Looking at the graph left me wondering just what it means exactly. I'm not well versed in statistics so "the standard deviation is 3.5°C" doesn't mean much. Also, what's up with that other line going down to -3.5°C? And what do the colors mean? In the sense that I'm not sure whether a darker blue means closer to or further from today.

▲

blaze33 4 hours ago | parent | next [-]

You can go to the source website https://climatereanalyzer.org/clim/sst_daily/?dm_id=nino3.4

There's is an interactive chart that's easier to understand

	▲	2b3a51 4 hours ago \| parent \| next [-]
		Thanks for posting that link. The graph has a key on the right hand side that clearly labels each colour of line, and the horizontal axis is scaled in months of each year. Scrolling down gets you notes and links to data sources. In answer to another poster in this thread, the dataset only reaches back to 1983, I'm assuming because that is when they started monitoring these temperatures?
	▲	camillomiller 4 hours ago \| parent \| prev [-]
		Check out 2015, it had way hotter temperatures in November, with higher temperatures than the average in this period, but I would like a climatologist to explain this, draw correlations etc. The original post is a weird LLM-mediated mix of vague scaremongering with some easy piling on journalism "just because". So what am I supposed to with it? Nothing, because it's written by an LLM, I guess.

▲

danhau 4 hours ago | parent | prev | next [-]

I'm not an authority on this, but here's is my understanding - I'd appreciate if someone could correct my mistakes.

The baseline of 0.0 represents the average of all years. Anything above / below the baseline is a (standard) deviation from the average. The blue lines are the individual years since 1991 [1] while the red line is the year 2026.

If a line is above the baseline, then the sea-surface temperature was hotter on that day than average. If below, it was cooler than average.

The year 2026 is an outlier, dwarfing all the others starting around June / July. The Nino 3.4 sea-surface temperature is significantly hotter than any previous year during that time. New record, I guess?

[1]: I'm confused about the two date ranges given: 1982-2026 and 1991-2020. I'm assuming this graph is based on measurements from 1982-2026 to calculate the average, but the lines shown are only from 1991-2020, for some statistical reason I don't understand.

	▲	plough 4 hours ago \| parent [-]
		I think it is the other way around: SD is calculated from 1982-2020, while all measurement readings in the plot are 1982-2026. I believe this is meant to not introduce an unwanted shift but compare to sort of a 'stable process'. However, that should have been described and argued somewhere.

▲

bestouff 4 hours ago | parent | prev | next [-]

> Each blue line represents a different year since 1982. The red line is this year. It doesn't just set a new record. It has departed entirely from the range of previous observations.

▲

bitter_michael 4 hours ago | parent | prev | next [-]

I had the same concerns and think the chart would benefit from color grading the individual years by age. If the other outlier in the opposite direction is equally likely then it should also be concerning (obviously it is not). My understanding is the deviation is from the 1991-2020 subset avg, so a warming trend would be indicated by relative drift towards positive in the std dev across years from 82-present

▲

ArnoVW 4 hours ago | parent | prev [-]

my layman understanding, a real statistician will surely intervene.

standard deviation is a measure that informs about the distribution. A high standard deviation means a "wide bell curve". A low standard deviation means that all values are closely clustered around the middle of the curve.

So if your value is 2 x standard deviation (for example) that means it is a relatively rare outlier, since 2 x standard deviation covers 95% of the bell curve. In particle physics I believe they require 5 standard deviations to confirm an observation.