david jon furbish

The roads in the pics were ice covered and littered with downed trees and power/data lines for about a week. They were mostly impassable for several days. Rowan and I still dodge downed lines on our local sniff-and-walks, and our favorite forest-access parks are still mostly closed.

Pic of side of road at intersection with power/data line on the ground next to stop sign

Pic of piles of tree logs and branches lining the side of a road in the neighborhood. Some of these piles are tens of meters long and 3-4 meters high, and have LARGE logs along with the slash

Pic of several piles of logs and branches lining the side of the road below our house. Foreground pile is one of several we have started

Pic of Rowan the intrepid woof standing in the front lawn near a wheel barrow checking out other woofs and hoomans doing yard work in the neighborhood

Our high-tree-density part of south Nashville suffered extensive damage to trees from the ice storm, w/ loss of power for nearly 2 wks. Most power/data lines are restored, but the cleanup has just started. Meanwhile, Rowan thinks yard work & stick chewing w/ Chief woof and Big woof is great fun.

Definitely have a look at Joshua Sokol’s thread regarding Hogg’s white paper on the role of LLMs in astrophysics.

This reminds me of extensive commentary by David Deutsch regarding predictions (in the sense of forecasting) without explanation versus the value of explanations that are “hard to vary” — an elaboration of Asimov’s idea of “relativity of wrong” as well as offering clarity on expectations of theory.

Why do we do astrophysics? At time of writing, large language models (LLMs) are beginning to obtain the ability to design, execute, write up, and referee scientific projects on the data-science side of astrophysics. What implic...

The issues raised should concern all of us in the sciences. A must-read.

“…LLMs are beginning to obtain the ability to design, execute, write up, and referee scientific projects on the data-science side of astrophysics. What implications does this have for our profession?”

arxiv.org/abs/2602.10181

lol yes

Your appeal to uniformly distributed angles reminds me of the famous Bertrand paradox in relation to the principle of indifference. If you have not seen it, check out Jaynes’ paper, “The well-posed problem,” a master class in clear thinking and writing — a classic.

en.wikipedia.org/wiki/Bertran...

Yes! The ratio of two zero-mean Gaussian variables is Cauchy with undefined mean and undefined variance. The reciprocal of an exponential variable, or the ratio of two exponentials, is equally weird! 🤓

Reciprocals and ratios of random variables can be problematic, often leading to pathological results. Regarding the example described in the excerpt, a rational formulation would instead pay attention to impulses in relation to the particle Stokes number. Instantaneous forces are mostly irrelevant.

Here we note that deterministic descriptions of sediment particle motions and transport often mix deterministic quantities with random variables and manipulate these according to the ordinary rules of algebra, not recognizing or acknowledging that the algebra of random variables is unlike the algebra of ordinary variables, nor properly averaging the algebraic expressions involved. As a consequence it is difficult to decipher the mechanical meaning and implications of such descriptions, as they can involve hidden or missing information, or algebraic relations among quantities that do not correctly represent the physical situation. Numerous examples exist in the literature. Here are brief summaries of three. Deterministic formulations of the onset of individual bed load particle motions have appealed to the ratio $R_F = F_\mathrm{L} / F_\mathrm{D}$ of the lift force and the drag force, $F_\mathrm{L}$ and $F_\mathrm{D}$, acting on a stationary particle. Within a strictly laminar flow the lift $F_\mathrm{L}$ and drag $F_\mathrm{D}$ are fixed, and the ratio $R_F$ is therefore fixed. In this situation the forces, and the ratio $R_F$, may be treated as deterministic quantities. Within a turbulent flow the lift and drag forces on a stationary particle fluctuate (e.g. \textcolor{blue}{Schmeeckle et al., 2007}),\footnote{Likewise the lift and drag forces on a moving particle fluctuate (\textcolor{red}{e.g. REFS}).} and these forces must be viewed as random variables. In turn, the ratio $R_F = F_\mathrm{L} / F_\mathrm{D}$ is a random variable. Now, whereas it is possible to form the ratio $R_F^* = \langle F_\mathrm{L} \rangle / \langle F_\mathrm{D} \rangle$ of the averages $\langle F_\mathrm{L} \rangle$ and $\langle F_\mathrm{D} \rangle$, this ratio differs from the expected value $\langle R_F \rangle$. In fact, $\langle R_F \rangle = \langle F_\mathrm{L} / F_\mathrm{D} \rangle$. Moreover, whereas the ratio $R_F^*$ is likely to be well-defined, the expected value $\…

A recent journal article reminded me that deterministic descriptions of sediment transport have led to some "famous" but decidedly flawed results that many uncritically cite as fact. To wit, here is an excerpt from my (unfinished) ebook under the heading, "The Algebra and Calculus of Expectations."

Here is a cool website that features a map of sandhill crane locations by month. The principal migration stopover site south of Nashville is the Wheeler National Wildlife Refuge.

sandhillfinder.savingcranes.org/distribution

The airport guy initially seemed dismissive. But Anne did not let up, and by the end of their conversation he actually seemed to engage with what Anne had to say about the cranes — their flight patterns in the Nashville area, etc. Or at least he gave this impression with his questions!

