Same core climate-analytics method as the 3D globe version, rendered natively in 2D as an equirectangular anomaly-field map (latitude on the vertical axis, longitude on the horizontal — the same projection real NASA GISTEMP / NOAA anomaly maps use) instead of a rotating sphere, plus a scrolling global-mean time series and a live z-score histogram underneath it. Every point on the field — not just the 320 discrete station markers overlaid on top — is evaluated from the same equations:
T_base(lat, doy) = 15 − 0.55·|lat| + A(lat)·cos(2π·(doy − peak)/365)
A(lat) = 3 + 15·(|lat|/90)^1.5
T_now = T_base + trend·(years/10) + patch(lat, lon, t) + noise
anomaly = T_now − T_base
z = anomaly / σ(lat), σ(lat) = 1.2 + 1.8·(|lat|/90)
- Field map (top) — a dense lat/lon grid evaluates the same anomaly equation continuously, so you see the ENSO-like patch as a moving blob and the poles running colder/more variable, not just at 320 sample points.
- Station markers — the same 320-station Fibonacci-distributed network as the 3D version, pulsing when |z| crosses your threshold.
- Time series (bottom) — global mean anomaly scrolling over simulated days, so the long-run warming trend and the patch's oscillation are visible as a trace instead of only an instantaneous globe.
- Histogram — live distribution of every station's z-score, with the flag threshold marked; a fatter tail beyond the threshold lines is exactly the stations currently flagged.
This is the same detect-and-flag workflow real climate-monitoring pipelines use on satellite and ground-station data: compare each location against its own historical normal, standardise to a z-score, and flag statistically significant departures rather than judging raw temperature alone.