Quick start: inspect a real transform on synthetic inputs

This exercise uses the installed CLI and production bathymetry pipeline. Its tiny source raster and H3 support are synthetic, so a passing run demonstrates software behavior rather than regional accuracy.

After installation, choose a fresh workspace outside the source checkout:

cd "$(mktemp -d "${TMPDIR:-/tmp}/seascape-first-result.XXXXXX")"
export SEASCAPE_WORKSPACE="$PWD/seascape-workspace"
seascape --workspace "$SEASCAPE_WORKSPACE" demo

The CLI reports Synthetic software acceptance: PASS (not a regional release) only after its checks and figures succeed. It prints exact paths beneath $SEASCAPE_WORKSPACE/.seascape/demo/ to:

Output	Why inspect it
output/BATHYMETRY.parquet	Positive-down depth statistics on exact H3 R8 support
output/bathymetry_manifest.json	Synthetic source identity and artifact checksums
report.json	Executed controls and PASS/FAIL evidence
figures/input.png, figures/output.png	Source and output visual checks

The fixture includes constant depth, gradient, sea-level, and nodata controls. Missing depth remains null; observed flat variation can legitimately be zero. Read the full demo guide for exact expected values, safe reruns, and validation limits.

Next steps

1. Read seascape concepts to distinguish cells, source evidence, and products.
2. Use the variable catalog to find fields and source limitations.
3. For real inputs, follow bounded real-data processing. Initializing a workspace and passing preflight do not by themselves create or validate a regional product.
4. For an existing audited schema-3 release, use the public product API to freeze an exact release ID and resolution.