Antarctic Land Surface Temperature (part 2)

In this episode I produce time series slides of mean annual temperature for the three main categories of base location using slightly more ‘robust’ data than offered by the raw series

In part 1 of this series I kicked-off by looking at data quality of the GHCN daily data series. What should be obvious by now is that there are not that many weather stations in Antarctica and that data collection among these is pretty patchy at the best of times. As a result it is quite possible to derive annual mean land surface temperatures that are rather wild and misleading.

In this article we’re going to get a handle on the impact of base location on land surface temperature and I shall start by plotting out the annual mean land surface temperature for those 29 coastal stations who’ve managed to maintain 50% or greater daily data capture over time. This is not exactly ‘robust’ by any means but if I push the data quality issue too far then we end up with nothing much to look at! Try this for size:

Coastal Bases (0 – 20km from nearest coast)

Isn’t that super interesting? We have four groups of bases and I can reveal the warmest of these that are crayoned in red and green are those that sit on the Antarctic peninsula. If you see an alarmist media headline chances are that the record-breaking/unprecedented temperature will have been recorded on the coast at one of these peninsula bases, a few of which are now small towns. The middle band of blue-tinged bases are those scattered around the coast elsewhere on the continent, with the purplish and grey colouring for Base Belgrano II, SANAE SAF-Base, McMurdo Sound NAF, Halley, Little America and Ellsworth IGY bringing up the chilly rear.

If this slide doesn’t get the message across that location is vitally important in determining the temperature record -  and especially so for a frigid Antarctica - then nothing will. These are all coastal bases yet we’re talking a twenty plus degree differential in average annual temperature on the same continent!

Mischief At Esperanza

Back on 7 February 2020 journalist Rosie McCall treated us to this alarming headline in Newsweek…

…the base in question was Esperanza (the dark green squiggle on my slide) and thermometers there allegedly hit 18.3°C, beating the previous record by “almost a whole degree centigrade”. I say ‘allegedly’ because whenever I check these things using source data I always find something different to what is claimed.

I doubt that many folk checked-out Esperanza base following that headline (many smartphoners these days don’t even read the article below the headline) but it’s a big one with a permanent civilian settlement and even a school house. Thus, we discover that they’re not recording the temperature of the raw and wild Antarctic continent per se, they’re recording the temperature of a developing coastal village on the peninsular. I think a frown is in order!

So what about that 18.3°C, was that real?

Yes it was, but despite SMN Argentina’s tweet of 18.3°C the official WMO/GHCNd record shows 18.4°C was reached on 7 February 2020. Rosie’s article (admittedly based on SMN Argentina’s tweet) goes on to claim that, “until yesterday, the highest temperature recorded in the area—and the entire Antarctic continent—was 17.5°C, which was reported on March 24, 2015, also at Base Esperanza”. This is a big fat porky pie for the WMO/GHCNd record indeed shows 17.5°C for 24 March 2015, but it also shows 18.0°C for 13 March 1979. One of those early warming records they’d like to conveniently ignore methinks. I think some growling is in order!

Then there’s them there holes to consider. It’s all very well saying things like, “since monitoring began in 1961” but this doesn’t mean much if data collection has been sporadic. If we count the number of daily maxima recorded between 1 January 1961 and 30 April 2023 we find 18,411 complete daily records out of a total of 22,276 days, this giving us a capture rate of 80.9%. This is not bad going but what if other record-breaking temperatures were sitting in that 19.1% of daily data they failed to capture at Esperanza?

Incomplete data capture is another of those fine details that climate scientists seem to overlook, so I guess I better wrap this section up by plotting out the hottest temperatures recorded at Esperanza each month:

This puts the 2020 record-breaker into context and displays the 1979 record-breaker that SMN Argentina have somehow ignored. It’s worth noting that five extreme maxima have popped up over the years and it’s worth considering why this is so for there’s a jolly good chance that sporadic extremes are a natural feature regardless of how much fossil fuel is being guzzled. If so, then we can forget the ludicrous notion of global warming from man-made emissions and instead think about ocean currents, volcanic activity, solar particle forcing and rather prosaic localised factors such as base development and incidence of kebab wagons – all depends on where they’ve stuck the thermometer.

The snaking green line - another of those Locally Estimated Scatterplot Smoothing (LOESS) functions – serves to remind us of the wiggly nature of temperature maxima. If I force a linear regression through this data cloud I discover a statistically significant warming trend of 1.8°C per century (p=0.012) but the big question here is what is causing that slight warming? Sure, we can go shouting fossil fuels, fossil fuels! if that takes our fancy but that cult-like approach to climate is missing more factors than I’ve had hot dinners (and I like my hot dinners).

Inland/Interior Bases (20 – 500km from nearest coast)

In trying to distil robust data from seriously patchy stuff my sample of inland and interior bases has dwindled from 14 to just 6, and so I’ve collapsed these two categories out of necessity. Plonking them on the same slide produces this sort of crazy:

Byrd Station is some 480km from the nearest coast, so we might expect that to be a rather cool place. Though I should say ‘was’ because the base closed in 1988. Equally as cool as Byrd was Eights Station, sitting at 149km from the nearest coast when it was doing its thing. In contrast, Neumayer Station is the ‘hot bod’ sitting up at a cosy -16°C. It should not come as a surprise to learn that Neumayer, SANAE (AWS) and Troll In Antarktis (named after a drinking session?) are grouped together in Queen Maud Land, though when I say ‘together’ this means within a 200km radius.

I am hoping the message about location, location and location is hitting home but another message that should be hitting home is how Spartan daily data records really are. In this regard we don’t have a single decent daily data series that stretches from 1954 through to 2022, and the best we can do is consider the Queen Maud Land mob as a merged sample.

I am also hoping that readers will note that, with mean annual temperatures as low as these, ice isn’t going to melt any day soon since the temperature isn’t going to rise above zero degrees. It might melt around the coastal stations during the summer months but that’s all the melt you are going to get.

Deep Interior Bases (>1000km from nearest coast)

There are only three players in this location category so let’s go see what they yield without delay:

Aside from the early daily series from Amundsen-Scott we’ve got very little to go on that could be construed as semi-reliable. It sure is cold in the deep interior so we can’t expect too much from equipment whose job it is to stand outside! One thing we can say is that the deep interior at Amundsen Scott didn’t show signs of global warming from 1957 to 1995. There are dollops of daily data from 1995 onward but it is very patchy indeed. If I throw caution to the wind and plot every scrap of data available we get this rather wild plot:

Those dirty great spikes have nothing to do with climate change and everything to do with sample bias – probes tend to work when it’s warmer! Anybody using this data to suggest climate change is affecting the deep interior of the Antarctic is either stark raving mad or is being paid handsomely to come up with any old garbage to suit a political agenda.

So Where Does This Leave Us?

If we are concerned about data quality and bias introduced by equipment failure then we really need to focus on the handful of semi-robust data points presented in three slides above. In doing so we find signs of warming at the coastal bases on the peninsula but that’s about it. There doesn’t appear to be any warming if we move slightly inland or if we go into the deep interior. The paucity of reliable data means we cannot generate a grand mean anomaly for the continent as a whole. Sweeping journalistic statements such as “Antarctica is warming” are thus utterly ludicrous and without foundation; at best we might declare that the peninsula is warming, but even then this only applies to the coastal areas adjacent to the ocean. In effect, what these coastal bases are doing is measuring ocean warming and/or the urban heat island effect. In the next article I’ll derive a few warming trends so we can look see; until then…

Kettle On!