Rooted in nature. Practiced in the real world. Built in public.
Tree Yoga School is an open-source, nature-centered yoga education project founded by Alex Julian. Its live school connects yoga, meditation, tree hugging, and outdoor exploration with practical ways to spend time among trees. Its foundation is the original Tree Yoga School book: a body of teaching that brings together yoga, meditation, time among trees, and the practice of learning from nature.
Nature is the teacher. Technology is the tool. Practice is the point.
Visit the live school · Begin Here · Read the original book · Explore the curriculum · Build with us
Project status: Live at treeyogaschool.com. The content-led Next.js school includes a four-practice introduction, seven native book chapter companions, the original pose reference, an 84-species Tree Library, a 50-state Champion Tree Map, and Trees to Visit guides. Return to Your Tree offers an optional browser-local journal. Map points indicate approximate areas; check source records and visiting guidance before traveling. There is no account system, database, certification platform, or AI teaching assistant.
Trees to Visit and Return to Your Tree are primary navigation destinations. The School menu connects About, Contact, Big Tree Volunteers, and the original book. The return practice includes My Tree: an optional species selection and dated digital journal saved only in the visitor’s browser, plus a four-visit handwritten journal and downloadable one-page PDF. Species profiles offer Make this my tree; the saved selection links to its outdoor practice and matching visit guides. Journal entries can be edited, deleted, and downloaded as JSON. Contact and Big Tree Volunteers invite visitors to email alexjulianyoga@gmail.com. Map correction links open an email draft with the record ID and state included. There is no website submission form.
Each supported state has a static /champion-trees/<state-name> destination with register dates and coverage, a published-height preview, matching Library species and practices, and available visit guides. State shortcuts open these pages; map filters retain their existing query URLs. Detailed selection and mapping notes are shared between the state pages and the map in lib/champion-register-notes.json.
The shared logo and photographic heroes now use responsive, content-hashed WebP variants with one-year immutable caching. Mobile image selection, repeat-visit cache reuse, and six 50-state map benchmarks are documented in Mobile images and map measurements. The map benchmark is browser emulation; physical-phone validation remains open.
- Champion Map: 50 states, 9,134 listed records, and 8,658 records with approximate map points.
- Tree Library: 84 species profiles and 8 photographic comparisons.
- Trees to Visit: 19 guides across 12 states.
Generated from the state registry and content collections. Run npm run sync:coverage after content changes; builds check for stale counts.
Tree Yoga is not simply Tree Pose performed outdoors. It is a practice of bringing attention, movement, breath, and contemplation into a relationship with actual, living trees.
The book introduces three ways to practice:
Yoga with a tree. Explore postures and movement beside a tree, using the trunk for appropriate support where the person, tree, and setting allow it. Stillness and simple movements belong here just as much as challenging poses.
Meditation with a tree. Sit, stand, or rest nearby. Observe the breath, notice the surroundings, and return to the present moment. The book includes chair-based, seated, reclined, and standing approaches.
Tree Yoga Hiking. Bring the practice into a nature walk. Rather than treating every tree as scenery on the way to a destination, pause, observe, and become acquainted with the place you are moving through.
These pathways come from “Who and How,” chapter 3. Underneath all three is the same invitation: stop passing nature by, and begin spending time with it.
You do not need to learn GitHub, use AI, or wait for a website to begin.
As a simple introduction, choose a safe, permitted, accessible place near a tree. Find a comfortable position. Put the screen aside, notice your surroundings, and let your breathing remain comfortable. Observe the bark, the canopy, the changing light, or the sounds around you. There is nothing particular you need to feel or achieve.
The book recommends beginning a new habit with five or ten minutes rather than an all-or-nothing commitment (chapter 4, printed p. 28). Start smaller when needed. Quiet observation is a complete first visit; physical contact, special equipment, and advanced postures are not prerequisites.
This introductory invitation is an adaptation for the public project, not a verbatim exercise from the book. Read Care and safety before exploring movement practices.
The source is Alex Julian’s Tree Yoga School (2023), ISBN 9798752094705. The book traces the school’s development through yoga practice and teaching, forest bathing, mentorship, and engagement with the Bhagavad Gita, Tao Te Ching, and Yoga Sutras of Patanjali (chapter 1).
This repository is a new home for that work—not a claim that AI invented it. The book remains the reference for what the original school taught. New lessons, revisions, and interpretations should be clearly identified as such.
| Chapter | What readers will find | Printed page |
|---|---|---|
| 1 · What is Tree Yoga? | The origin of the school, contemplation, and the invitation to practice. | 1 |
| 2 · Why Tree Yoga? | The author’s account of the practice’s purpose, benefits, and relationship with nature. | 4 |
| 3 · Who and How | Three approaches to practice, choosing a tree, equipment, and care for the surroundings. | 13 |
| 4 · When and Where | Daily habits, seasons, settings, and adapting practice to conditions. | 25 |
| 5 · Poses, Flows, and Meditations | A movement reference library, ten flows, and ten meditations. | 34 |
| 6 · Wisdom and Wonder | Ten guiding principles and reflections on the lives of trees. | 96 |
| 7 · Graduation | Taking the practice into daily life, journaling, and sharing what you learn. | 125 |
Chapter 5 is a reference library, not a beginner-to-advanced sequence. Its flows include Deep Relaxation, Mobility Energizer, Prime Balance, and Tree Trunk Strength. Its meditations include Chair Meditation, Tree of Life, Witness Meditation, and “Thank You, I Love You.” The presence of an exercise in the book does not make it suitable for every reader.
The links open the original PDF. Page numbers in the table refer to the book’s printed numbering; PDF page numbers include the front matter. Read the health and evidence note under Care and safety alongside the original text.
Trees are not merely a backdrop for this school. In the book, they become subjects of attention and sources of reflection on how to live.
The ten principles in Wisdom and Wonder provide a foundation for the new project. The applications below are new interpretations for this public edition, not quotations from the book.
The ten principles, applied to a living school
| Principle from the book | How we can put it into practice here |
|---|---|
| Acceptance | Begin with the body, understanding, and resources available today. |
| Adaptability | Offer different ways to participate instead of demanding one ideal form. |
| Balance and Optimization | Use technology deliberately; leave time and attention for life away from it. |
| One-Pointed Focus and Aspiration for Excellence | Give one practice—or one useful improvement—your full attention. |
| Perseverance | Return patiently. Sustainable growth includes rest and revision. |
| Presence | Notice what is actually happening, rather than performing for an imagined audience. |
| Release | Let go of unnecessary features, rigid expectations, and work that no longer serves the school. |
| Strength | Cultivate steadiness alongside kindness, not domination or display. |
| Unconditional Forgiveness | Make room for mistakes and repair while respecting healthy boundaries. |
| Unconditional Love | Welcome people with care, and extend that care to trees and their surroundings. |
Source: chapter 6, printed pp. 97–109.
AI should help us build a better school—not become the object of the practice.
The intention is to use AI thoughtfully in organizing material, improving explanations, developing software, and exploring new teaching formats. The original work supplies the roots; people remain responsible for what is published.
Our design commitments are:
- Keep the sources visible. Connect adapted lessons to the relevant book pages. Distinguish original text, editorial changes, personal reflections, and research-backed claims.
- Keep human judgment central. Review AI-assisted contributions for accuracy, safety, attribution, and tone. An AI response is not a medical assessment or a substitute for qualified instruction.
- Make the school understandable and accessible. Favor readable content, clear navigation, and tools that support different ways of learning. Future features belong on the roadmap until they work.
- Help people return to practice. Build useful stopping points, not reasons to scroll forever. A successful session may end with closing the device and going outside.
AI-native is a direction for development, not a claim that autonomous teaching features are already available.
The book’s final chapter encourages journaling, making nature a habit, and contributing insights. Building in public can be one more expression of that invitation:
Practice → Notice → Reflect → Make one useful change → Return outside.
A useful change might be a clearer explanation, a corrected reference, an accessibility improvement, or a carefully considered question. Small contributions count. Not every session needs a commit, and not every meaningful experience needs to become public content.
The measure is not a contribution streak. It is whether the work helps someone practice with greater attention, care, and independence.
The public school is live at treeyogaschool.com, with learning pages, tree discovery tools, visiting guides, and optional browser-local return practice. The original book remains linked as a PDF, alongside native chapter companions and the original pose reference. Original teaching retains its separate rights and attribution described below. The school is content-led: there is no database, account system, or AI API requirement.
The intended sequence is deliberately small:
| Stage | Focus |
|---|---|
| Foundation — complete | Establish the school’s identity, source material, scope, and contribution approach. |
| Website foundation — complete | A calm, accessible Next.js home connected to its Vercel production deployment. |
| Begin Here — available | Your First Five Minutes with a Tree, followed by meditation, gentle yoga, and hiking. |
| Living book and discovery — live | Seven native chapter companions, the original pose reference, Tree Library, Champion Tree Map, and Trees to Visit guides. |
| Further tools — only when useful | Consider guided learning or optional personal features after the core experience works. Accounts, Supabase, and AI APIs are not prerequisites for the first version. |
This README establishes the foundation. It does not announce a completed course platform, live classes, or a certification program.
Readers, practitioners, teachers, designers, and developers can all help. A thoughtful question can be as useful as a code change.
Use an issue to propose an improvement or discuss a larger idea before building it. For a focused change, submit a pull request that explains what changed, why it helps, and how it was checked. Contributions are subject to maintainer review.
For teaching material, include the relevant chapter and printed page, identify adaptations, and explain any safety or accessibility considerations. For scientific or health claims, provide appropriate primary sources and represent uncertainty honestly. Do not invent evidence, credentials, testimonials, or teaching outcomes.
Submit only material you have permission to share. Keep private journals, health information, access credentials, and sensitive location details out of public issues and commits. Obtain permission before contributing other people’s photographs or stories. Review AI-assisted work before submitting it and disclose material AI assistance in the pull request.
