It’s January 2023 and everyone is playing with language generating model ChatGPT. If you’re reading this post from the 2030s or later and want to reminisce about the quaint days of AI, before it caused and/or worsened all sorts of societal problems, this post is what you’re looking for. We don’t know what happens from here. I’m not even going to attempt to make predictions or discuss the possibilities (there’s already enough of that out there). I was curious and simply wanted to see what it was like. Below are several separate chats, all with a geoscience foundation. My questions/prompts are in bold italics and the program’s answers are in normal text. Some of the chats have follow-up questions. I’ve put some additional commentary (not part of the chat) in red italic text below each segment of the chat.

What are the most important uncertainties to consider when reconstructing past climate change from the sedimentary record?

There are several important uncertainties to consider when reconstructing past climate change from the sedimentary record:

1. Age control: Establishing the age of sedimentary deposits is critical for determining the timing of past climate events. This can be challenging, especially for older deposits, as the available dating techniques become less precise.
2. Stratigraphy: The layering of sedimentary deposits can be disrupted by geological processes, making it difficult to accurately reconstruct the original sequence of events.
3. Diagenesis: Physical and chemical changes that occur to sedimentary deposits after they are deposited can alter the original record of past climate conditions.
4. Sample size: The amount of sedimentary material available for study may be limited, which can make it difficult to accurately reconstruct past climate conditions.
5. Proxy data: Different types of sedimentary materials (such as fossil shells or pollen) can be used as proxies for past climate conditions. However, the accuracy of these proxies depends on the accuracy of the assumptions used to interpret them.
6. Spatial and temporal resolution: The resolution of the sedimentary record is limited by the spacing and duration of sedimentary deposits. This can make it difficult to accurately reconstruct fine-scale changes in past climate conditions.

First thing I notice is how the program repeats the question in the first part of the answer. This seems to be a fundamental part of the behavior, as it occurs in all the answers (in various ways). It reminds me of student answers to written exam questions. The list the program provides is really quite good!! I might edit this slightly, but it captures the main issues. Impressive. The below question/answer is a follow-up question to the above.

What are some examples of the layering of sedimentary deposits being disrupted by geological processes?

There are several geological processes that can disrupt the layering of sedimentary deposits:

1. Erosion: The removal of sediment by wind, water, or ice can disrupt the layering of sedimentary deposits.
2. Burial: Sedimentary deposits can be buried by newer deposits, making it difficult to accurately reconstruct the original sequence of events.
3. Deformation: Tectonic forces can deform sedimentary deposits, causing them to fold or fault. This can disrupt the original layering of the deposits.
4. Metamorphism: High temperatures and pressures can cause sedimentary rocks to undergo metamorphism, which can alter the original layering of the rocks.
5. Diagenesis: Physical and chemical changes that occur to sedimentary deposits after they are deposited can alter the original layering of the deposits.
6. Bioturbation: The actions of burrowing animals can mix and disturb sedimentary deposits, disrupting the original layering.

My follow-up question requested some more detail about one of the items in the list in the first answer. Again, a pretty good summary! However, check out #2 (“Burial”). The program is not comprehending the concept of stratigraphy. Stratigraphy is created by burial! A sequence of events (of sediment deposition at the surface of the Earth) is generated through burial. Layers being buried is, in fact, how we conceptualize the sequence of events (processes). Another nuance is how it uses “original layering” in each item. That phrase works perfectly well for some of those, but less so for others. For example, diagenetic change more commonly alters the composition (mineralogy, isotope geochemistry, etc.) of the sedimentary material, not really the “layering”. What does this all mean? 🤷🏼‍♂️

What was the Earth’s climate like in the Eocene?

The Eocene was a geological epoch that lasted from about 56 to 34 million years ago. During this time, the Earth’s climate was generally warm and humid, with temperatures ranging from subtropical to tropical. This warmth was due to high levels of atmospheric carbon dioxide, which acted as a greenhouse gas, trapping heat in the Earth’s atmosphere. The Eocene was also a time of high sea levels, as melting polar ice caps caused the oceans to rise. This led to the creation of shallow seas and the expansion of shallow-water habitats, which supported a diverse array of plant and animal life. The Eocene is also known for the proliferation of mammals and the evolution of modern whales and primates.

