Reopen and share a completed study

After a study finishes, choose Save complete study. The .aqstudy.json file contains the brief, plan, input tables, results, findings and artifact bytes. Use Tools → Open saved study to reopen it. Importing this file restores stored results; it does not run the plan or refetch a gauge. Runtime packages may still need to load from the network the first time this browser opens a study.

A checksum detects changed contents. It is not a signature, author authentication, or independent verification of the science. Keep the original file if you need to preserve its exact input and result identities. Complete files are limited to 50 MB. A lightweight workspace.json may lack artifact bytes; use the full bundle or the original complete study if that happens.

Copy plan link shares instructions to run an analysis. It does not carry a completed result. The recipient may get different observations from live agencies. Download bundle contains reports, tables, figures, the notebook and workspace.

Retained data and result tables

Station and flood-analysis steps retain their own full observation CSVs, including subdaily timestamps. Plotting decimation does not change these inputs. Check the step's input hash and observation count when reproducing a result.

Flood result CSVs use one row per estimator and return period. lower, upper, interval_method and confidence_level belong to that row's estimator and estimate. Empty bounds mean that estimator has no recorded interval; an interval from another fit is not substituted. This replaces the older wide table with generic GEV/LP3 columns and one ambiguous pair of bounds.

Workbook data tables over 100,000 rows continue in numbered sheets. Keep all parts when importing them elsewhere. The CSV retains the original serialized precision; spreadsheet applications impose their usual numeric precision limits.

Rerun the notebook

The notebook requests current agency data through the same execution path as Explorer. Its opening answer is labelled as the original answer; the run prints a new answer and regenerates its findings, checks, full input tables and figures. Uploaded tables are retained from workspace.json. Install the exact source revision shown in the notebook when one is recorded.

The last cell saves the new artifacts under rerun-<id>/, preserving the original bundle files. A failed step produces new limitations rather than reusing its old tables or claims. For the original result without a new agency request, reopen the complete study instead. Current-data reruns are distinct from reproducing the original analysis against its retained inputs.

Publish an explicit, versioned report

For a public result, review the complete file or self-contained HTML report for private inputs first. Then publish an immutable version through your chosen repository release or data repository. Include input hashes, source and agency credits, software version and build revision, method/interval identity, check outcomes and limitations. Name the release explicitly and retain its immutable URL or DOI. A mutable “latest” page is not a reproducible snapshot.

AquaScope does not upload these files when you save or reopen them. Anyone to whom you give the complete file can inspect the included inputs. Do not publish private uploads, locations or a participant's results without their permission.