Lightnews — Scholar-powered news

John Kennedy

@micefearboggis.bsky.social

3.2K followers 3.1K following 1.7K posts

Occasional climate scientist, diagram monkey, probabilistic historian, science anti-communicator. All views and opinions are my own. This is not, sadly, a promise of novelty: it’s a disclaimer. He/him. https://www.jkclimate.fr/ .. more

Videos

Environmental science 43%

Geography 15%

Posts Media Videos Starter Packs

John Kennedy @micefearboggis.bsky.social · 7h

My code has that effect on people.

John Kennedy @micefearboggis.bsky.social · 17h

Open Ocean #11

Playing whack-a-mole with some interesting artefacts in the data and interpolation.

diagrammonkey.wordpress.com/2025/10/09/o...

Open Ocean #11

Playing whack-amole with some interesting artefacts in the data and interpolation.

diagrammonkey.wordpress.com

1 3 5

Reposted by John Kennedy

Ruth Mottram @ruthmottram.bsky.social · 3d

🚨 New @oceaniceeu.bsky.social preprint🚨 why has Antarctica stopped (net) losing mass, in spite of increased discharge?
Declining sea ice is part of the answer but increasingly heavy and frequent atmospheric rivers are most important factor.

Lots of important subtleties:
arxiv.org/abs/2510.03590

Screenshot of figures from the paper: Above a graph showing wiggly lines depicting the cumulative mass budget of Antarctica: a dark line swoops and wiggles downwards before stabilising at the end, colourful dashed lines (depicting snow fall) wiggle along constantly before sloping upwards at the same tme stabilisation occurs.
Below 6 maps of Antarctica in blues and red depicting the mass change in different basins.

1 15 34

John Kennedy @micefearboggis.bsky.social · 3d

Open Ocean #10

In which I attempt maths and experience regret. But at least my mean ends up in about the right place.

diagrammonkey.wordpress.com/2025/10/06/o...

Open Ocean #10

In which I attempt maths and experience regret.

diagrammonkey.wordpress.com

2 4

Reposted by Dorothy Bishop, John Kennedy, Ben Bond‐Lamberty

Rodrigo Vargas @vargaslab.bsky.social · 4d

⭐️Nominate an Impactful Dataset

@agu.org invites you to nominate an impactful dataset!

📊Submissions will be considered for inclusion in an upcoming featured commentary published in AGU Advances.

Submit nominations by 20 October.

docs.google.com/forms/d/e/1F...

@agubiogeosciences.bsky.social

Nominate an Impactful Dataset

You are invited to nominate a dataset to be included in a commentary about the impact of Earth, space, and environmental data using the dimensions of people, planet, and prosperity. Selected submitter...

docs.google.com

7 5

John Kennedy @micefearboggis.bsky.social · 4d

Open Ocean #9

Lots of little things

diagrammonkey.wordpress.com/2025/10/05/o...

Open Ocean #9

Lots of little things

diagrammonkey.wordpress.com

2 3

John Kennedy @micefearboggis.bsky.social · 5d

I was supposed to drop the map of anomalies into this thread... It's quite warm in lots of places but there are some large anomalies over the NH.

bsky.app/profile/mice...

John Kennedy @micefearboggis.bsky.social · 5d

The anomalies locally look like this...

(from pulse.climate.copernicus.eu)

Map of surface air temperature anomales from ERA5 for September 2025. Red areas are warmer than the 1991-2020 average and blue areas are colder than it. There are four different logos along the bottom of the graph and one of them appears twice.

John Kennedy @micefearboggis.bsky.social · 5d

The anomalies locally look like this...

(from pulse.climate.copernicus.eu)

John Kennedy @micefearboggis.bsky.social · 5d

You can see the September bounce here too in a plot of all monthly averages. It will be interesting to see the final ERA5 figures.

Graph showing monthly global mean temperature from 2014 to September 2025. Seven datasets are shown: HadCRUT5, NOAAGlobalTemp v6, GISTEMP, two flavours of Berkeley Earth, JRA-3Q and ERA5.

John Kennedy @micefearboggis.bsky.social · 5d

September global mean temperature is going to take a big leap from August it looks like. I looked at the mean of the daily averages from ERA5 and on an anomaly basis it's much warmer than August was and way above anything before the huge 2023 lurch.

Graph showing global mean temperature for Septembers from 1850 to 2025. Seven datasets are shown: HadCRUT5, NOAAGlobalTemp v6, GISTEMP, two flavours of Berkeley Earth, JRA-3Q and ERA5.

3 2 4

John Kennedy @micefearboggis.bsky.social · 9d

Open Ocean #8

Extending the SST data further back in time and the wonderful horrors of ICOADS.

diagrammonkey.wordpress.com/2025/10/01/o...

Open Ocean #8

Extending the SST data further back in time and the wonderful horrors of ICOADS.

diagrammonkey.wordpress.com

2 5

Reposted by Peter Thorne

John Kennedy @micefearboggis.bsky.social · 10d

Open Ocean #7

Using a simple interpolation scheme to estimate biases in the ship data and create a bias adjusted dataset.

diagrammonkey.wordpress.com/2025/09/30/o...