The majestic sandhill cranes are starting their migration over our neighborhood, flying at astonishingly high altitudes. Only one person I know who would get on the phone to the Nashville control tower to insist that they keep their damn airplanes away from our beloved birds. Tell ‘em, Anne! 😄👍💥

My dear friend and brilliant colleague, Peter Haff, died two years ago. I miss Peter, and I think about him frequently. Here is a link to my 2024 statement.

“A legion of beautiful stars watching over Peter Haff’s beloved Mojave Desert twinkled extra last night…”

cdn.vanderbilt.edu/t2-my/my-prd...

pic of sun shining through ice covered tree, and surrounding area of trees, all ice covered. Damage to trees in the area of Nashville is extensive

pic of ice covered road with downed trees blocking parts/much of road. This is a “major” road in this part of Nashville. It was not passable, and we backed out to go a different route

Winter storm conditions in Nashville w/ continued freezing and power outages are grim. Needed to use crampons, for sure, just to take Rowan outside for her constitutionals, dodging downed trees and power lines everywhere. We did 2 days/nights w/ fireplace and camp stove before making our way south.

wow

what a sad, unserious fantasy of meaningful research

cartoon from @snail.and.snail with two snails conversing: English is difficult. When do we use “many” and when do we use “much”? “Many” is for nouns that are countable, like apples. “Much is for nouns that are not countable, like water. So we should say “many rational numbers” and “much real numbers”?

the English language meets Cantor’s continuum hypothesis lol

People have been trying (understandably) to dunk on the FDA about this, but I think there's real progress in their acceptance of Bayesian stats. When I wrote about pediatric COVID vaccines in 2022 the FDA was still requiring frequentist methods, delaying the approval.
www.nytimes.com/2022/03/01/o...

I occasionally show your videos to my 4 year old grandson, who lives in Charlotte, NC. He is a great fan of your work! If you ever grace Charlotte with your art, I hope I know and can tell him to find you. And if you ever visit Nashville, I’d definitely try to find you to watch your magic.

happy new year and wondering if the lesson is that we should agree to not be friends lol

tri-color Australian Shepherd comfortably laying in sunlight on a couch with a pillow beneath her head

When Rowan was a puppy she enrolled in several self-guided courses. After excelling in Basic Couch Use 101, she then took Principles of Pillowing 101, and Advanced Pillowing 240. Scored well in both. A genuine autodidact. 😊

#dogs
#dogsofbluesky
#AustralianShepherd

This chapter of the book is soooo fun. One of Earth's clearest examples of the elements and consequences of unsteady, rarefied, nonlocal transport conditions.

"Indeed, an explanation of how rain splash transport works requires a probabilistic description; no other possibility exists."

2/2

eBook Preview (continued): Statistical Physics of Sediment Particle Motions and Transport

Chapter 11: Rain Splash Transport

Rain splash transport has been described in the literature as the "simplest of all geomorphic processes."

sure ok lol 😊

1/2

cdn.vanderbilt.edu/t2-my/my-prd...

The inherent beauty of real paper maps!

22 year old stick shift Subaru

still doing great… CD player and all lol

will use paper maps if I ever do a Walden Pond pilgrimage

the bed stress is probably the wrong choice as a macroscopic state variable.) The preview below (Chapter 6: Kinematics of the Particle Flux) provides an explanation of key factors contributing to this behavior in measured transport rates, and much more.

3/3

cdn.vanderbilt.edu/t2-my/my-prd...

one or two orders of magnitude or more. To be sure, the data scatter in this type of plot will never, ever go away. It partly reflects the inherent stochastic qualities of rarefied bed load transport together with uncertainty of measurement procedures. (This is aside from the fact that...

2/3

Plot of (top) 50 realizations (black lines) of particle number N(t) versus time t showing increasing variance with t about expected value (yellow line) and +/- 2 standard deviations (red lines); and (bottom) 50 realizations (black lines) of time-averaged particle number flux q_{nx} versus averaging interval Delta t showing convergence to ensemble expected value <q_{nx}> (yellow line) with +/- 2 standard deviations (red lines). Plots are generated with Poisson rate lambda = 1 s^{-1}, and mimic measured times series of coarse gravel transport in flume experiments. See preview of Chapter 6: Kinematics of the Particle Flux, for further explanation.

Plot of 14 experimental realizations of the number flux q_{nx} with (black line) expected flux <q_{nx}> for D = 11.3-16 mm particles. The expected value <q_{nx}> is estimated as the time average of the full time series. Similar results occur for other sizes in the gravel mixture.

For those interested in sediment transport in rivers:

We've all stared at log-log plots involving a gazillion data points of bed load transport rate versus shear stress and noticed that the residual variance about some selected empirical flux-stress function typically spans...

1/3

tri-color Australian Shepherd wading in small creek looking under a wooden plank bridge for a footpath in an oak-hickory forest

Sometimes Rowan forgets to look, but today she responsibly and capably checked for trolls under all the bridges — and she’s not shy about chasing them out!

#dogs
#dogsofbluesky
#AustralianShepherd

6.4 Properties of Time Averaging
6.4.1 Uncertainty with Time Averaging
6.4.2 Moving Average
6.5 Local Versus Nonlocal Transport

Here is a link to the Preface and Chapter 1. These describe the material covered in the book, including its philosophical basis.

4/4

cdn.vanderbilt.edu/t2-my/my-prd...