Award: OCE-1558225

Simulations with Earth system models (ESMs) project that anthropogenic climate change will increase ocean acidification and affect phytoplankton productivity over the coming century, yet there is substantial uncertainty in these projections, particularly on regional scales. In this project, we have used ESM simulations and observations to carefully study the 3 main sources of uncertainty in ocean biogeochemical projections: (1) uncertainty due to the chaotic nature of the climate system, (2) uncertainty due to the mathematical representation of the Earth system in different climate models, and (3) uncertainty due to our future fossil fuel emission trajectories. We took advantage of multiple sets of ESM ensembles to study the chaotic nature of the climate system, so called ?internal variability?. These ensembles of ESMs were developed to mimic the ensemble approach to weather forecasting: begin each simulation with slightly different initial conditions and allow the chaos to create different climate states (aka, the ?butterfly effect?). The resulting spread across the ensemble tells us how our projections can be influenced by this chaos. We further took advantage of simulations of ESMs submitted to the 5th and 6th Coupled Model Intercomparison Projects (CMIP5 and CMIP6, respectively) to assess the differences in projections due to different models? mathematical representations of the climate system, so called ?model structure? and responses to a range of fossil fuel emissions, so called ?emission scenario?. Our publications resulting from this project show that projection uncertainty in air-sea carbon dioxide exchange differs on global and regional scales: on global scales, projection uncertainty grows with lead time and is primarily attributed to uncertainty in emission scenario, while on regional scales, projection uncertainty is high for all lead times and is primarily driven by internal variability and model structure. We use our ensemble framework to project that most regional changes in phytoplankton productivity are inevitable in a changing climate, even with emissions reductions. The ensemble framework also demonstrates where and when anthropogenic changes in the ocean biogeochemical system emerge from the background noise of internal variability in critical upwelling systems such as the California Current and Southern Ocean. Finally, we employ a novel statistical technique to create ensembles of observations using a single observational record and use this observational ensemble to comment on the likelihood of observed decadal trends given the background noise of climate variability. This work bridges an existing gap between Earth System Model and observational studies assessing impacts of climate change on ocean biogeochemistry through a careful analysis of the sources of prediction uncertainty. Better understanding of uncertainty in projections for ocean biogeochemistry will improve assessments of expected climate change impacts on marine ecology and the ocean carbon sink. Interpretation of observed change will also benefit from a robust assessment of these sources of uncertainty. Critical research needs for reducing uncertainty and enhancing predictability have been identified. Since its inception, this project has either directly or indirectly supported the work of 5 Ph.D. students, 1 Ph.D. student visitor, and 4 undergraduate underrepresented students at the University of Colorado, all of whom have conducted research with the ESM ensembles, received mentoring from the project PI, and traveled to scientific meetings to present their results to the broader community. Last Modified: 07/21/2021 Submitted by: Nicole S Lovenduski