Abstract
Background: Climate change is reshaping the marine environment through increasing sea surface temperatures, more intense and frequent marine heatwaves, decreased pH, deoxygenation, and concurrent alteration of many other environmental drivers. As these effects intensify, researchers are increasingly using laboratory studies to manipulate multiple drivers simultaneously. However, designing informative and tractable multiple driver experiments is challenging and has led to studies with limited comparability and predictive power. To advance this emerging field, the Scientific Committee on Oceanic Research infrastructural project Changing Ocean Biological Systems (COBS): how will biota respond to a changing ocean? developed a suite of resources to help researchers design, carry out, and analyse data from multiple driver experiments. To introduce these resources to the wider community, over 20 in-person and virtual workshops have been held, which has allowed the project to refine recommendations by uncovering key knowledge gaps.
Results: Across workshops, we identified four common issues with multiple driver experiments: (a) the design of overly complex experiments with low statistical power that are incapable of answering the question proposed, (b) the use of sub-optimal experimental designs, e.g. a full factorial design when experimental units are limited, (c) the design of experiments that confound the effects of treatment and experimental equipment, and (d) lack of confidence in applying the more advanced statistical techniques appropriate for some experimental designs. To address these challenges, we developed resources to help researchers define relevant questions, select the appropriate experimental design, and carry out the statistical analyses relevant to those designs.
Conclusions: Improved design and analysis of complex multiple driver experiments will increase the comparability of results across laboratories and disciplines, optimise limited resources, and increase the predictive power of marine climate change research.