Madison Chaos and Complex Systems Seminar

Spring 2006 Seminars

All seminars are Tuesday at 12:05 pm in 4274 Chamberlin except as noted.

Short List

Join us for lunch during the summer on the Union Terrace at noon each Tuesday, starting May 9th!


January 17, 2006

Is global warming for real?

Clint Sprott, UW Department of Physics

Probably, but if you're a skeptic, this talk will not likely convince you. What I will do is use Clay's ice core data and John Magnuson's Lake Mendota ice cover data from last semester in a case study of various time-series analysis techniques, both linear and nonlinear. I will show the limitations and weaknesses of the various techniques and the wide variety of predictions that they make. I will show the chaotic output of an artificial neural network trained on the data and demonstrate a new computer program that others can use to make time-series forecasts in fields as diverse as meteorology, ecology, and finance. Finally, I will make a prediction of when the ice will thaw on Lake Mendota this year.

This talk is available as a PowerPoint Presentation.

January 24, 2006

Investigating the dynamics of complex systems through the power law distributions they generate

Cristie Hurd, UW Department of Zoology

What do sand pile avalanches, earthquakes, stock market crashes, hits on Internet sites, extinctions of species, and social insect foragers have in common? All belong to complex systems characterized by a stable, scale-free distribution of event sizes. Though orders of magnitude less frequent, the sheer size of the most extreme events dramatically alters the system’s behavior. In some systems the consequences can be devastating. One example is the death toll from Pakistan’s earthquake last October. In other systems the self-similar distribution may indicate that the system has self-organized to an optimal trade-off between efficiency of energy flow through the system and system robustness. Current ideas on the emergence and classification of complex systems characterized by a power law distribution of event sizes will be discussed. Data on the allocation of labor in a social insect colony will be presented as an example of a scale-free system.

January 31, 2006

Brain waves on potential energy surfaces: relating particle dynamics to brain state

David Hsu, UW Department of Neurology
Abstract:  A working brain emits electromagnetic fields which can be detected at the scalp with electroencephalograms (EEG).  Since the invention of human EEG techniques in 1924, we’ve learned, on a phenomenological level, that certain waveforms of the brain represent an awake and alert state, other forms represent different levels of sleep, and yet others represent epileptic seizures.  Here we show how to go beyond a phenomenological description, by regarding EEG output as the coordinates of a hypothetical particle.  The hypothetical particle “moves on” an effective potential energy surface.  By extracting pieces of this potential energy surface from real EEG data, we find that seizure onset is associated with escape of the hypothetical particle from potential energy wells and crossings of potential energy barriers.  This picture of a hypothetical particle moving on a potential energy surface allows us to ask new kinds of questions about brain states and changes in brain state.  We discuss this particle-on-a-potential-energy-surface model in relation to chaos theory and persistent reports of widespread oscillations seen on different types of brain wave recordings.

February 7, 2006

Switch-like behavior in biochemical reaction networks

Gheorghe Craciun, UW Departments of Mathematics and Biomolecular Chemistry

Gene regulatory networks, metabolic and signaling networks in cells can be regarded as dynamical systems obeying mass-action chemical kinetics. These systems are usually high-dimensional, nonlinear, and depend on many unknown parameters. In general, chemical systems are capable of arbitrarily complex behavior, but, on the other hand, the range of possible dynamic behavior is often limited or even uniquely determined by the network structure. We will show that reaction diagrams, similar to those that biochemists usually draw, carry subtle information that relates the reaction network's topology with its capacity to exhibit switch-like behavior. We will also discuss implications for the interpretation of experiments in cell biology.