Like a good student, the program kicks off its answer with some basic info about the Eocene. I guess it’s programmed to provide definitions of some terminology before discussing it (?). Pretty good summary for only a paragraph, it covers the general state, the proposed cause, and then the state of various sub-systems of the Earth system. If I asked this question on an quiz in my Climate History (intro paleoclimate for non-science majors) course, I would mark it as correct! Follow-up question/answer below.

Why was atmospheric CO₂ so high in the Eocene?

There are several factors that may have contributed to the high levels of atmospheric carbon dioxide during the Eocene. One factor may have been the high levels of volcanic activity during this time, which released large amounts of carbon dioxide into the atmosphere. Another factor may have been the warm temperatures of the Eocene, which could have caused the oceans to release more dissolved carbon dioxide into the atmosphere through a process called outgassing. The warm temperatures may also have caused increased decomposition of organic matter, releasing more carbon dioxide into the atmosphere. It is also possible that there were other, unknown factors that contributed to the high levels of atmospheric carbon dioxide during the Eocene.

Nice. After talking about volcanic CO₂ it then gets into several amplifying feedbacks.

By how much would current CO₂ levels need to increase to match those of the Eocene?

The atmospheric carbon dioxide concentrations during the Eocene were much higher than they are today. It is estimated that atmospheric carbon dioxide concentrations during the Eocene ranged from about 1,000 to 2,000 parts per million (ppm), while current atmospheric carbon dioxide concentrations are around 415 ppm. To match the atmospheric carbon dioxide concentrations of the Eocene, current atmospheric carbon dioxide concentrations would need to increase by more than double their current level. This would have significant implications for the Earth’s climate and the life forms that inhabit it.

It is important to note that the high levels of atmospheric carbon dioxide during the Eocene were not caused by human activities, such as the burning of fossil fuels. Today, human activities are the primary driver of increasing atmospheric carbon dioxide concentrations, and the Intergovernmental Panel on Climate Change (IPCC) has warned that continued increases in atmospheric carbon dioxide concentrations could lead to significant and potentially catastrophic changes to the Earth’s climate.

Bonus points for the second paragraph. I asked a fairly straightforward question, but the program “decided” to provide additional information to further elaborate on how Eocene climate might be relevant to modern climate change.

How did AABW (Antarctic Bottom Water) formation respond to climate change over the past 4 million years?

Antarctic Bottom Water (AABW) is a type of cold, dense water that forms in the Southern Ocean around Antarctica and flows northward along the ocean floor. The formation of AABW is influenced by several factors, including the temperature, salinity, and density of the water in the region, as well as the mixing of water masses and the movement of the Earth’s crust.

Over the past 4 million years, the climate of the Earth has undergone many changes, including multiple ice ages and warmer periods. During ice ages, the formation of AABW may have been enhanced due to increased sea ice formation around Antarctica, which leads to increased brine rejection and increased density of the water. During warmer periods, the formation of AABW may have been reduced due to decreased sea ice formation and reduced density of the water.

It is difficult to say exactly how the formation of AABW has responded to climate change over the past 4 million years, as the climate has changed in complex ways and the mechanisms that control the formation of AABW are not fully understood. However, it is likely that the formation of AABW has varied in response to changing climate conditions, and that it will continue to do so in the future as the Earth’s climate continues to change.

So, this is a question we are actively working on! Our NSF-funded project is led by Ph.D. student Natalia Varela (2018-2023) and is focused on sediment samples and data from IODP Site U1524 in the Ross Sea (which I sailed on in 2018). Again, note how the program dutifully provides a definition and explanation of what AABW is before addressing our question. The third and final paragraph is a classic “it’s complicated” response that all of us are used to giving. We were happy that there wasn’t a clear answer (we’ve done our homework to identify a problem to work on).

What does the current scientific understanding say about how AABW is responding to very recent climate change (over the past few decades)?

The current scientific understanding is that the formation and properties of Antarctic Bottom Water (AABW) have been changing in response to recent climate change.

