TreeRings

treerings.howardalandman.org

An interactive view of individual subfossil and historical tree logs on a shared time axis. One rectangle per log, greedily packed into rows, coloured by taxon, with hover tooltips and click-through to each log's measured ring-width curve where the widths are open. A map view shows where the collections were taken, one pie per published coordinate, linked to the timeline.

The site draws expanded Europe from three sources: 123,910 logs · 8,498 sites · 12429 BC – 2025 CE (85,186 ITRDB, 38,184 PANGAEA, 540 Zenodo). A PANGAEA "site" is usually one object, and those logs are bars only: their dates are public, their ring widths restricted. Europe is where this started, not the limit of what it handles; see Scope.

Most tree-ring resources publish site chronologies — averaged curves. This shows the underlying measured series individually, which is what makes the overlapping ladder of chronologies visible: you can see how the record is assembled from them.

What a "log" is. One bar is one measured ring-width series — a core or a radius — not necessarily one tree. A tree sampled with two cores is two bars, because the archive has no field recording which series belong to the same tree. The count also includes redundant copies: an audit of an earlier build found about 2,000 series deposited byte-identically in more than one collection, and about 600 in blue-intensity files that remeasure trees already present. Merging both is planned but not done.

Why the archive looks the way it does

The corpus is overwhelmingly modern: 85 logs alive in 5000 BC, 81 at 1 CE, and 69,428 in 1971. Packed together that is a ~70,000-row wedge, so the viewer zooms on both axes and offers era filters. The "Before 1 CE" + "Packed together" view is the classic overlapping-chronology ladder, and the most legible entry point.

It also has a hole. No log is alive between 11746 BC and 6228 BC: before it there are only the anchored French Late Glacial pines, after it continuous coverage. The long Central European oak and pine chronology (Hohenheim) covers most of that interval, but its per-log measurements for those years are not publicly available.

Scope

ITRDB publishes measurements under regional directories: africa, asia, atlantic, australia, centralamerica, europe, northamerica, and southamerica. All can be mirrored for study (python -m ehd.fetch --all). The web build uses expanded Europe, the Western Palearctic, defined on site coordinates in ehd/region.py: all of europe, and africa, asia and atlantic by position.

The .rwl parser, the duplicate reconciliation, the dating classifier, the packing, and the viewer all operate on logs without reference to region. Going global means choosing the scope in corpus.scoped_files, and deciding how the viewer should present a corpus large enough that a single packed timeline stops being legible.

The pipeline is also source-agnostic beyond ITRDB: any per-log corpus that can be expressed as series with ring widths and a dating status drops in as another input.

Running the pipeline

uv venv --python 3.12 && uv pip install pytest brotli   # once
python -m ehd.fetch --templates   # download the corpus to the data root
python -m ehd.pangaea    # PANGAEA's public tree-ring metadata
python -m ehd.zenodo     # the Zenodo oaks, checked against their MD5
python -m ehd.harvest    # study metadata + corpus reconciliation report
python -m ehd.emit       # build web/data payloads
python -m ehd.basemap    # map outlines from Natural Earth (pinned, checksummed)
python -m ehd.validate   # cross-check parser against dplR + invariants
python scripts/serve.py  # view at localhost:8777

scripts/serve.py rather than python -m http.server, because ehd.emit writes the two first-paint payloads brotli-compressed under their plain names and web/_headers is what declares that to the browser; the dev server reads the same file, so local and deployed behaviour match. A plain static server hands the viewer compressed bytes and it draws nothing.

Data root resolves to $EHD_DATA, else /Volumes/Expansion/EHD, else ./data. ehd.validate additionally needs R with dplR.

The pipeline is source-agnostic: a new per-log corpus drops in as another input.

Correctness

The .rwl format has traps that silently corrupt results rather than failing loudly. Each of the following was found in real corpus files, and each has a regression test in tests/:

• RWL is fixed-width. IDs run straight into the year (0003401Q1890); a whitespace parser fails on ~32% of the corpus.
• BC years borrow column 8 for the minus sign (C002 -2894).
• -999 is a missing ring, not a terminator. It appears mid-series in long runs; treating it as terminal truncated one 1854–2010 log to 1854–1864.
• The terminator is per-file. In a 0.001 mm file, 999 is a real 0.999 mm ring, so terminators are trimmed from series ends rather than acted on inline.
• Trailing locally-absent rings (negative flags such as -8) must not extend a span; interior ones are real zero-width rings.
• Three kinds of duplicate — -noaa.rwl re-encodings, per-variable suffixes measured on the same trees, and series IDs repeated across sites. Left alone these inflate the log count by ~36%.
• A declared span that disagrees with the data is a metadata defect, not a site chronology narrower than its longest log — NCEI confirmed the study's age range should match the .rwl file. Disagreements are common, though (the first sweep found 152 files), so dating classification tests overlap rather than endpoint agreement, and ehd.sweep reports the mismatches upstream.

