Work Description

A Dataset of Great Lakes Extratropical Cyclone Tracks and Composites README

July, 2024

What is this?

Extratropical cyclones (ETCs) play a significant role in cold-season weather and climate for the Great Lakes Region (GLR). These cyclones are best identified by negative mean sea level pressure (MSLP) anomalies, with the lowest pressure located at the center of the storm. To identify and analyze historical trends in ETCs that have passed through the GLR, a tracking algorithm is performed on the ERA5 global reanalysis (Hersbach et al. 2020; doi:10.24381/cds.adbb2d47) to identify individual storms and their trajectories through space. Once the ETCs have been traced through the GLR, seasonal storm composites are created to analyze trends in cyclone characteristics such as temperature, moisture, and precipitation. Data is available from 1959 through 2021.

Data Packaging: Cyclone Tracks (storm_tracks.tar.gz)

Cyclone track data is analyzed and packaged using Python and the Numpy software, and files are saved in .npz format. Each individual .npz file contains the information for every storm tracked within a cold season (i.e., October through March). Each storm will have Data can be read in as follows (using Python and Numpy [version 1.23.3 is the latest version used for this dataset; https://numpy.org/] packages):

import numpy as np
data = np.load('storm_track_slp_1960.npz', allow_pickle=True, encoding='latin1')
data['storms']]

By calling each storm with its index (in order by occurrence during the season), the following information can be accessed (as defined by the variable name within the dataset):
lat: latitude of storm location
lon: longitude of storm location
amp: MSLP at the center of the storm (in Pascals)
type: 'cyclonic' or 'anticyclonic'.
year, month, day, and hour: Provide the date and time for each point that the storm is tracked.
exist_at_start: True or False. Determines whether the storm was already existing on October 1 of the season or not.
terminated: True or False. If False, the storm track did not end before March 31 of the season.
age: How long the cyclone was tracked. To get the ETC's age in hours, multiply age by 6 (which is the tilmestep of the ERA5 data used to track the cyclones).

All information above is given in numpy array format, the length of which is determined by the duration of the storm's life (length of the storm trajectory). Information for each storm has a temporal resolution of every 6 hours.

Data Packaging: Cyclone Composites (storm_composites.tar.gz):

Cyclone composites are created by identifying the time at which each storm is strongest (i.e., lowest MSLP) within the GLR, cropping a 20x20 degree square around the center of the storm, and then averaging all storms from one season together. Data is stored in the NetCDF file format. The internal structure of the files is as follows:
Spatial Variables: west-east, south-north
Data Variables: 2-m temperature, MSLP, 10-m U- and V- component of wind, 2-m dew point, total column integrated water vapor, evaporation, total precipitation, and precipitation in the form of snow, rain, freezing rain, and mixed precip.

The grid spacing is the same as ERA5. The number of storms used to make each seasonal composite is given in the file attributes.

Data Description:

Both the storm tracks and composites are saved one file per season. Storm tracks contain all tracked cyclones, regardless if they encountered the GLR. Storm composites only contain information for the storms that passed through the GLR.

The code used to create this data can be found in the zipped repository GLStormTrends_2024.zip (README included). Users can refer to the GitHub page https://github.com/abkenyon/GLStormTrends_2024/ for any updated versions of the code.

Filename Convention:

The naming convention for the storm track files is structured as follows:
storm_track_slp_YYYY.nc

The naming convention for the storm track files is structured as follows:
storm_composite_yyyy-YYYY.nc

Where:

yyyy: Year at the beginning of the cold season, defined as October-March.
YYYY: Year at the end of the cold season

Use and Access:

These data are made available under a Creative Commons Attribution Non-Commercial license
(CC BY 4.0).

Contact:

For any further questions or assistance with the dataset, please reach out to the corresponding data author (Abby Hutson) via email: hutsona@umich.edu

Date of last update:

July 9, 2024