← 論文一覧に戻る

土壌管理における人工知能：気候レジリエンスへの示唆—包括的レビュー

Artificial Intelligence in Soil Management: Implications for Climate Resilience—A Comprehensive Review (原題)

M.I.A. Barbosa, Pedro Miguel Rodrigues

Sustainability📚 査読済 / ジャーナル2026-10-01#気候科学Origin: EU対象セクター: agriculture
DOI: 10.3390/su181910057
原典: https://doi.org/10.3390/su181910057

🤖 gxceed AI 要約

日本語

本レビューは106件の査読論文を分析し、AI・機械学習・深層学習が土壌水分・養分・炭素動態・劣化指標の予測にどう用いられているかを整理する。精密灌漑や施肥管理、干ばつ・塩害の早期警報など適応的土壌管理への応用を検討し、データ不均質性やモデル説明性、気候・土壌型を超えた転用性といった課題を論じる。説明可能AIやマルチモーダル融合、デジタル土壌ツインを今後の方向性として提示する。

English

This review analyzes 106 peer-reviewed studies on AI, machine learning and deep learning for soil monitoring, modelling and decision support. It covers prediction of soil moisture, nutrients, carbon dynamics and degradation indicators, plus applications in precision irrigation, fertilization and early-warning systems for drought and salinity. Key challenges—data heterogeneity, explainability, sensor reliability and transferability—and emerging directions such as explainable AI, multimodal fusion and digital soil twins are discussed.

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

土壌炭素・劣化指標の計測高度化はScope 3や自然資本開示の基礎データになり得るが、本稿は開示制度に直接言及せず、日本のSSBJ対応への示唆は間接的である。農業・食品企業のサプライチェーン気候リスク評価に関心のある読者には参考になる。

In the global GX context

Soil carbon and degradation monitoring underpins nature-related and Scope 3 land-sector disclosure, though the paper does not engage TCFD/ISSB/TNFD frameworks directly. It is most useful as methodological background for quantifying agricultural climate resilience and land-based carbon, rather than as disclosure scholarship.

👥 読者別の含意

🔬研究者:AI・リモートセンシングによる土壌特性予測の手法動向と課題を俯瞰できる。

🏢実務担当者:農業サプライチェーンの気候リスク評価や土壌炭素モニタリング手法の選定に活用できる。

🏛政策担当者:気候レジリエントな農業政策におけるAI・センシング投資の根拠として参照可能。

📄 Abstract（原文）

Climate change is accelerating soil degradation, threatening agricultural sustainability and soil climate resilience. Key climate stressors include drought, altered precipitation, extreme rainfall, erosion and salinization. Artificial Intelligence (AI), combined with advanced sensing technologies, offers new opportunities for data-driven soil management under dynamic environmental conditions. This review aims to critically examine recent advances in AI-based approaches for soil monitoring, modelling and decision support, with an emphasis on soil management practices that can contribute to enhancing soil climate resilience. We analyze machine learning and deep learning methods applied to multisource data, from field observations, laboratory analyses, proximal sensing, remote sensing, meteorological records, geospatial information and agricultural management data, to predict key soil properties, including moisture, nutrient status, carbon dynamics and indicators of degradation. A total of 106 peer-reviewed studies were analysed. This review further explores AI-driven strategies for adaptive soil management, including precision irrigation, fertilization management and early-warning systems for drought, salinity and soil degradation. Key challenges, such as data heterogeneity, model explainability, sensor reliability and transferability across soil types and climates, are also discussed. Finally, emerging directions, including explainable AI, multimodal data fusion and digital soil twins, are highlighted, positioning AI-enabled sensing as a basis for climate-resilient soil management.

🔗 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。