Fire-driven dynamic mosaics in the Great Victoria Desert,Australia – II. A spatial and temporal landscape model

An explicitly spatial, large scale, high resolution model of fire driven landscape dynamics in the Great Victoria Desert is constructed and parameterized to simulate frequency distributions of fire size and shape obtained from previous analyses of satellite chronosequences. We conclude that probabilities of fire spread cannot be constant over time, and that realistic distributions of fire size and plausible rates of fire spread can be obtained by assuming that fire spread is conditional on observed durations of windy conditions. Landscapes subject to this form of disturbance show large scale correlation structure many times greater than the average dimensions of single fires, and exhibit low frequency quasi-periodic stochastically driven oscillations in proportions of the landscape at different successional states over spatial scales exceeding 100,000 km². Average fire return intervals are 30 yrs. Analysis of patch structure suggests that this landscape is composed of few large younger patches, embedded in a mature sea of surrounding habitat. Intermediate and late successional habitat must exist in more abundant patches somewhat smaller than young habitat. Numerous small patches of mature habitat are likely to be scattered throughout this younger habitat. The model predicts that fire size frequency distributions are relatively insensitive to changes of as much as ±50% of observed fire ignition frequency.     

Fire-driven dynamic mosaics in the Great Victoria Desert,Australia – I. Fire geometry

The dominant ground cover in the Great Victoria Desert is porcupine grass or spinifex, a fire-prone perennial grass that grows in hummocks or tussocks. Lightning sets hundreds of wildfires annually in inland arid Australia, generating an ever changing spatial-temporal patchwork of habitats that differ in their state of post-fire recovery. The spatial configuration of this patchwork is determined by the size, shape, frequency and inter-spatial relationships of fires, and is likely to play a vital role in the maintenance of the desert biota. Chronosequences of satellite imagery spanning the years 1972–1991 are used to extract and describe the geometry of over 800 fires from fire scars. In the imagery study area, an average of 43 fires occur annually, fire size frequency distributions are roughly log-normal with mild right skew, with average area of 28 km², burning between 2 and 5% of the burnable landscape each year. Average fire return interval is estimated to be at least 20 years. These empirical findings are an important prerequisite for developing a more sophisticated understanding of the dynamics of the fire cycle in this ecosystem.     

Validation of a novel estrogen receptor-based microtitration plate assay for the determination of phytoestrogens in soy-based foods.

A novel, nonisotopic microtitration plate assay based on the human estrogen receptor has been used to screen soy-based and soy-containing foods for their phytoestrogen content (measured as genistein equivalents). The validation of the assay for use with food extracts has been demonstrated by investigation of recoveries after acidic and enzymic hydrolysis, by investigation of matrix effects, and by comparison of results with HPLC analysis. Phytoestrogen levels in soy products analyzed ranged between 520 and 1872 microgram of genistein equiv/g of soy flour, 5-282 microgram/g of soy concentrates, 503-1292 microgram/g of soy-protein isolates, and 108-226 microgram/g of soy-based infant formulas. Samples of textured vegetable protein and bread containing soy and linseed gave values of 1114 and 68 microgram/g, respectively. Comparison of results for 12 samples analyzed both by the receptor assay and by HPLC showed good correlation (r(2) = 0.905). The assay, which is rapid and simple to perform, is suitable for screening phytoestrogen-containing foods in order to assess human exposure to these bioactive compounds. The assay sensitivity is 3.4 microgram/g, and 14 samples/plate can be analyzed in 4 h following hydrolysis.