Propagation & Channels

Cross-Polarization Discrimination (Prop)

Q: What is the difference between propagation XPD and antenna XPD?

Antenna XPD is a fixed hardware property: the cross-polar isolation an antenna provides on boresight in clear sky, typically 30 to 40 dB for a well-built feed. Propagation XPD is the additional, time-varying degradation the atmosphere adds. The two combine in power (their cross-polar contributions add), so the system XPD is always worse than either alone. In clear weather a link might show 35 dB system XPD set mostly by the antennas; during heavy rain the channel can drop propagation XPD to 15 dB or less, which then dominates the total. Link budgets must use the worst-case combined value for the required outage percentage.

Q: How does rain reduce XPD on a microwave or satellite link?

Falling raindrops are oblate (flattened) and, in wind, canted at an angle to the horizontal. An oblate drop presents different attenuation and phase shift to the horizontal and vertical field components, so a wave that starts purely co-polar emerges with a small orthogonal component, which is depolarization. Because the same drops also cause co-polar attenuation (CPA), XPD and CPA are strongly correlated. The ITU-R P.618 method exploits this with the equimembrane relation XPD approximately equals U minus V times log10(CPA), where U and V depend on frequency, elevation angle, and the rain canting-angle statistics. At 12 GHz a 10 dB rain fade typically corresponds to roughly 25 dB XPD.

Q: What XPD is needed to reuse the same frequency on two orthogonal polarizations?

Frequency reuse by dual polarization works only if the cross-polar leakage from the unwanted channel stays well below the wanted signal. The cross-polar interference is the negative of XPD relative to the co-polar carrier, so XPD acts as a carrier-to-interference ratio. For a link that needs 20 dB C/I to meet its modulation BER, the worst-case XPD must stay above about 20 to 23 dB for the target availability, often 99.99 percent of the worst month. Systems that cannot meet this with antenna isolation alone add an adaptive cross-polar interference canceller (XPIC), which can recover 15 to 20 dB and let the link ride through deep depolarization events.

/kraws-pol-uh-ruh-ZAY-shun dih-SKRIM-uh-NAY-shun/

Expressed in dB, this propagation parameter is the ratio of power received in the wanted (co-polar) sense to the power the channel couples into the orthogonal (cross-polar) sense. Unlike the fixed isolation of an antenna feed, propagation XPD varies with weather: oblate, canted raindrops and ice crystals impose different phase and amplitude shifts on the vertical and horizontal field components, so a clean wave acquires a depolarization component as it travels. Because the same hydrometeors also produce co-polar attenuation, XPD and rain fade are tightly correlated, which lets prediction methods derive one from the other. This metric sets the achievable channel isolation on dual-polarized frequency-reuse links from C-band satellite feeders to E-band terrestrial backhaul.

Category: Propagation & Channels

Clear-sky XPD: 30 to 40 dB

Heavy-rain XPD: 10 to 20 dB

How the Channel Depolarizes a Wave

A wave radiated in a single linear or circular polarization stays pure only in free space. Real atmospheric paths contain hydrometeors that are neither spherical nor randomly oriented. Raindrops larger than about 1 mm flatten into oblate spheroids, and aerodynamic torque plus wind shear tilts their major axes to a mean canting angle, often 5 to 25 degrees from horizontal. The horizontal and vertical components of the incident field then experience unequal specific attenuation and unequal phase velocity. When the two altered components recombine at the receiver, the result is no longer aligned with the transmit polarization: a fraction of the energy has rotated into the orthogonal channel. The dB ratio between the surviving co-polar power and this leaked cross-polar power is the propagation XPD.

Ice crystals high in the troposphere add a second, often subtler, contribution. Needle and plate crystals are nearly lossless but are aligned by atmospheric electric fields, so they produce differential phase shift with very little co-polar attenuation. The practical signature is XPD degradation during otherwise light precipitation, sometimes preceding the rain itself. Because ice depolarization can occur without a fade, it is harder to compensate by reading the co-polar signal level, and it is the reason measured XPD scatters around any single prediction curve.