One of the main factors that has been affecting the formation of AABW is the warming of the Southern Ocean, which is caused by anthropogenic greenhouse gas emissions. As the ocean warms, the sea ice around Antarctica melts, which leads to a decrease in the salinity and density of the water, making it more difficult for AABW to form. Additionally, research indicates that the amount of AABW being formed is decreasing as well as its volume and temperature of the water is also increasing.

The ocean circulation is also impacted due to the climate change, where it is expected that the density of the water in the Southern Ocean will decrease, resulting in the reduction of the sinking rate of the dense water and thus less of the deep and bottom water formation. This is expected to lead to a slowdown of the global ocean circulation and potential impacts on the ocean ecosystems and global climate.

Several studies have been conducted that support the above mentioned change, for example one study using observations and models suggests that the AABW formation is declining due to warming and freshening of the Southern Ocean. Another study using the ocean observation data since the 1960s also shows that the rate of the AABW formation has decreased.

In summary, current scientific understanding indicates that the recent climate change is having a significant impact on the formation and properties of Antarctic Bottom Water (AABW) and is leading to a decrease in the amount and density of AABW being formed, as well as an overall warming trend, which in turn is affecting the ocean circulation and having potential consequences for the global climate and marine ecosystem.

The above was a follow-up to the first question about AABW, so it did not need to provide me a definition of AABW, which makes it feel conversational. If I were reviewing/editing/grading this text I would recommend junking that first sentence, not necessary. I don’t focus on the modern behavior of AABW (I don’t keep up on all that literature), so I’m not able to rigorously evaluate what it’s saying, but it seems reasonable given what I do know. I was intrigued (and impressed!) when it mentioned “several studies” and “one study” in the example provided, so I asked the below follow-up:

Thanks … can you provide links or detailed citation information for these studies you mentioned above?

I’m sorry, but as an AI language model, I do not have the ability to provide specific links or detailed citation information for studies. However, I can help you with some general guidance on finding research papers that have looked at the impacts of recent climate change on the formation and properties of Antarctic Bottom Water (AABW).

Here are a few keywords that may be helpful in searching for relevant studies:

• Antarctic Bottom Water
• Southern Ocean
• Climate change
• Warming
• Freshening
• Ocean circulation

You can use these keywords to search for papers in scientific databases such as Web of Science, Scopus, or Google Scholar. It is also worth checking the websites of organizations that conduct research on oceanography, such as the National Oceanic and Atmospheric Administration (NOAA), the National Science Foundation (NSF), or the European Space Agency (ESA).

Alternatively, you can look into studies that were recently published on journals such as Journal of Climate, Nature Communications, or Global Biogeochemical Cycles which are the common venue for the Oceanography or Climate science research studies.

Please let me know if you have any other question or need help with your research.

Disappointing. I already know how to use GoogleScholar with keywords, we all do this everyday and have gotten pretty good at it. So, this program gets its information from what humans have written to guide it’s language generation, but it can’t connect to specific writing in a direct way. The text in the segment above says “for example one study using observations and models suggests …” yet it does not know what that one study is? Anyway, I presume the ability for this type of auto-generating text program to be able to integrate with specific info that is accessible (or, at a minimum provide a reference) will happen in the next few years 🤷🏼‍♂️. Okay, now for a couple of fun ones:

Can you describe the Paleocene-Eocene Thermal Maximum in the style of Kendrick Lamar?

I’ll note that it starts composing the text mere seconds after I hit ‘Enter’ to ask the question. It’s fast. I’ve listened to some Kendrick Lamar, but I must admit I haven’t listened to him so much that I have a solid feel for his lyrical style. The above seems pretty generic to me, but that’s consistent to all of the program’s answers in my small amount of interaction. Has a meteor impact been proposed to be the trigger for the PETM? I don’t think so, but I could be wrong, I’ve never focused on the PETM interval itself. I do appreciate the final two lines, a nice articulation of why it’s important and valuable to reconstruct past climates!

Can you describe the Pleistocene ice ages in the style of Shakespeare?

