About the UK HF Predictor

This dashboard combines real-time solar and ionospheric indices with observed amateur radio activity to give a per-band view of HF propagation conditions for the United Kingdom.

← Back to main dashboard

Quick reference

Tile borders: green = open, amber = fair, red = poor, grey = quiet (no recent spots). Band cards use the same colours based on the blended score.

Jump to: Controls · Matrix view · Indices · Observed data sources · Scoring · Propagation modes · Band characteristics

Dashboard controls

The dashboard has a few selectors above the per-band cards that let you customise what you see:

These settings are saved in your browser (localStorage) so they persist when you reload the page.

Matrix view (bands + indices)

Matrix replaces the card grid with a compact table so you can scan all bands and distance bins at once. It also condenses the top indices (SFI, Kp, Storm, MUF, LUF, Window) into a single row.

Band × Target bins

Each cell shows the current status label using the same thresholds as the cards: Open, Fair, Poor, Quiet, or Closed. Cells are colour-coded for faster scanning:

Hover (desktop) or tap (mobile) on any cell to see a compact summary: Pred / Obs / Blend, MUF/LUF hints, observed stats (countries, average distance, total spots), and a WSPR roll-up with alignment.

Indices row (top of the matrix)

The indices are shown as a one-row table. Each cell is colour-aware where useful and has a tooltip:

Interaction & accessibility

Headings & compact mode

On smaller screens, matrix headings switch to short labels (All, IG, EU, NDX, RDX) to preserve space. On larger screens, full headings are shown.

The matrix uses the same scoring and thresholds as the cards. “Overall” is the blended whole-band view (0.4·Pred + 0.6·Obs), while the bin columns reflect the specific distance ranges.

Indices

SFI (Solar Flux Index)

SFI measures solar radio emission at 2800 MHz (10.7 cm) and is a long-standing proxy for overall solar EUV output that ionises the F-layer. Higher SFI generally supports higher MUF and longer daytime openings on the upper HF bands.

Source & refresh

What we display

Rules of thumb

Why it matters: SFI isn’t the whole story—geomagnetic activity (Kp) and D-layer absorption also affect results— but it sets the “ceiling” for how high the MUF can climb.

Kp index

The Kp index measures geomagnetic disturbance, derived from a global network of magnetometers. Higher values indicate stronger geomagnetic activity, which generally degrades HF propagation, especially for polar and long-haul paths.

This site uses the NOAA SWPC 1-minute Kp feed (link). Each record includes both the official kp_index and an estimated_kp value.

To avoid spurious noise from single-minute spikes, the dashboard shows:

We also show the data age in minutes to highlight freshness. Kp values typically update every minute, but latency may vary depending on NOAA feed availability.

Why it matters: Quiet conditions (Kp ≤ 3) usually support reliable HF. Disturbed conditions (Kp ≥ 5) often cause fading, absorption, and total loss of high-latitude and polar paths.

Geomagnetic storms (G-scale)

NOAA classifies geomagnetic storms on a scale from G1 (Minor) to G5 (Extreme). These events are caused by solar wind disturbances and CMEs (coronal mass ejections) interacting with Earth’s magnetosphere.

Source & refresh

What we display

Rules of thumb

Why it matters: Even a G1 storm can degrade HF bands, especially at higher latitudes. Severe storms (G3–G5) may cause near-total HF blackout on many paths.

MUF (Maximum Usable Frequency)

MUF is the highest frequency that can support ionospheric propagation over a given path. If MUF is above a band’s frequency, that band is likely to support long-distance contacts.

Source & logic

What we display

Rules of thumb

How it’s used in scoring:

Why it matters: MUF sets the “upper limit” of what’s possible — high MUF means higher bands can carry DX.

LUF (Lowest Usable Frequency)

LUF is the lowest frequency that can propagate over a given path without being absorbed by the ionosphere (mainly the D-layer). If the LUF rises above a band’s frequency, that band’s short-skip/NVIS paths close.

Source & logic

What we display

Rules of thumb

Why it matters: LUF is most critical for low bands (160–40 m). High LUF means D-layer absorption is too strong for NVIS, cutting out reliable short-range HF.

UK Propagation Window

This is a simple time-of-day indicator showing whether the UK is in Daytime, Greyline, or Nighttime.

Source & logic

What we display

Why it matters: Day/night transitions strongly affect HF: low bands (160–40 m) improve at night, high bands (20–10 m) favour daytime, and Greyline can briefly enhance long-haul DX paths.

