Wildfire Size Disparity on Tribal Lands in California and Associated PM2.5 Burden, 2000–2018

Authors: Blind for review. Version 2.0 — June 2026. See revision notes below. Study Overview This dataset supports a quantitative investigation of wildfire size disparity between federally recognized tribal lands and non-tribal lands in California from 2000 to 2018, and the associated PM2.5 air quality burden borne by tribal communities. Using spatial analysis, panel econometrics, and four convergent causal identification strategies, the study documents that fires on tribal trust land are on average 6.4 times larger than fires on non-tribal land — a disparity that persists at 3.1× after excluding boundary-straddling fires and at 4.8× after excluding the catastrophic 2003 and 2007 fire years. All specifications are statistically significant (Mann-Whitney p < 0.0001). Tribal fire years are associated with a statistically significant increase in county mean PM2.5 of 0.845 μg/m³ (two-way fixed effects OLS, county + year FE, clustered SE, p = 0.0087). The Bureau of Indian Affairs — the only federal agency with statutory trust responsibility toward tribal nations — managed 7.0% of fires on tribal trust land during the study period. CAL FIRE and state agencies combined managed 54.6%, operating without trust obligations or mandates to integrate Traditional Ecological Knowledge. Data Contents This archive contains all outputs necessary to replicate the analysis. The complete replication pipeline (FIRE.ipynb) reproduces all results in a single execution and downloads source data directly from public APIs. Files included: FIRE.ipynb — Full integrated analysis pipeline (Sections 1–22) ca_fires_tribal_classified.csv — 4,496 California fire perimeters with tribal land classification tribal_fire_pm25_panel.csv — 816 county-year observations (PM2.5 + fire variables) tribal_nation_fire_burden.csv — Fire burden summary by tribal nation (top 20) event_study_coefficients.csv — Event study coefficients (t = −5 to +5) with 95% CI regression_results.txt — Full two-way FE OLS output including all fixed effects figure1_tribal_fire_analysis.png — Mean fire size by year; size distribution; top nations figure2_cedar_fire_aqi.png — San Diego County AQI 2000–2018; Cedar Fire 2003 figure3_regression_discontinuity.png — RD scatter plots at 10, 25, 50, 100km bandwidths figure4_psm_overlap.png — Propensity score overlap before and after matching figure5_psm_differences.png — PSM matched pair differences distribution figure6_event_study.png — Event study coefficients with pre-trend test figure7_causal_identification.png — Convergent causal identification evidence README.md — Full documentation including revision notes Methods Summary Fire perimeter data were obtained from the NIFC Historical GeoMAC Perimeters archive (2000–2018) via ArcGIS REST API. Tribal boundaries are from the Census TIGER/Line AIANNH 2025 vintage (185 California areas). PM2.5 concentrations are from EPA AQS daily FRM/FEM monitoring data (parameter 88101; 580,066 California daily observations). All spatial operations use the California Albers Equal Area projection (EPSG:3310) for area and centroid calculations. Causal identification employs four strategies: (1) regression discontinuity at tribal land boundaries (four bandwidth specifications, 10–100 km); (2) propensity