Preserve the spirit of the school: clear teaching, careful attribution, respectful collaboration, and care for the living world.
This is an educational and contemplative project, not medical care. Yoga practices need to fit the individual; advanced poses are not requirements. Seek qualified instruction, and discuss relevant health conditions or pregnancy with your healthcare professional. Do not push through pain or force breathing. See NCCIH’s yoga safety guidance.
Choose suitable conditions and follow local access rules. Avoid unstable ground, hazardous branches, storms, and unsafe temperatures. Protect yourself appropriately from sun, insects, and irritating plants. Leave bark, roots, wildlife, and habitat undisturbed; do not assume a tree is a safe anchor for suspended equipment. Keep needed mobility aids and emergency or navigation tools available.
Spiritual meaning and scientific evidence are distinct. The original book includes interpretations about energy and health. New teaching should not turn those interpretations into medical guarantees or present electromagnetic protection or universal healing as established facts. Preserve the source honestly while reviewing practical guidance against appropriate evidence and safety standards.
Repository code and associated documentation are covered by the MIT License, copyright © 2026 Alex Julian.
The linked Tree Yoga School book carries its own notice: copyright © 2023 Tree Yoga School. All rights reserved. This repository does not relicense that book. Its illustrations, photographs, and other externally linked materials are not automatically covered by the repository’s MIT license. Any future import or adaptation needs an explicit rights and licensing decision before publication.
Founded by Alex Julian. Rooted in the original Tree Yoga School. Growing in public.
Build the school. Keep the practice alive. Leave time to sit beneath a tree.
The /begin-here page connects the first five-minute practice to three introductory pathways: Meditation with a Tree, Gentle Yoga with a Tree, and Tree Yoga Hiking. Lessons use a shared structure: purpose, preparation, accessible options, practice, a screen-away stopping point, reflection, and book references. The new sequences are identified as adaptations developed with AI assistance.
Lesson content lives in lib/intro-lessons.ts; the shared lesson route is app/lessons/[slug]/page.tsx. Only the three authored slugs are generated. The original first lesson keeps its existing URL.
/trees introduces the Tree Library; /trees/pine is its first complete entry. It focuses botanical identification on Eastern white pine (Pinus strobus), with licensed photographs, observable features, seasonal observation, contemplative energy themes, a new five-minute practice, book references, and reflection. Constancy, clarity, and perseverance are new interpretations for this digital edition, not fixed traditional correspondences or claims about a measured healing field.
/book connects all seven chapters to native reading companions at /book/[slug]. These summarize selected themes and link to the full original PDF; they do not reproduce the entire book. New practice connections and adaptations are labeled. Chapter 6 retains the ten principle names and connects to Pine; the Tree Library links back to the relevant chapters and source pages.
Tree records live in lib/trees.ts; chapter records live in lib/book-chapters.ts. Both route families generate only authored slugs and return 404 for unknown entries. Add a tree record and any licensed assets to grow the library; keep botanical information sourced and distinguish new interpretation from original teaching or cultural tradition. Photo attribution and licenses are recorded in public/images/trees/ATTRIBUTION.md and on the Pine page.
The initial Tree Library included Pine, Oak, Birch, Maple, Willow, Beech, Hemlock, Cedar, Aspen, and Spruce. Each entry in lib/trees.ts supplies species details, licensed imagery, energy reflections, a distinct practice, and book connections to the shared static page template.
The following entries describe their original release snapshots. Cumulative counts and phrases such as “now” in these historical notes apply to that release, not current coverage. See the generated current coverage summary above.
/champion-trees explores all 139 rows of the supplied May 2026 DCR workbook across 61 municipalities. Leaflet is loaded in a client effect; normal browser requests load OpenStreetMap tiles with attribution. Records and approximate MassGIS town centroids are bundled locally, so no geocoding service or API key is needed. All markers represent towns, never precise tree positions. The complete searchable list works independently of map loading.
The original workbook is downloadable under public/data. Source values are preserved, including duplicate names, missing dates, zero crown measurements, spelling, and discrepant points. Detail cards flag point differences greater than one point and the Salix nigra/Black walnut name mismatch. Published locations do not imply public access.
To regenerate records, install Python openpyxl and run python scripts/import-champion-trees.py path/to/workbook.xlsx. The importer validates columns and the expected edition count before writing. Review schema and counts for any future edition. Coordinates were retrieved October 6, 2026 using returnCentroid=true, outSR=4326 from the MassGIS Massachusetts Municipalities FeatureServer layer 1 (https://services1.arcgis.com/hGdibHYSPO59RG1h/arcgis/rest/services/Massachusetts_Municipalities/FeatureServer/1). Manchester is mapped to Manchester-by-the-Sea; no address geocoding is performed.
The same /champion-trees explorer now includes 93 NH register records with Champion Status exactly State, retrieved October 6, 2026, alongside all 139 MA records. The combined map has 232 records in 122 municipalities. State, county, town, genus, search, location, and sort controls apply to the combined data. State-qualified town keys keep places such as Manchester, MA and Manchester, NH separate. Changing state clears dependent county/town selection. Reset returns to both states.
lib/data/new-hampshire-champion-trees.json preserves tree IDs, source place names, scientific labels, measurements, nomination dates, State designation, the source's national-status checkbox, and exact paginated source URLs. The national flag is not independently verified and is not displayed as a claim. Nomination dates are never presented as measurement dates. Circumference units are absent from the source header and remain unspecified. All state-designated entries are kept (77 scientific labels); multiple entries for a species are not deduplicated. County-only, dead, dethroned, disqualified, and other non-State entries are excluded.
Extraction reconciled 806 unique public register records: the date-sorted table can repeat rows across pagination boundaries. Two full sweeps plus reverse-order boundary reads reconciled all IDs with no conflicting row values. Source: https://www.nhbigtrees.org/trees. Future refreshes must reconcile unique IDs and validate status rather than assume ten distinct rows per page. The snapshot is bundled locally and does not fetch the register during page rendering.
lib/data/new-hampshire-town-points.json contains only the used municipality representative points from the Census Bureau 2025 County Subdivisions Gazetteer: https://www2.census.gov/geo/docs/maps-data/data/gazetteer/2025_Gazetteer/2025_gaz_cousubs_33.txt. These are not tree coordinates or village centers. Source aliases: East Swanzey → Swanzey (https://www.swanzeynh.gov/community/town_history.php); Rye Beach → Rye (https://www.ryenh.gov/197/Assessing-Department); Sanbornville → Wakefield (https://www.wakefieldnh.gov/about-our-town); W. Franklin → Franklin (https://www2.nhes.nh.gov/GraniteStats/SessionServlet?SID=1&city=000712&cityName=Franklin&page=Community.jsp). Boscowen is interpreted as a spelling variant of Census-listed Boscawen in Merrimack County; this inference is visible in the page and record note. Original source place names stay unchanged.
The full NH register provides town/county only and does not establish public access. Cards link to UNH's separate visiting map for a public-access subset: https://extension.unh.edu/resource/new-hampshire-big-tree-map. No street address, access claim, or individual-tree coordinate is inferred.
At overview zoom levels, town markers within 44 screen pixels group together to avoid overlap. Counts sum all matching records; selecting a group zooms to its towns without filtering away records. Individual town markers filter results. Both actions support Enter/Space. Groups redraw on zoom, with effect cleanup removing the zoom listener.
The explorer now contains 323 records across MA, NH, and VT, in 176 state-qualified municipalities. Vermont adds 91 confirmed champions in 54 municipalities. State choices derive from the shared state names and record counts. The new “Source confirms public access” filter selects only records whose source explicitly says yes; this classification is available for Vermont, with 24 yes and 67 no. MA/NH access remains unclassified.
Vermont's official app is https://experience.arcgis.com/experience/7637f5256b65454aa123e0c631f1f46a/page/Map/ (linked by https://vtcommunityforestry.org/projects/vermont-big-tree-program). Its app item points to web map 4baeb5cb5bd14459afe4f4cfd0911dbf, whose Champion Trees and table layers use BigTreeSurvey_ExperienceView/FeatureServer/0 with the filter Champion = 'yes'. The public view exposes only curated register fields. This October 6, 2026 snapshot matches the official layer's confirmed-record count of 91, even though the introduction still says nearly 80. The program announced a staffing-related hold in September 2026; the page and cards disclose this.
Run python3 scripts/import-vermont-champions.py from the repo to regenerate the two Vermont JSON files. The importer queries only public-view fields, omits geometry, checks the server count and unique IDs, and requires review if the count changes. Scientific names combine the published genus and species with whitespace normalized; spelling and capitalization remain unchanged. County suffixes are normalized for shared filters. Blank common name (record 188) and the Table Mountain Pine/Picea pungens discrepancy (record 125) are disclosed. All confirmed entries are retained, including multiple trees for a species. dateTime is labelled “Data collected,” not inferred as a remeasurement date; year listed is separate.
Main map markers remain approximate municipality points, sourced from the Census 2025 Vermont County Subdivisions Gazetteer. Barre Town and Rutland City are matched to their specific town/city entries. For the 24 public-access records only, coordinates are parsed from the explicitly published location_details text and offered as a separate map-search link. Private tree geometry is neither requested nor stored. Published road-visibility notes do not grant entry. The complete result list remains available when map tiles fail.
The map now contains 469 records across MA, NH, VT, and ME at 241 state-qualified town/township points. Maine adds all 146 rows of the official 2020 Register of Big Trees, downloaded October 6, 2026, at 65 points. The edition is explicit in cards and source notes; this is not a claim of current champion status. Original names (including typos/case/asterisks), measurements, points, and last-measured years are preserved. No day or month is invented. Each card links to its source PDF page. Duplicate species and the two English oak entries in Orland remain separate. The source prints 3012 points for Vassalboro's bur oak; the existing score discrepancy warning explains this without silently changing the score. The PDF contains no asterisk legend, so symbols are preserved without assigning designations.
