The research and use of archaeological predictive modelling has few examples in areas where the landscape's current use is most dominantly ploughing like in Hungary. As we think that such models can be a solution for the numerous problems generated by the low number of known sites in Hungary, our workgroup in the past years has investigated the possibilites of archaeological prediction in well-examinable areas. Archaeological prediction in mostly arable areas has some opportunities: as settlements can easily be detected by plough-walking, one can easily check the accuracy of input data, and such as easily verify the results. With these opportunity available nearly everywhere in the test areas we can have a good overview of the usability of our models. Our test areas are mostly low-lying wetlands, so the accuracy of the DEM is another major issue. We made the first models with DEM generated with ASTER, but it became obvious that in such areas this accuracy might not be enough. With the help of EUFAR Transnational Access we aquired the raw lidar data of one test area and learned to process it, therefore we had the opportunity to compare the accuracy of the models generated from the different DEMs. In our paper we would like to (1) present our current state of research with archaeological predictive models in Hungary; (2) draw attention to the opportunities and differences of the prediction opportunities of mostly ploughed areas; and (3) compare the models generated with different resolution DEMs with the field verifications of the aquired data.
