**5. Conclusions**

In this study, the University of Melbourne Frontal Tracking Scheme (FTS), a front tracking algorithm based solely on wind related criteria that was employed in Southern Hemisphere, was appropriately modified to identify cold fronts in the Mediterranean, a closed basin with complex topography. It was then used to compile a climatology of these features. The modified scheme (named MedFTS) employs two new criteria, i.e., total wind direction change and total wind magnitude, to better identify the position and tilt of a Mediterranean cold front. Different threshold values of the combined criteria were tested for 20 different cases of cold fronts in the Mediterranean and an optimum selection of critical values was obtained.

It was found that the total wind shift, along with the wind magnitude, accounts for good representation of different types of cold fronts in the Mediterranean throughout the year. Therefore, in order for the Mediterranean cold fronts to be identified, not only the wind shift is required but a su fficient wind magnitude as well. Since no thermodynamic criteria has been included in this scheme, it is implied that wind shift is a prerequisite for the transition of a baroclinic zone to an organised cold front in the Mediterranean, confirming the experience of operational forecasters that, if there is no wind component at the upper levels perpendicular to the low level baroclinic zone, the formation of the frontal zone is inhibited [34]. This wind shift is mostly related with upper level disturbances, approaching a pre-existing area of enhanced low level temperature gradients [35].

Considering the climatological component of the scheme, a statistical validation of its results referring to the frequency of cold fronts passing over Greece for a decade was performed against results derived manually from synoptic analyses. It was found that the total frequency of the identified cold fronts agreed very well with the frequency of the fronts identified from synoptic analyses over Greece. Furthermore, the scheme succeeded in capturing the inter-monthly variations of the frequency of cold fronts. The employment of statistical metrics, considering the front as a two-fold categorical variable, confirms the satisfactory performance of the MedFTS on a climatological basis.

We are exploring the value of including thermodynamic and moisture information into the MedFTS, although we are aware that these are not without problems and weaknesses [36]. However, we have seen that with the appropriate modifications for the Mediterranean region, the dynamically-based FTS can be successfully applied to cold front identification.

**Author Contributions:** Conceptualization, H.F. and J.K.; methodology, I.S. and I.R.; software, E.B., I.S., I.R.; validation, E.B. and J.K.; formal analysis, E.B., M.H.; investigation, E.B. and M.H.; data curation, M.H., J.K.; writing—original draft preparation, E.B., H.F., M.H.; writing—review and editing, E.B., H.F., M.H.; visualization, E.B.; supervision, H.F.

**Funding:** This work has been partly funded by the Greek Academy of Athens under a PhD scholarship.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.