Regenerate with python3 scripts/import-maine-champions.py /path/to/maine.pdf (Python standard library plus Poppler pdftotext). Official PDF: https://www.maine.gov/tools/whatsnew/attach.php?an=1&id=3332272. Import validates 42/41/42/21 rows on PDF pages 2–5, all numeric columns, bounds, and unique geography matches. The wrapped Japanese falsecypress name is joined into a single field.
Geography: 2025 Census Maine County Subdivisions Gazetteer, with Maine GIS Maine_Town_and_Townships_Boundary_Polygons/FeatureServer/0 centroids for Bradstreet and Moosehead Junction townships. South Paris maps to Paris; Milton Plantation maps to Milton UT; Bradstreet Twp. maps to Bradstreet Twp. Counties derive from official geography and are labeled accordingly. No precise tree position or public access is inferred. Maine access remains unclassified and is excluded by the public-access filter. The map's minimum zoom now permits the larger four-state extent to fit on narrow screens.
The map contains 624 records across five states at 269 state-qualified municipality points. Rhode Island adds 155 usable rank-one rows from RI Tree Council’s March 24, 2026 table, retrieved October 6, 2026: https://ritree.org/champion-tree/. Multiple entries for the same species remain separate. Original scientific/common labels, addresses, fractional measurements, scores and rank are retained; whitespace is normalized. Edition date is not a measurement date. Access remains unclassified. TR/HT/CR correspond to circumference in inches and height/crown in feet, corroborated by the named-tree measurement profiles on the same source page.
Regenerate with python3 scripts/import-rhode-island-champions.py (standard library); optional saved HTML and Gazetteer paths allow offline reproduction. Header, edition, row count, rank, final malformed row, unique geography matches and geographic bounds are validated before writing. Source row 156 has numeric species/name/place fields and is quarantined in lib/data/rhode-island-excluded-rows.json, not mapped. The public source notes disclose its exclusion. No replacement from an older edition is substituted.
The 28 municipality points and counties derive from the 2025 Census RI County Subdivisions Gazetteer. The empress-tree row lists Swan Point Cemetery as its town and Forest Street as its address; both original fields are preserved. mapTown: Providence is grounded in https://swanpointcemetery.com/ and disclosed in the card and source notes. Markers remain approximate municipality points; a published address is not public-access permission. Rhode Island location searches include the source site/address and the municipality. Tree Library links use the existing exact scientific-name matches.
The New England map includes all six states: 1,128 records at 366 state-qualified municipality points. Connecticut adds 504 explicitly designated state champions (400 champions and 104 co-champions) across 97 municipalities, retrieved October 6, 2026. Source: https://oak.conncoll.edu/notabletrees/. The common-name and scientific-name champion tables are reconciled by Tree ID, including all numeric fields, town and designation; individual public record pages supply measurement dates and published addresses. Other notable/historic trees are not imported. Species, hybrids and cultivated varieties retain their original labels. National designations are not promoted into independently verified claims.
Regenerate with python3 scripts/import-connecticut-champions.py COMMON.html SCIENTIFIC.html DETAILS_DIR GAZETTEER.txt, using the two ChampsByCommonName.jsp / ChampsByScientificName.jsp tables, one ViewTreeData.jsp?selected=ID HTML file saved as ID.html for each listed ID, and the 2025 Connecticut County Subdivisions Gazetteer. Alternatively, python3 scripts/import-connecticut-champions.py --download NEW_SNAPSHOT_DIRECTORY downloads those sources with six concurrent requests. Cached inputs are reused for reproducibility; use a new directory for a fresh snapshot. HTML is decoded as ISO-8859-1 as declared by the source. The importer validates 504 unique matching IDs, 104 co-champions, list/detail measurement agreement, dates, and unique geography matches before writing JSON. Refreshes that change those counts require review. Common names use the individual record wording; alternate labels in the alphabetized common-name table are retained in searchable notes. Named-tree labels are preserved. Photos and descriptive prose remain on the linked source records.
Connecticut's Census geography uses nine planning regions as county equivalents. county stores the region base name for shared filtering, while the UI labels the control “County or planning region” and CT detail cards say “Planning Region.” Region names and representative municipality points derive from Census geography, not from the champion register. Source: https://www2.census.gov/geo/docs/maps-data/data/gazetteer/2025_Gazetteer/2025_gaz_cousubs_09.txt and https://tigerweb.geo.census.gov/tigerwebmain/Files/acs26/tigerweb_acs26_county_ct.html.
Published addresses support the location filter and map-search link. Omitted addresses remain undisclosed; exact tree coordinates are not copied or inferred. No public-access permission is inferred from a published address, park name, or champion designation. CT remains excluded from the explicit public-access filter. Each tree links to its specific source record, including photographs and additional visiting context where published.
State register data is server-only and generated into six static /api/champion-trees/[state] snapshots. The explorer requests only states in the selected region/state and caches successful selections for the session. Failed selections show a retry control without presenting a partial list as complete. Data updates remain tied to deployments. Results render in batches of 50; all matching mapped records remain on the map.
A lightweight manifest reports listed and mapped record counts independently. Mapped means an approximate municipality point exists; records without a point remain searchable. Coverage cards distinguish edition dates from retrieval dates and link to each state's source notes. New England covers six states; Northeast adds New York and reserves New Jersey and Pennsylvania. Future regions should be exposed when they contain imported registers.
Adds 210 records (162 champions, 48 co-champions) from NYS DEC’s January 31, 2025 PDFs, retrieved October 6, 2026. Total: 1,338 records across seven states, with 366 municipality points and 37 New York county points. The importer reconciles both PDFs, accounting for two differently clipped scientific labels. Four county spelling/truncation aliases are documented in source notes. Original crown points remain a separate field; no full spread or precise tree location is invented. Run scripts/import-new-york-champions.py with both PDFs and the Census 2025 national county Gazetteer to reproduce. New York has no public-access classification.
Adds 206 NJDEP records with PERMISSIONTOLIST=YES and Champion (182), Co-Champion (9), Heritage_Champion (8), or National_Champion (7) status. Retrieved October 6, 2026; source data last edited March 19, 2026. Total coverage: 1,544 trees, eight states. Municipality matches require a unique Census name within the published county; 110 records use 59 town points, 96 use 17 county fallbacks. Original municipality and ranking fields stay in details/notes. No exact tree geometry is imported or public access inferred. Crown units are flagged as inconsistent and values are neither converted nor used to recalculate scores. Import with scripts/import-new-jersey-champions.py, the filtered non-geometry ArcGIS JSON, NJ 2025 Census subdivisions, and national 2025 county Gazetteer.
Adds 445 derived leaders from the user-supplied October 2026 PA Big Trees extraction, for 1,989 records across nine states. Compares the 1,520 public listing records by source botanical category; retains five pairs tied on published points and four explicitly legacy national flags. Excludes Malus spp., Unknown Fir, and the Northern Pin Oak explicitly marked removed, without promoting an unverified replacement. All records link to the original page and use 48 Census 2025 county points. No tree coordinates or public access are inferred. Of the included records, 384 have checked detail measurements and 61 retain listing fields with null missing measurements. London plane measurement and Chestnut Oak condition concerns remain visible. Run scripts/import-pennsylvania-champions.py EXTRACTION.json COUNTIES.txt to reproduce. Original extraction file: pennsylvania-champion-tree-extraction(2).json; no automated live refresh.
The October 6, 2026 snapshot of Delaware Forest Service's Big Trees Playground uses FirstMap Biota/DE_Big_Tree_Champs/FeatureServer/0. The official experience 43ab90c14951412eaa751004b5cc4597 links this exact layer; the state's Forest Service home page links the experience. Of 207 records, include all 79 with SPEC_RANK=1 and 12 lower-ranked records within five published points of the scientific-name group maximum. The latter are explicitly rule-derived co-champion qualifiers using printed page 3 of the 2019 fifth edition, not explicit GIS co-champion labels. Preserve equal scores, 3 multi-rank-one species, source ranks, dates, IDs, and original record links. One rank-three ginkgo has no scientific name and is excluded. The separate Unusual Big Trees layer is not imported. No national designation is inferred.
DATE_OF_RANK is a ranking date, never a measurement date; latest is 2026-10-06. Measurements and scores remain as published; discrepancies are surfaced by the existing source-warning function. Public access is unclassified. No individual tree coordinates or owner information are imported. 86 records use 31 Census Places Gazetteer points; five unmatched places (Canterbury, Rockland, Yorklyn) use two county points. 91 records, 76 scientific names, 33 mapped places. Total map: 2,080 records across 10 states. Mid-Atlantic shortcut includes NY/NJ/PA/DE, with MD reserved; Northeast remains the original nine states.
Reproduce with the complete FirstMap query JSON (where=1%3D1&outFields=*&returnGeometry=false&f=pjson), 2025 2025_gaz_place_10.txt, and 2025_Gaz_counties_national.txt:
python3 scripts/import-delaware-champions.py FIRSTMAP.json PLACES.txt COUNTIES.txtThe October 6, 2026 public State Champion Trees table, linked by Maryland DNR, contains 274 explicitly State-designated entries across 239 published botanical categories. Preserve every designation, including multiple trees in a category and 13 source-reported National labels. Do not derive champions from points or infer current national status. Public access is Yes for 127 and No for 147.
lib/data/maryland-champion-trees.json retains record links, IDs (including suffixes), names, measurements, dates, public-access labels, towns, and published addresses. It excludes GPS coordinates, images, personal owner/nominator details, and free-text comments. All records use 24 representative Census county/independent-city points, with Baltimore City distinct from Baltimore County.
