Risk-of-bias assessment (also known as the critical appraisal of internal validity) is a required component of systematic reviews. It evaluates how susceptible individual primary studies are to bias at the level of each effect estimate. The assessment results are then used in data synthesis, where subgroup analyses are conducted to draw overall conclusions about the effect. However, recent research has highlighted inadequate practices in risk-of-bias assessment in environmental evidence syntheses. The vast majority of these syntheses either omitted the critical appraisal step entirely or conducted it inappropriately. Additionally, there has been no standardised risk-of-bias assessment tool in the environmental sector. Although there are standardised tools in medicine, employing these tools in environmental evidence syntheses is not always appropriate. To address these issues and improve the standard of risk-of-bias assessment, we developed the first standardised risk-of-bias assessment tool in the sector: the Collaboration for Environmental Evidence Critical Appraisal Tool (environmentalevidence.org/cee-critical-appraisal-tool). Its prototype version was released in 2020, followed by real-world testing by environmental evidence synthesists, a survey to gather feedback from users, and the identification of action items for revision. All action items were addressed to produce a formal version, which this paper introduces. The tool can be used in quantitative evidence syntheses addressing causal questions in any environmental topic. It provides a standardised framework for risk-of-bias assessment, in which a study’s findings are first assessed within individual risk-of-bias criteria by answering a series of checklist questions. The results of these within-criteria assessments inform the overall judgement about risk of bias for the study’s findings. We hope that the release of this tool will lead to higher standards in risk-of-bias assessment. Future work is needed to further evaluate the tool and to create an external validity section to help evidence synthesists conduct critical appraisal of external validity objectively, comprehensively, and rigorously.
Land management practices have significant impacts on biodiversity and climate change processes. Conservation and rewilding practitioners are increasingly seeking to enhance ecosystem functions and carbon sequestration simultaneously, encouraged by climate policies such as the European Green Deal. In recent years, ecosystem restoration involving large herbivores (≥ 10 kg) has attracted growing interest. However, the climate impacts of large herbivores are complex, including direct and indirect impacts from greenhouse gas (GHG) emissions, carbon storage, wildfire and albedo. These impacts can have cooling or warming effects and vary with species, herd density, season and habitat. To date, the diverse research on this topic has not been collated and categorised in a systematic map. Here, we present a systematic map of research on the climate effects of large herbivores in European terrestrial habitats, highlighting research distribution, clusters and gaps, and the potential for further quantitative analyses. We address the primary question of what research exists on the climate effects of large herbivores in Europe. Following our pre-published protocol, we used a pre-defined search string to conduct a search of peer-reviewed and grey literature, using a range of bibliographic databases, search engines and websites. We screened search results for relevance in two stages: by title and abstract, then full text. We excluded articles that failed to meet the eligibility criteria of terrestrial habitats in Europe (Population), large herbivores (Exposure) and climate impacts (Outcome). We coded the included studies by multiple variables, recorded in a spreadsheet alongside the narrative synthesis. We identified 316 articles for inclusion, involving a total of 565 studies (different research comparisons within the same article). Most studies were conducted in Northern Europe (63
Abstract Background The global expansion of salmon aquaculture has transformed marine food systems, offering economic benefits while amplifying environmental and socio-economic challenges. The shift from wild-caught to farmed salmon has led to widespread concerns, including disease transmission, genetic introgression, pollution, and conflicts with Indigenous communities. Additionally, salmon feed supply chains have global implications, such as overfishing in West Africa. Although many studies exist, research is fragmented and dispersed across disciplines, and the influence of industry funding raises questions about bias. This overview of reviews synthesises existing systematic reviews on salmon farming to evaluate the scope, quality, and findings of the secondary literature on its environmental, economic, and social impacts. Specifically, we ask the question: “What systematic review evidence exists on Atlantic and Pacific salmon aquaculture, and what are the characteristics, thematic focus, and methodological quality of this evidence?”