Paleoelevation challenges: Insights from modern climate
As a reminder, arguably one of the most promising means of estimating paleoelevation (and one being used a lot) is to measure the isotopic composition of rainfall that has been stored in the rock record (in altered volcanic glasses, in bones and teeth, in clays, in soils, etc.). The idea is that as rain-bearing clouds rise up terrain, they rain on that terrain and the farther up they go, the more rain they have lost. Since heavy isotopes tend to rain out first, the ratio of heavy to light isotopes in water decreases. And if you look at river water or rainfall today, you broadly find that the most depleted water is falling in the highest areas–but it is a noisy record and so even if you were handed a modern water sample you might be hard-pressed to determine its elevation.
For the moment, let’s assume that the measurements of isotopic composition of paleo-rainwater is robust. Can we just use some regression of depletion vs. elevation to get paleoelevation? There are several who have argued no; at its heart, the basic problem is that it is not elevation that you are measuring but the amount of water that has been wrung out of the clouds. What else might control rainfall? These two articles point out two elements that are very challenging in the paleo-realm: air trajectories and rain nucleation. (We’ll leave out a lot of other issues for today).
First, air trajectories. While it rains on the side of the mountains where the air rises, it doesn’t rain nearly so much on the side where the air descends, so obviously which way the air goes is important. To first order, we have a pretty fair guess as to which way the air went in the past from the biota. But the angles and other obstacles along the way, interaction with broader topographic features (which can produce rainfall prior to the air encountering the terrain), interaction with jet stream dynamics and other issues are all problems. This turns out to also be important in making predictions of climate change, and that first article, Alex Hall’s Perspective on “Projecting Regional Change,” points out some real problems in trying to estimate climate change at a regional scale (like, say, a mountain range in the western U.S.). Simply embedding a regional model in a global model, Hall argues, can lead to grossly incorrect results if you don’t comprehend the limitations of the global model. His example is a known bias of global circulation models (GCMs) to put the jet stream in the wrong place; use that as your boundary condition and you will mess up precipitation in a big way. This is obviously a big concern for forecasting changes in precipitation in places like California, where lots of the water that has fallen historically is tied to “atmospheric rivers” which, while not the jet stream, are relatively localized and transient features originating far outside the region of interest. (The article goes on to note that in contrast the temperatures from GCMs are more robust as boundary conditions). What it means is that trying to estimate how precipitation developed in, say, the Oligocene in the western U.S. is going to be very tricky. Even if you have a decent GCM for that time (itself an open question), then embedding a regional model within that GCM might be prone to large biases. These then get translated into errors in the predicted isotopic shifts of precipitation.
The second story by Douglas Fox in High Country News documents work by scientists investigating the origin of dust reaching the western U.S. and its role in triggering rain. While this is somewhat removed from the primary research (some conducted here at CU Boulder), it nicely summarizes the recognition that dust and microbes are probably important in nucleating raindrops or ice crystals. Today a lot of that dust is coming off of Asia (other dust causing other issues is more locally derived). One statement deserves special mention:
[Marty Ralph] and [Kimberly] Prather recently teamed up to compare two atmospheric rivers that passed over California. They had nearly identical water content, temperatures and winds. But only one contained Asian dust — and it dropped 40 percent more precipitation than the other. That corresponds to a staggering 1.5 million acre-feet of water — greater than the amount of water currently sitting in California’s largest reservoir, Lake Shasta.
Presumably the storm which dropped 40% more precipitation ended up far more depleted than the other storm, so the isotopic content of rainfall at the same elevation from the two storms would be quite different. (This is actually seen in records of the changes during the winter season in the isotopic characteristics of the snowpack in the Sierra for instance). If changes in the average amount of nucleating material occurred in the past, wouldn’t that change the isotopic depletion at a given elevation?
Paleoelevation is one of the great challenges facing our understanding of the evolution of mountain belts. It is tricky both because of the relative paucity of material preserved from the higher elevations of orogens and our current inability to infer elevation from what materials we have. These points here just underscore the challenges we face in solving this.