An interactive model of the Team Spectro concept for the Tiangong Space Station: one wide-field camera watches the atmosphere continuously, YOLO_storm flags sprite-prone thunderstorms — persistent targets that last seconds or more — and four pan–tilt 2D spectrometers (diffraction grating + camera) slew to those storms in advance, then dwell, track and keep recording into a ring buffer. It never chases the sprite itself, which lasts only about 30 ms. When one fires over an assigned storm, the slew has already happened, so the delay at event time is zero — and YOLO_sprite’s job afterwards is only to flag and save the buffered event. Run the simulation, tune the hardware, and see exactly when a capture succeeds or is lost.
Earth is a real sphere and the ground visibly falls away on either side of the point directly below the station. This view defaults to Exaggerated ×5 — labelled, not to scale so the 400 km gap and the thin 50–90 km sprite layer are easy to see; switch to True scale for the real proportions. The storms and sprites here are the same live ones as in the flat Live simulation below — use the controls under that view, fire a sprite, and watch for the red glow in both views. Drag to orbit · scroll or pinch to zoom.
EXAGGERATED VIEW (×5) — default: all altitudes above the surface are stretched five times (station shown at 2,000 km-equivalent, sprite layer at 250–450 km-equivalent) so the thin layers are easier to see. Earth's radius is unchanged. This graph is labelled as not to scale — switch to True scale for the real proportions.
It detects a storm — a big, persistent target — and points at that in advance. The sprite is never chased. Split the timeline in two and the 30 ms problem changes completely:
The wide-field camera runs continuously. YOLO_storm flags a sprite-prone thunderstorm, which stays in view and keeps charging for seconds (in this sim, roughly 2.5–6.5 s of charge-up, and real storms persist far longer). That is ample time for a spectrometer to slew, settle, dwell and track it. No 30 ms deadline applies in this phase.
The spectrometer is already on the storm and already recording into a ring buffer, so slew delay at event time = 0 ms. The 30 ms sprite simply lands in the buffer as it happens. Only afterwards does YOLO_sprite flag the event and save that buffered window — it never starts a slew.
The table below is the corrected budget. Phase A is judged against seconds, Phase B against the 30 ms event. All numbers are this app’s design defaults, not measured flight performance.
So the honest claim for the paper is narrow and strong: pre-pointing eliminates the slew delay at event time for an assigned, tracked storm — it does not eliminate every delay, and it does not guarantee a capture.
When a spectrometer is locked on and a sprite fires, its grating spreads the flash into a spectrum. Choose an event type to preview the diagnostic emission lines a paper would report. After a live capture, this plot switches to the captured event.
Tap any labelled peak above — or any row in the table — to jump to that element’s enlarged zoom panel below.
| Emission feature | Wavelength | Origin | What it tells you |
|---|
Each panel isolates a single element / band from the spectrum above, magnified around its centre wavelength (±30 nm), so the line shape, exact wavelength and relative strength can be read without squinting at the full plot.
Each unit: camera + diffraction grating on a pan–tilt gimbal, station belly-mounted, nadir-looking.
A miss is counted when a sprite fires with no spectrometer locked onto its parent storm — wrong target, still slewing, below the YOLO threshold, an unassigned storm (only 4 units).
Computed live from the geometry — quote them, then state your assumptions.
Prototype ground truth from the poster: spectra of spark-gap “lightning” were recorded with both slitless and 2D-slit spectrometers. Slitless is lighter but hard to calibrate; the 2D slit is accurate and easy to calibrate but demands the precise, rapid alignment this simulator stresses.
Cold-plasma discharges in the mesosphere above strong thunderstorms, usually after positive cloud-to-ground (+CG) strokes. Red N₂ emission above, blue tendrils below.
Greenish Optical emission from Sprite Tops — the 557.7 nm atomic-oxygen green line glowing at sprite tops after the red flash fades. A prime target for this spectrometer.
Narrow blue cones propagating up from cloud tops into the stratosphere, dominated by N₂⁺ blue emission.
Expanding rings of light in the lower ionosphere driven by the electromagnetic pulse of an intense lightning stroke.
Science goals from the poster: mesosphere physics · sprites vs tropical-cyclone intensification · aviation-hazard remote sensing · ionospheric anomalies such as GHOSTs.
Simulation honesty note: storm behaviour, YOLO confidences, slew geometry and spectra are illustrative models with realistic orders of magnitude — suitable for exploring the design and drafting a paper's feasibility argument, not flight-qualified predictions. Spectral line positions are real; relative intensities are schematic.