SELF-ORGANIZED CRITICALITY IN CLOSED ECOSYSTEMS: CARBON DIOXIDE FLUCTUATIONS IN BIOSPHERE 2
✍ Scribed by CRONISE, RAYMOND J.; NOEVER, DAVID A.; BRITIAN, ANDREW
- Publisher
- John Wiley and Sons
- Year
- 1996
- Tongue
- English
- Weight
- 481 KB
- Volume
- 16
- Category
- Article
- ISSN
- 0899-8418
No coin nor oath required. For personal study only.
✦ Synopsis
A little understood question in climate and ecological modelling is when a system appropriately can be considered in statistical equilibrium or quasi-steady state. The answer bears on a host of central issues, including the ability of small perturbations to cause large catastrophes, the constant drift of unsettled systems, and the maximum amount of environmental control theoretically possible. Using Biosphere 2 records, the behaviour of carbon dioxide fluctuations was tested for correspondence with theories now known collectively as self-organized criticality. The signature of agreement with other large, composite systems, including forest tires, stock markets, and earthquakes, is a common frequency spectrum or power-law correlations. In this case, the largeand small-scale ends of the spectrum share a common driving force and consequently no single cut-off exists for excluding or ignoring small environmental changes. From the Biosphere 2 carbon dioxide data, the fluctuations in internal atmospheres vary in both small and large steps. The time fluctuations were examined as they varied over 2 years and over three orders of magnitude in fluctuation size, then binned into characteristic size classes. The statistics show a power-law scaling exponent of -1.3, compared with -1 for classical flicker noise (l/fspectrum) and -2.5 for analogous sand-pile experiments developed to test the predictions of a self-organized, critical system. For comparison with open ecosystems, the Byrd climatic record of global COz over the last 50 ka has a similar power-law relation but with -2.3 as the scaling exponent. For generalizing self-organized criticality, the design suggests that otherwise unrelated biological and physical models may share a common correlation between the frequency of small and large length-scales or equivalently exhibit temporal similarity laws. The results potentially have wide implications for environmental control in otherwise chaotic or difficult to predict ecological behaviour.