Meta-analysis on herbivore exclusion research and practice
Description
To assess the current state of knowledge on herbivory exclusion devices in seagrass projects in the southern Atlantic and wider Caribbean, we performed a meta-analysis on studies that incorporated caging and fencing for herbivory exclusion within their experimental designs. A modified version of the PICO (Population, Intervention, Comparison, Outcome) process was utilized to address our research question (Foster and Jewell 2017). The search string encompassed terms related to population (seagrass), variables (herbivory exclusion), and outcome (growth). Our review revealed a significant gap; herbivore exclusion is rarely addressed in published restoration studies, in part due to the absence of experimental design and limited publication of practitioner-led projects. Consequently, the meta-analysis was constrained by small sample sizes and potential publication bias, offering limited insight into broad-scale trends.
Files
Steps to reproduce
To assess the current state of knowledge on HEDs in seagrass projects, we reviewed studies that incorporated caging and fencing for herbivory exclusion within their experimental designs. In developing the search strategy, a modified version of the PICO (Population, Intervention, Comparison, Outcome) process was utilized to address our research question (Foster and Jewell 2017). The search string encompassed terms related to population (seagrass), variables (herbivory exclusion), and outcome (growth). Searches were performed on Ebsco Host, Google Scholar, and Web of Science to identify candidate publications (search date 07 August 2024 and 22 October 2024), using the search string (seagrass* OR seagrass meadow*) OR “thalassia testudinum” OR “halodule wrightii” OR “syringodium filiforme” NOT (Zostera* OR Posidonia*) AND ("cage" OR "herbivore exclusion device” OR “enclosure” OR “exclosure”) AND (“biomass” OR “productivity” OR “percent cover” OR shoot* OR “density” OR “height” OR “leaf area” OR “primary productivity”). We limited our review to studies conducted in tropical systems with canopy-forming seagrass species similar to those in the IRL (i.e., Halodule wrightii, Syringodium filiforme, and Thalassia testudinum) and excluded studies on Zostera spp. and Posidonia spp. to ensure geographic relevance and ecological comparability. We identified additional papers from in-text citations and personal communications. We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) standards for meta-analysis reporting (Moher et al. 2009). For inclusion, studies had to report experimental or observational field data on seagrass growth in both exclusion and control treatments throughout the wider Caribbean and/or the tropical Atlantic. We screened the titles, keywords, and abstracts of 487 candidate publications (dissertations, publications, theses, books, and gray literature) and identified 33 for full review. Of these, 21 were excluded for failing to meet the inclusion criteria. Six categories of seagrass growth metrics (primary productivity, aboveground biomass, shoot density, leaf area, shoot height, and percent cover) were extracted across HED treatments (full, partial, and control) as proxies for restoration success, yielding 165 individual peer-reviewed outcomes. For each study, means, sample sizes, and standard deviations or standard errors, were recorded for each metric and treatment. When responses were measured repeatedly at individual plots, only the final measurement was used to minimize temporal autocorrelation (Koricheva et al. 2013) and to capture the longest duration since treatment initiation. When data were not reported in the text or supplementary materials, values were extracted from figures using WebPlotDigitizer (Version 4.8) (Rohatgi 2023). Additional supporting information, such as study location and treatment details (e.g., mesh material, size, and shape) were also recorded.