Will the increase in global average temperature from 1978 to 2048 be above 1 degree Celsius?
➕
Plus
16
Ṁ503
2048
70%
chance

The increase in temperature will be calculated:

a) taking NOAA-STAR satellite measurements of the temperature of the lower troposphere (see https://www.star.nesdis.noaa.gov/smcd/emb/mscat/) OR the most similar such data if this dataset is discontinued,

b) calculating the slope and intercept of a least-squares affine fit of this data versus time (from November 1978 to November 2048),

c) using that fit to calculate the change in temperature between 1978 and 2048.

For reference, from Nov 1978 - Mar 2024, the anomaly has been 0.135 K/decade. The time elapsed has been 45.33 years. Multiplying these numbers, the increase to date (at the time of creating this market) is ~0.61 deg C. Therefore, to exceed 1 degree Celsius by 1948, a further increase of ~0.39 is required.

Get
Ṁ1,000
and
S1.00
Sort by:

1 - the title says 1948, not 2048

2 - meaningful comparisons are done with 30-year baselines, not a single year. Maybe you should benefit from your typo in the title and take the 1948-1978 average and then compare that to 2048.

Thanks @BrunoParga , appreciate it.

  1. Yeah, my blunder there. I've fixed it.

  2. I think the Q is ok here. If you look at the description you'll see the calculation method. It's a linear fit over the time period so there isn't the problem of having volatility at the start or end point. LMK

@Ronan I don't really understand the calculation method, but what little I understand seems to suggest that you are taking November 1978 as your starting point, rather than the average of a 30-year period as is usual. But maybe this is not so, I don't know.

@BrunoParga the method is to take all of the data from the period of 1978 to 2048 and fit a line to them. Using that line as a basis, a temperature increase over the period can be calculated.

This is the simplest (most robust) way to compute an increase over a given time period as it involves finding the closest fit to all datapoints over the period (as opposed to having to pick a specific time period for averaging at the start and at the end).

@Ronan sure, now I see how that smoothes out any big discrepancies in the endpoints from the trend line.

It is just not, to my knowledge, how this is generally done, and I think there is more value to predicting values that are usual.

@BrunoParga yeah I see your point. It can be good to stick to standards.

The issue with standards here is that to get a 30 year average for around the year 1950, we would need to wait on 15 years after that.

But yeah, maybe there's a separate worthwhile poll here where I compare the average over the years 2035-2065 (with the poll resolving in 2065) compared with the average over 1978-2008. Let me see what other good feedback I get before making another poll just yet. Thanks.

@Ronan AFAIK the 30-year average is only for the reference period; each individual year is compared to that.