Observed data sources

Reverse Beacon Network (RBN)

Live observations come from the Reverse Beacon Network, a worldwide cluster of automated receivers. These listen for CW and digital signals (e.g. FT8) and report them into a central database.

For this project we filter to focus on UK transmitting stations being heard by others worldwide. This allows us to measure:

Average distance is computed both for the whole band and per target bin; when you select a bin on the dashboard, the band card shows that bin’s average.

Why it matters: These real-world reports provide the best indication of what is actually open, balancing out the more theoretical predictions from solar indices.

WSPR summary (activity & alignment)

Each band card includes a compact WSPR roll-up derived from the WSPR.live ClickHouse global dataset over a rolling 30 minute window. It’s used as a secondary, independent cross-check alongside RBN:

How we derive the activity level

For each band we compute a lightweight “health” index from:

These are combined (roughly 60% volume / 40% reach) to produce: Strong ≥ 0.70 · OK ≥ 0.40 · Low ≥ 0.20 · Quiet < 0.20. The thresholds are intentionally simple and may be tuned as we gather feedback.

What the alignment line means

Scope & limitations

Future iterations may add a UK-weighted view and/or fold a calibrated WSPR signal into the scoring once we’ve validated the method.

Scoring

Note: The WSPR summary shown on the main page is reference-only and is not included in either the Predicted or Observed scores.

Predicted

A modelled score (0–10) that estimates how well a band should perform right now based on solar/ionospheric inputs.

Inputs

  • SFI – higher is better (especially 20–10 m). A mild curb is applied on 15–10 m when SFI is very low.
  • Kp – we use the 30-minute maximum to capture brief disturbances and apply a penalty at higher values.
  • MUF vs band – logistic “gating” against band edges: well below the band looks closed, just-below looks poor, above the band ramps smoothly to 1.0.
  • Time of day – small night penalty on higher bands (≥ 18 MHz). Low bands receive a night/greyline lift.
  • LUF vs band (low bands only) – LUF gating is applied for ≤ 5.6 MHz (160/80/60 m). For 40/30 m it is ignored unless LUF is implausibly high.

Simplified flow

pred ≈ base(SFI)
       × Kp_penalty
       × MUF_gate
       + low-band shaping (≤10.2 MHz)
       + twilight bump (40/30 m)
       − small night penalty on ≥18 MHz
then apply LUF gating for 160/80/60 m
→ clamp to 0..10

Low-band shaping (≤ 10.2 MHz)

  • MUF headroom uplift: when the MUF sits comfortably above the band, 60/40/30 m receive a smooth uplift (bigger headroom → slightly higher score). This helps avoid under-prediction on good low-band days.
  • Twilight bump (40/30 m): a small, time-weighted boost within ≈±120 min of local sunrise/sunset reflects typical evening/morning improvements.

Why it matters: Gives a physics-based expectation that doesn’t depend on who happens to be on the air.

Observed

A reality check (0–10) from live spots of UK stations being heard worldwide. We normalise activity so very busy periods don’t overwhelm quieter bands.

Components

  • Activity – total recent spots, normalised: min(1, total/50).
  • Country spread – unique DXCCs hearing the UK: min(1, unique/12).
  • Distance – average great-circle distance (km/mi), used as capped(avg_km/6000, 0–1). This is computed both overall and per target bin; the band cards show the per-bin average when a bin is selected.

Simplified formula

obs (0–1) = 0.5·activity + 0.3·country + 0.2·distance
→ scaled to 0–10

Targets / distance bins

  • We also compute per-target bins: Inter-G, Europe, Near-DX, Real DX.
  • When a target is selected on the main page, its observed/blended values are shown for that bin.

Why it matters: Reflects what’s actually open and being worked, not just what should be open.

Sanity guards (plausibility rules)

To avoid obviously unlikely outcomes, the model applies a few lightweight plausibility checks to the Predicted score before blending:

  • Kp penalty: uses the 30-minute max; higher Kp reduces predictions, especially for long-haul targets.
  • High-band night penalty (softer): small extra reduction at night for ≥ 18 MHz (previously stronger).
  • Low SFI curb on 15–10 m (softer): still present when SFI is very low, but less aggressive and primarily enforced via MUF gating.
  • MUF gating: logistic gate around each band’s edge prevents optimistic highs when MUF is below the band.
  • Low-band practicality: MUF-headroom uplift + twilight bump improve 60/40/30 m realism; LUF gating now applies only to 160/80/60 m (40/30 m ignored unless LUF is extreme).
  • Inter-G plausibility: for ≥ 14 MHz, Inter-G predictions are capped (except strong summer Es) to avoid unrealistic “always open” short-skip on high bands.

