6.24 A&G•December2006•Vol.47 T he aim of the meeting, held on 9 and 10 March 2006, was to take a fresh look at the scaling of Earth structure and dynam- ics, and its implications for practical problems such as natural-hazard prediction, subsurface engineering and climate change. The idea was: ● to examine new data critically for evidence of scale-free or scale-dependent behaviour; ● to suggest physical or numerical models that can explain the emergence of characteristic or scale-free geometries or time series; ● to address the issues raised in scaling up obser- vations for applications such as fluid flow or geo- physical observation; and ● to discuss the implications for predictability in Earth structure and dynamics. Invited speakers were Rachel Abercrombie (Boston University), Timothy Lenton (University of East Anglia), Patience Cowie (University of Edinburgh), Brian Berkowitz (Weizmann Insti- tute of Science, Israel), Michael Worthington (Imperial College), Massimo Cocco (National Institute for Geophysics and Volcanology, Rome) and Donald Turcotte (University of California at Davis). The meeting incorporated this year’s Bullerwell Lecture, “Understanding the Earth’s magnetic field through observation and theory”, by Andrew Jackson (ETH, Zurich). Over the two days, there were four oral ses- sions and one poster session. Discussion was made a priority by the organizers: each talk had ample time for audience questions and discus- sion, and each oral session included 30 minutes for discussion between the panel of speakers and the audience, which continued in the lunch and coffee breaks and at the wine reception. Begin at the beginning This opening session concentrated on “primary observations” in the Earth sciences in the broad sense, together with their scale-free or scale- dependent behaviour. Specific topics included earthquakes, wildfires, outcrops, sea ice, turbu- lence, fracture and folding. The session began with an extensive overview of scaling aspects of earthquakes, in space and time, and proceeded to the lack of scale dependence found for wild- fires with ecological and risk implications. After this, there was a presentation on 3-D methods for gathering extensive data at the outcrop level, and corollary of scaling issues for folds and frac- tures. The final presentation related fracturing observations and associated turbulence in the Arctic sea ice cover. The discussion on primary observation scaling issues was lively, including these common themes: ● whether primary data are best-fit by power- laws (in time or space) or something else, and how does one tell? ● using incremental vs cumulative distributions; ● how do power-laws (in time or space) termi- nate in nature? ● is the self-similar scaling that is observed in natural phenomena of any practical use? Modelling: all or nothing? Here speakers addressed a range of topics con- cerning the modelling and in particular the predictability of complex geophysical systems. Scale in structure Ian Main, Bruce Malamud, Chris Bean and John McCloskey summarize the presentations and lively debate at the British Geophysical Association’s annual British Discussion Meeting on Scale-Invariance and Scale-Dependence in Earth Structure and Dynamics. 1: The line of the San Andreas Fault across Carrizo Plain, California, looking northwest. This landscape is a result of repeated episodes of slip on a non-planar fault surface: the uplifted block is on the right in the foreground, and on the left in the distance. (Figure 22, USGS Open-File Report 83–98) 2: Scaling matters for natural hazards such as earthquake shaking. This map shows relative shaking hazards in the United States and Puerto Rico over a 50-year time period. The probability of strong shaking increases from very low (white), to moderate (blue, green and yellow), to high (orange, pink and red). Map not to scale. (USGS) MAIN ET AL. : MEETING REPORT