I worry that <snip> Piomas have had difficulties assessing the total volume of areas where we have scant floes and 'Slush Puppy'
Hi G-W,
WRT PIOMAS, their problem is a simplifying assumption made in the underlying model of the sea ice. PIOMAS uses one set of values for all type of sea ice: salinity, density, melting point. That means FYI and MYI both behave slightly differently than the median values chosen for the model parameters.
I think this is the reason why the Polar Science Center
validation effort found that:
Note that from the comparison with in situ observations it appears that PIOMAS tends to overestimate thin ice and underestimates thick ice.
Note that FYI is composed of about 10% air bubbles. That causes it's density to be lower. The liquid brine portion of sea water contained within FYI
IS dense, but it is
NOT sea ice, and should be subtracted from the volume estimate. Again, one set of parameters does not fit all species of sea ice!
I estimate that overall it takes about 10% less energy to melt FYI than PIOMAS modeled sea ice, and perhaps 10% more energy to melt MYI than modeled ice. This causes problems later on as we'll see.
Note that PIOMAS did a terrible job predicting the September 2012 avg SIE in its
Late Summer update to the SEARCH Project (
3.9 +/- 0.32 M km2). About 270,000 km
2 more ice melted in the event than PIOMAS predicted only 2 weeks earlier, almost all of it in FYI areas, while the MYI persisted.
So here's my take on PIOMAS's shortcoming for prediction: As the ratio of FYI/MYI increases, PIOMAS becomes increasingly inaccurate due to sea ice parameter mismatch.
The obvious fix is to model different types of sea ice, with persistence over time to age the ice.