Effect-size dataset and analysis code for a meta-analysis of airborne, soil and aquatic environmental DNA substrates

This repository contains the data and code for:Prasanna Kumar C (2026). Airborne environmental DNA outperforms soil and rivals water for terrestrial vertebrate monitoring: a meta-analysis of sampling substrates. Science of the Total Environment (submitted).Dataset citation: Prasanna Kumar C (2026). Effect-size dataset and analysis code for a meta-analysis of airborne, soil and aquatic environmental DNA substrates (Version 1) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.24106411Contact: micropras@gmail.comMethods summaryEffect sizes:Log response ratios (lnRR) of taxa counts, with Poisson variance.Log risk ratios of positive samples or sites, with a 0.5 continuity correction when a cell was 0 or n.Logit proportions of reference species detected, with binomial variance.Logit Jaccard overlap.Log ratios of first-order decay constants, with delta-method variance.Models: random effects with restricted maximum likelihood (REML) τ². The script reports 95% Wald intervals, Hartung–Knapp–Sidik–Jonkman (HKSJ) intervals, prediction intervals (k ≥ 3), Cochran's Q, I² and τ². Subgroup differences are tested with Q_between, and the analyses are repeated after study-level aggregation.Approximations: within-site correlation between paired substrates is ignored, and counts are treated as Poisson or binomial. See the `approximation` column and the article's Methods for details.