Rebuild using python3 scripts/import-maryland-champions.py PUBLIC_DOM_ROWS.json 2025_Gaz_counties_national.txt. The input is a read-only extraction of the publicly rendered Airtable grid: map record IDs to {id, href, cells}; cells uses published column indices documented in the importer. Extract all 274 rows and verify access and designation fields; the importer fails on incomplete required fields. The retrieval date is not a measurement date. Maryland joins the existing lazy-loaded API, state and Mid-Atlantic filters, source cards, and listed/mapped counts.
The Library now contains 16 species profiles. Red maple, northern red oak, yellow birch, balsam fir, tamarack, and American sycamore each have identification guidance, seasonal observations, original contemplative prompts, and a short accessible practice. Botanical observations are separated from creative interpretations.
lib/tree-library-expansion.ts contains the six profiles; lib/tree-images.ts contains their photograph metadata and identification galleries for the original nine non-pine profiles. Pine retains its existing gallery. The expansion adds 50 real photographs with individual credits and licenses; original sources and conversion details are recorded in public/images/trees/ATTRIBUTION.md. Galleries preserve full photographs rather than cropping identification features.
The six exact scientific-name mappings in lib/champion-trees.ts make the profiles available from existing Champion Map records. Existing location precision, access labels, and source-review warnings continue to apply.
Every Library profile links to an exact scientific-name selection on the Champion Map. Map records link back to the species practice. The address bar tracks region, state, search text, county, mapped place, genus, exact species, access/location filters, sorting, and selected record. Native browser history restores selections without reloading state datasets. Search typing replaces the current history entry; other filter changes and record selections add an entry.
Copy selection link shares the filters without an open record. Copy tree link shares the selected record and its state independently of the current filters. Clipboard restrictions reveal a selectable link instead. A missing or filtered-out record displays recovery options; selected records beyond the first 50 results are included automatically.
URL parameters: region, state, q, county, place, genus, species, public=1, location=1, sort, and tree. Record IDs retain their source-derived values. Unknown regions, state codes, and sort values fall back safely. Query parsing/serialization regression checks: node --test scripts/test-champion-links.cjs.
The visit directory uses compact cards and combined state, species, and visit-type filters. Selections persist in shareable query URLs and browser history; individual guide pages retain the full visitor guidance. When adding a state, aim to pair its register with one researched public visit, using official access and arrival sources before publishing. A state register alone does not establish permission to visit.
The initial /tree-visits release offered three source-checked starting points: the Pinchot Sycamore (CT), University Green dawn redwood (VT), and Smith College castor aralia (MA). lib/tree-visits.ts contains stable champion record IDs, arrival/access guidance, official sources, check dates, and related practices. Check dates describe online research, not field inspections. These guides do not alter source-register access flags or map coordinates. Links from the Library, Champion Map, and matching record details make the collection discoverable.
The Champion Map’s Tree species selector displays familiar names, scientific names, and counts within the loaded state/region. It filters exact published botanical categories; cultivars and hybrids remain distinct. lib/champion-species.ts normalizes inverted common names and prefers common names over botanical fallbacks. Choosing species clears genus, and choosing genus clears species. Existing species URLs, sharing, and browser history continue to work. Regression coverage is in node --test scripts/test-champion-links.cjs.
Virginia adds 369 records from Virginia Tech’s live State Champions list, retrieved October 7, 2026. Each detail page was checked for alive status and explicit Virginia Champion designation, and reconciled with the list’s names, locality, and score. Original measurements, measurement years, national labels, and stable source IDs are retained. National labels are source designations, not independently verified national titles.
All Virginia markers use 2025 Census county or independent-city points, keeping cities distinct from similarly named counties. No private owner/contact fields or tree GPS are imported; public visiting access remains unclassified. Virginia is available in the state selector, Mid-Atlantic region, familiar species selector, sorting, and shared URLs. The per-state static endpoint is /api/champion-trees/VA.
Reproduce the import with scripts/import-virginia-champions.py SNAPSHOT_DIR (requires Beautiful Soup). The directory must contain the current list as list.html, each detail page as <database ID>.html, and the 2025 national Census counties Gazetteer as counties.zip. The importer checks every detail/list match before writing either dataset. Browser and regression checks include city/county distinction, map coverage, stable IDs, and shared Virginia selections.
South Carolina adds 196 records explicitly marked State Champ in Clemson University's official database, bringing the map to 3,424 records across 15 states. The Southeast shortcut includes North and South Carolina. All South Carolina records map to 35 representative Census county points; public access remains unclassified.
The official Clemson program page links Experience Builder item 25dbba11925a4bf8b22d6430581a02b2, whose web map 9c0fd85e35614c42a9a0cbec58707852 uses https://services1.arcgis.com/x5wCko8UnSi4h0CB/arcgis/rest/services/South_Carolina_Champion_Tree/FeatureServer/0. The complete query contains 269 records: 196 State Champ, 66 State Runnner-Up, and seven other entries. Retain all designated champions across 190 botanical categories, including two entries in each of six categories; do not derive extra co-champions from scores. Ten National Champ and one National Co-Champ annotations remain source labels, not independently verified titles.
Use the layer's /query endpoint with where=1=1, returnGeometry=false, orderByFields=OBJECTID, f=json, and the following outFields allowlist:
OBJECTID,tree_county,tree_common_name,tree_science_name,tree_cond,statechamp,champion,submission_date,Score,Circumference,Height,Crown,Reassessed_Date,Needs_Reassessment,Reassessed
Save the complete response as RECORDS.json and download the 2025 Census county Gazetteer ZIP, then run:
python3 scripts/import-south-carolina-champions.py RECORDS.json COUNTIES.zip
node --test scripts/test-champion-links.cjsThe importer rejects incomplete or changed snapshot totals, duplicate source IDs, unknown counties, and invalid measurements before writing outputs. Source names, scores, and condition text are preserved with whitespace normalized. Tungoil tree sc-250 retains null circumference and score. Twelve records lack dates. Clemson's web map labels submission_date as Measurement Date, while layer metadata calls it Today's Date; the UI uses “Date shown by Clemson” and explains the field. The layer's last data edit is November 25, 2025; snapshot retrieval is October 7, 2026. Condition descriptions and dates are historical source observations, not current inspections. No owner/contact fields, addresses, coordinates, attachments, or remarks are imported. Source OBJECTIDs provide stable sc-<id> share links.
The University of Tennessee Current Champion Trees page adds 183 entries across 164 published scientific names. Preserve every published entry, including repeated species; do not infer co-champion or national labels from scores. The page gives no measurement dates, edition date, exact tree coordinates, or access classification. 182 records map to 38 representative Census county points. One entry says Lafayette in the county field: retain it with a review note, without guessing a county or map point.
Reproduce from the complete HTML of https://naturalresources.tennessee.edu/trees/ and the 2025 national Census county Gazetteer ZIP:
python scripts/import-tennessee-champions.py TREES.html COUNTIES.zip
node --test scripts/test-champion-links.cjs scripts/test-champion-expansion.cjsThe importer requires Beautiful Soup, checks source totals and numeric fields, and uses hashes of source names/place/image URL as stable snapshot identities because UT supplies no record IDs. Measurement changes and source row reordering do not change these identities; changes to the identity fields require reconciliation before a refresh. No source photos, owner contacts, or inferred locations are included in the map data.
Tulip tree, sassafras, blackgum, and dawn redwood bring the Library to 20 species. Each has botanical guidance, credited real photographs, original contemplative prompts, and a short practice. Exact scientific-name links connect profiles and map records. The University Green dawn redwood visit now links to its own species practice.
State registers publish independently as downloads complete. A failed request leaves successful records and map markers usable, names missing states, and offers a retry that skips cached successes. Requests time out after 20 seconds. Partial result counts are labeled; empty-state messaging does not claim there are no matching trees while registers are missing.
American hornbeam, shagbark hickory, silver maple, and black cherry bring the Library to 24 profiles. Each includes three identification notes, seasonal guidance, three original contemplative themes, a four-step practice, and NC Extension botanical sources. Thirteen licensed real photographs include individual credits and source links in the page and image attribution file. These are new practices, not species teachings attributed to the original book.
Exact scientific-name connections add 79 existing champion records (21 hornbeam, 20 hickory, 20 silver maple, 18 black cherry), bringing direct Library coverage to 386 of 3,607 records. No champion data changes.
The Library filters its existing profiles by common name, scientific name, selected alternate names, and one practice theme. Name terms are case-insensitive, ignore punctuation, and combine with the theme. Theme choices derive from the current profiles. Matching shortcuts, result counts, empty-state recovery, reset controls, and direct practice links update without a network request. The server sends only card fields to the interactive component; complete practices and galleries stay on species pages. An Illustrated/Compact toggle preserves the current search and theme selection; Compact shows text cards with profile and outdoor-practice links. Species pages suggest at most three related trees, prioritizing the same botanical genus and then shared contemplative themes, with reasons shown beside each link. Search logic and current profile integration are covered by node --test scripts/test-library-search.cjs.
Georgia is available in the State selector, Southeast region, shared record URLs, and static /api/champion-trees/GA snapshot. The official source is https://gatrees.org/learn-explore/champion-trees/.
The 407 summary rows contain 403 unique published IDs. IDs 1950 and 1673 repeat; each is counted once using the first linked record. Pecan ID 1950 has conflicting spelling, county and designation variants, retained in lib/data/georgia-import-audit.json and explained in its record notes. The import preserves all unique register IDs, including blank/ambiguous designations, without deriving new champion titles from scores. Individual record values and measurement dates are used when available; unavailable detail pages fall back to the summary row with an explicit note.
398 entries map to 99 approximate Census county points; five entries have no county and remain searchable without markers. Source county capitalization is retained separately (for example, Dekalb → DeKalb for mapping). Simple species-epithet capitalization is normalized and documented; source spellings and varieties are otherwise preserved. GFC's nomination form confirms circumference in inches and height/crown spread in feet. Published scores, zeros and missing values are retained. Public visiting access is unclassified; exact tree positions, owner and nominator details are not imported.
