Some 2600 bog oaks have been dated from German, Dutch and Irish bogs covering the period 6000 bc to ad 1000. The ring patterns of these 'bog oaks' are characterized by recurrent, long-term growth depressions. In addition, obvious changes in the temporal distribution of the bog-oak trunks throughout the Holocene are found. Both features were probably caused by unfavourable growth conditions, which are most likely linked to changes in site hydrology. We use a new variable, 'annual mean age', as a tool to analyse the population dynamics of bog oaks in more detail, enabling the detection of synchronous intersite and interregional changes. It is calculated as the arithmetic mean of the age of all trees in each calendar year. We performed the calculation on regional (Germany, The Netherlands, Ireland) subsets of the bog-oak series. Abrupt changes in annual mean age are taken to indicate periods of generation change. We nd good agreement for the interval from 5000 bc to 2000 bc between the continental (combined German and Dutch) and the Irish mean-age chronologies. Most changes in population dynamics correspond with contemporary changes in the associated regional tree-ring chronologies. It is concluded that the observed changes in population dynamics and growth activity are responses to common environmental forcing, most likely related to climate.

No abstract

No abstract

During the Roman and early-medieval period in the Netherlands, an extensive network of routes connected settlements on the local, regional and supraregional scale. The orientation of these route networks in part was determined by settlement locations, and in part by environmental factors (e.g. soil type, relief). Therefore these route networks provide a key in understanding the dynamic interplay between cultural and environmental factors.This study focuses on modelling Roman and early-medieval routes using a multi-proxy approach. By combining network friction with archaeological data representing settlements, burial sites and shipping-related finds we wish to investigate the possibilities of using these large-scale datasets for modelling Roman and earlymedieval route networks in the Netherlands. Data representing past infrastructure and isolated archaeological finds were used to validate the model output.Results show that in geomorphologically diverse lowland regions, such as the Netherlands, network friction is extremely useful for modelling historical route networks. We found a clear relationship between environmental conditions, settlement locations and the spatial distribution of infrastructure. Using evidence-based modelling, we were able to correctly predict the location of 89% of the currently identified Roman infrastructure, and 85% of the known early-medieval infrastructure in the Netherlands within a 1000 m buffer. Additionally, despite only roughly covering a surface area of 13% in the Roman and 11% in the early-medieval period of the Netherlands, 82% and 72% of all known isolated finds were located within the same buffer.