February 14, 2006

Thoughts on communicating global changes issues

Jon Martin, UW Department of Atmospheric and Oceanic Sciences 
Study of global climate change has become a major endeavor in the atmospheric and oceanic sciences within the last 25 years.  Given the considerable potential for societal impact from global climate change, the atmospheric sciences have been increasingly called upon for sound, reasoned guidance in the shaping of public policy.  Unfortunately, the public policy arena is unfamiliar ground for most scientists and is significantly influenced by politics.  Consequently, the central issues in climate change science are often obfuscated by political rhetoric.  This circumstance compels the scientific community to not only continue the basic scientific research on global change, but to also identify means by which the complicated interplay between physical processes and boundary conditions in the climate system can be explained, in their essence, to non-experts.  In this talk I suggest a baseball analogy that may provide an example of the extreme sensitivity of complicated systems to small changes in boundary conditions.

February 21, 2006

Role of deer behavior in transmission of CWD prions

Nancy Mathews, Gaylord Nelson Institute for Environmental Studies 

Chronic wasting disease (CWD), an infectious and fatal disease of white-tailed deer, was first detected in south-central Wisconsin in 2001. Understanding movements of deer is an essential aspect of modeling CWD transmission and for managing the disease in the wild. Home range sizes of individual female deer and social groups are a result of individual movement, social interactions, and behavioral responses to habitat characteristics. While there is substantial evidence that home range size is influenced by factors such as sex, age, body weight, and season, the influence of deer density, harvest, and landscape pattern is not as well understood. My students and I examined individual and social group home range sizes, and movement patterns in relation to habitat fragmentation (forest edge density and the ratio of agriculture to forest), local deer density, and deer harvest intensity. Forest edge density, the only predictor of home range size for both individual females and social groups, was inversely related to home range size. Further, forest edge density was greater within natal gestation home ranges of yearling male dispersers than non-dispersers, but there were no differences in year, the ratio of agriculture to forest, deer density, or harvest intensity. Perhaps the most significant finding of the research, however, is preliminary evidence suggesting that clusters of infected deer may be more linked to geographic location than to deer density. This suggests the potential existence of hotspots of environmental contamination. I will discuss the implications of direct transmission among deer, verses indirect transmission through environmental contamination, and how this might affect future management.

February 28, 2006

Improving short-term numerical weather forecasts using observation from satellites
Robert Aune, NOAA's Advanced Satellite Products Branch, UW Madison
Computer models that generate forecasts of future atmospheric states (1-3 days) have evolved into highly complex systems over the past 30 years.  Increases in computer power allow modelers to include additional realism in their models with the goal of improving forecast accuracy.  Changes to any forecast model must be objectively validated against measured atmospheric parameters.  Historically this has been accomplished using point-source observations with poor spatial and temporal coverage.  In the late 1980's scientists at the Cooperative Institute for Meteorological Satellite Studies (CIMSS), began using observations from geostationary satellites to validate numerical models.  What resulted was the CIMSS Regional Assimilation System (CRAS), a forecast model unique in that its development was guided by validation against truth as viewed by satellites.  I will present a summary of numerical weather prediction concepts, focusing on limitations set by non-linearity, error growth and  predictability.  I will conclude with a brief history of the CRAS forecast model and how satellite observations were used to improve its accuracy.  Real-time CRAS forecasts can be viewed at

March 7, 2006

The N-dimensions of global climate change “complexity”

David Houghton, UW Department of Atmospheric and Oceanic Sciences

A broad overview of the climate change issue reveals the complexity of the physical system that must be considered and the long list of uncertainties facing those trying to make predictions for the future. I share my perceptions arising from writing an education text on this subject and making many presentations to public groups. The relevant physical system includes the land, air, oceans, and biosphere (including most importantly – people). Science hurdles include both observing the system and predicting its changes. It is shown how uncertainties build on each other so that predictions for future climate conditions for local communities along with resulting impacts can be considered still to be quite uncertain. Examples of these uncertainties document the “complexity” of the system and the challenges to both the scientific community and the world community.

