Bridging Standards and Sensors: A Systematic Literature Review and Quantitative Evidence Synthesis of Carbon Accounting Methods for Energy Efficiency Across Terrestrial, Marine, And Built-Environment Ecosystems

carbon accounting; systematic literature review; quantitative evidence synthesis; energy efficiency; LiDAR; greenhouse gas reporting; PRISMA 2020

Authors

August 11, 2026
August 12, 2026

Downloads

ABSTRACT

Background: Carbon accounting has moved from a specialist scientific discipline into a mainstream requirement of corporate governance, financial regulation, and ecosystem management. The literature measuring it, however, remains divided between two largely non-communicating traditions: administrative and disclosure-based accounting on one hand, and physically measured, sensor-based accounting on the other.

Objective: This study systematically reviews the governance and standards literature on carbon accounting and synthesises reported measurement-accuracy metrics from primary studies spanning forest, soil, marine, agricultural, built-environment, and corporate domains, with the aim of characterising the degree of alignment — or misalignment — between these two traditions and their implications for energy-efficiency-linked policy.

Methods: A systematic literature review with quantitative evidence synthesis was conducted following the PRISMA 2020 framework. Searches in Scopus and Web of Science used the Boolean string, restricted to English-language peer-reviewed sources from 2011 to June 2026. After deduplication and screening, 79 unique records were retained: 35 governance and standards sources synthesised narratively, and 55 quantitative primary studies (60 statistical records) pooled descriptively by ecosystem and sector domain. Formal inverse-variance-weighted meta-analysis was not performed due to outcome-measure heterogeneity; accuracy statistics (R2, RMSE, rRMSE) are reported as domain-level ranges. A five-criterion quality appraisal was applied to all quantitative records, and a sensitivity analysis examined the effect of excluding China-based studies on domain-level accuracy estimates.

Results: Model accuracy varied sharply by domain. Engineered and semi-controlled systems — UAV-LiDAR forest biomass inversion (R2 = 0.93-0.95) and residential building carbon-intensity modelling (R2 = 0.91) — consistently outperformed open-ecosystem reconciliation models, where regional forest-carbon inversions fell as low as R2 = 0.147. Sensitivity analysis showed that excluding China-based forest studies reduced the domain upper bound from R2 = 0.95 to R2 = 0.558. No primary study in the marine and blue-carbon subset reported a comparable accuracy statistic, despite well-quantified carbon-stock densities (mangrove density = 937 t/ha; Liu et al., 2024). Narrative synthesis identified persistent reliance on estimate-based reporting, unresolved tension between attributional and consequential accounting logics, and a documented concentration of unverified exposure in Scope 3 emissions.

Conclusions: Administrative carbon-accounting frameworks and physically measured, sensor-based accounting are developing along separate trajectories with materially different accuracy standards. Carbon claims tied to built-environment retrofits and grid-connected systems rest on firmer empirical ground than those derived from open-ecosystem reconciliation models. The marine and blue-carbon domain — where accuracy benchmarks are absent and credit issuance is expanding — represents the most consequential governance gap identified in this review. Regulatory extension of mandatory disclosure into ecosystem-based carbon domains should be sequenced against demonstrated measurement readiness.