System designers care about XPD because polarization is a reuse resource. Sending two independent data streams on orthogonal polarizations doubles spectral efficiency, but only if leakage from the unwanted stream stays far below the wanted carrier. The instantaneous cross-polar interference equals the negative of XPD relative to the co-polar carrier, so XPD behaves directly as a carrier-to-interference ratio for the reused channel and must be predicted at the worst-case availability, not the median.

The XPD-to-CPA Relationship

The dominant prediction tool is the empirical link between XPD and co-polar attenuation (CPA), formalized in ITU-R P.618 and the related terrestrial methods. Because the same canted drops cause both effects, a measured or predicted fade depth maps to an expected XPD through frequency-, elevation-, and tilt-dependent coefficients. Circular polarization is more vulnerable than linear aligned to the canting axis, and tilt mismatch can cost several dB. The governing relations below are the form an engineer evaluates for a given path and percentage of time.

Propagation XPD definition:
XPD (dB) = 10 log₁₀ ( P_co / P_cross )

Equimembrane XPD-to-CPA relation (ITU-R P.618 form):
XPD ≈ U − V × log₁₀(CPA)

Frequency and tilt coefficients:
U ≈ 30 log₁₀(f) − 10 log₁₀(0.5 − 0.5 cos 4τ) − 40 log₁₀(cos θ) V ≈ 20 for 8 to 35 GHz

Where P_co = co-polar power, P_cross = cross-polar power, CPA = co-polar attenuation in dB, f = frequency in GHz, τ = polarization tilt angle, θ = path elevation angle. Example: f = 12 GHz, θ = 30°, τ = 45°, CPA = 10 dB → XPD ≈ 25 dB.

XPD Across Bands and Conditions

Scenario	Frequency	Co-polar fade (CPA)	Typical XPD	Dominant mechanism	Mitigation
Clear sky	4 to 40 GHz	~0 dB	30 to 40 dB	Antenna feed only	Good feed / OMT
Light rain	12 GHz Ku-band	3 dB	30 to 33 dB	Mild rain canting	Margin only
Heavy rain	12 GHz Ku-band	10 dB	22 to 25 dB	Oblate drop tilt	XPIC canceller
Intense rain	20 to 30 GHz Ka	20 dB	12 to 18 dB	Rain + path length	XPIC + adaptive coding
Ice / dry storm	4 to 12 GHz C/Ku	0 to 2 dB	15 to 25 dB	Aligned ice crystals	XPIC (no fade cue)

Common Questions

Frequently Asked Questions

What is the difference between propagation XPD and antenna XPD?

Antenna XPD is a fixed hardware property: the boresight cross-polar isolation a feed delivers in clear sky, typically 30 to 40 dB. Propagation XPD is the time-varying degradation the atmosphere adds. Their cross-polar contributions add in power, so system XPD is always worse than either. Clear weather may show 35 dB set by the antennas; heavy rain can pull channel XPD below 15 dB, where it dominates. Link budgets must use the combined worst-case value for the required outage percentage.

How does rain reduce XPD on a microwave or satellite link?

Falling drops are oblate and, in wind, canted off horizontal. An oblate drop gives different attenuation and phase shift to the horizontal and vertical field components, so a co-polar wave emerges with an orthogonal component. The same drops cause co-polar attenuation, so XPD and CPA correlate, captured by XPD ≈ U − V log₁₀(CPA) with U and V dependent on frequency, elevation, and canting. At 12 GHz a 10 dB fade maps to roughly 25 dB XPD.

What XPD is needed to reuse the same frequency on two orthogonal polarizations?

Cross-polar leakage from the unwanted channel must stay well below the wanted signal, so XPD acts as a carrier-to-interference ratio. A link needing 20 dB C/I requires worst-case XPD above about 20 to 23 dB at the target availability, often 99.99% of the worst month. When antenna isolation alone cannot hold this, an adaptive cross-polar interference canceller (XPIC) recovers 15 to 20 dB and lets the link survive deep depolarization.

Related Terms

← Cross-Polarization Discrimination Cross-Polarization Interference →