2004 LOM Workshop Wednesday 1:10 - 1:30 p.m.
Performance of mixed layer models in HYCOM on the onset of the 1998 La Nina
BIROL KARA and ALAN WALLCRAFT
NAVAL RESEARCH LAB.
birol@nrlssc.navy.mil
ABSTRACT
The three K-profile vertical mixing models in HYCOM (KPP, Mellor-Yamada level 2.5, and GISS) all perform well enough that it is difficult to choose between them in a typical year or on typical individual events. The onset of the 1998 La Nina is one of the largest short term events (7 degC change is SST over 30 days) on record, and it is certainly the best observed with five equatorial buoys between 180W and 125W and several more at 2S or 2N. It has also proved difficult to replicate with ocean models, and is therefore an excellent test case for HYCOM and its mixed layer submodels. Our standard 26-layer 0.72 degree fully global (Arctic bipolar patch) HYCOM configuration, with KPP, was first spun up from Navy GDEM3 climatology for five years using ECMWF Reanalysis climatological forcing and then extended 1979-2001 using 6-hourly ECMWF forcing. No data assimilation or temperature relaxation was used, but SSS was relaxed to monthly GDEM3 values as a correction to the standard E-P and riverine surface salinity forcing. The June 1995 to December 1998 portion of this simulation was then repeated using Mellor-Yamada and GISS mixed layer submodels to study the El Nino/La Nina cycle. Note that the latitudinal grid size of this model is halved to 0.36 degrees near the equator to better resolve equatorial dynamics. We present the results of comparisons between the three twin simulations and observations for 1996-1998, and in particular for the 1998 La Nina onset in the central equatorial Pacific. The model--data comparisons were also extended to cover the time period from 1996 to 2001. Comparisons with 166 year--long daily SST buoy time series showed that while and KPP and GISS give median median RMS differences of 0.70 degC and 0.76 degC, respectively, MY2.5 gives median RMS difference of 0.85 degC during 1996-2001. TOPICS: e. mixed layer dynamics and mixed layer/layered model interactions h. model/model or model/data comparisons
LOM Users' Workshop, February 9-11, 2004