather than directly prescriptive for amateur or portable systems: FCC rules describe an excellent ground system for ground-based AM radiators as 120 radials, 0.35 to 0.4 wavelength long, spaced 3 degrees; they require at least 90 in that regulatory setting.
A practical HF installation often cannot use broadcast-scale wire. A defensible approach is to lay as many radials as the site supports, use thin inexpensive wire if that enables substantially more coverage, and prioritize reliable base bonds. ARRL/N6LF guidance reports 32 radials of 33 feet as working very well from 40 through 10 meters, while cautioning that low-band requirements differ.
3.4 Salt-water ground
Salt water has much higher electrical conductivity than ordinary soil and supports an efficient image for a nearby vertical. That benefit can appear both in the antenna near-field return and in low-angle propagation over water; the two effects should not be confused. A station on a seawater shoreline may have excellent launch conditions toward water and different behavior inland, depending on radial layout, beach width, tide, and terrain.
The correct lesson is not “no radials are needed at the beach.” With a substantial number of quarter-wave radials, ground losses can be low; one cited case found that 60 radials changed signal strength by only 0.1 dB. That is evidence of a well-developed system at a particular installation, not proof that every salt-water vertical is insensitive to radials. Use corrosion-resistant connections and inspect them: a low-resistance design can be defeated by a few poor marine-environment bonds.
4. Quantifying loss without overclaiming
Ground loss can be estimated four ways, each with different confidence. First, use a validated electromagnetic model with realistic terrain parameters and the actual wire geometry; then sweep plausible conductivity and permittivity values. Second, measure far-field strength or received signal changes while adding a controlled radial set, maintaining transmitter power and frequency. Third, infer changes from base current and input resistance, with careful accounting for matching and conductor loss. Fourth, use the simple resistance estimate in this article to communicate what a plausible Rg means in watts and decibels. The last method is useful but is not a substitute for the first two.
| Assumed Rg | η, ground only | Ground loss | Of 100 W accepted power | Meaning |
|---|---|---|---|---|
| 60 Ω | 37.5% | 4.26 dB | 62.5 W to earth; 37.5 W radiated before other losses | Severely compromised return system. |
| 25 Ω | 59.0% | 2.29 dB | 41.0 W to earth; 59.0 W radiated | Sparse field or poor soil can live here. |
| 10 Ω | 78.3% | 1.09 dB | 21.7 W to earth; 78.3 W radiated | Reasonable but improvable terrestrial system. |
| 4 Ω | 90.0% | 0.46 dB | 10.0 W to earth; 90.0 W radiated | Strong radial field; other losses now matter. |
| 1 Ω | 97.3% | 0.12 dB | 2.7 W to earth; 97.3 W radiated | Excellent return path; no system is perfectly lossless. |
The arithmetic explains why ground systems matter. A 3 dB improvement doubles radiated power for the same accepted power; it is not a minor adjustment. It also explains diminishing returns: reducing Rg from 60 to 25 ohms gains almost 2 dB, while reducing it from 4 to 1 ohm gains about 0.34 dB. Both may be worthwhile, but their priorities differ. Field-strength measurements remain the strongest way to decide whether additional wire is justified at a particular site.
5. Design and measurement guidance
5.1 Ground-mounted radial systems
- Start with the base: make a weather-protected radial hub with low-resistance, mechanically secure bonds. Keep feedpoint, choke, and return connection physically short and intentional.
- Use as many radials as the site allows, distributed over all azimuths. A target of 0.25 wavelength is well established; more shorter wires are often preferable to a small number of long wires when space is constrained.
- Place radials at or just below the surface. Deep burial generally adds labor without RF benefit at HF; shallow placement commonly protects wire from people and mowing equipment.
- Keep the feedline from becoming a radial. Put an effective common-mode choke at the feedpoint and, where useful, verify shield current with a current probe.
- Model actual geometry, including a missing wedge, sloping terrain, nearby fencing, and the waterline. Then validate with an A/B field-strength or reciprocal receive test when the result matters.
5.2 Reading measurements correctly
SWR and input impedance are diagnostics, not performance certificates. Adding radials can change both resistive and reactive feedpoint components, so the matching network may need retuning before any A/B comparison. Keep frequency, accepted power, antenna configuration, receiver gain, and propagation window as constant as practical. On HF skywave paths, use many samples or a nearby calibrated receive point; on MF ground-wave paths, field-strength measurements are especially relevant. A change of less than a few tenths of a dB can be hard to separate from ordinary uncertainty unless the test is carefully controlled.
