PALAEOASSOCIA as a methodological tool for phytosociological analyses is further developed

Vegetation History and Archaeobotany https://doi.org/10.1007/s00334-023-00928-y ORIGINAL ARTICLE PALAEOASSOCIA as a methodological tool for phytosociological analyses is further developed Otto Brinkkemper1 · Mans Schepers2 · Onno Van Tongeren3 Received: 29 December 2022 / Accepted: 2 April 2023 © The Author(s) 2023 Abstract The earlier version of PALAEOASSOCIA involved a considerable input of manual labour in sorting species tables with association data to identify plant communities that could have been present. A large archaeobotanical dataset from the site of Best (The Netherlands) was used to judge whether this manual sorting results in subjective results. As these were found, we developed a fully automatic version of PALAEOASSOCIA, including this sorting process. Likelihood clustering with prior probability yielded the highest number of associations recovered from four samples, and was therefore chosen as the optimal clustering method. The sorted tables are automatically converted to syntaxonomical groups. The hierarchical level of these groups can be pre-defined by the user of the program. Syntaxa that are highly improbable geographically cannot be ruled out a priori, but need to be removed manually. PALAEOASSOCIA is not meant to replace other methods of ecological interpretations of archaeobotanical data, but instead as a tool to obtain a more detailed result. Keywords Phytosociology · Methodology · Syntaxonomical groups · Likelihood clustering Introduction Detailed understanding of the environment surrounding an archaeological site has always been a major goal in archaeobotanical analyses. Various methods have been applied that take the individual species as a starting point. Indicator species can be defined to make a judgement upon specific environmental properties (e.g. Salicornia europaea: salinity, Sphagnum spp: acidity). Other approaches apply indicator values (e.g. Ellenberg et al. 1991) to all species in a spectrum. From an ecological perspective, however, a disadvantage of this individualistic approach is that it largely ignores the Communicated by M. Ptáková. Mans Schepers 1 Cultural Heritage Agency of the Netherlands, Rijksdienst voor het Cultureel Erfgoed, PO Box 1600, Amersfoort 3811 BP, The Netherlands 2 University of Groningen, Centre for Landscape Studies, PO Box 716, Groningen 9700 AS, The Netherlands 3 Retired from Data-Analyse Ecologie, Vrij Nederlandstraat 57, Arnhem 6826 AW, The Netherlands interplay, or sociology, between plant species. The spatial manifestation and co-occurrence of plant species is known as vegetation. A basic understanding of vegetation is often acquired by grouping species in ‘ecological groups’ based on individual labelling of species such as, for example, ‘arable weed’, ‘grassland species’, or ‘ruderal’ (e.g. Arnolds and van der Maarel 1979). Another approach is phytosociology, the study of plant communities (syntaxa). Plant communities are defined on the basis of field observations of co-occurrence of species. Although substantial variation in field methods and research density exists, numerous systematic vegetation recordings (relevés) are available in most countries (Schaminée et al. 1995a). In the Netherlands, this adds up to more than 600,000 at present. The ASSOCIA software package (van Tongeren et al. 2008) assigns plots from field observations to pre-defined and well-established plant communities. Schepers et al. (2013) developed a method to divide an archaeobotanical sample into overlapping species groups. PALAEOASSOCIA more or less treats archaeobotanical datasets as modern plots. A major challenge to overcome is the fact that the vast majority of archaeobotanical assemblages contain plant species from various environmental origins. Essentially, what the method aspires to, is to split this environmental mixture 13 Vegetation History and Archaeobotany In this paper we describe the developments in the methods and the results of the different methods are compared. Site and sample description Fig. 1 Simplified visualisation of the three main steps being used in PALAEOASSOCIA (after Schepers 2014) into various groups of taxa that may have grown together, and then assign these ‘subgroups’ to one or more vegetation types in varying degrees of ‘fitting’, which may result in more than one possible syntaxon for one group of species (Fig. 1). However, only the most likely syntaxon for each species group has been used in this study. The manual formation of groups as seen in the centre of Fig. 1 proved very time-consuming in the 2013 version of the computer program, required substantial practice and experience, and was therefore potentially subject to interpersonal differences. Thus, we felt the need to develop an automated, user-friendly, and faster version and compare the results with manual formation of groups by two experienced and one unexperienced person. A rich dataset from the Southern Netherlands village of Best provided an excellent test set for this further development. Fig. 2 Location of Best on a soil map of the Netherlands 13 The village of Best is located in the province North Brabant (Fig. 2). The samples were obtained in preparation of major restoration works on a monumental farm building, ‘De Armenhoef’. Dendrochronological research revealed that the oldest parts of the building date back to 1263, the oldest farm still standing in the Netherlands and probably its wider surroundings. Detailed corings in the barn area revealed the presence of dung-rich layers at various depths (de Kort et al. 2016). Sampling A total of 15 samples for archaeobotanical research were obtained from these dung-rich layers in seven corings by archaeologist J.W. de Kort (Cultural Heritage Agency, The Netherlands). Methodological details can be found in de Kort et al. (2016). As was expected considering the context of the samples, the species found in each individual sample originated from various vegetation types and became mixed during deposition of material from various sources in the sunken byre. The dataset with a large number of species from a large variety of vegetation types was ideally suited for a more detailed analysis by means of an improved automated version of PALAEOASSOCIA. Vegetation History and Archaeobotany Methods Archaeobotanical analyses The analyses resulted in a rich and diverse dataset, representing 130 unique taxa. Statistical analyses The species lists of the 15 samples, different in species richness, were analysed with PALAEOASSOCIA. ASSOCIA, the original program (van Tongeren et al. 2008), as well as PALAEOASSOCIA are based on Bayesian statistics. These are used to estimate the conditional probability that a species list is a random sample from a certain syntaxon. The basic assumption is that this conditional probability is sufficient to discriminate between syntaxa, thus linking a species list to a certain syntaxon. The so-called prior probability of finding a syntaxon in a (...truncated)