The paper deals with the definition of a conceptual model for groundwater flow systems in Apennine watersheds mainly in siliciclastic turbidites. In general this lithology has been poorly studied and the scarcity of data has not allowed a "whole watershed" systematic study of this kind of fractured aquifer/aquitard. The boring of a series of tunnels and the related collection of new hydrogeological data for the High Speed/High Capacity railway which connects Florence with Bologna, has led to a renaissance in the topic. Four watersheds are taken into account. They are located within a large outcropping area of the Marnoso-Arenacea Formation across the main Apennine divide in the Tuscan-Emilian Apennines. The Marnoso-Arenacea is a Middle Miocenic turbiditic formation pertaining to the Umbro-Marchigiano-Romagnola succession and is formed by a more or less regular alternation of siliciclastic arenaceous and marly hemipelagic beds (Ricci Lucchi, 1981). The unit was deposited in a deep marine basin as a siliciclastic cone fed from the sedimentary material eroded and transported away from the Alpine and Apennine orogens. Different lithostratigraphic members of the unit are particularly defined according to the arenite/pelite ratio. Since 1996, the western edge of the most extensive outcrop in the Northern Apennines has been affected by the boring works for the tunnels. The project is characterised by a total underground length of 73 km (92% of the entire track of 79 km) subdivided into a series of tunnels separated by short open-air portions. The tunnelling was completed in 2005. The huge and rather constant amounts of groundwater locally drained by the tunnels, together with the effects on the hydrogeologic system (Canuti et alii, 2002), bear evidence to an "aquifer-like" behaviour of the Marnoso-Arenacea, giving a renewed interest in the study of Groundwater Flow Systems (GFS), sensu Toth (1963, 1999), in such a lithology. The tunnel boring was a chance to improve our knowledge in the hydrogeology of Northern Apennines turbidites. The huge amount of data collected by the hydrogeological monitoring performed since 1995 (Agnelli et alii, 1999) is of great interest to our study. Raw data were collected during 9 hydrological years from 1995 to 2003 both on natural hydrological systems and on tunnel waters. Hydrogeological data on springs and creeks are related to 47 and 11 monitoring points respectively during a time interval of hydrological years from 1994-1995 to 2003-2004. The chance to record spring hydrographs on many successive hydrological years not only shed light on the typical Apennine subtropical hydrologic regime but also on some specific differences between distinct Groundwater Flow Systems. In order to define spring hydrogeological properties annual average discharge (Q <sub>m</sub> ), recession period average discharge (Q <sub>s</sub> ), Variability Discharge Index (VDI) and Maillet recession coefficient (alpha ) (Civita, 2005) have been calculated, based on a significant number of available measurements. The behaviour of a spring during the recession period is the most important hydrographic feature which provides evidence on the typology of Groundwater Flow Systems discharging at that point. The recession coefficient of creeks is related to the whole groundwater contribution to the recharging of surface water, including both point-type springs and linear ones and also the fractured zones intersecting the creek bed. During summer the base flow discharge of creeks is generally higher than the summation of the discharge of all the point-type springs, so the discharge is strongly affected by a direct hydraulic connection between the aquifer and the creeks; the aquifer recharges creeks through relatively higher permeability fractures, related to either large tectonic structures (in regional flow systems) or to de-tensioned and weathered rock mass (watershed or local flow systems). As far as hydrochemistry is concerned, water analyses showed that most of the samples (springs, creeks and some tunnel water inrushes) belong to the alkaline-earth-bicarbonate facies and this indicates an active shallow groundwater circuit and high renewal rate of the system. Some of the tunnel samples presented an enrichment in Na (super +) ions and a higher electrical conductivity, giving evidence of a deeper and "geochemically confined" water circulation system where an extensive rock/water interaction takes place. As the comparison between chemical composition and occurrence of water inrushes clearly showed, ionic concentration trends depend on the spatial/temporal distance between inrush occurrence and sampling collection. Na-rich samples are generally collected at a short distance from the advancing front of the tunnel boring and/or just after water inrush occurrence. A groundwater flow conceptual model for turbidites based upon a hierarchy of at least 4 groundwater flow systems is proposed.