Reproduce with saved register HTML, individual pages named by URL slug, and the Census 2025 Counties Gazetteer ZIP:
python scripts/import-georgia-champions.py REGISTER.html DETAILS_DIR COUNTIES.zip
node --test scripts/test-*.cjs
npm run buildThe import audit records duplicate variants, summary/detail differences and unavailable detail pages. Georgia regression checks cover county precision, missing locations, stable source IDs, shared selections, measurements and duplicate handling.
Indiana adds 93 populated records from Indiana DNR’s current champion table, including 10 source-labeled co-champions, bringing the map to 4,210 entries across 19 states. Two empty rows (sugar maple and Canada plum) are excluded and recorded in lib/data/indiana-import-audit.json. The source contradicts itself about umbrella magnolia: both populated co-champion rows remain with a warning. Published measurements and scores are unchanged; no measurement dates or access permissions are inferred.
All 93 records map to 40 approximate Census 2025 county internal points. St Joseph and Laporte are matched to St. Joseph and LaPorte, retaining original county wording. Local IDs hash scientific name, canonical county, and common name without the co-champion suffix; measurement updates and row reordering do not change them. Name or county changes require reconciliation. Indiana is available through the state selector, Midwest shortcut, species filters, shared record URLs, and /api/champion-trees/IN.
Reproduce with saved register HTML and the Census 2025 national counties Gazetteer ZIP (Python standard library only):
python scripts/import-indiana-champions.py REGISTER.html COUNTIES.zip
node --test scripts/test-*.cjs
npm run buildAlabama includes all 145 tree rows on pages 3–9 of AFC’s 2025 PDF, retrieved October 8, 2026, with 49 approximate Census county markers. Names and published scores are preserved; year crowned and remeasurement due are separate from unknown measurement dates. Visiting access is unclassified. Repeated species remain separate records without inferred co-champion labels.
Regenerate using python scripts/import-alabama-champions.py REGISTER.pdf COUNTIES.zip (requires pdfplumber and the 2025 national Census county Gazetteer ZIP). The importer records source SHA256 and audit counts. Run node --test scripts/test-alabama-champions.cjs.
Florida includes 311 unique champion/co-champion IDs from the Forest Service summary table retrieved October 8, 2026, mapped to 47 approximate county points. Of 600 rows, 313 match the four included designations; repeated IDs 155 and 1270 are identical and deduplicated. Exclude challengers, emeritus, discontinued, nominees, and unknown designations. Preserve national labels as source designations, without independent verification. Summary-only import: dates, ownership, coordinates, and access from detail pages are not inferred.
Regenerate with python scripts/import-florida-champions.py REGISTER.html COUNTIES.zip (requires lxml and the 2025 Census national county ZIP). Audit includes source hash, exclusions, and duplicate rows.
Illinois Extension's official program page links Experience Builder item df5e7296d76a4c8ba133c9e8adf1f85a. Its web map 32415ec76af74ff8b50165d7b1e360e1 uses https://services.arcgis.com/GL0fWlNkwysZaKeV/arcgis/rest/services/IBTR_masterfile/FeatureServer/0. The unfiltered layer has 202 records (also confirmed with returnCountOnly=true): 124 populated entries across 106 botanical categories, including 18 name-labeled co-champions, and 78 empty species slots without counties or measurements. Include every populated entry; do not derive extra titles or discard designated co-champions based on points. Dataset last edited July 30, 2026.
All 124 records map to 47 representative Census 2025 county points. Saint Clair maps to St. Clair while preserving source wording. Exact tree coordinates, street locations, nominator details, and images are not imported; public visiting access remains unclassified. Stable il-<ObjectId> IDs support shared selections. Circumference is displayed in source feet; the internal inches field is feet × 12 for existing score checks. Published points are unchanged. The combined Year listed/Last measured text remains separate from the null measurement date. Missing silverbell circumference and the cottonwood's non-typical form are retained with notes. Scientific-name synonym annotations are preserved as published.
Illinois joins the Midwest shortcut, state/species filters, shared URLs, source dates, and /api/champion-trees/IL. Total coverage is 4,790 records across 22 states. The Library's map teaser now derives its state count from available states.
Reproduce using the layer's /query endpoint with where=1=1, returnGeometry=false, orderByFields=ObjectId, f=json, and this outFields allowlist:
ObjectId,GlobalId,name,Species,Common_name,County,Year_listed_Last_measured,Diameter_at_breast_height__in__,Circumference__ft__,Height__ft__,Spread__ft__,Total_Points,shortlist_id,tab_id
Save the complete response as RECORDS.json, download the 2025 national Census county Gazetteer ZIP, then run:
python scripts/import-illinois-champions.py RECORDS.json COUNTIES.zip
node --test scripts/test-*.cjs
npm run buildThe audit stores source SHA256, inclusion counts, and excluded slot IDs. Refreshing a changed source requires reconciling counts, IDs, and schema before updating importer assertions.
python scripts/import-michigan-champions.py REGISTER.xlsx COUNTIES.zip imports
the Michigan Botanical Society public spreadsheet linked from
https://michiganbotanicalsociety.org/big-tree-on-google. Download the XLSX via
https://docs.google.com/spreadsheets/d/1x0l0BRXxdGzV6kxMLtvNBb88x0jqlTJfNLpQGZtjrB8/export?format=xlsx
and use the Census 2025 national county Gazetteer ZIP.
The October 9, 2026 snapshot has one sheet, Big_Tree_Data_2026-07-14, with 667
populated rows. It has no state-champion designation field. We include only the
highest published Points per scientific-name species category, including exact
ties: 122 leaders across 120 categories at 38 county points. The 545 smaller
entries are excluded. Variety metadata does not create extra categories. The two
spellings of Magnolia x soulangeana are grouped for ranking while the original
scientific names stay on cards. No additional near-score co-champions are inferred.
The import audit retains every candidate ID, source row, category, score, and selection decision, plus the XLSX SHA-256. Names, measurements and scores follow the source; common-name fallbacks are disclosed. Verification dates are preserved, including five older than ten years and one missing. County names are normalized against Census geography, including Gd. Traverse → Grand Traverse. GPS coordinates, addresses, owners and location field notes are omitted. Visiting access remains unclassified. National Y flags are source labels, not independently verified titles.
Ohio includes all 127 entries from the Native Ohio Champion Trees table, transcribed and checked against the six supplied screenshots captured October 9, 2026 (pages 1–6, ending at 127 of 127). All 126 non-native entries from six additional screenshots (pages 1–6, ending at 126 of 126) are included. There are 253 Ohio records: 252 mapped to 67 approximate Census county points and one Gray Birch with a blank source county, retained in results without an invented location.
The import audit records screenshot hashes, page ranges, published totals, and coverage for each list. All source rows are preserved, including repeated botanical categories and decimal measurements/scores. Category follows the source table. Existing record IDs remain stable; same-species/same-county rows use common names to distinguish separate trees. No measurement dates, precise locations, or visiting permission are inferred. American elm’s published score discrepancy is retained and flagged. County points use the 2025 Census Gazetteer.
Non-native source rows are retained independently, including repeated Common Apple, Hedge Maple, and both Western Red Cedar rows (identical measurements and county, differing scientific-name capitalization; possibly the same tree). The Oriental Arborvitae scientific label has replacement-character encoding artifacts normalized to spaces. Published fractional scores and source classifications are preserved. All previous Ohio record IDs remain valid.
Wisconsin adds all 57 records exposed by both public layers behind DNR’s official map and data table: 47 with Map_Display = 'Exact location' and 10 with Map_Display = 'Township'. Case-insensitive ArcGIS matching retains one Exact Location capitalization variant. No rank or score selection is applied. Records use stable global IDs and preserve repeated species. Crown width averages the two verified diameters, matching the web-map expression; raw diameters remain in notes. Measurements and published scores remain unchanged.
All 57 map to 21 approximate Census county points; milwaukee is normalized to Milwaukee while the original county stays available. No exact geometry or GPS fields, owner names, access instructions, or private geometry are imported. The 28 source-labelled private records use publicAccess=false; 29 others remain unclassified. A public exact-location label does not authorize a visit. DNR’s staffing-related program pause is disclosed. Measurement dates are unknown; dataset edit dates are separate.
Reproduce using python3 scripts/import-wisconsin-champions.py /path/to/2025_Gaz_counties_national.zip. The importer checks both server counts, transfer completeness, unique global IDs and county matching. Review count changes before accepting a new snapshot. Total map coverage: 5,222 records across 25 states.
All 62 current trees from Minnesota DNR’s var species array are included across 51 populated native species. Two of 53 species slots have no tree (Northern mountain ash and Mountain maple). DNR’s page explicitly displays Co-Champions for species containing multiple records; retain every record. Total points reproduce the official getTotalPoints formula exactly. Measurements, location descriptions, ownership and source year remain as published; the year appears beside the nominator and is not assumed to be a measurement date. National flags remain DNR labels.
All records map to 28 approximate Census county points. St Louis/St.Louis normalize to St. Louis, and Fairibault to Faribault; original counties are preserved. The 24 Private records use publicAccess=false, while Public/Federal/Tribal ownership remains unclassified for visiting. No exact coordinates or nominator names are imported. Stable IDs hash the published species/location/year/measurements. Reproduce with python3 scripts/import-minnesota-champions.py PAGE_HTML COUNTY_GAZETTEER_ZIP. Total: 5,284 records across 26 states.
Green Ash (Fraxinus pennsylvanica), Black Oak (Quercus velutina), Sweetgum (Liquidambar styraciflua), and Bur Oak (Quercus macrocarpa) bring the Library to 36 profiles. Each has botanical identification and seasonal notes, three original contemplative themes, four outdoor practice steps, a reflection, and reciprocal champion-map links. The existing profile and practice routes derive these additions from trees.
