1. Data Description

This document provides a comprehensive overview of the HENS Tropical Cyclone Tracks Dataset, including data description, directory and file structure, file contents, and references.

2. Directory and File Structure

The HENS tracks are organized chronologically. All ensemble members generated for a specific day are aggregated into a single directory named after the ensemble's initialization date.

Directory Naming Convention

HENS_summer23_YYYYMMDDT000000/

• YYYYMMDD: Initialization date, where
  • YYYY is the four-digit year
  • MM is the two-digit month
  • DD is the two-digit day

File Naming Convention

Within each daily directory, individual files correspond to the tracks of each ensemble member for that specific initialization date:
ensemble_out_0NNNN_YYYY-MM-DD-00-00-00.txt

• NNNN: The HENS ensemble-member number (ranging from 0 to 7423).
• YYYY-MM-DD is still the initialization date in a different format.

Example Hierarchy

data/
└── HENS_summer23_20230831T000000/           # Directory for Aug 31, 2023
    ├── ensemble_out_00000_2023-08-31-00-00-00.txt
    ├── ensemble_out_00001_2023-08-31-00-00-00.txt
    └── ... (up to member 07423)

3. File Contents

Each file contains the tracks of all storms detected within the corresponding ensemble member.
Each track begins with a start line, followed by multiple subsequent data lines describing the storm evolution.

Start Lines

Start lines indicate the beginning of a new storm track.

• Example: start 15 2023 8 31 0
• Format: start [Number of Points] [Year] [Month] [Day] [Hour]
• The number of points indicates how many subsequent data lines will follow for that storm.
• [Hour] is in 24-hour format (00–23) and expressed in UTC

Subsequent Data Lines

These lines provide the physical and temporal attributes of the storm at specific time steps.

• Example: 1117 230 279.250000 32.500000 9.893917e+04 2.033962e+01 2023 8 31 18