Joint OSSE experiment Conventional data Yuanfu Xie Jack

Joint OSSE experiment - Conventional data Yuanfu Xie, Jack Woollen Michiko Masutani, Yucheng Song Daryl Kleist

Conventional data OSSE test • Jack generated the conventional obs using the new EC nature run; • This OSSE is tested for the time period of May 1 -8, 2005; • T 62 L 64 is used in the experiment; • NCEP SST and ICE analysis are used, interestingly, snow does not seems necessary.

Initial results • Some temperature ingest problem was found and Daryl fixed the routine for temperature (sensible/virtual temperature); • Wind and temperature are examined; • Comparison of NR, NO (no observation), and CO (conventional observation) runs.

Data ingest by prepqc • • sst: spd: q: t: uv: pw: ps: 0; 17, 835; 64, 458; 48, 104; 477, 874; 0; 53, 574

Time sequence of RMS and mean at 500 mb (wind) red: NR-NO; green: NR-CO

Time sequence of RMS and mean at 200, 500, and 850 mb (temp)

Wind (U) comparison (700 mb) at 12 Z May 8, 2005 ECWMF nature run OSSE without obs

Wind (U) comparison (700 mb) at 12 Z May 8, 2005 ECWMF nature run OSSE with convention obs

Wind (U) differences (700 mb) at 12 Z May 8, 2005 NR-NO NR-CO

OSSE wind (U) impact: CO-NO (700 mb) at 12 Z May 8, 2005

Temperature comparison (700 mb) at 12 Z May 8, 2005 ECWMF nature run OSSE without obs

Temperature comparison (700 mb) at 12 Z May 8, 2005 ECWMF nature run OSSE with convention obs

Temperature differences (700 mb) at 12 Z May 8, 2005 NR-NO NR-CO

OSSE Temp impact: CO-NO (700 mb) at 12 Z May 8, 2005

RMS average over time (wind)

RMS average over time (temp)

Remarks • Wind has the most observations; • OSSE impact (conventional obs) on wind and temperature seems good; • Other field will be also examined; • With conventional observations generated for May 2005 -2006, this OSSE can be run through the whole year.