mirai-ml/climate_hazard_index

Climate Hazard Index Dataset

This directory contains a comprehensive dataset of environmental and climate hazard indicators extracted from various sources, primarily focused on climate change scenarios and natural hazard assessments.

Dataset Overview

The dataset consists of 113 total datasets with 109 successfully downloaded files, representing a total download size of approximately 105 GB. The data was extracted on September 19, 2025 from the source directory containing data layers and their corresponding download instructions.

Data Sources

The datasets are sourced from multiple authoritative organizations including:

• UNEP Grid (https://unepgrid.ch/en)
• CIMA Foundation (https://www.cimafoundation.org/en/)
• Various climate modeling institutions

Dataset Composition

File Structure

• indicator_data.json - Metadata and catalog of all datasets (168KB)
• archive.tar.gz.part-* - Multi-part compressed archive containing the actual dataset files
• .cache/ - Caching directory for temporary files

Multi-Part Archive

The dataset is distributed as a multi-part tar.gz archive split into 21 parts:

• archive.tar.gz.part-aa through archive.tar.gz.part-at
• Individual part sizes range from 1.6GB to 4.9GB
• Total archive size: ~100GB

Indicator Data Summary

The indicator_data.json file contains structured metadata for each dataset with the following key variables:

Metadata Fields

Field	Description	Example
dataset_id	Unique identifier for the dataset	dataset-MX-0LGY6-ZYFR1-SKBOJ
title	Human-readable title	Drought hazard SSI 5-year return period - SSP5 Upper bound
description	Detailed description of the dataset	Full scientific description of the hazard type and methodology
download_url	Source URL for the data file	https://hazards-data.unepgrid.ch/ssi-01_SSP585_sev3_GF.tif
keywords	Comma-separated keywords	hazard, drought, SSI, SSP5 RCP8.5
release_date	Publication date	2023-05-01
data_value_starting_date	Start of data coverage	2050-01-01
data_value_ending_date	End of data coverage	2100-12-31
source_homepage	Organization homepage	https://www.cimafoundation.org/en/
downloaded_file	Local file path	/home/nacim/workspace/serenia/CAF/proto_caf/downloads/dataset-MX-0LGY6-ZYFR1-SKBOJ.tif
is_downloaded	Download status	true
file_size_bytes	File size in bytes	5122267
file_size_mb	File size in megabytes	4.88

Dataset Categories

The datasets cover various environmental and climate hazard types:

1. Drought Hazards - Standardised Streamflow Index (SSI) based drought assessments
2. Storm Surge - Coastal storm surge hazard maps with various return periods
3. Flood Hazards - Riverine and coastal flood risk assessments
4. Climate Change Scenarios - Data for different Shared Socioeconomic Pathways (SSP) and Representative Concentration Pathways (RCP)

Climate Scenarios

The datasets include projections for multiple climate scenarios:

• SSP1 RCP2.6 - Low emissions scenario
• SSP5 RCP8.5 - High emissions scenario
• SSP585 - Alternative high emissions scenario

Temporal Coverage

• Present Climate: 2017-2025
• Future Projections: 2050-2100
• Return Periods: 5, 10, 25, 50, 100, 250, 500, 1000 years

Tutorial: Extracting Multi-Part Tar Files

Prerequisites

Ensure you have sufficient disk space (at least 200GB free) and the following tools installed:

• cat (for combining parts)
• tar (for extraction)
• gzip (for decompression)

Step 1: Combine the Multi-Part Archive

The archive is split into multiple parts that need to be combined before extraction. Use the following command to concatenate all parts:

# Navigate to the hfdataset directory
cd /path/to/hfdataset

# Combine all parts into a single archive
cat archive.tar.gz.part-* > archive.tar.gz

Note: This command will create a large file (~100GB), so ensure you have sufficient disk space.

Step 2: Verify the Combined Archive

Before extraction, verify the integrity of the combined archive:

# Check if the combined file is valid
file archive.tar.gz
# Should output: archive.tar.gz: gzip compressed data

# Optional: Test the archive without extracting
tar -tzf archive.tar.gz | head -10

Step 3: Extract the Archive

Extract the contents to your desired location:

# Extract to current directory
tar -xzf archive.tar.gz

# Or extract to a specific directory
tar -xzf archive.tar.gz -C /path/to/extraction/directory

Step 4: Clean Up (Optional)

After successful extraction, you can remove the combined archive to save space:

# Remove the combined archive (keep individual parts for backup)
rm archive.tar.gz

Alternative: Extract Directly from Parts

If you prefer not to combine the parts first, you can extract directly:

# Extract directly from parts (requires all parts to be present)
cat archive.tar.gz.part-* | tar -xz

Troubleshooting

Issue: "No space left on device"

• Solution: Ensure you have at least 200GB free space before starting

Issue: "Unexpected end of file" or "gzip: stdin: unexpected end of file"

• Solution: Verify all part files are present and not corrupted. Check file sizes match the expected values.

Issue: Permission denied

• Solution: Ensure you have write permissions in the target directory

Issue: Parts are missing or corrupted

• Solution: Re-download the missing or corrupted parts

Expected Extraction Results

After successful extraction, you should see:

• Individual dataset files (primarily .tif and .gpkg formats)
• Organized directory structure matching the original data organization
• Total extracted size of approximately 100GB

Usage Notes

• The datasets are primarily in GeoTIFF (.tif) and GeoPackage (.gpkg) formats
• Most files are suitable for use with GIS software like QGIS, ArcGIS, or programmatic tools like GDAL
• The data covers global and regional scales with varying spatial resolutions
• Temporal data includes both historical and future climate projections

Data License and Citation

Please refer to the individual dataset metadata in indicator_data.json for specific licensing terms and citation requirements. Most datasets are from UNEP Grid and CIMA Foundation and may require proper attribution.

Support

For technical issues with the dataset or extraction process, please refer to:

• The source organization websites listed in the metadata
• The original DOWNLOAD-INSTRUCTION.md files referenced in the dataset paths