Using ARM Measurements to Improve the Simulated Vertical Distribution of Arctic Aerosols in the Community Atmosphere Model

Principal Investigator(s):
Mark Flanner, University of Michigan

The overarching objective of this project is to identify strategies for improving the simulation, by global models, of the vertical distribution of aerosols in the Arctic. Motivation for this is two-fold. First, the response of Arctic surface climate to light-absorbing aerosols has been shown to depend strongly on their vertical distribution, with surface cooling caused by aerosols in the upper troposphere, and strong warming associated with near-surface heating by light-absorbing aerosols. Second, relative to a limited number of aircraft measurements, global models poorly simulate vertical profiles of Arctic black carbon (BC), the dominant light-absorbing aerosol species in most environments. Nearly all models simulate excessive BC at higher altitudes and too little BC near the surface of the Arctic.

To achieve our objective, we will assemble a suite of observations of vertical aerosol content in the Arctic, using High Spectral Resolution Lidar (HSRL) measurements from the ARM Barrow site as the primary dataset. With proper cloud screening, the HSRL measurements provide the vertical profile of aerosol extinction optical depth. HSRL data from Barrow are now available for an entire year, offering the potential to evaluate a full annual cycle of simulated aerosol profiles above Barrow. We will supplement these data with measurements throughout the Arctic that distinguish contributions to aerosol absorption, including vertical measurements of BC from aircraft campaigns, near-surface aerosol measurements from the AMF-3 facility at Oliktok Point and other Arctic monitoring stations, measurements of BC in snow, and total column aerosol extinction and absorption optical depth.

We will then conduct a series of sensitivity studies with the Community Atmosphere Model (CAM), embedded with the Modal Aerosol Model (MAM), to explore how the quality of vertical aerosol simulation is affected by perturbed aerosol and cloud microphysical parameters, model resolution, and perturbed emissions from different regions. Leveraging knowledge gained from recent studies, we will alter parameters that determine: the aging rate of carbonaceous aerosols (governed by hygroscopicity and a threshold of condensed monolayers), aerosol removal efficiency by stratiform precipitation, aerosol activation at different stages of convection, dry deposition velocity over snow and ice surfaces (governed by surface resistance), and relative distributions of liquid and ice-phase clouds. We will use an iterative process, and possibly a Latin hypercube technique, to identify one or more optimal sets of parameters, and we will then apply these parameter sets in simulations at four different horizontal grid resolutions, ranging from roughly 2 degrees to 0.25 degrees, to identify how model performance is affected by grid size. We will also conduct “tagged” aerosol runs to identify the contributions of emissions from different regions and sectors to aerosol burdens at different altitudes above Barrow and throughout the Arctic, and we will apply this information to quantify the maximum range of impact on simulated vertical aerosol profiles associated with uncertainty in emissions from different regions and sectors. Finally, we also plan to explore how finer model vertical resolution and use of a new spectral element dynamical core impact simulated Arctic aerosol distributions.

A broad impact of this project is that it will result in improved simulation of Arctic climate impacts from anthropogenic aerosols. A more fundamental outcome will be the identification of parameter combinations in CAM that produce optimal simulations of Arctic vertical aerosol distributions. We will identify optimal parameter sets for multiple grid resolutions, thus benefiting a broader range of CAM users, and our sensitivity studies will also shed light on minimum grid resolutions that may be needed for realistic simulation of Arctic aerosol distributions. Identification of parameter changes that improve the simulation of remote aerosol distributions will inform on strategies for the development of more physically realistic, self-consistent, and scale-independent representations of aerosol processes and aerosol-cloud interactions, germane for an Earth System Model like ACME. Finally, this project will enable the training of a Ph.D. student.