Glossary term

Spectrum Occupancy

Engineering definition of spectrum occupancy covering RF channel occupancy, thresholding, duty factor, time-frequency measurement and coexistence validation.

Branch: Telecommunications Engineering
Glossary type: metric
Content: Glossary term
Updated: Jun 26, 2026
Revision: v1.0.0 · reviewed

Definition

metric

Spectrum occupancy is the measured fraction of time, frequency bins or channel observations in which RF energy exceeds a stated threshold inside a specified measurement bandwidth.

Spectrum occupancy is used in wireless site surveys, coexistence reviews, channel planning and interference monitoring. It is not only a percentage. A useful occupancy statement also includes threshold, detector, resolution bandwidth, time window, antenna state, reference plane, observed amplitudes and uncertainty. Low occupancy can still be harmful if the occupied bursts are strong, synchronized with service traffic or inside a sensitive receiver bandwidth.

Spectrum occupancy is the measured fraction of observations in which RF energy is present above a defined threshold. It is used to decide whether a channel, band or time-frequency resource is quiet enough for a service, congested enough to require coordination, or intermittent enough to need longer monitoring.

An occupancy number is meaningful only with its measurement boundary. The statement should name the frequency span, channel bandwidth, resolution bandwidth, detector, antenna, reference plane, threshold, observation window and whether the value is time occupancy, frequency-bin occupancy or time-frequency occupancy.

Thresholded Occupancy

For a sequence of N_tot observations, define an occupied observation when measured power exceeds threshold:

b_k=\begin{cases}1,&P_k\ge P_{th}\\0,&P_k<P_{th}\end{cases}

The channel occupancy is:

\displaystyle O_{ch}=\frac{N_{occ}}{N_{tot}}

where:

N_{occ}=\sum_k b_k

If measurements cover both time and frequency bins, a time-frequency occupancy metric is:

\displaystyle O_{TF}=\frac{1}{N_tN_f}\sum_t\sum_f b_{t,f}

This metric depends strongly on threshold and bin size. Changing the resolution bandwidth, detector mode or threshold can change the reported occupancy without any change in the environment.

Occupancy Margin

For a release rule with maximum allowed occupancy O_limit, a guarded margin is:

M_O=O_{limit}-O_{meas}-U_O

where U_O is the allowance for measurement uncertainty, sampling representativeness or day-to-day variation. The occupancy screen passes only when:

M_O\ge0

Occupancy is a screening metric. It does not replace SINR, adjacent-channel rejection, co-channel interference or receiver overload checks.

Worked Example

A site survey records 600 spectrum sweeps on a candidate channel. The occupancy threshold is:

P_{th}=-85\ \text{dBm}

Energy exceeds the threshold in:

N_{occ}=42

sweeps, so:

\displaystyle O_{ch}=\frac{42}{600}=0.070=7.0\%

The project allows:

O_{limit}=10.0\%

and assigns an uncertainty allowance:

U_O=1.5\%

The guarded occupancy margin is:

M_O=10.0-7.0-1.5=1.5\%

The occupancy percentage passes.

However, the strongest observed burst on that channel is:

P_{I,max}=-82\ \text{dBm}

The desired carrier at the receiver is expected to be:

P_C=-73\ \text{dBm}

The instantaneous carrier-to-interference ratio during that burst is only:

C/I=-73-(-82)=9\ \text{dB}

If the service requires:

(C/I)_{req}=15\ \text{dB}

the burst margin is:

M_{C/I}=9-15=-6\ \text{dB}

The channel passed occupancy but failed the amplitude screen. The engineering decision should not accept the channel without mitigation, a different channel, scheduling proof or a more robust operating mode.

Amplitude Distribution

Occupancy should normally be reported with amplitude statistics, not only a pass/fail count. Useful summaries include the maximum observed level, percentile levels, burst duration and recurrence interval. A channel with 3% occupancy from very strong bursts may be worse than a channel with 15% occupancy from weak distant signals. The receiver consequence depends on when the bursts occur, how strong they are inside the receiver bandwidth and whether the wanted service can tolerate retries or dropouts.

Measurement Boundaries

Spectrum occupancy can be measured with a spectrum analyzer, receiver logs, FFT captures, channel-state logs or dedicated monitoring receivers. Each method has bias. A slow sweep can miss short bursts. A wide resolution bandwidth can hide narrow carriers. A low threshold can classify noise as occupancy. A high threshold can miss harmful signals near a sensitive receiver.

The antenna state matters. Occupancy measured with an omnidirectional survey antenna may not match the installed directional antenna. Occupancy measured at a rooftop may not match the receiver input after feeder loss, filtering, antenna gain and polarization.

Validation Evidence

A defensible occupancy record includes site location, antenna, polarization, height, gain, cable loss, analyzer or receiver model, calibration state, frequency span, resolution bandwidth, detector, sweep time, threshold, observation duration, time of day, traffic condition, weather if relevant and raw or summarized amplitude distribution.

For release decisions, store both percentage occupancy and amplitude evidence. Useful summaries include maximum observed power, percentile power, burst duration, recurrence interval, occupied bandwidth, candidate channel ranking and comparison with SINR or receiver-sensitivity requirements.

Common Mistakes

Common mistakes include accepting a low occupancy percentage without checking burst amplitude, comparing occupancy values measured with different thresholds, surveying during quiet hours only, using a spectrum plot without time history, ignoring adjacent-channel leakage, and assuming an empty-looking waterfall proves future availability.

The practical rule is to treat spectrum occupancy as one measurement layer. A channel is usable only when occupancy, amplitude, receiver margin, time variation and service consequence are all acceptable.

REF

Disciplines