Verification:

• pytest tests/ — the parser fixtures are the files that broke earlier parsers.
• python -m ehd.validate — 100.0000% span agreement with dplR across 76,397 series, 0 disagreements, 0 invariant failures, and 0 series found by dplR that this parser misses. It additionally reads 12,349 series dplR cannot (run 2026-11-01, dplR 1.7.9). This covers ITRDB only: PANGAEA's bars have no ring widths to read.
• ehd.emit asserts the packing is overlap-free before writing.

Payload size

The two files first paint waits on are written brotli-compressed at quality 11 and served with Content-Encoding: br. logs.bin is application/octet-stream, which Cloudflare's auto-compression allowlist excludes, so it shipped raw at 84% of first load; index.json was compressed, but by an edge encoder tuned for speed rather than ratio. Pre-compressing both takes first load from 2,192,561 to 678,710 bytes — 2.09 MiB to 0.64 MiB, 69% smaller — measured live rather than locally, since the ratio a CDN serves is not the ratio a build produces (2026-08-20, on that day's build). The current build's first load is 1,263,649 bytes: logs.bin 740,181, index.json 497,736, and the page, app.js and style.css 25,732, measured live the same way. (The 2026-10-22 build, 73,390 logs, was 655,918 bytes.) The map's coastline file (438 KB raw, 177 KB as served) is not part of first load; it is fetched the first time the map is opened.

Clients that do not accept brotli are not stranded: Cloudflare decodes and re-encodes at the edge, so an Accept-Encoding: gzip request still gets gzip and one with no Accept-Encoding gets the raw bytes.

The compression and the _headers rule break the site if separated, so tests/test_headers.py pins them together and app.js checks logs.bin's length against n_logs and names the likely cause if it ever mismatches.

Dating status

Logs are labelled rather than silently mixed:

• absolutely dated — calendar-dated by the depositors (123,556). For ITRDB the dating is consistent with the archive's metadata; for PANGAEA's bars the metadata is the only evidence available, because the widths are restricted. The Zenodo oaks carry the depositors' span, checked at both ends against the record.
• anchored — a floating collection placed on the calendar by a reviewed anchor taken from its depositors' own dating statement, carried with its uncertainty (191, across 5 collections: the French Late Glacial pines fra053–055 and the Fenland yews brit099–100). The build fails if a file stops reproducing its declared span, so an anchor cannot silently go stale.
• unverified — dated, but no declared span exists to check against (86, across 3 collections)
• disputed — plotted over a declared span the data contradicts (77, across 2 collections whose records end in 1928 while their data runs to 2025; each case is reviewed by hand)
• floating — withheld from the timeline entirely; the year column holds ring ordinals, not calendar years (136, across 5 Swiss Late Glacial collections)

Logs from studies NCEI has withdrawn are parsed and validated but never drawn (3,424 logs, 25 studies).

The Swiss Late Glacial pines have been placed on the calendar in the literature by radiocarbon wiggle-matching, but no per-file anchor appears anywhere in the archive, so they stay floating until a placement can be tied to each file.

Data source and credit

Ring widths come from the NOAA/NCEI International Tree-Ring Data Bank, and the bars without curves from the public metadata of the Hohenheim archive on PANGAEA. The 540 Neolithic oaks are from Zenodo. All of it was contributed by the many investigators credited in each site record and reachable through the study links in the viewer. This project neither collected nor re-measured any rings; it parses, reconciles, and displays their published data.

Metadata defects found during reconciliation have been reported to NCEI, which has been correcting them — see docs/ncei-metadata-report.md and the study-by-study sweep in docs/ncei-sweep-2026-08.txt (python -m ehd.sweep reruns it).

License

Code: MIT. The underlying ITRDB measurements are NOAA public-domain data and are not redistributed by this repository — ehd.fetch downloads them. The 540 Main-region oaks are from Land, Remmele & Küppers 2018 (10.5281/zenodo.1326474), CC BY 4.0; python -m ehd.zenodo downloads them, and ehd.emit stops if they are missing.