March 21, 2006

The role of terrestrial snow cover in the climate system

Steve Vavrus, Gaylord Nelson Institute for Environmental Studies

Snow cover is known to exert a strong influence on the overlying atmosphere and underlying soil, but quantifying this impact is difficult.  Besides its well-accepted ability to cool locally, snow cover can also force climate remotely in complex ways by inducing changes in the atmospheric circulation.  Most research on the impact of snow cover has focused on the regional rather than global scale.  By contrast, this study investigates the global impact of terrestrial snow cover in the present climate by comparing a pair of numerical climate model simulations run with prognostic snow cover (control case) and with all snow cover on land eliminated (NOSNOWCOVER).  In this experiment all snowfall over land was converted into liquid water upon reaching the surface.  Compared with the control run, NOSNOWCOVER produces mean-annual surface air temperatures up to 5 K higher over northern North America and Eurasia and 8 to 9 K greater in these regions during winter.  The global-mean warming of 0.8 K in NOSNOWCOVER is 1/3 as large as the model’s response to a doubling of carbon dioxide and larger than the observed warming trend since accurate measurements began in the late 19th century.  This pronounced surface heating propagates throughout the atmosphere, causing changes in circulation patterns aloft.  Despite the large atmospheric warming, the absence of an insulating snow pack causes soil temperatures in NOSNOWCOVER to fall throughout northern Asia and Canada, including extreme wintertime cooling of more than 20 K in Siberia and a 5 to 10o equatorward expansion of simulated permafrost.  The absence of local melt-water percolation causes significantly drier soils over northern boreal regions and a consequent decrease in cloudiness.  The removal of snow cover also drastically affects extreme weather in middle latitudes.  Extreme cold-air outbreaks (CAOs), defined relative to the control simulation, essentially disappear in NOSNOWCOVER.  The loss of CAOs appears to stem from both the local effect of eliminating the chilling influence of snow cover in mid-latitudes and a remote effect over CAO source regions in the Arctic, where –40oC air masses are unable to form in NOSNOWCOVER.

March 28, 2006

Social Security--in chaos or all as predicted?

Karen Holden, La Follette School of Public Affairs

It is often argued that the U.S. Old Age Security, Disability and Health Insurance (OASDHI) program--known as "Social Security"--is unsustainable due to the underlying characteristics of the program and economic changes that were unpredicted by its framers. This seminar attempts to clarify the purpose and nature of the program and what exactly was expected by the framers of the program (who were actually surprisingly close in predicting today's current beneficiary to worker ratios). It will review what are the data being presented to the public on the financial status of Social Security, why we uniquely have long-term projections of OASDI (in contrast to most other government programs), and what are the issues being raised in current discussions of Social Security reform. Even though OASDI may no longer be a major public policy initiative of the White House, the issues and discussions do continue among economists and policy makers. What are they saying about the future? What can the audience at the seminar expect to see over their lifetime? I won't answer that question, but hopefully after the seminar you may be a bit more informed in shaping your own expectations.

April 4, 2006

On polyclonality of intestinal tumors: statistical and biological analysis indicates short range interaction of initiated clones.

Michael Newton, UW Department of Statistics

The working hypothesis in cancer research is that cells within a tumor have descended from a single initiated aberrant cell. In contrast to this clonality hypothesis, data have emerged which support the polyclonality of intestinal tumors. I will discuss statistical and biological aspects of polyclonality from a recent study of mouse aggregation chimeras. Three important issues are: (1) Random collision: Is it possible that observed polyclonality can be explained by the chance proximity of initiated clones? A hypothesis test is derived using some classical results from stochastic geometry. (2) What is the spatial extent of interaction among initiated clones? Techniques from Bayesian image analysis are used to combine image data on the chimeric patch structure of the intestinal wall with tumor count data. (3) What fraction of tumors are polyclonal? The lower bound proposed by Novelli and colleagues is shown to be flawed owing to a misinterpretation of conditional probability.

This is a joint project with A. Thliveris, R. Halberg, L. Clipson, R. Sullivan and W. F. Dove from the McArdle Laboratory, and S. Stanhope from Statistics.

This talk is available in PDF format.

