Hi-
I'm working on using ROMS to model interannual variability in freshwater lakes that form ice seasonally, like the Great Lakes. I'm currently working on a simple 1-D test platform with seasonal shortwave forcing, and all other forcings held constant. I've noticed two very odd behaviors, and I'm not sure where to go from here. I'd like to discuss the SeaIce routines with anybody else who might be having more success.
1. Ice persists long after the surface water has begun warming- in fact, the ice does not clear away entirely until well after the spring overturn, and surface temperatures often reach 6-7C before the ice goes away. While I expect that some shortwave radiation must get through the ice and into the water, it appears that the surface water and ice layer are not well-coupled. It works better at the other end, with ice not forming until the surface reaches ~0C.
2. There's something odd in the heat balance of the system. If I compare the heat content of the water column plus the latent heat of fusion of the ice to the total surface heat flux, things work fine until ice starts to form. Then roughly 3x as much ice forms as one would expect from the surface heat loss. Additionally, there is more heat in the water column after the ice goes away than one would expect given the surface heat flux. This balance works fine in the absence of ice.
Any ideas? I know that the SeaIce package is developmental at this stage, but there must be other people out there having some of the same experiences.
SeaIce issues
1D ice models act in unexpected ways.
1) It is very hard to melt ice with solar radiation, especially with a snow layer. Most ice melts by shortwave heating the water and the water melting the ice. Limit your ice concentration to 90 or 95% and you will see faster melt in spring.
2) We had a problem with excess ice in some cases due to humidity. Be careful with units (I forget the details). If humidity is high, then you will get a lot of snow which will accumulate as ice. You should not be able to freeze more than 3/4 to 1m of ice thermodynamically due to the insulating effect of ice on heat exchange. There could also be an issue with frazzle ice formation at zero salinity, which may not have been tested before.
Good luck,
John
1) It is very hard to melt ice with solar radiation, especially with a snow layer. Most ice melts by shortwave heating the water and the water melting the ice. Limit your ice concentration to 90 or 95% and you will see faster melt in spring.
2) We had a problem with excess ice in some cases due to humidity. Be careful with units (I forget the details). If humidity is high, then you will get a lot of snow which will accumulate as ice. You should not be able to freeze more than 3/4 to 1m of ice thermodynamically due to the insulating effect of ice on heat exchange. There could also be an issue with frazzle ice formation at zero salinity, which may not have been tested before.
Good luck,
John