Impact Areas and Dynamical Features associated with Mediterranean Cyclones (1980-2019)

The dataset includes NetCDF files of Impact Areas and Dynamical Features associated with Mediterranean Cyclones (henceforth MedCyclones).

MedCyclone tracks correspond to confidence-level 5 tracks from Flaounas et al. (2023), https://doi.org/10.5194/wcd-4-639-2023.

Temporal frequency: 6h (00, 06, 12, 18 UTC)Years: 1980 – 2019Spatial resolution: 0.5 degGrid extension:  0-70N, 40W-65E

Dynamical Features

Files "dynfeats_bool_rmax2000_YYYY.nc" include the following list of variables, describing connected boolean objects:

r_500, r_1000: central areas of fixed 500 or 1000 km radius around MedCyclone centres;

WCB: warm conveyor belts related to MedCyclones (i.e., overlapping with r_500 in at least one grid point). Each WCB is eventually separated into inflow (WCBin, up to 800 hPa) and ascent (WCBout, between 800 and 400 hPa) regions. Ref. at https://doi.org/10.1175/JCLI-D-12-00720.1, https://doi.org/10.5194/wcd-5-537-2024;

fronts: cold fronts related to MedCyclones (i.e., overlapping with r_500 in at least one grid point). Ref. at https://doi.org/10.5194/gmd-17-6137-2024;

DI: dry instrusions related to MedCyclones (i.e., overlapping with r_1000 in at least one grid point). Ref. at https://doi.org/10.1175/JCLI-D-16-0782.1;

r_1000_Nodynfeat: the central 1000 km area excluding regions of MedCyclone WCB, fronts and DI objects.

The criteria for the identification of WCB, fronts and DI objects are described in Section 2.3 of Portal et al. (2024), https://doi.org/10.5194/wcd-5-1043-2024.

Additionally, we note that :i. a weaker overlap constraint was used to associate DI objects to MedCyclones (r_1000 compared to r_500 for WCB and fronts objects) because of the relatively large distance of the DI airstream from the cyclone centre;ii. in this dataset, all connected objects related to MedCyclones are cropped within a 2000 km area circle from the cyclone centre for two reasons. Firstly, the dynamical-feature related surface impacts usually weaken with the distance from the cyclone centre. Secondly, to cut connected objects composed by multiple overlapping features of the same kind - this often happens for fronts in summer because of their high detection density. Far from the cyclone centre, these objects are usually unrelated with the MedCyclone circulation.

Impact Areas

Files "IAs_bool_rmax2000_YYYY.nc" include boolean impact areas, combining a central area (r_1000 or r_500) and cyclone-related WCB, CF and DI objects. The three types of impact area are described in the following :

IA01 is composed by a 1000 km radius circle around the cyclone centre (r_1000) extended by cyclone-related WCB, fronts and DI;

IA02 is composed by a 500 km radius circle around the cyclone centre (r_500) extended by cyclone-related WCB, fronts and DI

IA03 is composed by a 500 km radius circle around the cyclone centre (r_500) extended by cyclone-related WCB and fronts (DI is neglected).

As discussed in Section 3.1 and Appendix A of Portal et al. (2024) (https://doi.org/10.5194/wcd-5-1043-2024), IA01, composed by a central area of 1000 km, is adequate for intercepting long-range wind impacts associated with MedCyclones. IA02 and IA03, on the contrary, are better devised for detecting impacts expected at shorter distances from the cyclone centre, such as rainfall, thunderstorm and storm surges. In particular, IA03 neglects the DI region, which is normally of little interest for cyclone-related moist processes, involved in producing precipitation. Noetheless, DI remains relevant for the identification of strong cyclone-related winds.

Case Studies

A pdf file providing the visualisation and description of impact areas and dynamical features of all MedCyclones occurring in 1980 is available at the link https://boris.unibe.ch/192315/. Note that in the examples the dynamical features are not cropped at 2000 km from the cyclone centres, as for the present dataset. 

Note that many of the "Annotations and Limitations" listed below derive from the attentive analysis of these study cases.

Annotations and limitations

In the case of more than one MedCyclone centre per timestep, the dasaset does not distinguish the impact areas / dynamical features associated with each centre.

Because of the automated criteria for associating WCB, fronts and DI objects to MedCyclones, at times objects close to the centre but unrelated to the MedCyclone's circulation, are considered to be cyclone-related and included in the impact area.

Elaborating on the point above, at times fronts responsible for Mediterranean cyclogenesis (and not produced by the cyclonic circulation itself) are included in the MedCyclone impact area.

When computing statistics over a long time interval (e.g., climatology), the effects of erroneous associations of dynamical features to MedCyclone impact areas are attenuated by the aggregation of large quantity of data. 

Over a long time interval (e.g., climatology) the choice of a 1000 km fixed-radius impact area provides similar statistics to IA01, although in the first case it is not possible to isolate the role played by the different features composing the MedCyclones.