Fantastic stuff. I would’ve liked to have seen a stanza about orbital forcings (Milankovitch cycles) but I suppose I didn’t ask for the cause of the ice ages, just a description. To what extent I’ll actually use this to help me work (instead of merely experimenting), I’m not sure yet. If this “evolves” to be able to help me find specific information within journal articles then I would use all the time. For example, “Briefly summarize and point me to a handful of papers that include measurements of velocity for dense shelf water cascading processes in high latitudes.” Even if it didn’t return all or the most relevant papers, this would be a useful way to narrow down the search. We shall see what happens next!

What is rare? Is a recurring event that happens once every decade rare or common? Is an event that occurs once a century rare? Looking to the dictionary we can find that “rare” is “an event, situation, or condition that does not occur very often.” It depends, of course, on the temporal frame of reference. Someone who has well-developed timefulness skills may respond “Duh, of course it depends on the timescale of investigation.” But, would the student new to geoscience (or other areas of inquiry that work in longer timescales) or the person in your family or community who rarely thinks beyond the timescale of their lived experience consider this obvious and trivial?

In our lived experience we can naturally shift our understanding of what is ‘frequent’ versus ‘rare’ based on the timescale. For example, if you do some activity (go for a jog, a bicycle ride, etc.) once a week, this might be quite frequent for some people or contexts but infrequent for others. We’ll add phrases like “which is a lot for me” or “which is pretty rare for us” as a way to communicate the temporal reference. Or, if we don’t add that, someone will ask “So, is that often for you?”. The point is that, in the context of the range of timescales we experience, we are pretty good at incorporating the temporal reference without thinking about it too much.

Coupled with how common or rare an event is over time is the significance or ‘size’ of the event (i.e., frequency-magnitude relationship). Humans have internalized this in numerous contexts. For example, as I write this post (April 2021) the pandemic is still affecting the globe, to varying degrees, and a phrase we find ourselves saying is “this is a once-in-a-century event!”. Implicit in such a statement is that this event we are experiencing is both rare and of significant magnitude. In turn, we often link this to some vague idea of probability — because an event is significant and infrequent it’s likely not to happen again for a long time.

We can consider these ideas for timescales longer than the human lived experience and for ‘prehistorical’ times (before the collective memory or spoken/written record of humans). While this may not come as naturally to us, it’s a skill we can develop. While thinking about this, I was reminded of this 1967 paper from Gretener that I read many years ago (as part of a grad-level stratigraphy course). I won’t summarize the details of this paper here, I encourage you to read it.

Gretener’s main point, which I’ve highlighted in the image of the Abstract above, is that, given enough time, an event that is exceedingly rare becomes probable if not certain. We could apply this notion to the huge range of timescales represented in Earth history. For example, a rare event at the scale of 100,000 years might be considered ‘common’ when viewing 10,000,000 years. How can this stretching of time and the associated frequency-magnitude relationships help us think about and cope with the ‘rare’ events in our society?

Another thought-provoking paper along these lines is this 1989 paper from Kenneth Hsu. This paper focuses on extinction events but the message is very similar to the Gretener paper.

Why is this important? Related to these ideas is the notion of shifting baselines, or a gradual change in what is considered a “normal” range of conditions due to lack of memory and/or knowledge of past conditions. The idea of shifting baselines has been discussed in the context of environmental issues for decades and, more recently, for climate change. And the relationship to hazards and how we collectively assess risk (e.g., “a 100-year flood”) is particularly relevant. As geoscientists, we make the case that the enhanced temporal perspectives gained from studying Earth history provides insight into contemporary challenges. In addition to the technical insights, I think we can also highlight the development of temporal reasoning itself.

As the previous posts in this series (clock time and event time and what is an event?) discussed, my motivation in thinking about timefulness is to move towards developing activities/assignments in my courses that are more explicit about temporal reasoning. One way to do this is develop specific activities centered on temporal reasoning as the primary learning objective. And/or it could be embedded within activities/assignments that have other learning goals, but it would be highlighted and intentionally discussed as part of the activity (a temporal reasoning ‘check’).

In the context of this post, a bunch of potential ideas — e.g., analyze the frequency/magnitude relationships of well-dated sedimentation events and include a class discussion (even just 5 minutes) or reflection question about the dependency of terms like ‘rare’ versus ‘frequent’. If such temporal reasoning ‘checks’ occur many times throughout the course there could then be a more lengthy summary-style activity near the end of the semester to tie them together. 🤔