These guards don’t replace physics or observations; they just keep the predictions sensible while the Observed and Blended scores tell you what’s actually happening now.

Blended

A combined score (0–10) that balances model expectations with real-world activity. We weight observations a bit higher to favour reality.

Formula

blended (0–1) = 0.4·pred + 0.6·obs
→ scaled to 0–10 and mapped to: Closed / Poor / Fair / Open / Quiet

Why it matters: Keeps the dashboard useful when models lag or activity spikes, while still using physics to guide expectations.

Status labels & colours

  • Open – green border/pill: reliable openings are likely/observed.
  • Fair – amber: workable with some limits (shorter paths, timing, or reduced rates).
  • Poor – red: openings are marginal or brief.
  • Quiet – grey: no recent UK spots in the selected time/bin; not an impossibility verdict, just that nothing is being reported right now.
  • Closed – neutral border: conditions or plausibility rules indicate it’s effectively shut.

Thresholds used by the scorer: Open ≥ 7.5, Fair 2.5–7.4, Poor < 2.5. “Closed” reflects gating (e.g., MUF near-closed or LUF closure), and “Quiet” indicates no recent UK spots in the selected bin.

“Hide Closed” hides Closed bands. Quiet bands remain visible (useful to watch for the first signs of life).

Target bins (what the cards reflect)

The per-band cards respect your Target selector:

  • Observed and Blended numbers are computed for the selected bin: Overall, Inter-G, Europe, Near-DX, or Real-DX. If a bin has no spots in the recent window, the card shows Quiet.
  • Predicted is physics-based (SFI, Kp, MUF/LUF, day/night). It’s global to the band, but may receive small bin-dependent sanity nudges (see below) to remain plausible for your selection (e.g. Inter-G on 20–10 m).
  • The “UK heard by” country list is an overall per-band snapshot from the same time window (useful context even when focusing a specific bin).

When you select a Target bin, the Observed and Blended figures on the band cards are computed for that bin. If there are no spots in that bin for the recent window, the card is marked Quiet.

How to read: Overall vs Target bins

The dashboard can show scores for the whole band (Overall) or for a specific distance bin (Inter-G, Europe, Near-DX, Real-DX). These views are related but not identical:

  • Overall — a weighted blend of all bins (roughly: Inter-G 20%, Europe 25%, Near-DX 30%, Real-DX 25%). It answers: “Is this band generally worthwhile right now from the UK?”
  • Target bins — the same scoring formulas, but filtered to just that distance range. They answer: “Is this band working for this kind of path?”

Why they can disagree

  • Overall can be lower than Europe/Real-DX if Inter-G is weak (e.g., daytime on 80 m) or if multiple bins are quiet.
  • Inter-G guardrails cap optimistic predictions on the high bands (≥14 MHz), except during strong summer Es. This keeps “Inter-G on 12/10 m” from appearing routinely open.
  • Quiet means no recent UK spots in the selected time/bin. It’s not a verdict of impossibility—just a lack of current reports.

Example flow

  1. Predicted (physics: SFI, Kp, MUF/LUF, day/night) → apply light plausibility guards (e.g., Inter-G on high bands, high-Kp penalties).
  2. Observed (RBN spots) → activity, country spread, average distance (per bin and overall).
  3. Blended = 0.4·Predicted + 0.6·Observed → label: Closed / Poor / Fair / Open.

Tip: If a bin shows Quiet but your Overall is Fair/Open, ...

Propagation modes

HF radio signals reach their destinations in several distinct ways. Understanding these propagation modes helps explain why some bands favour short-range contacts while others excel at long-distance DX.

Ground wave

Travels directly along the Earth’s surface and is strongest on the lower bands (especially 160 m and 80 m). Range is typically up to 50–100 km depending on frequency, power, and terrain. Ground wave dominates daytime local coverage when the ionosphere is absorbing too much for skywave.

NVIS (Near-Vertical Incidence Skywave)

NVIS occurs when signals are radiated almost straight up and reflected nearly vertically back to Earth by the ionosphere, filling the gap between ground-wave and the first long-skip hop. It provides strong, reliable coverage across 50–400 km, ideal for UK-wide or regional nets.