Open Ocean #7

Using a simple interpolation scheme to estimate biases in the ship data and create a bias adjusted dataset

diagrammonkey.wordpress.com

3 6

John Kennedy @micefearboggis.bsky.social · 10d

Invincibly ignorant is a thing.

John Kennedy @micefearboggis.bsky.social · 11d

Open Ocean #6

Implementing a pattern-based reconstruction technique that works by – I swear – pure magic.

diagrammonkey.wordpress.com/2025/09/28/o...

Open Ocean #6

Implementing a pattern-based reconstruction technique that works by – I swear – pure magic.

diagrammonkey.wordpress.com

3 5

Reposted by Peter Jacobs

John Kennedy @micefearboggis.bsky.social · 12d

Open Ocean #5

Interpolating anomalies in the simple gridder to produce globally complete fields.

diagrammonkey.wordpress.com/2025/09/27/o...

Open Ocean #5

Interpolating anomalies in the simple gridder to produce globally complete fields.

diagrammonkey.wordpress.com

3 13

John Kennedy @micefearboggis.bsky.social · 13d

Open Ocean #4

In which I estimate sampling uncertainties for my simple gridder.

diagrammonkey.wordpress.com/2025/09/26/o...

Open Ocean #4

In which I estimate sampling uncertainties for my simple gridder.

diagrammonkey.wordpress.com

2 4

John Kennedy @micefearboggis.bsky.social · 15d

That’s *the* cropolis.

1 2

John Kennedy @micefearboggis.bsky.social · 15d

There's another fault line: people who really look at their data and people who don't.

John Kennedy @micefearboggis.bsky.social · 15d

Open Ocean #3

Estimating ship-buoy biases with a simple gridder.

diagrammonkey.wordpress.com/2025/09/24/o...

Open Ocean #3

Last time, I did some comparisons between HadSST4 (unadjusted data) and the output from the simple gridder. They were pretty close to each other. I ended up by showing that the number of occupied g…

diagrammonkey.wordpress.com

2 2

John Kennedy @micefearboggis.bsky.social · 15d

Yes, it is bad teaching. The "right" way, you propose, postpones the punchline but it is admirably chaotic.

John Kennedy @micefearboggis.bsky.social · 16d

Thanks! That's nice to hear.

At the risk of saying something I will come to regret, the world needs more blogging scientists.

1 1

John Kennedy @micefearboggis.bsky.social · 16d

That is exactly how I used git for years.

John Kennedy @micefearboggis.bsky.social · 16d

You don't use git? Oh wow. OK. It's really cool. What's that? You heard it's kind of difficult. Nah, it's really simple. How does it work? Well, you just need to know a little graph theor... Hey! Where are you going?

2 1 13

John Kennedy @micefearboggis.bsky.social · 16d

Gah. These days I'm only interested in solving climate problems with a comically large mallet.

1 3

Reposted by John Kennedy, Peter Thorne, Ed Hawkins

Doug McNeall @dougmcneall.bsky.social · 16d

Come and join us developing AI for solving climate problems. careers.metoffice.gov.uk/join-us/sear...

Met Office Careers

careers.metoffice.gov.uk

2 3 3

John Kennedy @micefearboggis.bsky.social · 16d

More steps

diagrammonkey.wordpress.com/2025/09/23/o...

Open Ocean #2

Yesterday1, I posted the results from a very basic gridder. Today, I’ve been processing some more data, tidying up the code a bit, looking at the output and doing some initial comparisons. Th…

diagrammonkey.wordpress.com

1 1 2

John Kennedy @micefearboggis.bsky.social · 17d

It depends what you mean by non-ML AI I guess.

Neural nets have been used for interpolation of global temperature datasets.

John Kennedy @micefearboggis.bsky.social · 17d

It might turn out to be how we finally digitise the majority of paper records

brohan.org/AI_daily_pre...

AI Data Rescue: Daily Precipitation — AI data rescue: Daily precipitation

brohan.org

1 1

Reposted by John Kennedy

Tim Osborn @timosbornclim.bsky.social · 17d

For regular monthly updates to CRUTEM (and hence to HadCRUT) we rely on monthly data provided by National Met Services via the WMO's CLIMAT system. Unusually, the USA data didn't yet make it into the CLIMAT data feed for Mar-Jun 2025. Note the USA-sized gap in the non-infilled version of HadCRUT:

2 2 10

Reposted by John Kennedy, Peter Thorne, John K. Pinnegar , and 3 more John Kennedy, Peter Thorne, John K. Pinnegar, Du Toit, Ed Hawkins, Flavio Lehner

Tim Osborn @timosbornclim.bsky.social · 17d

After a brief (ahem) hiatus, the HadCRUT5 global temperature dataset has now been updated to June 2025 by @metoffice.gov.uk @climateuea.bsky.social and @ncas-uk.bsky.social

1850-2025 time series of global mean temperature anomaly from the HadCRUT5 dataset.

6 44 120