. Methods This overview of reviews follows guidelines from the Collaboration for Environmental Evidence (CEE) and the ROSES reporting standards. Systematic reviews (broadly defined as reviews either claiming to or actually using some form of systematic methodology) were identified via searches in Scopus, Web of Science, and Lens.org using a structured, validated search strategy. Inclusion criteria focused on reviews explicitly targeting Atlantic or Pacific salmon aquaculture (including rainbow trout). Metadata were extracted on study methods, topical focus, production stages, and funding sources. Review quality was assessed using the CEESAT critical appraisal tool, and visualisations were produced to identify gaps in coverage and assess methodological rigour. Review findings Out of 327 unique records, 37 reviews were included. Most reviews focused on production-related topics (e.g. fish health, sea lice, feed supply), with very few addressing environmental or social impacts. No reviews met the threshold for high methodological rigour, and only four were rated as moderate quality; the remainder were judged as low or exceptionally low quality. Common weaknesses included absence of a priori protocols, lack of critical appraisal, poor reporting of screening and extraction methods, and inappropriate synthesis techniques such as vote counting. Industry co-authorship was present in 7 reviews, and 8 were industry-funded. Research on sustainability was hindered by methodological inconsistency and limited transparency. Conclusions Despite a growing body of literature on salmon aquaculture, high-quality evidence syntheses remain rare. Most systematic reviews are methodologically flawed and overly focused on productivity, with limited assessment of environmental or socio-economic externalities. Greater attention is urgently needed to unbiased, interdisciplinary, and high-quality evidence synthesis—especially on the broader impacts of salmon farming. Decision-makers should interpret existing evidence syntheses on salmon aquaculture with caution, given the generally low methodological rigour and high risk of bias identified across reviews. Apparent consensus or patterns in the literature may reflect methodological artefacts rather than robust underlying evidence. We therefore call for enhanced rigour in review methods, transparent reporting, and increased scrutiny of industry influence on research agendas.
Abstract Background Tropical areas, particularly Brazil, are home to a significant portion of the world’s biodiversity, with the Brazilian Atlantic Forest (BAF) being a crucial biome recognized as a global biodiversity hotspot. However, the BAF has been subject to severe degradation due to anthropogenic activities, resulting in extensive habitat loss and fragmentation. These changes have prompted numerous studies on their effects on biodiversity, focusing on forest fragments and protected areas. Despite this research, there is a notable lack of meta-data on biodiversity in human-modified landscapes within the BAF- where only 31% of the land remains as native forest. Most of the biome has been transformed by human activities, primarily for pasture and agriculture. By recognizing these knowledge gaps, researchers have begun investigating biodiversity in anthropogenically altered landscapes, to better understand species’ responses to human-induced changes and inform conservation policies. A comprehensive literature search was conducted to answer the following primary question: “ What is the diversity of birds and mammal species in agricultural fields and their adjacent areas of natural vegetation ?” We aimed to address this question by creating a systematic map summarizing current knowledge about avian and mammalian species’ presence in the BAF’s agricultural matrix. This research also aimed (i) to identify common species in agricultural landscapes, (ii) to compare species richness in-crop (inside the agricultural field) and off-crop (non-cultivated areas outside the agricultural field), (iii) to determine crops with high species visitation, and (iv) to highlight areas for future research. Methods We conducted a systematic literature map on bird and mammal biodiversity in agricultural landscapes within the BAF and adjacent natural vegetation areas. The search covered Web of Knowledge, Scopus, the National Library of Medicine, and the “ Biblioteca Digital Brasileira ” (Brazilian Digital Library) database for grey literature (dissertations and theses). Citation records were surveyed based on predefined criteria through title/abstract screening and full-text analysis. Eligible publications were used for meta-data coding and systematic map construction. Review findings Our systematic map identified 207 eligible publications included in the final evidence base. Research effort was distributed across taxonomic groups, with 78 studies focusing on birds, 126 on non-flying mammals (NFM), and 22 on bats. The most frequently studied agricultural systems were pasture and eucalyptus, followed by sugarcane, coffee, cacao, rubber tree, pinus, grape, araucaria, and corn. Extracted meta-data showed that studies reported species occurrences in both agricultural fields and adjacent natural areas. The map revealed significant knowledge clusters and gaps, particularly regarding the distribution of research across different crop types and the inclusion of species with varying conservation statuses. Conclusion This systematic map identifies 846 species across 48 crop types in BAF agricultural landscapes, potentially indicating their use by wildlife; however, further quantitative synthesis is necessary to assess the relative importance of these areas. By mapping knowledge clusters and gaps—such as underrepresented regions and specific crop-taxa associations—this work provides a foundation for future systematic reviews and informs the development of more targeted environmental risk assessments and conservation policies.
We welcome Hodgson et al. (2026) empirical evaluation of ontology-grounded large language models (LLMs) for data extraction in environmental evidence synthesis. The study makes a valuable contribution by quantifying performance across attribute types and by openly documenting where current approaches struggle.