score matching with caliper 0.05 (93 matched fire pairs); (3) event study with county and year fixed effects and pre-trend joint F-test (F = 1.720, p = 0.206 — no pre-trend); (4) two-way fixed effects DiD (county + year FE, clustered SE). Key Results Fire Size Disparity Specification Tribal mean (acres) Non-tribal mean (acres) Ratio p-value Full dataset 17,265 2,706 6.4× < 0.0001 Excl. 2003 14,845 2,782 5.3× < 0.0001 Excl. 2003 + 2007 13,545 2,808 4.8× < 0.0001 Excl. boundary-straddlers 8,253 2,706 3.1× < 0.0001 PM2.5 Health Burden Estimator Coefficient p-value Two-way FE OLS (primary) 0.845 μg/m³ 0.0087 Event study post-period (F-test) F = 3.350 0.0320 Cedar Fire 2003 San Diego max AQI 328 (Hazardous) +124% spike Managing Agency Distribution (Tribal Trust Lands) Agency Fires (N) Share CAL FIRE (CDF) 39 27.3% State Agency 39 27.3% U.S. Forest Service 33 23.1% Bureau of Indian Affairs ✦ 10 7.0% Bureau of Land Management 6 4.2% Other / Unknown 16 11.2% ✦ Only agency with statutory trust responsibility toward tribal nations. Revision Notes (v2.0 — June 2026) Version 2.0 corrects several methodological issues identified after the initial deposit: 1. Spatial join deduplication (critical fix) The v1.0 pipeline used sjoin with predicate="intersects" without deduplication, producing inflated tribal fire counts. Two issues were identified and corrected: (a) 12 NIFC duplicate polygon records (same fire geometry entered twice in the source data) were identified and removed; (b) 8 fires whose perimeters straddled multiple tribal boundaries were retained as single records, assigned to the tribal nation with the largest intersection area. After deduplication: 111 tribal fires (from 143 pre-dedup). Primary ratio: 6.4× (from 15.7× in v1.0). 2. Two-way fixed effects (Section 15) The primary DiD estimator now includes county + year fixed effects. The v1.0 year-only specification (p = 0.37) is replaced by the two-way FE specification (p = 0.0087). 3. 2SLS removed Both available lightning instruments have weak first stages (F < 10). v1.0 reported inflated 2SLS estimates; these are removed entirely in v2.0. IV-related output files (tribal_fire_pm25_panel_iv.csv, lightning_county_year.csv) are removed from this deposit. 4. Boundary-straddler sensitivity Section 6 now correctly reports the fire size ratio excluding boundary-straddling fires (3.1×) using content-key matching robust to dataframe index resets. Computational Environment Python 3.12.13 · geopandas 1.1.3 · pandas 2.2.2 · statsmodels 0.14.6 · scipy 1.16.3 · scikit-learn 1.6.1 · netCDF4 1.7.4 · Executed on Google Colab CPU runtime · Approximate runtime: 20–30 minutes License Creative Commons Attribution 4.0 International (CC BY 4.0). Users are free to share and adapt the material for any purpose provided appropriate credit is given to the authors. Citation Wildfire Size Disparity on Tribal Lands in California and Associated PM2.5 Burden, 2000–2018 [Data set]. Zenodo. https://doi.org/10.5281/zenodo.20665230作者：匿名评审。 版本2.0——2026年6月。详见下文修订说明。 # 研究概述 本数据集用于支持一项定量研究，探究2000年至2018年间加利福尼亚州联邦认可部落土地与非部落土地的野火规模差异，以及部落社区所承担的相关PM2.5空气质量负担。 本研究采用空间分析、面板计量经济学方法与四种一致性因果识别策略，研究发现：部落信托土地上的野火平均规模是非部落土地的6.4倍；排除跨边界野火后，该差异仍保持3.1倍；排除2003年与2007年灾难性火灾年份后，差异为4.8倍。所有设定均具有统计显著性（曼-惠特尼检验p < 0.0001）。部落火灾年份与县级平均PM2.5浓度显著上升0.845 μg/m³相关（双向固定效应OLS模型，县级+年度固定效应，聚类标准误，p=0.0087）。 