← 論文一覧に戻る

<b>High-Resolution Urban Carbon Footprint Maps of the Metropolitan Area of Mexico City (MCMA)</b>

メキシコシティ都市圏（MCMA）の高解像度都市カーボンフットプリントマップ (AI 翻訳)

Rogelio O. Corona‐Núñez

Figshareデータセット2026-08-14#AI×ESG対象セクター: urban_planning
DOI: 10.6084/m9.figshare.33260739
原典: https://doi.org/10.6084/m9.figshare.33260739

🤖 gxceed AI 要約

日本語

メキシコシティ都市圏の世帯カーボンフットプリントを100m解像度で推定したオープンデータセット。992世帯の調査データとランダムフォレストを用いて、食・住・交通の排出を空間モデル化し、平均・標準偏差・変動係数を提供する。都市構造と社会経済的不平等が排出パターンに与える影響を明らかにする。

English

This dataset provides high-resolution (100m) spatial models of household carbon footprints for the Mexico City metropolitan area, using Random Forest on survey data from 992 dwellings. It maps diet, housing, and transport emissions with statistical uncertainty, revealing how urban structure and socioeconomic inequalities shape carbon footprints.

Unofficial AI-generated summary based on the public title and abstract. Not an official translation.

📝 gxceed 編集解説 — Why this matters

日本のGX文脈において

日本の都市計画やカーボンニュートラル施策において、高解像度の排出マップは地域脱炭素の基礎データとして有用。SSBJ開示や自治体の排出インベントリ精緻化に応用可能。

In the global GX context

This work contributes to global urban climate scholarship by demonstrating a replicable ML-based approach to fine-scale carbon mapping, relevant for cities pursuing climate action and for ISSB-aligned disclosure of urban emissions.

👥 読者別の含意

🔬研究者:Methodological template for combining survey data with ML to produce high-resolution carbon footprints with uncertainty metrics.

🏢実務担当者:Spatial emission data can inform urban planning and targeted decarbonization interventions.

🏛政策担当者:High-resolution maps support evidence-based climate policy and resource allocation in cities.

📄 Abstract（原文）

This dataset provides high-resolution spatial models of household carbon footprints across the Metropolitan Area of Mexico City (MCMA), structured for open-access sharing. The repository contains raster surfaces mapping annual per capita greenhouse gas emissions at a <b>100-m spatial resolution</b>. To ensure robust statistical transparency for environmental analysis and spatial planning, each mapping product includes the <b>mean, standard deviation (SD), and coefficient of variation (CV)</b> derived from the underlying predictive modeling framework.All emission values are expressed in <b>metric tons of CO2-equivalent per year per person (tCO2eq / year / person)</b>.<b>Associated Publication Reference</b>The data and modeling frameworks contained within this repository are associated with the following study, currently under consideration in the journal <i>Sustainable Cities and Society</i>:<b>Manuscript Title:</b> Urban Structural and Infrastructural Inequalities Shape Fine‑Scale Carbon Footprints of Diet, Housing, and Transportation: Evidence from the Metropolitan Area of Mexico City.<b>Status:</b> Under review / consideration in <i>Sustainable Cities and Society</i>.<b>Methodological Summary &amp; Data Generation</b>The spatial carbon models were constructed by integrating primary household survey data with high-resolution environmental, socioeconomic, and infrastructural predictors using machine learning spatial modeling:<b>Sampling &amp; Survey Design:</b>Primary data collection was conducted via a face-to-face household survey across 992 randomly approached dwellings (houses and apartments) within the MCMA.The survey captured the full metropolitan density gradient, ranging from low peripheral densities (1.6 inhabitants/ha) to dense urban cores (362 inhabitants/ha), mapped across a 100-m population density grid spanning up to 461 inhabitants/ha.Exact GPS coordinates were logged for each surveyed dwelling to enable spatial modeling. Questionnaires were adapted from the Greenhouse Gas Protocol for Project Accounting.<b>Emission Domains:</b><b>Dietary Emissions:</b> Estimated using a food-frequency approach capturing weekly consumption across major food groups and retail outlet types. Local and international life-cycle assessment emission factors were applied. Production-phase impacts dominate (&gt;94% of variance), while transport distances were excluded due to product-origin data limitations and to prevent spatial bias.<b>Housing Component:</b> Captured direct and indirect household consumption of electricity (via billing data and water distribution energy requirements) and liquefied petroleum gas (LPG) for cooking and water heating, evaluated using official Mexican government emission factors for 2022.<b>Transportation Emissions:</b> Calculated from routine mobility tracking (work, education, and daily travel) across public transport, private vehicles (incorporating vehicle types and fuel specifications), and point-to-point commercial aviation distances.<b>Predictive Modeling &amp; Spatial Mapping:</b>To scale survey responses across the MCMA, predictive spatial models were trained using the <b>Random Forest</b> algorithm on 683 spatially aggregated observations (70% training, 30% independent validation).Random Forest was selected for its capacity to handle high-dimensional urban data, accommodate multicollinearity, and model complex non-linear relationships.Model explainability and feature evaluation were performed using the DALEX library, applying permutation-based variable importance and Ceteris paribus partial dependence profiles to map how urban structure, accessibility, and socioeconomic drivers influence household carbon footprints across the metropolitan gradient.

🔗 Provenance — このレコードを発見したソース

🔔 こうした論文の新着を逃したくない方は キーワードアラート に登録（無料・3キーワードまで）。

gxceed は公開メタデータに基づく研究支援データセットです。要約・翻訳・解説は AI 支援で生成されています。 最終的な解釈・検証は利用者が原典資料に基づいて行うことを前提とします。