1.19 Degrees

The record as color

Each stripe is one year. Blue sits below the 1951–1980 average, red above, and the palette is fixed before the argument starts: the same ramp will color every figure on this page. Read left to right. A cool half-century, a mild rise through the 1940s, a long pause, then two decades of saturation.

Color is honest about direction and mute about magnitude. A stripe cannot tell you how much warmer. For that, each year has to become a measurement.

After Ed Hawkins’s climate stripes (2018), redrawn here from the raw series. The color scale is printed on the figure; nothing is rescaled for effect.

The record as measurement

The stripes fall to their values. Every point is a global annual mean anomaly: the difference, in degrees Celsius, between that year and the 1951–1980 baseline. Zero is not “normal.” It is a reference line, and it stays on the chart from here to the end.

Anomalies travel better than absolutes. A mountain station and a valley station disagree about temperature but agree about departure, so a sparse network can still constrain the globe. The scatter is real weather: Krakatoa’s aerosols shade 1884, Pinatubo’s cool 1992, El Niño tips other years warm. Weather scatters the points. Climate is the level they scatter around.

What survives the smoothing

A locally weighted regression, computed in your browser from the same 146 points, which stay on the chart, dimmed but not deleted. Smoothing is an editorial act. It decides which wiggles you get to see, so it gets a label instead of a mystery: lowess, tricube weights, fifteen-year window.

What survives: a cool start deepening to about −0.38 °C near 1910, a rise into the 1940s, three flat decades while postwar sulfate haze offset the early greenhouse push, and then, near 1975, a bend that has not unbent since.

The last fifty-five years

The axis re-draws for the close-up and the zero line comes along, because a zoom that hides its baseline is an argument, not a chart. Since 1970 the least-squares trend is +0.20 °C per decade, drawn dashed and labeled.

1998 was the outlier once: an El Niño year so warm it anchored a decade of claims that warming had stopped. Every year since 2012 has been warmer than 1998. Then 2023 cleared the standing record by 0.16 °C, and 2024 raised it again, to +1.28. The points do not know about the argument. They keep landing higher.

You are here

Pull back. The whole record returns, faded to context, and one point stays lit: 2025, the most recent complete year, at +1.19 °C above the 1951–1980 baseline. It is not the warmest. 2024 holds that, at +1.28. The ten warmest years in the record have all come since 2015.

Next year the chart gets one more point. The record does not argue. It accumulates.