印第安事务局（Bureau of Indian Affairs）是唯一对部落国家负有法定信托责任的联邦机构，在研究期间负责管理7.0%的部落信托土地野火。加州消防局（California Department of Forestry and Fire Protection, CAL FIRE）与其他州级机构合计管理54.6%的野火，其运作无需承担信托义务，也无强制要求整合传统生态知识。 # 数据内容 本归档文件包含复现分析所需的全部输出内容。完整复现流水线（FIRE.ipynb）可单次执行复现所有结果，并直接从公开API下载源数据。 包含文件： FIRE.ipynb —— 完整集成分析流水线（1-22节） ca_fires_tribal_classified.csv —— 4496条加利福尼亚州野火边界数据，附带部落土地分类标签 tribal_fire_pm25_panel.csv —— 816条县级-年度观测数据（含PM2.5与野火相关变量） tribal_nation_fire_burden.csv —— 按部落统计的野火负担汇总（前20大部落） event_study_coefficients.csv —— 事件研究系数（t=-5至+5）及95%置信区间 regression_results.txt —— 完整双向固定效应OLS输出结果，包含所有固定效应 figure1_tribal_fire_analysis.png —— 年度平均野火规模、规模分布、前20大部落野火情况图 figure2_cedar_fire_aqi.png —— 2000-2018年圣地亚哥县空气质量指数、2003年雪松火灾情况图 figure3_regression_discontinuity.png —— 10、25、50、100km带宽下的断点回归散点图 figure4_psm_overlap.png —— 匹配前后的倾向得分重叠分布图 figure5_psm_differences.png —— 倾向得分匹配配对差异分布直方图 figure6_event_study.png —— 包含事前趋势检验的事件研究系数图 figure7_causal_identification.png —— 一致性因果识别证据图 README.md —— 完整文档，包含修订说明 # 方法摘要 野火边界数据来自国家跨部门消防中心（National Interagency Fire Center, NIFC）历史GeoMAC边界档案（2000-2018），通过ArcGIS REST API获取。部落边界数据来自美国人口普查局TIGER/Line AIANNH 2025版数据（涵盖185个加利福尼亚区域）。PM2.5浓度数据来自美国环境保护署（Environmental Protection Agency, EPA）空气质量系统（Air Quality System, AQS）的每日FRM/FEM监测数据（参数代码88101；共580066条加利福尼亚州每日观测记录）。所有空间运算均采用加州阿尔伯斯等面积投影（EPSG:3310）进行面积与质心计算。 因果识别采用四种策略：（1）部落土地边界处的断点回归设计（四种带宽设定，10-100km）；（2）卡尺为0.05的倾向得分匹配（共93对匹配野火样本）；（3）纳入县级与年度固定效应并进行事前趋势联合F检验的事件研究法（F=1.720，p=0.206——无显著事前趋势）；（4）双向固定效应双重差分模型（Difference-in-Differences, DiD，县级+年度固定效应，聚类标准误）。 # 核心结果 ## 野火规模差异 | 设定组别 | 部落土地平均规模（英亩） | 非部落土地平均规模（英亩） | 倍率 | p值 | |-------------------------|--------------------------|--------------------------|-------|--------------| | 全数据集 | 17,265 | 2,706 | 6.4× | < 0.0001 | | 排除2003年火灾 | 14,845 | 2,782 | 5.3× | < 0.0001 | | 排除2003+2007年火灾 | 13,545 | 2,808 | 4.8× | < 0.0001 | | 排除跨边界野火 | 8,253 | 2,706 | 3.1× | < 0.0001 | ## PM2.5健康负担 | 估计方法 | 系数 | p值 | |-------------------------|-----------------------|--------------| | 双向固定效应OLS（核心） | 0.845 μg/m³ | 0.0087 | | 事件研究法事后阶段（F检验） | F=3.350 | 0.0320 | | 2003年圣地亚哥雪松火灾峰值AQI | 328（危险级） | +124% 涨幅 | ## 管理机构分布（部落信托土地） | 管理机构 | 野火数量（N） | 占比 | |-------------------------|--------------|----------| | 加州消防局（CDF） | 39 | 27.3% | | 州级机构 | 39 | 27.3% | | 美国林务局 | 33 | 23.1% | | 印第安事务局 ✦ | 10 | 7.0% | | 土地管理局 | 6 | 4.2% | | 其他/未知 | 16 | 11.2% | ✦ 唯一对部落国家负有法定信托责任的机构。 # 修订说明（v2.0——2026年6月） 版本2.0修正了初始提交后发现的若干方法学问题： 1. 空间连接去重（关键修复）：v1.0版本的分析流水线使用`sjoin`函数且谓词设为"intersects"时未进行去重操作，导致部落野火计数被高估。本次修正了两处问题：（a）识别并移除了12条NIFC重复多边形记录（源数据中同一野火几何被重复录入两次）；（b）将8个边界跨越多部落区域的野火保留为单条记录，并将其归属到交集面积最大的部落。去重后，部落野火样本量为111条（去重前为143条），核心倍率修正为6.4倍（v1.0版本为15.7倍）。 2. 双向固定效应模型（第15节）：核心双重差分估计器现已纳入县级+年度固定效应。v1.0版本仅采用年度固定效应的设定（p=0.37）已替换为双向固定效应设定（p=0.0087）。 3. 移除两阶段最小二乘模型（Two-Stage Least Squares, 2SLS）：现有两个闪电工具变量均存在弱一阶阶段问题（F<10）。v1.0版本报告了被高估的两阶段最小二乘估计结果，本版本已完全移除该部分内容。与工具变量相关的输出文件（tribal_fire_pm25_panel_iv.csv、lightning_county_year.csv）已从本归档中移除。 4. 跨边界野火敏感性分析：第6节现已正确报告排除跨边界野火后的野火规模倍率（3.1×），采用与数据框索引重置兼容的内容键匹配方法。 # 计算环境 Python 3.12.13 · geopandas 1.1.3 · pandas 2.2.2 · statsmodels 0.14.6 · scipy 1.16.3 · scikit-learn 1.6.1 · netCDF4 1.7.4；运行于Google Colab CPU运行时，预估运行时长：20-30分钟。 # 使用许可 采用知识共享署名4.0国际许可协议（Creative Commons Attribution 4.0 International, CC BY 4.0）。使用者可自由共享、改编本数据集用于任何用途，但需为原作者标注合适的引用信息。 # 引用方式 《2000-2018年加利福尼亚州部落土地野火规模差异及相关PM2.5负担》[数据集]。Zenodo。https://doi.org/10.5281/zenodo.20665230