Reproducibility snapshot v4.23.0 for "The anti-maladaptation filter hypothesis: cohort-scale empirical test that NAM-classification-filtered portfolio allocation delivers higher expected avoided-loss NPV than five naive baseline allocations for climate adaptation of 725,462 electricity substations across 39 OECD countries"

再現性スナップショットv4.23.0：「反不適応フィルター仮説」—NAM分類フィルターによるポートフォリオ配分が、39か国725,462変電所の気候適応において5つのナイーブベースラインより高い期待回避損失NPVを実現するコホート規模実証 (AI 翻訳)

Bérard, Cedric

Zenodoプレプリント2026-07-27#AI×ESGOrigin: Global経営インパクト: コスト削減対象セクター: power

DOI: 10.5281/zenodo.21620260

原典: https://zenodo.org/records/21620260

🤖 gxceed AI 要約

日本語

本スナップショットは、39か国72万以上の変電所を対象に、NAM分類フィルターを用いたポートフォリオ配分が5つのベースラインを上回る期待回避損失NPVを達成することを実証。参照セルで全ベースラインを支配し、多重比較補正後も有意。LPと整数解のギャップは小さく、再現性スクリプトを提供。

English

This snapshot demonstrates that NAM-classification-filtered portfolio allocation outperforms five naive baselines in expected avoided-loss NPV for climate adaptation of 725,462 substations across 39 OECD countries. At the reference cell, all baselines are strictly dominated, with robust results after multiple-correction. LP-integer gaps are small, and reproducibility scripts are provided.

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

日本の電力インフラ老朽化と気候変動適応策の優先順位付けに示唆。SSBJ開示やTCFD対応で気候リスク評価が求められる中、データ駆動型の適応投資判断は投資家対応にも有用。

In the global GX context

Contributes to global climate adaptation finance by providing a scalable, data-driven method for prioritizing infrastructure investments under climate risk. Aligns with TCFD/ISSB emphasis on climate resilience and transition planning.

👥 読者別の含意

🔬研究者:Provides a rigorous empirical framework for testing adaptation portfolio allocation methods at scale.

🏢実務担当者:Offers a decision-support tool for prioritizing substation upgrades under budget constraints.

🏛政策担当者:Informs infrastructure investment planning and climate adaptation policy with evidence-based prioritization.

📄 Abstract（原文）

  Frozen v4.23.0 reproducibility snapshot accompanying the P4 manuscript submitted   to Energy Economics (Elsevier). Contains: main manuscript v3 (12,947 words) with   §4 empirical results from the Phase P4-5 39-country × 725,462-substation cohort   solve (2026-07-26T16:28:03Z); Supplementary Materials §S1 per-country cost matrix   + §S3 LP-vs-integer 0/1 + CVaR extended + shared-model τ² caveat + Cochran's Q   refresh note + uniform-CV limitation + §S8 pooled DerSimonian-Laird per-country   weight tables; five main-text SVG figures (Wong 2011 accessibility palette);   full cohort JSON outputs from the 5,265 LP solves; one-command reproducibility   scripts (cohort_solve.py + cohort_extended.py) with orchestrator runbook; cover   letter + author declaration for the Energy Economics submission; audit response   memo documenting cumulative 25-finding Beta remediation closure + Part 3 R19.1   CLOSURE ADDENDUM.   Empirical headline: 5-of-5 baselines strictly dominated at reference cell   (SSP2-4.5 × 2050 × 15% budget) with cohort-mean reference-cell dominance   margins +145.08% (population-weighted), +152.65% (GDP-weighted), +97.50%   (substation-count-weighted), +55.14% (capacity-MW-weighted), and +2.04%   (hazard-weighted). Multi-correction on 135-cell test family: Bonferroni   117/135 (86.7%); Holm 126/135 (93.3%); BH-FDR 126/135 (93.3%). DerSimonian-   Laird random-effects pooled effects strictly positive at 95% CI on all 5   baselines. CVaR-α at α ∈ {0.95, 0.99} preserves dominance direction. LP-vs-   integer 0/1 gap ≤ 0.5% of theoretical ceiling in 38 of 39 countries.   Companion papers in the SSI Index publication slate:   - P1 (Bérard 2026, Climate Risk Management, under review) — NAM operationalisation     as first-stage screening classifier at cohort scale   - P2 (Bérard 2026a, Environmental Research: Energy, DOI 10.1088/2753-3751/ae87a5,     published August 2026) — atmospheric-corrosion damage-function library   - P3 (Bérard 2026b, Journal of Infrastructure Preservation and Resilience,     DOI 10.1186/s43065-026-00193-z, published July 2026) — Markov degradation prior

🔗 Provenance — このレコードを発見したソース

Zenodo https://zenodo.org/records/21620260first seen 2026-08-07 04:16:55

🔔 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。