GFDL - Geophysical Fluid Dynamics Laboratory

Skip to main content

Isaac Held's Blog

1. Introduction

http://www.youtube.com/watch?v=UhTEhhZeMfE

Infrared radiation emitted to space simulated by an atmospheric model under development at GFDL. (1 frame/3 hours for one full year, starting in January).

 

My goal in this blog is to provide a forum for discussion of climate dynamics, with an emphasis, but not an exclusive focus, on climate change.  The level of discussion is meant to be appropriate for graduate students in atmospheric and oceanic sciences, but I hope that this type of discussion is also useful to students in other fields with good applied math, physics and/or engineering backgrounds, to practicing scientists in other fields, and to some of my own colleagues.  Different threads will probably focus on different parts of this intended readership.

Comments will be heavily moderated to maintain a tone and a level of discussion appropriate for the intended audience.  Moderation will likely be slow. Comments must be closely related to the topic under discussion. I  hope to post something every other week, on average.

I am employed by NOAA (and also lecture and advise graduate students at Princeton University).   The opinions that I express are mine and not official positions of NOAA.  However, I consider working on this blog to be fully consistent with NOAA’s outreach and communications policies.

I call myself an atmospheric or climate dynamicist/theorist/modeler.  I am sure that there are philosophers of science who distinguish between the terms “theory” and “model”, but I don’t.  I work with a range of theories of different kinds; when these reach a certain level of complexity they are typically referred to as computer models. The most relevant distinction relates to the purpose of the model.  Some models are meant to improve our understanding of the climate system, not to simulate it with any precision.  I like to talk about building a hierarchy of these models designed to improve and encapsulate our understanding.  The most comprehensive models can be thought of as our best attempts at simulation, limited by available computer resources and our understanding of the effective governing dynamics on space and time scales resolvable with those resources.

Here is an example of a very simple model consisting of two coupled linear ordinary differential equations:

c  \,dT/dt \, = - \beta T - \gamma (T - T_0) + \mathcal{F}(t)

 c_0 \, dT_0/dt = \gamma (T - T_0)

 T and T_0 represent the perturbations to the global mean surface temperature and deep ocean temperature resulting from the radiative forcing  \mathcal{F}. This model is used in a recent paper by myself and several colleagues to help frame the discussion of what we refer to as the recalcitrant component of global warming.

The animation at the top is a small part of the output from another model that a group of us have been analyzing lately, a global atmospheric/land model living on a grid with approximately 50km spacing in the horizontal.  (One can think of the atmospheric component of this model as 37,519,200 coupled ordinary differential equations — not that this is a good measure of the complexity of the model.)  Shown in the animation is a full year of the infrared energy emitted to space  (black is high emission, white is low emission.)  What one sees mostly are the simulated high clouds that provide cold weakly emitting surfaces, but if one looks carefully one can see the diurnal cycle in the emission from the surface, which provides a feeling for the rate at which time is passing.  Notice the sharp distinction between the mid-latitude atmosphere (dominated by non-linear waves) and the tropical atmosphere (dominated by smaller scale moist convection).

The model is introduced in this paper.  It is initialized at some point in the past (about 20 years before this animation loop) and is constrained only by imposed boundary conditions over the ocean and sea ice.   In a full climate model, the state of the oceans and sea ice would evolve freely as well.  Comparing this particular simulated space-time field with observations in ways that are most informative about model deficiencies and the reliability of the model for various applications is a formidable challenge.

The two-box model and this high resolution atmospheric model illustrate two very distinct elements in the hierarchy of climate models. I’ll discuss both models in the next few posts.  My own work seems to gravitate towards creating models intermediate in complexity between these two limits, in an attempt to both increase our understand of the climate and provide ideas on how to improve our high-end models.  See this essay for a discussion of the importance of model hierarchies.

[The views expressed on this blog are in no sense official positions of the Geophysical Fluid Dynamics Laboratory, the National Oceanic and Atmospheric Administration, or the Department of Commerce.]

