When we approach the problem of archaelogical landscape reconstruction, often we have to deal with a complex context, that can be hardly interpreted, even by researchers. In our case study, in which we are currently investigating and reconstructing the thermal landscape typical of the Euganean Hills, around Montegrotto Terme (Padova, Italy), within the project Aquae Patavinae VR (in coorperation with University of Padova, funded by MIUR and Veneto Region), the known elements  are different and spread along a wide area: there is only a main/primary archaeological site, that is open to the public; others are still under excavation and cannot yet be seen; others are recognizable only through little scattered evidences identified through archaeological and geological surveys, historical studies and remote sensing of the whole area.

It’s quite evident that we are facing a well-known scientific problem, quite recurrent in archaeology, that regards the interpretation and reconstruction of an archaeological landscape as a single global entity. This issue becomes more and more relevant in the case of a project directed/targeted to  non-expert users, through on line virtual archaeology tools. The problem has therefore two faces: one directed to the “user-researcher” and the second one directed to the “user-visitor”.

We are dealing with this issue adopting two approaches, that take into account the London Charter (especially for data transparency) and Seville Principles:

1. developing an ad hoc workflow whose goal is to build reality-based models, acquired on the field,interpretative/simulation models (connected to the first ones) and not-reality-based models (connected with the second ones), through procedural and manual modelling, both based on historical sources and excavation results. All models, including the 3d paged terrains, are stored on a server together with metadata, accessible only by registered users through a back-end (a CMS accessible on line through a normal browser);

2. evolving the 3d on line exploration plug-in, OSG4WEB (based on the open source library OpenSceneGraph), already developed by CNR ITABC and CINECA, adding new specific features such as: 

   navigation system personalized in accordance with the type of exploration, the scale of the landscape, the visulisation device (fly, walk; touchscreen, natural interaction); 

   specific user interface through which the user can upload on the landscape (all data are GIS-based) both interpretation/simulation models (transparent volumes above the original 3d archaeological remains) and reconstructed hypotethical models;

   loading system for information in 3d related to used sources and level of reliability;

   dynamic real time system to add plants directly on the scene, useful in case of gardens reconstruction;

   optimization system for on line 3d exploration.