GIS and the application of a model of pollen deposition and dispersal: a new approach to testing landscape hypotheses using the POLLANDCAL models

The recent advances in computer-based modelling of pollen dispersal and deposition through the POLLANDCAL (POLlen-LANDscape CALibration) network have enabled landscape simulations to be constructed and tested, both to validate the theoretical models, and to validate modelled landscapes using empirical data. This paper presents the first attempt to develop simulations based on real landscapes, by placing distinct vegetation communities using reclassified digital elevation models (combined slope and altitude data). The approach is tested by modelling a palynological signature for the expansion of mixed agriculture into an upland margin, and the results for 49 simulated pollen deposition sites, across an altitudinal range for a series of landscape scenarios, is compared with existing empirical data. The broad trends within the two data sets are in general accordance, and variation between them is attributed to local variation in vegetation heterogeneity. The results show that the POLLANDCAL models offer a new and exciting method by which sub-fossil pollen assemblages can be examined, theoretical models of landscape use tested, and will offer a significant tool in the interpretation of environmental data in archaeology, in particular when combined with GIS-based models of the physical environment.