April 11, 2006

Heating the solar corona: A hot topic in plasma astrophysics

Christopher Watts, Electrical & Computer Engineering, University of New Mexico

The surface or photosphere of the sun is a blackbody with a temperature of about 7500° C, and the basic mechanism that heats the sun, nuclear fusion, is well understood. However, there is a disconcerting paradox: The temperature of the solar atmosphere or corona starts to rise away from the surface to about 1,000,000° C. It's like walking away from a fire S<caron> and you suddenly feel hotter. The energy that heats the corona is almost certainly stored in the magnetic field of the sun. There are two main competing models for how this energy is released: 1) Magnetic waves and 2) Tearing and reconnection of the magnetic field. Both models are probably valid in different regimes. In this talk, I will present an overview of the coronal heating paradox and the two heating models. Then I'll talk about current research by plasma physicists, using both remote observations and laboratory simulations, focused on substantiating these models.

April 18, 2006

Complexity in disease emergence from global climate and ecological change
Jonathan Patz, Gaylord Nelson Institute for Environmental Studies
Health effects of global climate and ecological change will come via multiple exposure pathways, many of which are complex and non linear.  Climate-related health outcomes, for example, heat-related mortality or morbidity, air pollution-related illnesses, and infectious diseases may also be exacerbated by concomitant landscape changes that influence local climate.  Every environmental perturbation impacts the ecological balance and context within which disease manifests itself within populations.  These changes affect the hosts or vectors of disease and the pathogens and parasites that breed, develop, and transmit disease.  In addition to climate change, landscape impacts such as deforestation, human settlement sprawl, industrial development, road construction (e.g., linear disturbances), large water control projects (e.g., dams, canals, and irrigation systems, reservoirs) have been accompanied by the spread of pathogens into new areas. For example, the competence of different anopheline mosquitoes to transmit malaria varies between species, and anopheline species occupy a variety of ecological niches. Moreover, biological systems (e.g, mosquitoes and disease-causing agents within) can amplify the effects of subtle environmental and climatic changes.  These complexities of disease risk stemming from global environmental change will be presented.

April 25, 2006

Central force motion and the physics of running

Jim Reardon, UW Department of Physics

The motion of a human runner is complex yet also familiar. In this talk I discuss a mathematical "toy model" of running humans that is founded on four observations of runners: all successful runners have a cadence above 180 steps per minute; the resultant ground force during stance phase passes through the runner's center of gravity; heel strike is much more energetically costly than leg swing; and energy cost is proportional to the time integral of muscle tension. At issue is the optimal solution of the "runner's dilemma": should one run so that one's center of gravity stays at a constant height above the ground, or should one try to minimize one's ground contact time? Implications of the toy model for real-world training will also be discussed.

May 2, 2006

Self-organization in magnetized plasmas

Michael Brown, Swarthmore College
It is a common process in the universe for plasma and magnetic fields to evolve together in a turbulent way but then rapidly relax to simple, self-organized structures. Solar flares erupt from the photosphere tangled and chaotic, but via a process called magnetic reconnection, they relax and straighten. This process releases energy in the form of superheated plasma and rapidly flowing jets. On a much larger scale (millions of light years), galactic disks collapse, rapidly shedding angular momentum and in the process generate extended, magnetized jets along their axes. On human scales, laboratory experiments are underway seeking self-organized magnetic structures that would be suitable “bottles’’ for a fusion reactor. We present recent experimental results from the merger of two rings of hot, magnetized plasma in the Swarthmore Spheromak Experiment (SSX). During the merging process, the plasma self-organizes to generate a single, large scale (r = 0.2 m, L = 0.6 m), three-dimensional magnetic structure called a field-reversed configuration (FRC). The rate at which the merging proceeds is governed locally by magnetic reconnection in which magnetic fields associated with each ring become shared. The magnetic reconnection rate is fast and fully three-dimensional. Magnetic reconnection converts magnetic energy to heat (up to Te ~ 106 K), energetic particles (Ei > 100 eV), and flow (up to 100 km/s). See for more information.