Record Details
Modeling Arctic melt ponds using a resolved ice model with GCM forcing
ScholarsArchive at Oregon State University
| Field | Value |
|---|---|
| Title | Modeling Arctic melt ponds using a resolved ice model with GCM forcing |
| Names |
Collins, Lee E.
(creator) Shell, Karen (advisor) Skyllingstad, Eric (advisor) |
| Date Issued | 2013-10-23 (iso8601) |
| Note | Graduation date: 2014 |
| Abstract | The albedo of Arctic sea ice depends greatly on the formation of melt ponds. These ponds form in depressions on the ice as surface snow melts during the summer months, and their location is determined mainly by the initial snow topography. Using a high resolution sea ice model forced with data taken from the Atmospheric Radiation Measurement (ARM) site in Barrow, AK, we investigate how specific factors, both internal model parameters and initial conditions, affect the evolution of melt ponds on Arctic sea ice. We also use forcing data taken from output of the Community Earth Systems Model (CESM) to investigate the differences in melt pond parametrization between our model and CESM. The resolved model uses a unique and innovative approach in pond modeling, the "trigger depth" method, to initiate pond drainage. Results from sensitivity analysis on the trigger depth show the validity of this new approach, suggesting it could be useful in other ice models. The initial snowpack has a very large role in pond formation and extent. We use surface topography gathered from LiDAR scans from the ARM site to provide a realistic snowpack surface. For our sensitivity analysis of the total initial amount of snow in the model, we alter only the minimum thickness of the snow on top of the ice, retaining a consistent surface topography for each simulation. The LiDAR topography from the ARM site provides a more realistic approach to the pond model, as opposed to a randomly generated method of creating snow topography. Large initial snowpack inhibits the formation of deep channels in the ice, reducing pond fraction at the end of the melt season. Finally, we force the resolved model simulations with data from CESM and compare the pond behavior to that of CESM. CESM does an unrealistic job of representing melt ponds, partially due to the way melt ponds are parametrized in the model, using a "thickness-class" method for creating and categorizing melt ponds. CESM pond formation occurs over a much broader time span compared to observations and our resolved model. Results from this work will be used to investigate and possibly improve the melt pond parametrization in CESM. |
| Genre | Thesis/Dissertation |
| Topic | Arctic |
| Identifier | http://hdl.handle.net/1957/44613 |
