Collaborative Research: Combining Arctic Observing Network Observations and Remote Sensing Data to Understand Sea Ice Mass Balance and Albedo Feedbacks in a Changing Arctic
Arctic sea ice is undergoing significant and accelerating change, which has been observed in increasing detail over the past several decades. Stakeholders in the Arctic and beyond are impacted by the cascading effects that these changes have on resource accessibility, ecosystem health, and earths physical climate system. Planning mitigation and adaptation strategies requires improvements in the ability to predict the arctic climate system. Confounding predictive model development are large gaps that remain in the understanding of the role albedo feedbacks are playing in sea ice loss. This project will improve the understanding the impacts of solar energy absorption and partitioning on ice mass balance by tracking Arctic Observing Network (AON) sea ice sites as they move through space and time, using the rich datasets being collected at the sites as case studies. The approach integrates AON data with measurements from prior field campaigns, atmospheric reanalysis products, and high resolution remote sensing data using a 3D resolved sea ice-ocean mixed layer coupled model. The objectives are to quantify the deposition of solar energy within a 10x10km study domain around the sites with meter-scale resolution, identify the fate of that solar energy over time (e.g. ocean storage vs. ice bottom melt vs. lateral melt), and improve the ability of our resolved scale model to represent the processes controlling solar absorption and fate. The model explicitly represents critical ice and ocean processes such as melt pond formation, brine drainage, freshwater balance, and upper ocean stratification, providing a tool for data integration that can account for all first order processes impacting radiative transfer and heat storage. Important gaps in initialization datasets will be addressed by an ensemble modeling approach. Iterative comparison of site observations with model states will inform the selection of poorly constrained initial conditions, evaluate system sensitivity, and allow testing of improved model parameterizations for processes such as pond evolution.