The spatial distribution of tests of conservation interventions does not align with global conservation needs

Aligning research effort to address major threats to biodiversity is key to bending the curve of biodiversity loss with limited resources. Recent global threat mapping for terrestrial amphibians, birds, and mammals enabled us to assess whether there was alignment between the spatial distributions of key threats (logging, hunting, agriculture, pollution, Invasive Non-Native Species, and climate change) and published tests of conservation interventions addressing these threats (1025 English and 147 non-English language studies from 15 languages collated in the Conservation Evidence database). We ran Generalised Linear Models (GLMs) to determine the key predictors (including socio-economic factors) of the number of studies that test interventions on amphibians, birds, and mammals per country. We found poor spatial alignment between studies and the impact probabilities of threats for each taxon. Studies were distributed across 92 countries with 64 % of all studies conducted in the United States, United Kingdom, Canada, and Germany. Most studies focused on interventions against agricultural threats (60 %, 79 %, and 58 % for amphibians, birds, and mammals, respectively) – next most common were interventions tackling climate change (18 % of bird studies), hunting/trapping (29 % of mammal studies), and logging (20 % of amphibian studies). Countries with the highest threat levels typically had few or no studies, although several countries had both high threat levels and numbers of studies, including: Germany for amphibian interventions against agriculture, New Zealand for bird interventions against invasive species, and Brazil for mammal interventions against hunting. However, our modelling suggested that certain socio-economic factors, not levels of conservation risk posed by threats, were associated with more studies testing interventions. These factors included countries with higher levels of Gross Domestic Product (GDP) (amphibians and birds) and greater levels of government effectiveness (mammals). We call for rapid action to remove barriers to (and incentivise) reporting tests of interventions, such as through specific funding to test conservation actions and a global reporting database in both English and non-English languages to fill evidence gaps in underrepresented regions. Building ‘testing partnerships’ to link Global North and Global South institutions through networks of funders, local community, governmental, non-governmental and research organisations would also help to better coordinate testing of interventions to address conservation needs. This will require international coordination to build capacity for testing interventions and reporting their effects, whilst avoiding parachute science. • Research effort needs to address threats to bend the curve of biodiversity loss. • We analysed the predictors of the spatial distribution of evidence on conservation actions. • We found poor spatial alignment between studies and threats across taxa. • Other factors, such as higher GDP and government effectiveness, were better predictors of the distribution of evidence. • We need global testing partnerships and coordinated funding to prioritise research to address these literature biases.