27-Year Dataset of European Warm-Season Mesoscale Convective Systems

This Version is not yet Peer-Reviewed. 
This dataset provides a 27-year climatology of Mesoscale Convective Systems (MCSs) across the European domain (30N to 70N, 20W to 40E).

It covers the warm seasons (May through September) for the period 1998 to 2024. The data was generated using the Evolution-based Mesoscale Convective System Model Assessment-Tracker (EMMA-Tracker), an algorithm designed to identify and track organized convection using variables available in standard climate model output.

The primary purpose of this dataset is to serve as a process-based observational benchmark for evaluating Regional Climate Models (RCMs) and Convection-Permitting Models (CPMs). Unlike general precipitation datasets, these tracks have undergone physics-based post-processing filters (Track Straightness, Environmental Instability, and Area Volatility) to explicitly separate self-sustaining, propagating MCSs from synoptic-scale frontal precipitation and stationary orographic convection.

This dataset is not only restricted for model evaluation but itself represents a robust observational record that enables in-depth
climatological analysis of organized convection over Europe.

Dataset Specifications:

• Temporal Coverage: 1998-2024 (Warm Season: May-September)

• Spatial Domain: Europe (30N-70N, 20W-40E)

• Spatial Resolution: 0.1 degree x 0.1 degree (matching GPM IMERG native grid)

• Temporal Resolution: Hourly

• Input Sources: Precipitation from GPM IMERG V07 Final Run; Atmospheric Stability (Lifted Index) from ERA5 Reanalysis.

Content and Variable Information:

The dataset includes tracking information identifying the lifecycle of individual MCSs. Key characteristics retained for identified systems include:

• • robust_mcs_id: Contains only the mature phase (periods where the system exceeds 3500 km2 for at least 4 consecutive hours).

  • mcs_id: full lifecycle track of an MCS.

  • mcs_id_merge_split: Information on all merging and splitting events associated with a mature MCS.

  • System location (precipitation-weighted centroid lat and lon)

  • A boundary touching flag. 

The individual IDs are only unique within each year.

Usage Notes and Limitations:

1. This dataset is optimized for climatological analysis and model evaluation rather than individual meteorological case studies. Specific events should be cross-referenced with other meteorological data.

2. The domain extends to 70N; however, data poleward of 60N is subject to higher uncertainty due to the limitations of the input satellite observations at high latitudes.

3. Users comparing this data to model output should map their data to a common grid before tracking to ensure consistent detection of morphological features.

4. The EMMA-Tracker code used to generate this dataset is available on GitHub (https://github.com/DavidKneidinger/emma-tracker/).

Data Extraction Instructions:

The hourly tracking files are compressed into tar archives. To extract the files for usage, run the following command in your terminal, replacing 'filename.tar' with the specific archive name:

tar -xvf EMMA_MCS_Climatology_Europe_YYYY_vX.tar.gz

config.yaml setup used to produce this Dataset:

# config.yaml

# Detection parameter
min_size_threshold: 10  # min number of grid cells for an object to be detected
heavy_precip_threshold: 6.82  # mm/h
moderate_precip_threshold: 1.0  # mm/h
min_nr_plumes: 1  # number of heavy precipitation plums
lifted_index_percentage_threshold: 0.1   # % 
lifted_index_threshold: -2  # K

# Tracking parameter
main_lifetime_thresh: 4  # min number of hours for an MCS in the mature state
main_area_thresh: 3500  # min area for a system to be in the mature state
nmaxmerge: 5  # maximum number of objects to merge in a single timestep

# Post-processing / Filtering
# This section controls the physics-based filtering logic
postprocessing_filters:
  lifted_index_threshold: 1.5          # Keep systems with mean LI < 1.5 K
  track_straightness_threshold: 0.4    # Keep systems with straightness > 0.4
  max_area_volatility: 90000.0        # Keep systems with max volatility < 0.9e5