Anthropocene snapshot assets

This directory contains the authoritative profiles and normalized snapshots used by ordinary offline Anthropocene generation:

The old unqualified anthropocene snapshot is intentionally gone. Both accepted standard families now use year-bearing pass IDs and filenames so a complete year cannot be confused with a partial snapshot or overwrite it. Each profile’s literal duration.year is the sole year authority; no generator reads the host clock.

The particulate snapshots contain one point at the center of each resolution-4 H3 cell with one or more positive unique-source-reporting-day counts. Complete 2025 contains 49,470 cells; partial 2026 contains 43,895. These are observations with uneven source coverage, not complete global fields. An omitted property is unavailable or unobserved and never an asserted zero. The normalized SHA-256 values are:

Pass Status GeoJSON SHA-256
anthropocene-particulate-2025 Complete through 2025-12-31 ecd4b11bab4faa9895522dbfe75436ef66820f53620bf0c5f6fc7404e9f5426b
anthropocene-particulate-2026 Partial; source-specific dates in profile a3b2fcb3a809710d278ce88aef52ea938fa67e8b900313cf2e5c2ed9c6ddde42

The Stage 8b enrichment plan records the CPC temperature/FIRMS-gate increment implemented here and the remaining CAMS, PurpleAir, and ocean work.

Particulate observation atlas

The two particulate products retain nine source-separated metrics: temperature record highs and lows, precipitation records, heavy precipitation, active-fire hotspots, analyst-observed smoke, located flood/heavy-rain events, located severe-weather events, and regulatory PM2.5 exceedance days. They compute no composite severity or causal-attribution score. Their source-specific marker opacity is logarithmic, and the primary title uses the common doubled-title rule.

Year Temperature high/low Precipitation record/heavy Active fire Smoke Flood Severe weather PM2.5
2025 4,129 / 714 1,757 / 1,491 127,392 1,794,158 7,135 21,694 1,240
2026 partial 1,567 / 434 424 / 426 35,940 1,088,261 1,000 4,195 129

These are H3 cell-day totals, not event, person, area, emission, or damage totals. Both checked snapshots record zero NASA FIRMS rows. Their fire layer therefore reflects the checked regional CWFIS input and is not a global active-fire census.

Stage 8b temperature fields

The temperature profiles use NOAA CPC Global Unified Temperature V1.0 daily TMAX/TMIN on its 0.5-degree global grid. The preparer maps valid source-grid centers to resolution-3 H3, averages multiple grid centers in one cell for each 06Z-to-06Z reporting day, and counts strict daily records against all available same-calendar-day values since 1979 once at least 30 history years exist. Every one of the 41,162 global H3 cells is serialized. Positive valid_days means an analyzed cell even when both record counts are zero; zero valid days means missing or outside CPC’s land domain.

The 2025 product is complete through December 31 and compares with 1979–2024. The 2026 product is partial through August 4 (216 source days) and compares with 1979–2025. The checked normalization audit is:

Year Status Covered H3 cells Record-high days Record-low days TMAX valid cell-days TMIN valid cell-days
2025 complete 11,945 178,896 59,806 4,359,841 4,359,925
2026 partial through August 4 11,945 105,041 40,943 2,579,998 2,580,120

The shared covered domain includes 2,069 North American, 1,553 South American, 936 European, 2,836 African, 1,580 Siberian, 1,132 China/Japan, and 951 Oceania bounding-box audit cells. These boxes overlap and are coverage gates, not additive regional statistics. The display saturates record-high counts at 32 days and record-low counts at 16 days, approximately the complete-2025 95th-percentile values; the raw counts remain embedded and unchanged. Antimeridian H3 rings are split before projection. Cahill-Keyes uses an SVG-audited centered-hex fallback for 21 covered cells at its outer topology cuts; the other five projections use no fallback cells.

The 96-file CPC source manifest has SHA-256 9e8507f1ad63332b3af55b5a5ab76e26209baabd97f6f644df55b5cd2faa0bee. The normalized 2025 and 2026 GeoJSON SHA-256 values are respectively 7dbb1858e357ff87b8fd4b7ee9874d0a9978a55ccfe7943e16ec7563e9a520c6 and 5953eba7ed3e73ecc0d77961a17759336f1833118d7850a5891a316369c9bde2.

Generate their SVG families or all exported artifacts with:

make generate-anthropocene
make generate-anthropocene-artifacts
make generate-anthropocene-2025
make generate-anthropocene-2026
make generate-anthropocene-year-artifacts

The unqualified targets build both 2025 and 2026. The source-separated particulate pair is available explicitly through generate-anthropocene-particulate and generate-anthropocene-particulate-artifacts; the former atlas aliases point to that same pair and do not recreate an unqualified artifact.

The ignored raw archive is staged and candidates are prepared explicitly:

make fetch-anthropocene-cpc-data
make prepare-anthropocene-temperature-data

The raw .raw/cpc/SHA256SUMS file is itself pinned by both profiles. Candidate preparation verifies every yearly TMAX/TMIN file and never overwrites the checked GeoJSON.

Checked particulate acquisition

The partial-2026 snapshot was normalized on 2026-08-05 from:

Raw input Captured SHA-256 or set digest
NOAA GHCN GSN archive 3cb70cbd4af615154bbd3513c3b258f9913082b91a32cf4c58d9b7328b053954
NOAA GHCN station inventory 91e4bd9b4c991c3ea9c19062eae73bca118cad0f4a436e33149ff345bffc8f9c
EPA AirData daily PM2.5 archive c6900edda13ee5d451abb393f961ae427e76506bb87931967db54150143c91d5
NOAA HMS annual smoke archive 42f89a34af1a3a6f4fcaf08c0e87b51b0741a220ecec271e9f7856ba3a9350c5
NOAA Storm Events details 2afe1359fe06a94bc0fcd099f186f0e7f8aba8243cc37665ab8d2c87e4b42f19
NOAA Storm Events locations 36960089c2ce0a491ac5010eeea51fd7a9747e876cc128eaf14eb92087b785f3
167 CWFIS daily CSVs, filename-sorted SHA-256 manifest 0d9e1656a3778e5c166db8cf520dfdf50a9445305c6d38d40e67a0770ae1432a

The raw archives are intentionally not checked in. They are mutable annual feeds and total hundreds of megabytes. make fetch-anthropocene-particulate-YEAR stages one year under ignored .raw/YEAR/; make prepare-anthropocene-particulate-YEAR produces its candidate under ignored .prepared/ without replacing a checked snapshot. The *-data aggregate targets run both 2025 and 2026. Promotion is deliberately manual because every refresh changes source coverage and must update profile dates, statistics, and the normalized digest together.

NASA FIRMS requires the free FIRMS_MAP_KEY. Request it on the official NASA FIRMS key page and keep it only in the acquisition process environment. A new global refresh fails when the key or staged FIRMS rows are absent unless the explicit ANTHROPOCENE_REGIONAL_DEVELOPMENT_ONLY=1 debugging override is set. The fetcher reads FIRMS’ advertised date ranges, joins standard S-NPP processing to its NRT tail in five-day chunks, and avoids logging the key. Preparation audits at least 95% of expected dates and nonzero rows in North America, South America, Europe, Africa, northern Asia, East Asia, and Oceania. Both checked particulate snapshots have no FIRMS rows and therefore make no global fire claim; CWFIS is regional QA/fallback only. The complete secure key procedure is recorded in the Stage 15 plan.

See the Anthropocene implementation notes for current source roles, formulas, coverage boundaries, refresh commands, and interpretation limits, and the Stage 8b enrichment plan for the full global-coverage design and its remaining phases.