Background

Ocean-related options (OROs) to mitigate and adapt to climate change are receiving increasing attention from practitioners, decision-makers, and researchers. In order to guide future ORO development and implementation, a catalogue of scientific evidence addressing outcomes related to different ORO types is critical. However, until now, such a synthesis has been hindered by the large size of the evidence base. Here, we detail a protocol using a machine learning-based approach to systematically map the extent and distribution of academic evidence relevant to the development, implementation, and outcomes of OROs.

Method

To produce this systematic map, literature searches will be conducted in English across two bibliographic databases using a string of search terms relating to the ocean, climate change, and OROs. A sample of articles from the resulting de-duplicated corpus will be manually screened at the title and abstract level for inclusion or exclusion against a set of predefined eligibility criteria in order to select all relevant literature on marine and coastal socio-ecological systems, the type of ORO and its outcomes. Descriptive metadata on the type and location of intervention, study methodology, and outcomes will be coded from the included articles in the sample. This sample of screening and coding decisions will be used to train a machine learning model that will be used to estimate these labels for all the remaining unseen publications. The results will be reported in a narrative synthesis summarising key trends, knowledge gaps, and knowledge clusters.