Botanical sources and photograph credits are recorded in lib/tree-library-map.ts and lib/tree-library-map-images.ts, using NC State Extension's species profiles. Ten photographs use CC BY or CC BY-SA licenses, with photographer, license, and source links on each gallery image. Local WebP assets preserve image proportions. These themes and practices are new interpretations, distinct from botanical facts and the original book's teachings.
The visit collection now has 11 guides across eight states. October 9, 2026 additions: Angel Oak at Charleston's public park (SC), Lady Liberty at Big Tree Park (FL), and the bur oaks in The Morton Arboretum's Oak Collection (IL). Each guide records official visitor sources, arrival and parking, walking and access limitations, hours/fees where published, a pause, and linked outdoor practice.
These three destinations are not asserted to match a current champion record. TreeVisit.championId is optional: the eight original record-linked guides retain their exact links; destination-only guides link to the corresponding state in the Champion Map. No champion data or public-access classification is changed. The Morton guide explicitly describes a collection rather than one designated champion. Coverage counts and discovery copy now derive from shared visit data. Source check dates remain per guide; older source editions, such as Seminole County's 2020 accessibility report and 2025 parks brochure, are labelled and do not assert a field inspection or current route accessibility.
lib/site-seo.ts defines the production origin and discoverable routes. All public content pages have a self-canonical, page-specific Open Graph and Twitter metadata, and a 1200×630 branded PNG preview served from /social-preview?path=…. Champion filter URLs canonicalize to the main map. /sitemap.xml lists content routes without API, query, fragment or image URLs; /robots.txt advertises it. Sitemap dates are omitted rather than invented. Visit guides now have individual /tree-visits/[slug] pages, sharing the overview's exact visitor details and source links.
All 26 states use lib/champion-states.ts for names, provenance, source dates, saved dataset paths, precision rules and reviewed import baselines. lib/champion-state-data.ts normalizes records and coordinate prefixes once for the server loader and validation. Existing source aliases and state-specific warnings retain their behavior.
Run npm run validate:champions for every state, or npm run validate:champions -- --state MN for one source refresh. npm run build runs validation automatically before building. See docs/champion-state-expansion.md for the complete import and release workflow.
All 148 rows from the Kansas Forest Service table are included, with four explicit co-champions and every repeated species. Published names, measurements, scores, nomination years and last-measured dates are retained. Source spelling is not silently corrected, and scores do not determine additional champion designations. No nominator names, private addresses or individual-tree coordinates are imported; visiting access remains unclassified.
134 records match 60 approximate Census place points. Fourteen records in six unmatched places remain searchable without markers. The source does not provide counties, and none are inferred. Census points represent places, not tree locations. Kansas joins the Midwest filter, state selection, species search, shareable links and /api/champion-trees/KS. Total coverage: 5,775 records across 30 states.
Reproduce with python scripts/import-kansas-champions.py REGISTER.html PLACES.txt, using the official register and 2025 Kansas Census place Gazetteer URLs in the state registry. The audit records both source hashes, reviewed counts and explicit unmapped IDs. Changed counts, identities or units require source review before refreshing. Run npm run validate:champions, npm test, and npm run build.
All 90 Nebraska Forest Service register rows are included, preserving numbered entries, source spelling, published scores, and separate nomination/measurement dates. Year-only dates remain years. Circumference feet convert to inches, except Peachleaf Willow’s explicit 124" entry. No titles or visiting access are inferred.
83 records map to 42 approximate Census places. Seven remain searchable unmapped: Sparks, Marsland and five records without an unambiguous municipality. Published location descriptions remain available; exact tree coordinates and personal steward names are omitted. Coverage is now 5,865 records across 31 states.
Reproduce with python scripts/import-nebraska-champions.py REGISTER.html PLACES.txt using the official register and Census URLs in the state configuration. The audit records source hashes, reviewed place labels, counts, unmapped IDs and the unit exception. Regression tests compare every name, date, measurement and score with a reviewed source fixture. Changed source structure, row counts or place labels require review before refresh.
Includes 232 unique Tree IDs from the union of 229 state and 54 national champion designations (51 overlap). Codes 1 and 2 retain champion/co-champion status in separate state and national fields for a future national filter. The other 429 records in the 661-tree registry are excluded. No titles are inferred from scores.
All records map to 79 approximate Census county points. Private owner names and all source tree coordinates are omitted; 152 private, 72 public and eight unclassified ownership labels are preserved without inferring visiting permission. Month/year date precision, source condition labels and unmodified Tree Index values are retained. Total coverage: 6,097 records across 32 states.
Reproduce with python scripts/import-texas-champions.py REGISTRY.json COUNTIES.zip. Obtain the official registry using POST Home/GetAllTrees with JSON {"species":"all"} under the register origin. Use the 2025 Census county Gazetteer listed in configuration. Reviewed fixture comparisons and audited counts require explicit source review if refreshes change names, measurements, dates, ownership or titles.
Adds Eastern Redcedar, Northern Catalpa, Cucumber Magnolia, Eastern Cottonwood, Bitternut Hickory and Flowering Dogwood. The Library now has 46 authored profiles, each with identification, seasonal observations, three contemplative themes and a four-step outdoor practice. Botanical references are NC State Extension profiles; practice associations are original interpretations. Eleven local WebP photographs retain source, photographer and CC BY/CC BY-SA credits, recorded in scripts/fixtures/tree-library-six-photo-sources.json and on the species pages.
New profiles generate six tree routes and six outdoor practice routes automatically, including canonical/social metadata and sitemap entries. Library-to-map species links and map-to-Library links match scientific names without regard to capitalization; botanical synonyms and varieties remain separate. Search aliases support common names such as red cedar, cigar tree and cucumber tree. Tests cover route discovery, attributed local images, aliases and reciprocal matching against all register records.
All 846 entries from both Colorado Tree Coalition 2026 workbooks reconcile across source fields except one cultivar label (Hot Wings versus GarAnn Hot Wings), documented on the record and in the audit. There are 354 position-1/1T listings, further positions through 7, and one unranked entry. Position codes and T suffixes are preserved; scores never generate replacement titles, and “Nat’l Pts.” is not a national champion flag.
All records map to 45 approximate Census county points. Moffatt maps to Moffat with original spelling retained. One street address is omitted; visiting access and measurement dates remain unclassified. Source names, cultivar strings, null measurements, DBH, circumference, crown spread and scores remain intact. County/alphabetical row references provide provenance; hashed full-row fingerprints distinguish entries without inventing source tree IDs. Source changes may change those fingerprints.
Reproduce with python scripts/import-colorado-champions.py COUNTY.xlsx ALPHABETICAL.xlsx COUNTIES.zip, using the URLs in lib/champion-states.ts. The importer compares both lists and every sanitized row against the reviewed fixture before writing. Builds run common import validation; source regression tests cover all 846 records, position codes, missing values, geography and omitted addresses. Colorado joins the West filter and /api/champion-trees/CO. Total coverage: 6,943 records across 33 states.
115 records across 34 approximate parish points, from all six table pages retrieved October 10, 2026. Source page displays “2021 listing” and omits subsequent 2025/2026 additions. Preserve 37 co-champions and one historical national flag without inferring current status. Exclude the deceased East Feliciana Winged Elm; the deceased Sassafras is already absent. No scientific names or measurement dates are supplied. Source raw measurements remain in record notes; scores remain unchanged. DeSoto is matched to Census De Soto Parish, preserving original spelling. Tallowtree’s 71-inch spread is interpreted as 71 feet because that agrees with its score, with an explicit note.
Regenerate with python scripts/import-louisiana-champions.py COUNTIES.zip. The reviewed nine-field fixture preserves all 116 source rows, including the excluded row. Pages 2–6 match the browser extraction checksum (864 fields, 6663 serialized characters, FNV-1a 713304292); first 20 rows come from the official page’s SSR table data. Audit retains exclusions and freshness evidence. Louisiana is in South Central, the state selector, shared URLs, and /api/champion-trees/LA.
The Library now includes 52 tree profiles. Osage Orange, Chinkapin Oak, Post Oak, Kentucky Coffeetree, Slippery Elm, and Butternut add identification notes, seasonal observations, original four-step outdoor practices, and reciprocal Champion Map links. Common-name aliases support discovery without merging botanical synonyms or changing source spellings in champion records. The shared route generators publish all six Library pages and all six practice pages, including canonical metadata and sitemap entries.
Twelve local WebP photographs have verified attribution and licenses from NC State Extension and Wikimedia Commons. Original URLs, dimensions, and credits are retained in scripts/fixtures/tree-library-plains-photo-sources.json and public/images/trees/ATTRIBUTION.md. Photos with NC/ND restrictions and the Commons file Quercus stellata.jpg (now identified by its source as another species) are excluded. ShareAlike adaptations retain their source licenses; the USDA post-oak detail photograph is public domain.
Fourteen visit guides now cover eleven states. Three new public destinations connect the recent map expansion with existing Library profiles and outdoor practices: Chatfield Farms' identified cottonwood, the cypress habitat of LSU's Burden Woods, and Keystone Ancient Forest's post oaks. These are garden/forest visits, without inferred champion IDs. Official visitor sources checked October 10 supply directions, hours, parking, route conditions, and access qualifications. LSU's current trail closing time takes precedence over its older map; boardwalk accessibility is not assumed. The city of Sand Springs' mobility support is qualified by current route and equipment availability. Barataria Preserve is excluded because its visitor center and most trails remain closed for construction.
Each new guide includes two existing licensed Library photographs explicitly labelled as species references photographed elsewhere. Destination photographs on the official sites were not copied without a reusable license. Library links, outdoor practice links, state map links, standalone metadata and sitemap entries are generated through the existing shared routes.
