**6. Conclusion**

This chapter presented an experimental study of sediment gravity flows in which six types of flows were distinguished based on a comparison of hydrodynamic, depositional and rheological properties. A phase diagram was created, showing the boundaries between these flow types in terms of rheological behaviour, bulk volumetric concentration and clay concentration. The main characteristics of the flow types are summarized below:

Type I: Low density flow; Newtonian; grains supported by upward component of turbulence; no hindered settling; segregation of grains and normally graded beds; Type II: Newtonian; grains supported by turbulence; turbulent flow with gently undulating highconcentration near-bed layer; partial hindered settling and partial size segregation forming partially graded beds; Type III: Newtonian; fully turbulent flow with strongly undulating high-concentration near-bed layer; hindered settling resulting in rapid deposition and generation of partially graded beds; Type IV: non-Newtonian (plastic); viscous flow; formation of plug and shear flow (mud layer close the bottom); viscous forces cause freezing of the flow and forming graded beds of muddy sand and; Types V and VI: non-Newtonian (plastic); viscous flow with thick mud layer; grain support by matrix strength; weakly undulating internal mud layer (type VI show no undulations); cohesive freezing forms an ungraded muddy sand with coarse-tail grading on top.

The six types of flow/deposits classified represent the transition between the two most known types of sedimentary gravity flows: from turbidity currents (low-concentration, low clay and Newtonian behaviour) to debris flow (high-concentration and high clay content and non-Newtonian behaviour). The experimental study allows the comparison and extrapolation of the results obtained from physical model to natural environments. However it must be considered the experimental simplifications adopted. Apart from that, the rheological properties of mixtures and some hydrodynamic (e.g. cohesion effects) and depositional (e.g. settling velocity) properties are scale-independent and can be applied for further interpretation.

The experiments simulated a single catastrophic event and do not consider a continuous sediment supply from rivers (for instance, plumes and hyperpycnal flows among others) which can change some properties of the flow along time and space. Moreover, the limit of maximum value of volumetric concentration was 35% by volume. In this case, regions III (see Amy et al., 2006) and region VI (see e.g. Hampton, 1972; Ilstad et al., 2004; Marr et al., 2001; Mohrig et al., 1999; Mohrig & Marr, 2003) were left with an open boundary to further experiments and perhaps the creation of a complementary experimental-derived classification of sediment gravity flows.
