Hemispheric Power & the HP Forecast

Hemispheric power (HP) is one of the most practically useful numbers in Lumina. It tells you, in plain gigawatts, how much aurora-producing energy is raining down on each hemisphere right now — and where it's headed in the next 90 minutes.

What actually is hemispheric power?

NOAA estimates hemispheric power from its OVATION model, which combines satellite measurements of precipitating particles with a global aurora model. The number you see — say, 45 GW — is the estimated total power being dumped into the southern hemisphere's upper atmosphere by auroral particles at that moment.

To put that in perspective: 45 GW is about 15% of the total electricity generation capacity of Australia. It's a lot of energy, but spread over a huge area — which is why you still need dark skies to see it.

What HP levels mean for you

Lumina classifies hemispheric power into bands, and these thresholds are location-aware — they're calibrated to your magnetic latitude. Here's the Adelaide baseline:

If you're in Hobart or Invercargill, the thresholds are about 6–8% lower — you need slightly less HP to reach the same activity band because you're closer to the oval.

A quick note on Kp and why it's not great

The Kp index is the old standard — you'll see it everywhere in aurora forecasting. But for real-time aurora chasing from Australia and New Zealand, it has some serious problems:

• It's a 3-hour average. Kp tells you what the magnetosphere was doing over the last three hours. A substorm can fire, peak, and fade entirely within one Kp window — and the index won't budge.
• It's northern-hemisphere. Kp is derived from a network of ground magnetometers at mid-to-high northern latitudes. It's measuring northern electrojet activity, not what's happening over Macquarie Island.
• It's global, not local. Kp gives you one number for the whole planet. It can't tell you whether the auroral oval has expanded enough to clear your southern horizon.

Lumina shows Kp (hidden away in the advanced analytics section) because it's still the lingua franca of space weather, but the app's own metrics — solar wind, real-time magnetometer signals, and the probability model — are much better tuned to what actually matters for southern hemisphere watchers.

How HP relates to Kp

HP and Kp are correlated but not the same thing. Kp is a 3-hour average of ground magnetometer deflections — it tells you how disturbed the magnetic field is globally. HP tells you how much energy is actually reaching the atmosphere. Lumina uses this approximate mapping:

Lumina's own southern-hemisphere Kp estimate

Rather than relying on the official 3-hourly NOAA Kp (with all the lag and northern bias described above), Lumina calculates its own near-real-time Kp estimate tuned to the southern hemisphere. This is what drives the G-level display you see on the dashboard and visualiser.

It's a weighted blend of three independent signals:

• Hemispheric power (60% weight). HP is the strongest empirical predictor of Kp — the correlation is about 0.95 at 3-hour resolution. Crucially, HP is Earth-measured, so it has none of the L1 propagation lag that affects IMF-based estimates.
• IMF coupling field score (40% weight). Derived from real-time B_z, B_t, and solar wind speed. Provides a forward-looking signal when HP is still catching up to freshly-arrived solar wind.
• Ground magnetometer confirmation. The current ΔH from Casey and Mawson acts as a lower-bound sanity check — a large negative bay at a polar station directly confirms ionospheric energy dissipation and can pull the estimate upward if the HP/IMF blend looks too low.

The estimate updates every few minutes, maps to the standard G1–G5 scale, and comes with a confidence label (high/medium/low) depending on how many sources contributed. When you see a G-level in Lumina, it's this southern-aware estimate — not the official NOAA Kp from three hours ago.

The HP forecast — what's coming next

NOAA provides a 90-minute HP forecast that Lumina pulls and displays on the dashboard. It shows you whether HP is expected to rise, fall, or stay flat over the next hour and a half.

When Plan mode says something like "HP forecast peaks near 85 GW — activity likely around 11:30 PM" , here's what that means:

• Lumina scanned the 90-minute HP forecast and found a peak of ~85 GW (which is elevated-to-major territory).
• The forecast predicts the best viewing window based on when that peak occurs, combined with darkness and moon timing.
• The trend signal (rising/falling/flat) tells you whether things are improving or deteriorating.

HP's limitations

Hemispheric power is useful, but it's not the whole story. Two things to keep in mind:

It's calibrated mostly on the northern hemisphere

NOAA's OVATION model — which produces the HP number — is primarily calibrated against northern hemisphere satellite and ground data. The same hemispheric asymmetries that affect substorm detection (IMF B_y tilt, ionospheric conductivity differences, and Earth's asymmetric internal field) also mean the southern hemisphere's actual auroral energy deposition can differ from what a northern-calibrated model predicts. Lumina uses the southern hemisphere HP value specifically, and cross-checks against real-time southern ground magnetometers, but the raw HP number itself comes with an inherent northern bias.

To mitigate this, Lumina's aurora oval model uses a probabilistic approach: instead of relying on a hard geometric boundary (which amplifies calibration errors), the OVATION flux field is sampled at your exact location and converted to a continuous probability. Flux values are more robust to calibration bias than boundary positions — a small systematic offset shifts a probability by a few percent rather than flipping a binary inside/outside determination.

Additional corrections address specific physical asymmetries:

• IMF B_y oval tilt — the solar wind's east-west magnetic field tilts the oval, shifting it equatorward on one side and poleward on the other depending on hemisphere. Lumina applies this correction to the geometric elevation calculation.
• Ground magnetometer cross-validation — real-time data from Casey, Mawson, and Macquarie Island stations provides an independent ground-truth signal that can be compared against model predictions.

It's a hemispheric total — not your local view

HP tells you the total energy raining down across the entire southern hemisphere at that moment. It doesn't tell you where the oval is relative to your location. You can have 80 GW of hemispheric power with the oval sitting squarely over Antarctica — great for penguins, not much use for someone in Melbourne. Lumina accounts for this by separately modelling the oval boundary and your elevation angle (see the Aurora Oval Geometry page) — the raw HP number on its own shouldn't be read as a simple "higher = better for me" metric without checking your local geometry.

Reading the HP chart

The Hemispheric Power chart on the dashboard shows the last 24 data points (covering about 2 hours) plus the 90-minute forecast. The colour coding matches the activity bands above — green for quiet through to purple for major. You'll also see location-specific threshold markers (star icons) showing where the active and elevated bands kick in at your latitude.