WARMING
How much California has warmed, where it has warmed fastest, and why it matters that the coast warms at night.
The series, and the line through it
ANNUAL MEAN TEMPERATURE, 1895–2025This is the whole argument in one chart: every complete year's statewide mean temperature, with an ordinary least-squares line fitted through it and a shaded 95% interval around that line. The interval is the part worth looking at. A trend quoted without one is a number pretending to be a fact, and on a noisy series it is often not distinguishable from no trend at all.
—
It is not a straight line
DECADE BY DECADEFitting one line to 131 years produces one number, and that number quietly assumes the warming has been steady. It has not been. Averaging by decade removes the year-to-year noise without imposing a shape, and what emerges is a long flat stretch through the middle of the century followed by an almost uninterrupted climb from the 1970s on.
The plateau from roughly the 1940s to the 1970s appears in the global record too, and is generally attributed to industrial sulphate aerosols — which cool — offsetting part of the greenhouse warming until air-quality regulation removed them. That is an explanation from the wider literature, not something this site's 131 California numbers can demonstrate on their own, and it is flagged as such.
Day and night are warming differently
DAILY MAXIMUM AGAINST DAILY MINIMUM, BY DIVISIONA mean temperature is the average of a daily high and a daily low, and those two have not moved together. Splitting them apart by division produces the most interesting result on this page — and it splits California along a line the statewide average erases completely.
—
Faster night-time than daytime warming is the signature usually associated with more water vapour and more low cloud holding heat in after dark, and with urbanisation changing what the ground does with the day's heat overnight. The interior divisions show the opposite sign, with daytime maxima climbing faster than nights. This site reports the split because it is clearly in the data; it does not adjudicate the cause, which needs work this dataset cannot do.
The warming, in energy
COOLING AND HEATING DEGREE DAYSDegree days are the same temperature record re-expressed as demand: how far, and for how long, the outside air sat above or below 65°F. NOAA publishes them alongside the temperatures, and they turn an abstract fraction of a degree into the quantity that actually shows up on a grid operator's chart.
These are outdoor-temperature integrals, nothing more. Actual energy use also depends on how many buildings there are, how well they are insulated, how many have air conditioning at all, and what people set the thermostat to — every one of which has changed enormously over this record. The trend below is the weather's contribution in isolation, which is the only part this data can isolate.
What could be wrong with this
THE HONEST CAVEATSThree real objections to a century-scale temperature trend, and what this record does about each.
The thermometers moved
They did — sites relocated, shelters changed, and the time of day observations were taken drifted from afternoon to morning across much of the network, which alone can shift a station's apparent mean by more than a degree. nClimDiv is built from homogenised GHCN-Daily data precisely to remove these, which is the reason this site uses it rather than raw station series.
Cities are hotter
Urban heat islands are real and large. They are also local, and the divisional analysis is a gridded average over an area the size of a small country — the South Coast division contains Los Angeles but is mostly not Los Angeles. Urbanisation is a legitimate contributor to the coastal night-time signal above, which is why that section names it rather than claiming a clean greenhouse fingerprint.
The trend is just noise
This is the objection with a real test, and it is applied here. Every interval on this site is widened for the fact that a warm year tends to follow a warm year, which shrinks the effective sample size below the number of years. The statewide temperature trend survives that adjustment comfortably. The precipitation trend does not — and the precipitation page says so.