| dc.description.abstract | The Greenland ice sheet is the largest ice volume in the Northern Hemisphere. Current data suggests
that the ice volume is decreasing with 142 Gt/yr. Using physical ice sheet models like
ANICE we may see the evolution of the ice sheet over multiple glacial cycles. An important part of
these models is the surface mass balance, but global mass balance parameterizations based on insolation
and temperature have not yet resulted in an accurate model. This is especially true for the
parameterization of ablation. The three parameters are initially a1 = 0,513, a2 = 10, a3 = 45. Using
RACMO2/GR, a high resolution atmospheric climate model, as reference, we will optimize our ablation
model. Using an iterative optimization method to decrease the model error to match the reference data,
we find a single minimum for the error, with the a resulting parameterset of a1 = -0,103, a2 = 1,99,
a3 = -44,3. The spatial distribution of the ablation increases drastically, with correlation coefficient
between reference and model increasing from 0,19 initially to 0,88 with the optimized parameterset.
The results raise questions about the physical background of the parameters. With a1 being a parameter
for the transmissivity the resulting negative value can not appear physically. We conclude that
the insolation-temperature method does increase accuracy over the positive degree day model. The
resulting ablation stays high in glacial periods with an even higher refreezing fraction making sure the
ice volume seems relatively stable. Solutions are tried to be found in using runoff as parameterization,
making the refreezing fraction obsolete. This runoff parameterization does not change the signs on the
best parameters, but final results do not differ physically from the ablation plus refreezing model. | |
