ground- water, geo- statistics, environmental- engineering, earth- science

AGU, Day 2 & 3

without comments

There are so many things going on over here, it’s hard to keep up. Additionally, the first time in my life, my body has issues adjusting to a new time zone. I don’t understand… 🙂

Langbein Lecture

This lecture, presented by John Schaake, was one of the better lectures I’ve attended in a while. He was humorous, had great examples illustrating the points he made, and was well organized. John Schaake was employed by the Hydrologic Department of NOAA. It’s hard to summarize his talk, but here are a few points:

  • understanding of basic math and how to apply it was the key thing he learned in his education
  • he emphasized that you have to use both for understanding of natural phenomena: statistics and physics
  • he mentioned a typical problem of downscaling a couple of times: he said you could use a big IPCC cell for precipitation from Ethiopia, downscale, and it’d work in California. It’d work, despite the fact that most of the precipitation in Ethiopia occurs in summer, and most precipitation occurs in California in winter.
  • if you have to or want to make adjustments to non-linear errors, it’s hard to say how to do this (I would probably have to draw a little sketch here, but that’ll have to wait)
  • John Schaake is instrumental in the HEPEX initiative
  • He said something about community based planning, I wrote down CHPS, and all google finds seems to be related to “community based health planning”. I’m not sure where this is going
  • He used an empirical copula, and I knew what he was talking about. 🙂
  • I learned what a “Continuous Rank Probability Score” (CRPSS) is;
  • it’s always good to be reminded of the total probability law
  • he asked this interesting question: “Can we produce ensemble analyses that represent spatial and temporal scale-dependent analysis uncertainty?”
  • he thinks that several of the best imperfect models may be better than any one
  • this might be an interesting document to read: the strategic science plan of NOAA’s hydrologic department

Talk by Yoram Rubin

Yoram Rubin gave an excellent overview of the development of stochastic hydrogeology. He divided this development into four phases:

  1. Field scale experiments and related analysis (Borden, Cape Cod, MADE); The learning-effect was that solute transport is controlled by spatial variability (of K);
  2. Establishment of a physical basis for geostatistics with major contributions from Fogg, Ritz, Weissman;
  3. Adoption of probabilistic concepts by “bottom line” oriented organizations;
  4. New methods of data acquisition, which he refers to as “Hydrogeophysics”, which is also the title of one of his books;

things I picked up

  • I haven’t heard of the Anscombe’s quartet before, but sure I will use it at least for teaching purposes!
  • somebody said a macroscope is a tool to see the infinitely complex;
  • the CUAHSI site looks interesting;

papers maybe worth to read

Written by Claus

December 17th, 2009 at 7:27 pm

Posted in

Leave a Reply