The snow depletion parameters (mfsno and scffac) in the IMS converter have mfsno truncated to 3.0, as mfsno > 3 results in the model not being able to attain full snow cover, even for very high snow depth. In the assimilation of IMS where the obs indicate 100% snow cover this can lead to excessively high snow density and snow depth, as the DA keeps adding snow depth to try to reach 100% snow cover.
Compare to the model parametsr.
This mismatch appears to be causing problems in the snow DA. Long term the snow depletion curve in the model needs to be improved. In the short term, QCing the IMS obs out for these vegatation classes is likely the best solution. @yuanxue2870 is running some tests.
This discrepancy exists in v17.
The snow depletion parameters (mfsno and scffac) in the IMS converter have mfsno truncated to 3.0, as mfsno > 3 results in the model not being able to attain full snow cover, even for very high snow depth. In the assimilation of IMS where the obs indicate 100% snow cover this can lead to excessively high snow density and snow depth, as the DA keeps adding snow depth to try to reach 100% snow cover.
Compare to the model parametsr.
This mismatch appears to be causing problems in the snow DA. Long term the snow depletion curve in the model needs to be improved. In the short term, QCing the IMS obs out for these vegatation classes is likely the best solution. @yuanxue2870 is running some tests.
This discrepancy exists in v17.