  • Typical bands: 160 m, 80 m, 60 m and sometimes 40 m.
  • Best times: night and greyline (sunrise/sunset) when D-layer absorption is weak.
  • Daytime: high D-layer absorption often raises the LUF, closing NVIS paths.
  • Antenna: low horizontal dipoles (≈ 0.15–0.25 λ high) maximise high-angle radiation.

On the dashboard, NVIS is considered open when LUF < band < MUF. If the LUF rises above the band, D-layer absorption dominates and NVIS paths close.

Skywave (single and multi-hop)

The classic long-distance mode. Signals leave the antenna at shallow angles and are refracted back to Earth by the F-layer, sometimes multiple times. One F2-layer hop covers roughly 2000–4000 km; multi-hop paths can reach worldwide.

  • Typical bands: 20 m, 17 m, 15 m, 12 m, 10 m.
  • Best times: daylight hours when the F-layer is well ionised (high SFI/MUF).
  • Night: upper bands fade as ionisation drops; lower bands (40–80 m) take over.

Greyline enhancement

Around sunrise and sunset, the boundary between night and day creates a narrow region where ionisation gradients in the D and F layers allow exceptionally low absorption. Signals that follow this “terminator” path can travel thousands of kilometres with little loss.

  • Best bands: 40–20 m are most affected; higher or lower bands may also benefit.
  • Timing: roughly ±30 min around local sunrise or sunset.

On the dashboard this corresponds to the Greyline phase of the UK propagation window.

Sporadic-E (Es)

An irregular ionospheric phenomenon caused by dense patches of ionisation in the E-layer, usually in summer. It can open short-skip paths (500–2000 km) on frequencies up to 50 MHz and occasionally into 10 m.

  • Seasonal: peaks May–August in the northern hemisphere.
  • Effect: often enables strong but fleeting Inter-G and European paths on 12 m – 10 m.

F2-layer propagation

The most important mechanism for global DX. The F2 layer, 200–400 km high, supports long single- and multi-hop paths on the higher HF bands when solar activity (SFI) is high. During solar minimum, F2 reflection may only support 20 m and lower bands.

High solar flux and low geomagnetic disturbance (low Kp) favour strong, stable F2 propagation.

Band characteristics

160 m (Top Band)

Primarily a night-time band. Ground wave can support local/regional contacts; NVIS is possible for UK-wide coverage after sunset. Longer skip (DX) opens on dark winter nights when absorption is low. Strongly affected by storms, QRN, and daylight absorption.

80 m

Mainly night-time but more forgiving than 160 m. Excellent for Inter-G (NVIS and ground wave). Supports European DX at night and occasionally longer paths in winter. Daytime D-layer absorption usually kills longer skip, leaving only NVIS/ground wave.

60 m

Very effective for UK and near-European NVIS work, especially in the evening. Useful for short-to-medium paths when 80 m is noisy or 40 m is too long. Sensitive to LUF changes.

40 m

The classic all-rounder. Strong NVIS for Inter-G during the day/early evening. At night it opens to Europe, Near-DX, and worldwide DX depending on solar conditions. Can be crowded.

30 m

Works day and night with fewer contest/QRM issues (CW & digital only). Good for mid-distance DX into Europe and Near-DX. Often stable long-path opportunities. Less absorption-sensitive than lower bands, less MUF-sensitive than higher ones.

20 m

Daytime workhorse. Usually the most reliable DX band: Europe/Near-DX single-hop and long-haul multi-hop worldwide when SFI is moderate/high. Often open from morning to evening, sometimes 24 h at solar peak.

17 m

Quieter band (less contest activity). Daytime openings often parallel 20 m but close earlier. Good for Europe, Near-DX, and long-haul DX under good solar conditions. Sensitive to MUF drop-offs.

15 m

Daytime DX band, strongly dependent on SFI/MUF. Excellent for Europe and worldwide DX when SFI is high; can be quiet at solar minimum. Responds quickly to solar events.

12 m

Similar to 10 m but opens less often. Excellent long-haul DX when SFI is high. Sporadic-E can provide short-skip UK–Europe, especially in summer.

10 m

Highly influenced by the solar cycle and MUF. In high-SFI periods it supports everything from short-skip (Es/NVIS) to global DX. At solar minimum it can be largely closed except during sporadic-E events.

The model now explicitly includes logistic MUF gating (a smooth, gradual adjustment as MUF approaches or falls below each band, rather than a hard cutoff), a stronger Kp penalty, a high-band night penalty, and low-band shaping (MUF headroom + time-of-day) before applying LUF-based penalties/closure on ≤10.2 MHz.

← Back to main dashboard