N-LDAS Meeting Minutes

Hosted by NWS/OHD

08 January 2002

 

A - ATTENDEES:

Ken Mitchell, Dag Lohmann, John Schaake, Qingyun Duan , Dennis Lettenmaier, Paul Houser, Brian Cosgrove, Alan Robock, Lifeng Luo, Christa Peters-Lidard

 

B - DATES/TIMES OF NEXT MEETINGS:

-- Next phone conference: Jan 25, 11:30 AM

- Next in-person meeting: Feb 26, Noon, at OHD in Silver Spring

 

C - N-LDAS ABSTRACTS FOR APRIL EGS MEETINGS:

-- Dennis Lettenmaier will attend spring EGS and volunteered to submit/present both an N-LDAS Project overview abstract and a VIC-focus LDAS abstract. Ken Mitchell will email Dennis a couple of past N-LDAS project overview abstracts that Dennis can draw upon. Dennis will draft a strawman overview abstract and email out for comment.

-- Alan Robock will likely submit/present an EGS abstract on the Rutgers validation of N-LDAS forcing and model output.

 

D - N-LDAS ABSTRACTS FOR MAY GAPP/GCIP CONFERENCE:

-- abstract deadline is Feb 15

-- group agreed LDAS project needs high visibility there

-- group agreed to submit around 6 distinct abstracts, titles and lead authors to be assigned during Feb 26 phone conference

-- group agreed that GAPP Program Managers should be contacted with a proposal to have a special N-LDAS session at the May meetings. Ken Mitchell propose this to Rick Lawford.

 

E - COAUTHORSHIP CONVENTION FOR N-LDAS PROJECT ABSTRACTS/PAPER

-- number and order of authors on N-LDAS project abstracts has proved problematic in the past

-- group discussed various conventions and decided upon the following N-LDAS project convention: A "core" group of 13 co-authors will always be given, listed in the following order after the two lead authors (PI/Co-I pair) that are writing/composing/presenting the given abstract or paper

 

Kenneth Mitchell, Paul Houser, Eric Wood, John Schaake, Dennis Lettenmaier, Alan Robock, Dag Lohmann, Brian Cosgrove, Justin Sheffield, Qingyun Duan, Justin Sheffield, Lifeng Luo, Wayne Higgins, Rachel Pinker, and Dan Tarpley

 

(5 core PIs listed first, followed by 5 core support Co-Is listed in same institutional order as the respective PIs, followed by the 3 PIs associated with key LDAS input forcing products.)

 

-- In N-LDAS papers/abstracts that make explicit use of OU Mesonet validation data, co-authors Ken Crawford and Jeff Basara will be listed at the end.

-- Example: if Lifeng Luo and Alan Robock are submitting an abstract on LDAS validation using OU Mesonet, then Lifeng and Alan get moved up as lead authors, followed by remainder of author list in given order, followed by Ken Crawford and Jeff Basara.

 

F - TEMPORARY 1-2 MONTH SUSPENSION OF NCEP REALTIME N-LDAS RUNS

 

-- the robotic cartridge-tape 1 Tb mass storage device on the SGI-108 LDAS platform at NCEP is starting to fail, owing to very heavy use and mechanical breakdown from wear and tear. The installation of the replacement system (PC-Linux-driven 1-Tb RAIDZONE cluster, whose hardware has been delivered to NCEP) awaits completion of prior higher priority tasks by the NCEP IT Systems/Network support group. While NCEP awaits this latter installation, and given the competing computer processing demands of the concurrent executions of the N-LDAS 3-year retrospective runs and reruns (Oct 96 - Sep 99), NCEP must temporarily suspend the realtime execution of the four LDAS land models. During this period, NCEP will continue to generate and archive the realtime N-LDAS forcing. When NCEP resumes realtime execution of the four N-LDAS land models, the group agreed that NCEP will 1) run realtime scripts that execute the four N-LDAS land models (NOAH, VIC, MOSAIC, SAC) in a unified NCEP-controlled sequential order, kicked off each day by the successful daily production and quality checking by NCEP of the realtime N-LDAS forcing, and followed immediately by the daily update of the streamflow from all 4 models and daily water and energy budget computations, and 2) will cycle forward the catch-up of these realtime runs from the 4-model restart files/states valid at the end of the 3-year Oct 96 to Sep 99 retrospective forcing.