6. Elevated-radial vertical systems
An elevated-radial system should be treated as a counterpoise or ground plane, not as a few convenient wires above a ground-mounted antenna. When radial conductors are raised high enough and adjusted to resonate with the vertical, most return current can remain on wires rather than diffuse through soil. This can deliver high efficiency with a small number of conductors, a major advantage at poor-soil sites, roofs, portable locations, and low bands where a 120-wire buried field is impractical.
The classic arrangement uses four horizontal radials, approximately quarter-wave long, connected at the feedpoint and symmetrically placed. The radial ends may slope down for clearance or impedance transformation, but the system must preserve symmetry. Their electrical length is affected by diameter, insulation, height above ground, nearby objects, and angle. A vertical tuned for four elevated radials will not necessarily remain tuned when the radials are lowered onto soil, or when one wire runs close to a metal roof, fence, or tree.
6.1 Why four can work when four buried wires usually do not
Buried or ground-laid radials are coupled to soil and are primarily a distributed conductivity improvement. They need many wires because the goal is to screen a broad region of lossy earth. Elevated radials are conductors separated from soil and can form a deliberate resonant return structure. In the cited IEEE Transactions on Broadcasting study, four elevated horizontal radials and a quarter-wave monopole at 1 MHz produced more ground-wave field in the computer models than a conventional ground-mounted monopole with 120 buried radials; average, very good, and very bad ground constants were examined.
That result is powerful but conditional. It does not license any four wires above ground. The modeled geometry used quarter-wave radial length and quarter-wave monopole height, and performance is sensitive to installation. Elevated systems can suffer imbalance, unwanted feedline current, detuning from structures, and an impedance different from the nominal 36-ohm ground-plane value. A current choke and a tuning/measurement process are essential.
6.2 Elevated-system checklist
- Use four radials as the normal minimum for an omnidirectional system; space them symmetrically and connect all at the feedpoint reference plane.
- Design radial length and vertical length together. Trim or tune the entire system in situ rather than assuming textbook wire length will be resonant after installation.
- Raise the feedpoint and radials enough to reduce coupling to soil and objects. More height normally reduces soil involvement, but exact height is not a universal threshold.
- Use a feedline choke at the feedpoint. Check for common-mode current and for a pattern change when the feedline is moved; either can reveal an unintended fifth radial.
- Expect a different feedpoint impedance from a buried-radial vertical. Sloping radials can conveniently move impedance toward coax values, but should not mask asymmetry or poor tuning.
7. Conclusions
For a ground-mounted vertical, the ground system is an active RF component and often the dominant controllable efficiency term. Poor soil combined with sparse radials can dissipate roughly 40% to 60% or more of accepted power in the earth under plausible scenarios. A moderate field may cut that loss to about 20% to 40%; a well-developed terrestrial field can move it below about 10% to 20%; and a salt-water or properly implemented elevated system can reduce it to only a few percent. Those ranges are engineering illustrations, not site guarantees.
The highest-confidence recommendations are durable across antenna scale: install many evenly distributed buried radials for a ground-mounted vertical; put low-resistance conductor where current density is highest; use field strength or validated modeling rather than SWR as the performance test; and select a tuned, symmetrical elevated counterpoise when physical constraints make a large buried field unrealistic. The largest gains generally come from repairing a compromised system, not from chasing the last fraction of a dB on one that is already good.
References
- G. H. Brown, R. F. Lewis, and J. Epstein, “Ground Systems as a Factor in Antenna Efficiency,” Proceedings of the IRE, vol. 25, no. 6, pp. 753–787, June 1937. doi:10.1109/JRPROC.1937.228255.
- J. Sevick, W2FMI, “Ground Image Vertical Antenna,” QST, July 1999, pp. 16–20. ARRL technical-information PDF.
- ITU-R Recommendation P.527-6, Electrical characteristics of the surface of the Earth, 2021.
- R. Severns, N6LF, “Seawater Grounds,” ARRL Antenna Book supplemental files, 22nd edition.
- A. Christman, R. Radcliff, D. Adler, J. Breakall, and A. Resnick, “AM Broadcast Antennas with Elevated Radial Ground Systems,” IEEE Transactions on Broadcasting, vol. 34, no. 1, pp. 75–77, March 1988. doi:10.1109/11.1415.
- Federal Communications Commission, 47 CFR 73.189(b)(4), AM broadcast antenna-system ground requirements.
- ARRL, “Antenna Ground,” technical article index, including “The Effects of Real Ground on Antennas,” QST, 1984, and “Verticals, Ground Systems and Some History,” QST, July 2000.