The Library and the six relevant profiles link to three statically rendered comparison pages: white oak/bur oak, red maple/sugar maple, and eastern redcedar/northern white-cedar. Each compares foliage, bark, and acorns/samaras/cones with credited photographs, observable clues, seasonal and age variation, botanical reference links, and profile/outdoor-practice links. Photos retain their full frame with object-fit: contain; mobile cards stack within each feature. No client bundle, external image requests, or additional dependencies are required. Sources were checked online, not through a field identification.
Shared Champion Map metadata, homepage totals, Library coverage counts, and region descriptions follow lib/champion-states.ts. npm run sync:coverage refreshes the README current-coverage block; npm run check:coverage runs before builds and in regression tests. Historical import entries retain their original release counts.
Six cold-cache Chromium runs apply mobile emulation, network throttling, and 4× CPU slowdown to a local production build with Brotli state responses. Results and limitations are in Champion map loading. Downloads stay capped at four; measured long tasks and cached filter latency warrant a focused rendering profile before further substantial expansion. These runs are not a physical phone or production CDN benchmark.
Montana: the complete DNRC 2024 printable register contributes 174 entries (81 native, 93 non-native), mapped to 15 approximate Census county points. Both official PDF layouts reconcile. Regenerate with python scripts/import-montana-champions.py REGISTER.pdf COUNTIES.zip (requires pdfplumber); reviewed non-personal table fields are in scripts/fixtures/montana-champions-2024.json. Published co-champions, urban/wildland categories, national badges, scores and year histories are retained. The ambiguous “7.” crown cell remains null and the Scarlet Oak / Quercus rubra mismatch is flagged.
python scripts/import-new-mexico-champions.py REGISTER.pdf COUNTIES.zip reproduces the reviewed April 14, 2020 Forestry Division database import. Only 37 of 144 entries explicitly marked STATE CHAMPION are included. The sanitized fixture omits names, addresses and tree coordinates. County points are approximate; submission dates are separate from measurements, and national labels remain dated source designations. Nomination consent notes and the Catron/Socorro conflict are visible on records.
python scripts/import-oregon-champions.py REGISTRY.html COUNTIES.zip reproduces the reviewed Oregon literal Google Charts table without executing source JavaScript. 132 SC/SC-C/NC/NC-C records are included; 35 PV entries excluded. The sanitized fixture omits nominators. 131 records are searchable without markers. One Port-Orford cedar exactly matches the Oregon scientific name and all four measurements/points in the 2012 national register (PDF page 58), supplying Coos County for an explicitly historical county marker. Edition and measurement dates remain unknown; current titles and access are unconfirmed.
python scripts/import-washington-champions.py REGISTRY.html COUNTIES.zip reproduces the reviewed literal Google Charts table without executing source JavaScript. All 279 SC/SC-C/NC/NC-C records are included; 13 PV entries excluded. Sanitized fixtures omit nominators. 278 entries remain searchable without markers. One Pacific silver fir exactly matches the Washington species and all three measurements plus score in the 2020 national register (PDF page 1, record 2915), supplying Clallam County for an explicitly historical marker. The other silver fir co-champion remains separate and unmapped. Edition and measurement dates remain unknown; current titles and visiting access are unconfirmed.
Adds Scarlet Oak, Red Mulberry, Black Locust, Pawpaw, Ginkgo, and American Holly, bringing the Library to 58 species. Each has identification and seasonal notes, three original contemplative themes, a four-step observation practice, two locally hosted licensed photographs, and exact scientific-name links to matching Champion Tree records. Botanical references and photographic attribution link to NC State Extension’s Plant Toolbox; source URLs, credits, dimensions, and originals are recorded in scripts/fixtures/tree-library-next-photo-sources.json. Practices are original interpretations, not botanical or therapeutic claims.
Adds Southern Magnolia, Honeylocust, Loblolly Pine, American Persimmon, and Pitch Pine, bringing the Library to 63 species. Each includes identification, seasonal observations, three original contemplative themes, a four-step outdoor practice, credited local photographs, search aliases, and reciprocal exact scientific-name Champion Map links. Pitch Pine’s practice explores resilience, renewal, and belonging, with Massachusetts pine barrens context and a prompt to return with care. Photographic originals and licenses are recorded in scripts/fixtures/tree-library-five-photo-sources.json; converted WebP versions retain the indicated licenses. Botanical references are NC State Extension Plant Toolbox profiles; the Pitch Pine page also links to MassWildlife’s barrens overview.
Idaho adds 121 state-champion records across 117 published species and varieties, retaining four pairs of state co-champions and the European white birch’s historical combined state/national co-champion label. All STATE rows from PDF pages 2–18 are included; national comparison rows from other states are excluded. The University of Idaho’s current program page links this dated register and warns that some trees may no longer be alive. Current titles, condition and visiting access are unverified.
All records map to 29 approximate Census county points; no tree coordinates, detailed locations, owner/nominator names or addresses are published. The source county “Adam” maps to Adams. The Cambridge / Idaho County golden weeping willow entry retains its published county, with a visible conflict note. Source botanical spellings and varieties remain unchanged.
The source’s quarter-average crown spread in feet is retained as crownPoints and multiplied by four for full average crown spread. DBH/circumference are in inches and height in feet. Published scores remain unchanged. Nomination years are shown as nominated; measured remains null. All 121 measurement rows reconcile independently with Poppler layout text.
Source: https://sites.stat.washington.edu/jaw/CAW/june-2019-id-registry-big-trees.pdf
Program: https://www.uidaho.edu/extension/natural-resources/forestry/big-tree
Retrieved October 11, 2026. Reproduce using python scripts/import-idaho-champions.py REGISTER.pdf COUNTIES.zip and the 2025 Census national county Gazetteer. The importer checks the sanitized fixture before writing; audit hashes preserve source identity. Idaho is available through state and West filters, shared record URLs, and /api/champion-trees/ID.
Nevada adds all 223 records from the Division of Forestry’s six-page 2023 register: 211 champions and 12 co-champions at seven approximate Census county points. Source names, full average crown spread, rounded crown points, and scores remain as published. The program page’s 303-tree description differs from the linked PDF; the importer adds only printed records. NCTP lists the program on hiatus; current titles, conditions, and access are unverified.
128 incomplete botanical labels are preserved and flagged. They remain searchable, but are excluded from species/genus grouping; no full identity is inferred from potentially inconsistent PDF hyperlinks. “Pine” is interpreted as White Pine County with a source note. Ownership/personal details and exact positions are omitted. Run python scripts/import-nevada-champions.py REGISTER.pdf COUNTIES.zip with the published PDF and Census 2025 national counties ZIP; it checks against the reviewed fixture and emits the records, geography, and import audit.
All 60 rows from DANR’s undated online “Current Register of Big Trees” table are included at 23 approximate Census county points. Published nearest communities remain searchable. The overview’s 57-species/234-tree counts and cottonwood example differ from the linked table; table values take precedence. No missing rows, measurement dates, row-level champion/challenger labels or national titles are inferred.
Original names, decimal measurements and scores are preserved. Explicit feet/inches circumference values are converted to inches. Four measurement/score discrepancies are flagged; American sycamore retains its printed 11-inch circumference rather than guessing a missing feet mark. Current condition, titles and visiting access are unverified. Run python scripts/import-south-dakota-champions.py CURRENT.html COUNTIES.zip with the official table and Census 2025 county Gazetteer; it checks all rows against the reviewed fixture and writes records, county points and an audit with source hashes. Retrieved October 11, 2026.
Contact and Big Tree Volunteers use alexjulianyoga@gmail.com, defined once in lib/school-contact.ts. Map records offer an email correction link with the record ID and state in the draft. Visitors review and send messages or photographs through their own email application; the website does not collect submissions. Contributions are reviewed privately before site updates, and publication permission and credit must be stated in the email.
The direct submission form and its API have been removed. Email-provider configuration is not required. A future form can be introduced when the school is ready.
Mississippi includes 163 historical source rows plus one preserved 2024 national-register record, at 51 approximate Census 2025 county points. The historical source is the coordinator-supplied scan of MFC Publication #35, revised October 2000, received October 21, 2026 and audited October 28. A clearly labeled reviewed transcription is served at /data/mississippi-champions-2000-transcription.pdf, with matching source-page numbers and page-specific links on each record. The original scan is not published. Loren Erickson said a new register is being prepared for after the start of 2027; no firm release date was given. The official program page lists MS CHAMPION TREES as Coming Soon. The source-review tracker records this completed review and a January 12, 2027 follow-up target; this does not create an automation.
The introduction reports 164 champions, but pages 3–10 contain 163 table rows (20, 23, 19, 25, 25, 20, 11 and 20 per page). Page 11 lists species without a champion and is excluded. The tables designate 147 unstarred state champions, four state co-champions, ten national champions and two national co-champions, all historical designations. Rusty Blackhaw on page 3 and Rusty Viburnum on page 10 appear to repeat one tree under different names: county, measurements, score, nominators and owner agree. Both source rows are cross-referenced and retained as records, not asserted to be distinct living trees. No missing row is invented to meet the introduction’s count.
Historical IDs use page and row (ms-2000-p03-r01), separate from national ID ms-5821. Every historical card says “Historical champion record — October 2000”; state overview and filtered-result counts distinguish historical rows from the newer national record. Source-selection controls use the existing shareable search. Historical measurement dates remain null: the publication edition is not a measurement date. Original feet/inches strings and printed points are preserved alongside circumference converted to inches and crown spread to decimal feet. Common names combine each printed row name with its section heading. Printed botanical anomalies, the Durand Oak score mismatch, the Spruce Pine/introduction height conflict, the White Oak table/caption discrepancy and the possible duplicate listings have visible record notes. Pearl river is normalized to Pearl River only for the county match. Ownership, nominators and individual tree coordinates are omitted from imported fields; the public PDF contains only reviewed species, measurements, designations and counties, with no original personal columns. No current survival, title, ownership or public access is inferred.
