SciGroveBeta
Environment

Scope3Trace: Evidence-Based Identification and Extraction of Scope 3 GHG Emissions from Sustainability Reports

Siyuan Zheng, Yifan Duan, Chao Xue, Flora D. Salim

Featured July 22, 2026

Original Paper

This analysis was generated by SciGrove. Upload your own PDFs or enter a DOI — and get the same AI breakdown on any paper.

Get started

AI-generated analysis — This is SciGrove's AI interpretation of the paper, not peer-reviewed content. Always refer to the original paper.

Simply

A new system called Scope3Trace uses smart computer programs to read messy company reports, find out how much pollution they make, and show exactly where that information came from.

In depth
The paper introduces Scope3Trace, an evidence-grounded framework that extracts interpretable and traceable Scope 3 GHG emissions from corporate sustainability reports. It combines large language models (LLMs) for semantic understanding (like finding relevant pages and interpreting tables) with rule-based components for precise tasks (like numeric parsing and unit conversion), ensuring high accuracy and explicit links to source text. This framework also generates a unique dual-level dataset connecting organizational disclosures to physical building assets.

Key Takeaways

  • 1
    Scope3Trace offers an evidence-grounded framework for extracting Scope 3 GHG emissions, combining LLMs with rule-based logic for traceable and verifiable results.
  • 2
    The framework constructs a novel dual-level, multimodal dataset linking organization-level disclosures to building-level assets, complete with explicit textual evidence.
  • 3
    Its hybrid rule-LLM design ensures high accuracy and robustness by delegating deterministic tasks to rules and semantic tasks to LLMs, making it efficient and reliable.

Conceptual Flow

HIGH LEVEL
1
How Scope3Trace Extracts Emissions Data

Smart computer programs read company reports, find pollution numbers, and check them against rules to make sure they are correct and traceable.

Messy Company Reports
Read & Understand
Clean Pollution Data
Proof from Report
2
High Accuracy and Detailed Data

The system finds pollution numbers very accurately and creates a detailed map of emissions for both companies and individual buildings.

Old Extraction Methods
Scope3Trace
Compare Performance
Lower Accuracy
Higher Accuracy
Company Emissions Map
Building Emissions Map