Reconstructing the heterogeneity of past woodlands in anthracology using the spatial distribution of charcoals in archaeological layers: Applied to the postglacial occupation of the Abeurador cave (Hérault) in the South of France.

Previous methodological developments in the field of anthracology have led to the definition of rigorous sampling methods in archaeological contexts, which provide highly accurate and reliable taxonomic frequency spectra as a basis for understanding past woodland vegetation. However, the fuelwood taxon frequency spectra represent a synthetic picture of forest composition in the environs of an archaeological site. Our aim in this paper is to describe the heterogeneity of woodland distribution. The Abeurador cave (Hérault) in the south of France, has previously provided an exceptionally rich palaeoecological study. We further explore the spatial frequency distributions of the charcoal taxa found in charcoal scatters from two layers dated to the Mesolithic (Preboreal). The frequencies of abundant taxa have a unimodal distribution in the sampled grid squares, close to the normal distribution. The ratio s2/m (variance/mean) termed the 'distribution index', estimates the regularity of distribution of each taxon in the grid squares. We hypothesized that the frequency distribution of a taxon in a layer could be considered to be its statistical sampling distribution (i.e. probability distribution) in the firewood collection catchment at a given time, since the quantity of this taxon fluctuates in each firewood assemblage according to its probability of collection, relating to its distribution heterogeneity (or accessibility) in the past environment. The distribution index was then used to estimate the heterogeneity of the palaeoenvironmental distribution of each taxon and thus its random, uniform or clustered distribution. On this basis, we propose a mapped restitution of the evolution of the main woodland species in the surroundings of the Abeurador during the Early Holocene, integrated with the whole sequence dynamics. [ABSTRACT FROM AUTHOR]