 

G - 01 FEB 02 SWITCH TO PRIZM-BASED PRECIPITATION ANALYSIS FOR N-LDAS

 

-- At various of our N-LDAS meetings over the past year, and most recently at the Fall 2001 AGU meetings, John Schaake has presented examples of his tests with Wayne Higgins and Wei Shi of his proposed precipitation analysis using PRIZM-based topography adjustments and least-squares distance weighting, as a replacement to the present Cressman-scheme based analysis used in Wayne Higgin's realtime daily gage-only CONUS precipitation analysis.

-- Wei Shi of Wayne Higgin's team has tested John's proposed scheme successfully in extensive retrospective and realtime pilot demonstrations and has the software ready to implement permanently in realtime

-- at the present meeting, the N-LDAS group agreed to a proposal by Ken Mitchell to direct Higgins/Shi to switch the precipitation analysis scheme that feeds the realtime N-LDAS forcing to the PRIZM-based least squares distance weighting scheme on 01 Feb 02. This scheme is also much more consistent with the scheme used by Dennis Lettenmaier team in the 50-year retrospective N-LDAS forcing

 

H - REVIEW OF NOAH MODEL PERFORMANCE IN PILPS-2E AND RHONE EXPS

-- Dag Lohmann presented results of the NOAH model performance in his NOAH submissions to the recent PILPS-2e and Rhone Experiments of GLASS. Dag attended the separate modeling intercomparison workshops of both experiments this past Fall. The PILPS-2e experiment was spearheaded by the team of Dennis Lettenmaier at U.Washington. The Rhone experiment was spearheaded by Aaron Boone and Joel Noilhan of Meteo France.

-- the NOAH model performed reasonably well in both experiments, though it shows a continued tendency, as did many of the participating land models, to melt/sublimate off the snowpack somewhat too early in the spring snowmelt season. The imminent public-use release of the next substantial upgrade of the NOAH model (Version 3.0) includes a new treatment of the melting/sublimation of snowpack, which embraces a sub-grid patchiness approach that will retard the spring melt/sublimation rate. Moreover, the NCEP NOAH developers will collaborate with Victor Koren of OHD in his recent success in NOAH snowpack simulations in SMIP experiments by his addition of a snowmelt water retention capability in the NOAH model, which allows, for example, for nighttime refreezing of daytime snowmelt water held in the snowpack.

-- Dag pointed out the noteworthy success in the Rhone experiment of the sub-grid "elevation banding" approach of the VIC model of retaining a robust representation of snowpack evolution in the Rhone mountainous terrain across a wide range of grid resolutions employed in the simulations.

 

DRY RUNS OF N-LDAS TALKS FOR THE ORLANDO AMS MEETINGS

-- Dag Lohmann presented his Powerpoint presentation of the N-LDAS talk that Ken Mitchell will present for him at the Orlando AMS meetings. Highlights of the presentation were intercomparisons of the 4-model a) snowpack depths and cover at 3 or 4 times during the cool season and b) surface versus baseflow partitioning or runoff.

-- Alan Robock and Lifeng Luo presented their Powerpoint presentation of their AMS N-LDAS talk, including comparisons of LDAS-forcing to OU Mesonet observed forcing and LDAS 3-model soil moisture versus OU Mesonet observed soil moisture.

-- Dennis Lettenmaier reviewed his AMS presentation on the forcing sources and VIC model results from the 50-year retrospective N-LDAS performed by his team.