8 Responses to “1. Introduction”

  1. Jessica Kleiss says:

    Hey Isaac,

    Great idea for a new blog! And I love the animations. I teach courses in Environmental Studies (although I’m a Physical Oceanographer), and I think resources like this online will be a huge benefit to the exposure of concepts in our fields to students and educators!

    Keep up the great work with your blog!

    - Jessica

  2. Chris Colose says:

    Hello Dr. Held,

    There’s virtually no blogs on the web aimed at your target audience on climate science, so this is very exciting, and I look forward to your posts.

  3. Alexander Harvey says:

    Dr Held:

    This video series may be of some interest to those that pass by here:

    Mathematical and Statistical Approaches to Climate Modelling and Prediction

    http://www.sms.cam.ac.uk/collection/870907;jsessionid=3ACCC619D2789C97963B065333655F64

    It is a series of presentations mostly by modellers to modellers including various approaches to modelling, data assimulation, statitical emulation of simulators, exploring parameter space, how to construct climate model experiments (ensemble design as opposed to ensembles of “missed” opportunity {their joke not mine}), how to build parameterisations, and much more.

    Alex

  4. Edwin Kite says:

    Thank you for taking the time to write this blog. It’s a important service to explain things at a level that can be understood by scientists working in neighbouring fields.

  5. Ron Cram says:

    Dr. Held,
    You write “Some models are meant to improve our understanding of the climate system, not to simulate it with any precision.”

    True, and these models have value. Models which attempt to represent the climate precisely and from which researchers make claim about the future do not have any predictive value. The faith researchers put into these models is unwarranted and dependent on muddle-headed thinking. I have seen modelers talk of computer runs as “experiments.” Experiments are only performed in nature and in the lab, not in computer runs. I have even read modelers write words to the effect they were working “on something real.” I’m sorry to be blunt, but this is delusional.

    I used to invest in the stock market on the basis of computer models. Stock prices move on the basis of known laws including the law of supply and demand. (A great deal of computer trading still goes on, but it is mainly arbitrage – not longer term trading.) I was able to pick a number of variables and could perfectly hindcast the broader stock market or major market segments. The problem, of course, is that conditions change. The stock market is a chaotic system. Much like climate, the number of forces affecting the stock market is still unknown and future changes of the known forces (both short and longer-term) is impossible to know.

    To anyone who thinks computer modeling can foretell the future, I highly recommend the book “Useless Arithmetic” by Orrin Pilkey of Duke University and his daughter. Pilkey is an environmentalist who has extensive experience with computer models of shorelines. While computer models of shorelines are interesting, they are always wrong in the long-term.

    • Isaac Held says:

      Ron, with respect:

      I am not interested in comments like “the stock market is so-and-so therefore climate models are etc.” I am interested in comments that address my arguments about the climate system directly. Several of my posts are precisely concerned with what I refer to as the “argument from complexity” that you seem to support. I am totally serious in post #9: when I am confronted with this argument my initial response is “Well, summer IS warmer than winter — maybe climate isn’t all THAT complicated”. Several of the other posts are meant to introduce my view that the forced climate response to increasing CO2 is likely to be quite simple and linear in large part, just like the seasonal cycle, despite the complex and chaotic internal variability superposed.

      I am also not interested in getting into arguments about semantics. Computational “experimentation” and model-generated “data” are standard terminology in a lot of fields, with no implicit implications about the realism of the underlying model. If you don’t like this terminology that’s fine but there are far more interesting things to worry about.

      Trying to understand a model often requires you to put everything else aside, willingly suspending your disbelief and treating the model as your universe, much as if you were trying to enjoy, and maybe even understand, a novel. My wife and I were reading Don Quixote out loud recently, and after a session we would talk of the characters as if they were real. I would not call our state of mind “delusional” (never quite reaching the level of Woody Allen’s Kugelmass.)

  6. John Puma says:

    Dr. Held,

    Thanks for the great blog.

    Would the simulated, emitted-IR clip be enhanced by a superimposed Day/Month counter?

    John Puma

Leave a Reply

Your email address will not be published. Required fields are marked *

*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>