Download files

The prepared files for each dataset are also available for direct retrieval at data.bafu.admin.ch/download/. Access is via an S3-compatible client through a custom endpoint, with download serving as the bucket name. No AWS account or credentials are required.

The prefix follows the folder structure <theme>/<dataproduct>/<version>/… (e.g. water/nawa-trend/v2026-02-25/).

AWS CLI

The endpoint is passed per command via --endpoint-url; --no-sign-request suppresses request signing, so no credentials are transmitted.

ENDPOINT="https://data.bafu.admin.ch"

Listing

# Contents of a dataset:
aws s3 ls s3://download/water/nawa-trend/ --endpoint-url "$ENDPOINT" --no-sign-request

# Recursively, all files below a prefix:
aws s3 ls s3://download/water/nawa-trend/ --recursive --endpoint-url "$ENDPOINT" --no-sign-request

Retrieving

# A single file:
aws s3 cp s3://download/water/nawa-trend/v2026-02-25/stations/water_nawa-trend_stations.csv.gz . \
  --endpoint-url "$ENDPOINT" --no-sign-request

# All files under a prefix:
aws s3 cp s3://download/water/nawa-trend/ ./nawa-trend/ --recursive \
  --endpoint-url "$ENDPOINT" --no-sign-request

# Recurring sync (downloads only new or changed files):
aws s3 sync s3://download/water/nawa-trend/ ./nawa-trend/ \
  --endpoint-url "$ENDPOINT" --no-sign-request

rclone

rclone is a suitable open-source alternative. Configure a remote once in ~/.config/rclone/rclone.conf:

[bafu]
type = s3
provider = Other
endpoint = https://data.bafu.admin.ch
list_version = 2

list_version = 2 is required: otherwise rclone uses an older listing method that the endpoint does not support.

The usual commands are then available:

# Listing:
rclone lsf bafu:download/water/nawa-trend/
rclone lsd bafu:download/water/

# Retrieve a single file:
rclone copy bafu:download/water/nawa-trend/v2026-02-25/stations/water_nawa-trend_stations.csv.gz .

# Sync an entire dataset:
rclone sync bafu:download/water/nawa-trend/ ./nawa-trend/