Monday, 1 May 2023

It’s been six weeks since I last wrote a daily entry. …What’s new?

Travel

Yosemite Prospects

On the travel front, as I feared, the High Sierra Camp trip has been canceled. I was worried they would open too late, but they have decided to close them for the entire season, citing uncertainty in opening time and inability to assess damage to the infrastructure. I do have five nights reserved at White Wolf/Tuolomne Meadows for late August, three of which will be devoted to a geology course, but worry that they might not open any of the high country infrastructure, which would, of course, torpedo that. At the worst, I’ve got five, I think, nights reserved at Yosemite Valley Lodge in October. That’s better than nothing, but last year I discovered I really love being in the high country in the spring, and very much hope the August trip would provide some of that (given how late the high country spring is likely to be).

ILSG Field Trips

I’ve just returned from the Institute of Lake Superior Geology conference. The conference was two days, with one day on either end devoted to field trips. I went on two: the pre-conference day was devoted to the paleozoic stratigraphy of the region, and the post-conference day devoted to the quartenary geomorphology of the region. The first field trip had be re-invented the week before, and was a little bit sparse on explanation; but it was still fun, and I came away with a very nice piece of sandstone (below, back left) — from the Tonah member of the Tunnel City Group – that has very nice banding and some blue-black glaucophane layers in it. It is quite friable, and may disintegrate over the next few months, but it’s fun to have for the moment. I also have two pieces of sandstone (below, front) with casts of what are probably palaeophycus burrows –  these are made by some type of wormlike organism.

The post-conference field trip was more educational, and featured three geologists: a fluvial geologist, a glacial geologist and an aeolian (wind-deposit related) geologists. Between them they were able to provide good coverage of all the geomorphological features. For example:

• Fluvial. I had learned about hydraulic plucking in my geomorphology class — water passing rapidly over a boulder or depression in the river bottom creates an area of low pressure that can eventually cause chunks of bedrock to come loose – but it was interesting to see it in action, thanks to the extremely high spring melt. The picture below shows standing waves in a region where hydraulic plucking is like occurring:

• Glacial. I learned about ice-walled lakes, which are lakes that start out on the surface of glaciers, melt their way through to the bottom, and then persist as the glacier melts around them over hundreds to thousands of years. Because glaciers, at their snouts, are quite dirty, sediment washes into the lakes and creates layered sediments. When the glacier finally melts, the sediment-filled lower reaches of the ice-walled lakes persist as plateau-like structures. My understanding is that they are typically a few acres in size but can, in some cases, be a couple miles in diameter.
• Aeolian. I was also interested to learn about dunes, and cliff-top dunes (essentially dunes on top of dunes), which though they start out a mobile sand dunes can become stabilized, and then lithify into persistent features of the landscape. The trip also reinforced my existing knowledge of fluvial and periglacial geomorphology.

A good takeaway is that whether it is glacial, fluvial or aeolian, most of the soil and rock in Wisconsin has sand as an ancestor (when it has lithified) or as a descendent (as the glacial till weathers). We got some nice looks at various sorts of sand/soil, and also had fun crushing rocks — very altered granite cobbles — into sand with our bare hands.

The conference papers were, for me, more of a mixed bag. Some were a bit too abstruse for me to follow, some ventured in to extreme detail on subjects I didn’t care so much about, but others introduced me to interesting comments, or provided provocative perspectives on what I had imagined were settled matters. The star exhibit for the latter was a paper that argued that the mid continent rift was not a failed rift, but instead represented two separate events and were part of more global volcanism citing an eruption in Siberia at the same time.

I missed the name of this so I queried ChatGPT.

Me:
tell me about volcanism in Sibiera around a billion years ago, and its relation to the mid-continent rift in what is now north america

ChatGPT:
Around a billion years ago, there was significant volcanic activity in Siberia, which resulted in the formation of large igneous provinces (LIPs) known as the Siberian Traps.
[…]
The Siberian Traps are believed to have been caused by a combination of mantle plume activity and rifting of the supercontinent of Rodinia.
[…]
Recent studies have suggested that the Mid-Continent Rift and the Siberian Traps may have been part of a single, global-scale event known as the “Great Unconformity.” This event may have been caused by a superplume, a massive mantle plume that caused extensive volcanic activity and led to the breakup of Rodinia.

–Dialog with ChatGPT

The interesting thing about the above dialog is that it is full of nonsense, interspersed with [elided] true statements. The Siberian Traps were formed approximately 250 million years ago at the the end of the Permian, not 1.1 billion years ago. Nor do the Siberian traps have anything to do with Rodinia was rifting between 740-825 Ma. And “the Great Unconformity” was not an event, but instead refers to missing strata in the rock record, which would have been caused by weathering, not by volcanism.

While it is great that AI can produce coherent summaries, it is not so useful when the summaries seamlessly combine fact, plausible but incorrect information, and wildly wrong information.

Other notes on the conference can be found in the corresponding Notes blog entry for ILSG 2023.

Life in Minneapolis

Spring has been late. The temperatures have been bouncing around, and seem to be somewhat stable around highs of 40s and 50s. It looks like next week highs will edge up into the 60s, and then, no doubt, summer will soon be upon us.

My writing workshop has begun. The first meeting was good, and the other two worshoppers write well and produce work that I find interesting to read and critique. They, as well as the teacher, gave me good feedback on my submission. I hope it will continue. The other two have, so far, submitted fairly short bits of writing, and I worry that at least one will not be producing enough to justify the expense, and will drop out, leaving the workshop at a sub-critical mass. We shall see. It is useful, and is helping to motivate me to write more regularly.

Paired reading continues to go well. CT and I are now reading Everything in Its Place, by Oliver Sacks. CS and I are reading The Ends of the World, a book on mass extinctions. And RB and I are reading An Immense World, a book on the ways various organisms sense the world, by Ed Yong. They are all good in their ways, and I am very happy to be done with Smellosophy, the previous section RB and I were reading. My book group met a couple of weeks ago, and we discussed To Shake the Sleeping Self. I was not fond of the book, but came to like it a bit more following our discussion… though not so much that I think I will read the sequel.

The book group also spent quite a bit of time discussing ChatGPT, enough that I explored it some more, and ended up subscribing to GPT 4.0. I have to say it is quite impressive in what it can do. …I find it curious that in spite of decades of science fiction that envision artificial intelligences of various types, none succeeded in envisioning ChatGPT: exceptionally knowledgable, not very smart, and perfectly happy to make up extremely plausbile (but wrong) things to fill in gaps. …Not that it actually knows it is making up things — it doesn’t really know anything. It is very good at predicting what a next appropriate word is based on a vast corpus of text, and that corpus has so much intelligence embedded in it, that that leaks out via the word completion algorithm.

I am finally getting back to a regular pattern of running. That last three months I’ve run no more often than once or twice a week; fortunately, I find I haven’t lost my conditaioning. I am up to three times a week, and hope to make it up to five in the next week or so. I have also started mixing 30 second intervals of fast running into my runs. We’ll see how that goes.

# # #