Grid observations: 19–20 October 2026

This fixed analysis uses the exported observations supplied on 20 October 2026: 2026-10-19 17:31:03 UTC to 2026-10-20 17:43:02 UTC. It covers about 24 hours and 12 minutes, not a complete calendar day. It is not a live report.

What the recorded samples show

Demand: sample minimum 26.71 GW at 2026-10-20 01:45:02 UTC; maximum 40.96 GW at 2026-10-20 17:43:02 UTC. Arithmetic sample mean: 33.59 GW, from 1,440 numeric records.

Wind: sample minimum 8.36 GW at 2026-10-19 17:35:02 UTC; maximum 13.71 GW at 2026-10-20 11:25:02 UTC. Arithmetic sample mean: 11.44 GW, from 1,440 numeric records.

Gas-fired generation: sample minimum 4.30 GW at 2026-10-20 01:45:02 UTC; maximum 15.53 GW at 2026-10-19 17:40:02 UTC. Arithmetic sample mean: 8.80 GW, from 1,440 numeric records.

Carbon intensity: sample minimum 88.00 gCO2/kWh at 2026-10-20 04:42:02 UTC; maximum 194.00 gCO2/kWh at 2026-10-19 18:12:02 UTC. Arithmetic sample mean: 126.82 gCO2/kWh, from 1,440 numeric records.

What this comparison means

The ranges show variation in this exported window. They do not establish its cause, identify an outage or measure reserves. The lowest wind and highest demand need not occur together: inspect the times rather than assuming coincident extremes. These means give equal weight to every record rather than weighting by elapsed time.

Coverage and uncertainty

There are 1,440 unique timestamps, with a median spacing of about one minute. The first and last are more than 24 hours apart. Missing samples and changes between samples remain possible. The legacy collector substitutes zero for some unparsed fields and 50 Hz for unparsed frequency. Numeric validity does not establish successful parsing. Carbon values can be forecast when actual is unavailable, without that distinction being saved per row.

The collector reads the Gridwatch Templar page and separately requests carbon intensity from the Carbon Intensity API. This report summarises those saved observations; it does not independently verify upstream accuracy. Later readings are not included and this page does not update automatically.

Repeat the analysis

Choose a defined period in History, inspect coverage, export its CSV and compare sample extrema with their timestamps. State the sampling resolution. Hourly snapshots can conceal short-lived peaks, so a different maximum from an hourly view does not automatically indicate a calculation error.

How to compare charts fairly · Source and collection details · All guides