The newer dataset is unchanged: Roughleaf Dogwood, Cornus drummondii, Tree ID 5821 in Tunica County, selected from the State column of the supplied 2024 National Register CSV (547 national records). It was nominated and measured in 2016; blank NCTME verification is not treated as verified. Its 39-inch circumference, 30-foot CSV height, 34-foot crown and 78 points are retained, with the companion PDF’s 28-foot height discrepancy disclosed. The historical Roughleaf Dogwood is in Washington County and is not merged with it.
Render the public transcription with python scripts/render-mississippi-transcription.py (requires ReportLab); it reads only the non-personal PSV fixture and never opens the original scan. Reproduce the historical import with python scripts/import-mississippi-historical.py SCAN.pdf COUNTIES.zip. The versioned, visually reviewed PSV transcription drives extraction; OCR alone is not imported. The importer checks the exact scan hash, per-page row totals, original units, star designations and county matches, and writes a separate historical audit. Run python scripts/import-mississippi-champions.py REGISTER.csv COUNTIES.zip for the newer national record; it preserves county points used by the historical layer. The original national audit is retained independently.
All 87 first-place records in the North Dakota Forest Service official StoryMap are included; its separate 50 second-place trees are excluded. The public story was last published May 27, 2026 and retrieved October 11, 2026. Later item modifications flag unpublished changes; draft resources are not used. The older 2024 table (updated April 30, 2025) has 82 champions and 47 runners-up. It is a reconciliation source, not a list to union with current champions. All 13 champions announced May 1, 2026 are present with matching scores; replaced champions stay excluded.
80 trees use 15 approximate Census county points. Named municipalities are matched to county subdivisions, while coordinate-only records receive historical town/county labels only from exact botanical-name and three-measurement matches. The matching evidence appears in record notes and the audit. Miyabe maple’s Fargo location is independently supported by the official announcement. Seven unsupported general locations remain searchable without markers; no tree coordinates, street addresses, owners, nominators, contacts, or photographs are imported. Access stays unclassified.
The current StoryMap controls names, measurements, scores and dates. Norway spruce crown differs from the older table; river birch circumference differs from the announcement; Miyabe maple retains its June 2023 measurement date despite being announced among 2025 additions. Genus-only and misspelled scientific labels, missing dates and score discrepancies are preserved. Run python scripts/import-north-dakota-champions.py STORY.json OLD.html META.json COUNTIES.zip COUSUBS.zip NEWS.html. The importer checks the reviewed 87-record fixture, the published tour graph, explicit champion labels, excluded second-place records, all 13 announced scores, and source hashes. Personal/location details in raw inputs must not be committed.
All nine Utah Community Forest Council directory pages (250 distinct IDs, retrieved October 11, 2026) are reconciled to their detail pages. The map includes 248 entries with published State Champ Status 1; IDs 313 (code 0) and 95 (code 2) are excluded without inferred titles. Multiple designated trees, explicit co-champion special notes and ten source national flags are preserved. National flags do not independently verify current national standing. Measurement years and entry/update dates remain distinct; published names, scores and measurements are unchanged.
225 records use approximate Census 2025 Utah place points. The remaining 23 stay searchable without markers: forests, county-only locations, ambiguous labels, a misspelling and the ambiguous Enterprise city/CDP are not resolved by guessing. Public access is explicitly marked for 159 included records; 89 are marked no public access. Personal fields, street addresses, images, individual tree coordinates and location directions are omitted. Markers are city/place points, not individual tree locations.
Run python scripts/import-utah-champions.py RAW.json PLACES.txt with the ordered nine-page snapshot and Census Utah place gazetteer. The importer validates a reviewed nonpersonal 250-record fixture, unique tree IDs, page/detail reconciliation and source hashes; changed rows require review. Raw detail pages contain personal and precise location data and must not be committed.
All 57 current Wyoming State Forestry Division register entries are included, retrieved October 11, 2026, with every individual page reconciled against the listing. Older Google Sites pages are not unioned into this current directory. Names, cultivars, genus-only labels, counties and published measurements/scores are retained. No edition, measurement dates or public visiting classifications are displayed; none are inferred.
All records map to 15 approximate Census 2025 county points. These points do not identify individual trees or grant visiting access. Personal fields, addresses, photographs and individual coordinates are omitted. Eastern Cottonwood’s published 497-inch circumference, 70-foot height and 100-foot spread calculate to 592 points, while the source publishes 388 points; the source score remains unchanged with a visible warning. Other score discrepancies are likewise retained and flagged.
Run python scripts/import-wyoming-champions.py REGISTER.html DETAILS_DIR COUNTIES.zip with the current register, all 57 slug-named individual HTML pages and Census county Gazetteer. The importer checks the reviewed 57-record fixture, unique links, all listing/detail values, explicit units, county geography and source hashes.
Adds the eleven HI rows explicitly designated Champion in the official 2024 National Register CSV. All circumference, height, crown spread and score values were compared with the accompanying PDF and agree. Last-measured years (2016–2017) and nomination years remain distinct from the register edition; current condition and visiting access are unverified. This is limited national coverage, not a complete state register. No older DLNR records are merged into this selection.
Eight records map to Census 2025 Hawaii County and three to Honolulu County. Lonomea’s source O‘ahu county label is explicitly normalized to Honolulu; the original wording is retained. The CSV’s K?lea Lau Nui spelling follows the PDF’s Kōlea Lau Nui with a visible note. Other published common names are retained. Exact tree coordinates and personal measurer names are omitted. Run python scripts/import-hawaii-champions.py REGISTER.csv COUNTIES.zip to reproduce the checked import. The fixture contains the eleven independent PDF measurement checks and the audit records source hashes.
Hawaiʻi-only searches automatically show Hawaiʻi. For searches spanning both geographies, the map opens on the contiguous US with a Hawaiʻi switch; the result list retains all matching records. Selecting a Hawaiian record reveals its map view. The Pacific region shortcut filters to Hawaiʻi.
One UAF campus black spruce is included from an account first published in 2003 and republished in 2010. It is explicitly labeled a historical state champion, with height 71 feet and circumference 45 inches. Measurement date, crown spread, points, current condition, current title and visiting permission remain unknown. The marker uses the Census 2025 Fairbanks city representative point (GEOID 0224230), not tree coordinates. This curated selection has an internal record ID and source audit; it does not constitute a complete Alaska register. Alaska has its own map view and region shortcut.
Chapter 5 now links to all 87 entries in the book’s pose section, including Yoga
Nidra. Each statically rendered page uses its own root URL, retains the original
instructions, names its printed pages, links to the exact PDF page, and connects
back to the complete directory and adjacent entries. /reclined-wrist-circles
is restored as a page. All pose pages have canonical/social metadata and sitemap
entries. The directory is alphabetically grouped and needs no client JavaScript.
lib/book-pose-records.json contains the source text; scripts/fixtures/book-poses-source.json
records the PDF hash, page range and all headings. Regenerate with
python scripts/import-book-poses.py BOOK.pdf (PyMuPDF). Book text remains the
original author’s copyrighted material, rather than code relicensed under MIT.
This is the pose reference; flows and meditations retain their chapter reading.
Five source-checked guides cover Bretzfelder Park and The Rocks (NH), Gifford Woods (VT), Hemlock Hill (MA), and Vaughan Woods Memorial State Park in South Berwick (ME). Each has walking/terrain/fee summaries, arrival and seasonal access details, official sources, and an outdoor practice. The homepage links all five. These are online source checks, not field inspections; unpublished distances and fees remain explicitly unknown. Gifford's current hiking table lists a 0.10-mile loop while an older regional map lists 0.3 miles; the guide explains the difference and recommends checking locally.
Two actual Gifford Woods photographs by Victoria Stauffenberg / National Park Service (October 5, 2024) are US public-domain federal works. The guide retains source and license links, full-frame dimensions, and attribution. The 1000-pixel WebP files are 297,598 and 85,132 bytes and use the existing immutable hashed-image cache. Other new destinations link their managers' photographs; species reference photographs are labeled as photographed elsewhere. Source and photo metadata live in lib/new-england-visit-records.json.
The source review process covers all 50 state registers with an editable review tracker, older/limited coverage priorities, actual completed-review outcomes and explicit next targets. Run npm run review:sources for an offline batch of up to five due states. The public tracker at /champion-trees/source-review distinguishes snapshot evidence from source editions, tree measurement dates and planned maintenance. Existing email corrections remain the contribution channel. No scheduled scraping or GitHub Actions workflow is added.
Use Node.js 24 (nvm use reads .nvmrc), then npm ci, npm test, and npm run build.
Dependabot checks npm dependencies weekly; Next.js major upgrades require a
separate migration decision. Next.js 15 pins an older PostCSS, so the scoped
PostCSS 8.5.23 override addresses its published source-map and stringify
advisories. Review the override when Next.js updates that dependency and keep
manifest/lockfile changes together. Validate dependency changes with the
regression tests and production build before merging.
The Quality / Tests and production build GitHub Actions job runs these commands on pull requests and pushes to main; it can also be started manually. The build already includes coverage freshness and champion-import validation. Actions are pinned to commit SHAs, use read-only repository permissions, and do not need deployment secrets. Vercel's existing Git integration continues to handle deployments.
Champion card labels, common measurement/date semantics, location headings, location guidance and result summaries live in lib/champion-presentation.ts. Register/import metadata remains in lib/champion-states.ts. State-specific source citations and specialized detail markup remain in the explorer. Presentation regression fixtures were captured from the previous renderer at c5623fa and protect all 8,971 records across 50 states; review wording/source changes before updating those fixtures. Common cream callout styling is grouped in app/globals.css.
The workflow reports checks; requiring them before merge is a separate branch-protection setting. It does not by itself block direct pushes or Vercel production deployments.