CHRONOLOGY OF MAJOR CHANGES TO NCEP EARLY ETA/EDAS SUITE (SINCE OCT 95) The term "Early" in "Early Eta" refers to the fact that the Eta model is the earliest or first model to execute in both the 00 GMT and 12 GMT NCEP NWP model run suites. Hence the Early Eta executes before the NGM, AVN, MRF, GFDL, or RSM models. The Early Eta run suite and its associated 4DDA Data Assimilation System (EDAS) is the suite that "populates" the NCEP Eta GCIP archive maintained at NCAR. June 1993: Baseline - Eta model first enters formal operational production at NCEP - 00 GMT and 12 GMT runs out to 48 hours - 80-km/38-lyr resolution - OI analysis - no Eta-based assimilation (no EDAS) -- rather OI analysis 6-hour guess from GDAS/global model !! NOTE: The operational implementation given below for 12 Oct 95, began realtime parallel testing in spring of 1995. This parallel system was used to launch the NCEP Eta GCIP archive on 01 Apr 95. Hence the Eta GCIP archive began with the characteristics given below for 12 Oct. !! 12 Oct 95, 1200 GMT: A MAJOR IMPLEMENTATION - horizontal resolution increased to 48-km - model grid domain size is increased - EDAS is installed: the 12 hour Eta-based 4DDA assimilation system to initialize each 00 and 12 GMT Early Eta run -- a string of 4 3-hour Eta guess/analysis cycles starting at T00-12 from a GDAS analysis -- reduces spin-up in mainline 48-hour forecast run -- improves inherent resolution of initial state - explicit cloud microphysical scheme is added -- cloud/ice mixing ratio becomes an explicit state variable - SSM/I total precipitable water retrievals (sea only) are assimilated - the Eta GCIP archive begins to draw from this operational suite 09 Nov 95, 0000 GMT: - fixed glitch that was erroneously setting cloud mixing ratio to zero at beginning of mainline 48-hourforecast (this glitch negated the EDAS produced initial cloud field) !! 31 Jan 96, 1200 GMT: A MAJOR IMPLEMENTATION, IMPORTANT TO GCIP (see doc at http://nic.fb4.noaa.gov:8000/research/eta.ntim.0196.html) - the new land-surface scheme developed in the NCEP/OH GCIP Project is implemented (close predecessor scheme implemented in GDAS in Jan 95) -- two soil layers (10 cm, 190 cm): matched the GDAS --- soil moisture and soil temperature predicted --- initialized from GDAS, but an empirical adjustment employed to clip the moist bias often present in GDAS soil moisture -- land-surface skin temperature from surface energy balance -- explicit vegetation canopy -- NASA ISLSCP satellite-derived annual cycle of greenness fraction -- explicit snowpack physics -- 12 SSiB vegetation types, 8 soil texture classes (both 1-deg res) -- total evaporation is sum of three components --- direct evaporation from top soil layer --- transpiration through vegetation --- canopy evaporation of dewfall or intercepted rain/snow -- a host of new land-surface variables are added to the Eta output suite and GCIP archive -- see Secs. 3.1.1-3.1.2 of Chen et al., 1996, JGR, 101, Page 7251 - significant changes to the turbulence scheme in surface layer and PBL -- Paulson surface layer exchange coefficient formulation adopted --- Zilitinkevitch method for separate heat/wind roughness lengths -- Mellor-Yamada Level 2.5 scheme PBL above surface layer refined -- see doc in http://nic.fb4.noaa.gov:8000/research/eta.turb.txt -- see paper by Chen et al., 1997, Boundary-Layer Meteor, 85, p 391 24 Jul 96, 1200 GMT: - two changes to mitigate warm season low-level warm and dry bias noted by field forecasters since Jan 96 implementation of new land scheme -- 1) glitch fixed in empirical adjustment of GDAS moist soil bias -- 2) empirical increase of about 20 percent applied to NASA greenness fraction database to mitiagate perceived low bias !! 18 Feb 97, 1200 GMT: A MAJOR IMPLEMENTATION, important to GCIP (see doc in http://nic.fb4.noaa.gov:8000/research/eta.ntim.0796.txt) MUCH OF THIS IMPLEMENTATION TARGETED AT THE SUCCESSFUL MITIGATION OF THE LOW-LEVEL WARM/DRY BIAS NOTED IN 1996 WARM SEASON (see paper by Betts et al., 1997, MWR, 125, p 2896) - significantly reduce large surface solar insolation bias by about 50 percent by --- using eccentric (rather than circular) earth orbit --- correcting ozone climo that was inadvertently zero --- adding poorman aerosol absorption --- changing cloud/radiation interaction to work layer by layer (rather than over thick low, mid, hi cloud regimes) --- changed cloud cover fraction algorithm - changed algorithm for direct evaporation from top soil layer - replaced NASA vegetation greenness annual cycle database with a much improved NESDIS database developed by G. Gutman - improved melting snow physics - reduced albedo increase in presence of snow, employ veg fraction - increased the backround level of PBL turbulent mixing - new surface form drag scheme to improve behavior of storm tracks out of the Rocky Mountains - improve numerical accuracy of water vapor and cloud water advection - new land-surface output fields were added to the GRIB fields (MORDS) - numerous GRIB (MORDS) and BUFR (MOLTS) output glitches were fixed: (see bottom of http://nic.fb4.noaa.gov:/8000/research/eta.ntim.0796.txt) 29 Jul 97, 1200 GMT: - an error in the post-processing of cloud water from Eta to pressure surfaces was corrected - minor changes to the GRIB (MORDS) and BUFR (MOLTS) postprocessing (see http://nic.fb4.noaa.gov:8000/research/FAQ-eta.html#POST11) 19 Aug 97, 1200 GMT: - fix a glitch in the computation of the master length scale in the PBL physics, affected mainly sea points, negligible over land -- this error reduced the vertical moisture transport out of the lowest model layers and tended to reduce desired heavy precip 09 Sep 97, 1200 GMT: - numerous changes to BUFR (MOLTS) station list (see details in http://nic.fb4.noaa.gov:8000/research/eta.log.html) 25 Sep 97, 1200 GMT: - GOES satellite-derived retrievals of low, middle, and high precipitable water PW (non-cloudy areas) begin ingest and use in EDAS !! 09 Feb 98, 1200 GMT: A MAJOR IMPLEMENTATION - 3-D VAR variational analysis replaces Optimum Interpolation - with new 3-D VAR, EDAS begins ingest of ACARS temperature obs, surface winds over land, VAD winds from NEXRAD, SSM/I oceanic surface winds, tropical cyclone bogus data - horizontal resolution increased from 48-km to 32-km -- postprocessed AWIP Grid 212 resolution remains at 40-km - vertical resolution increased from 38-layers to 45-layers 27 Apr 98, 1200 GMT: - use of GOES precipitable water (PW) retrievals turned off due to a NESDIS retrieval glitch that was using wrong SST fields -- this glitch led to low PWs, which was drying out the Eta initial fields, especially over Gulf and in Gulf coastal states !! 03 Jun 98, 1200 GMT: A MAJOR IMPLEMENTATION, important to GCIP - EDAS begins uninterrupted fully continuous cycling -- GDAS no longer used to start EDAS -- virtually no model spin-up remains - new NESDIS 23-km daily snow cover and ice cover analysis begins daily ingest in 00Z cycle of EDAS -- Air Force snowdepth analysis still used for snow depth