Participants: UW: Dennis P. Lettenmaier and Kaiyuan Y. Li
Princeton: Eric F. Wood and Justin Sheffield
During this meeting Princeton and UW both presented brief summaries of progress to date by task, and discussed how we will proceed in the early next year, particularly with respect to incorporation of VIC soil physics and snow parameterizations into CLM. We agreed that by the next CCPP PI meeting (Camridge, April 2006) we need to have a working version of CLM upgraded with VIC soil physics, tested over most or all of the evaluation sites for which UW presented results of VIC and CLM comparisons. Those results will need to show that the new version of CLM significantly improves hydrologic prediction (i.e., resolves most or all of the problems that are apparent in the results that Kaiyuan showed).
UW: (1)Finish incorporation of VIC soil hydrology into CLM; (2) Evaluate new version of CLM and show that its hydrologic performance is significantly improved; (3) begin incorporation of VIC snow model (especially snow bands into CLM). It is understood that items 1) and 2) will consume about 80% of Kaiyuan's time during this period, and the snow code modifications will account for the balance.
Princeton: (1) consolidate data sets (forcing, soil and veg parameters) shown by Kaiyuan, which will also be used to test upgraded CLM. Augment the data sets prepared by Kaiyuan to date with a mid-latitude forest site, and possibly CART ARM; (2) Develop an implicit scheme for solution of soil (and possibly thermal) profiles for revised CLM/VIC soil physics (initial version to be coded by Kaiyuan will be explicit(probably backward difference).
Based on Kaiyuan's preliminary proposal and the discussions between Princeton and UW, the methodology of incorporating VIC hydrology into CLM is as follows:
Below when we refer to 2-Layer VIC, it means the upper layer (actually the first and second layer in 3-layer scheme) and lower layer (the third layer in 3-layer scheme).
In the time order during a model time step:
(1) Surface runoff is calculated based on the 2-layer VIC scheme;
(2) Infiltration is based on 10-layers (CLM layer scheme) using information for
VIC layers interpolated to specified CLM layers;
(3) Soil evaporation, root-water-uptake and water redistribution (upward and
downward) will be based on 10 CLM methodology layers;
(4) Soil thermal states and fluxes will be computed based on post-evaporation
moisture in 10 CLM layers (Kaiyuan, you need to think about whether you want to
do this before or after the evaporation computation);
(5) 10 CLM layers are aggregated back to 2 VIC layers and base flow is
calculated using VIC drainage formulation; modified (lower VIC layer) soil
moisture is carried to the next time step, and the loop repeats with step 1.