LDAS Meeting Minutes
June 18th
2001
NCEP/EMC.
1. The LDAS group discussed the common surface air temperature criteria ( < 0.0 C ) that each LDAS LSM driver should be applying to decide if the input precipitation forcing is snowfall.
ACTION: OHD volunteered to confirm that the rainfall and snowfall forcing agree across the four LDAS LSMs. (Note the common LDAS output document calls for each LDAS LSM to echo out its rainfall and snowfall forcing in its hourly LSM output.)
2. On 06 Apr 01, the expanded LDAS land mask to accommodate the NASA/GSFC Catchment LSM was implemented into NCEP's realtime LDAS forcing software by Dag Lohmann. This expanded land mask wholly contains the slightly smaller original mask. (Note: This was a change to the land mask in the LDAS input forcing fields only. The slightly smaller common land mask used in LDAS LSM GRIB output was left unchanged.)
3. The LDAS group confirmed that the Higgins precipitation forcing will not be used outside the 48-state CONUS domain. Each LDAS LSM driver should be employing the 16-row blending weights produced and provided by Brian Cosgrove of NASA to blend EDAS and Higgins precipitation forcing beyond the northern and southern borders of the CONUS land domain. Beyond the 16 blending rows in Canada and Mexico, only EDAS precip forcing should be used.
4. John Schaake discussed and illustrated the 188 CONUS test basins he has chosen as candidates for validation of LDAS streamflow validation. From these he had chosen 26 basins for initial scrutiny. These 26 basins all lie east of 100 W longitude, and most lie in the southeast U.S., both to avoid mountainous areas and winter snowpack in our initial phase of LDAS streamflow validations. In follow-on discussions, the LDAS group urged John to add a few more basins near the 100 W western extremity, plus a few in Illinois (motivated by the readily accessible Illinois statewide soil moisture monitoring network), plus one basin from the OHD DMIP project, and last 1-2 added basins from the PILPS-2c Arkansas-Red River domain.
5. Jesse Meng is developing validations of various solar insolation products (e.g. GOES, EDAS, AGRMET) against the 6-station SURFRAD network in CONUS. He showed preliminary but strong evidence that there is a substantial low bias in the GOES-based solar insolation retrievals over snow cover. Jesse will pursue cause of candidate fixes of this bias further with Istvan Laszlo, Dan Tarpley, and Rachel Pinker.
6. The LDAS group resumed discussions from recent meetings of the thorny issue of how to provide common and reasonable initialize land states for the 30 Sep 96 start date of the 3-year retrospective LDAS LSM reruns. It was decided to initialize all water stores, except soil moisture to zero (e.g. zero canopy water and, since Sep is the month of annual minimum in snowpack, zero snowpack). The group was strongly leaning toward using the NCEP Global Reanalysis land states as the source of initial soil temperature and soil moisture. Dag Lohmann and Ken Mitchell will explore this path for the next meeting. The group rejected the strawman proposal of initializing the soil moistures stores at saturation, since the deep soil spin-down of such a wet state in vegetation-sparse arid and semi-arid regions could be several years.