1. Overview
Goal:
To develop TAMDAR data assimilation capability in WRFDA system;
To assess the impact of TAMDAR observations on the mesoscale numerical forecast.
Period:
1 June 2008 - 31 May 2009
People:
PI: Hans Huang
Co-PI: Yongsheng Chen, Xin Zhang
WRFDA participants: Hans Huang, Xin Zhang, Yongsheng Chen, Hongli Wang, Ruifang Li, Don Stark
2. Task
Four stages of work:
1) To develop TAMDAR data assimilation capability in WRFDA system;
2) To evaluate the impact of TAMDAR observation using 3D-Var;
3) To evaluate the impact of TAMDAR observation using 4D-Var;
4) To estabilish an operational data assimilation system to assimilate TAMDAR data.
For details please see Statement of work.
3. Progress
3.1 In 2008
June
Project meeting:
June 10 2008: meeting memo
Main work:
1) Develop the TAMDAR obervation operator in WRFVAR;
2) Work on the performance optimization and some bugs fix issue on the WRF tangent linear and adjoint models;
3) Get the first example of TAMDAR data in Litter_R format, and it is used to test the TAMDAR obervation operator
July
1) Develop the TAMDAR obervation operator;
2) Work on the performance optimization of WRF-4Dvar system
August
Project meeting:
Aug 6 2008: meeting memo
Main work:
1) Develop the TAMDAR obervation operator;
2) Work on the performance optimization of WRF-4Dvar system
September
1) Develop the TAMDAR obervation process code in OBSPROC;
2) Work on the performance optimization of WRF-4Dvar system
October
1) Work on the performance optimization of WRF-4Dvar system;
2) Obervation operator in WRFVAR, and correspongding data process code in OBSPROC are inished;
3) Move the WRFVAR and OBSPROC onto Airdat IBM Machine, compile successlly;
4) Preliminary work to show how the TAMDAR data works in WRF-VAR is done;
5) begin to prepare for a real case study
November
Project meeting:
Nov 4 2008: meeting memo
Main work:
1) Test the performance optimization of WRF-4Dvar system;
2) Results to show how the TAMDAR data works is summarized; PPT
3) Real tropical cyclone IKE case study: a) Model domain setup; b) Experiment configure; c) Data format issue ;
4) WRF-Var job scripts issue in IBM at Airdat;
5) There are some mistakes in the observation provided by Airdat, we fixed it.
December
1) Test the performance optimization of WRF-4Dvar system;
2) Real tropical cyclone IKE case study:
a) 6 hourly 3Dvar cycling experiment;
b) Cycling period is from 07/12Z to 11/00Z Sep. 2008;
c) Two data assimilation and forecast experiments are carried out. One is Control experiment (with conventional data only), and the other is TAMDAR experiment ( conventional data + TAMDAR data) ;
3) Verifications against best track, sea level pressure and maximum wind speed are carried out. PPT
3.2 In 2009
January
1) Project progress review. PPT
2) Merge TAMDAR observation operator to the latest WRF-Var code;
3) Prepare an experiment setup for 4D-Var run.
February
1) Merge TAMDAR observation operator to lastest WRF-Var and test its validation;
2) Test the computation efficiency of WRF-4Dvar system on NCAR IBM machine;
3) Begin to rework the multiple concurrent executable 4DVar system into a single executable sequential code;
Please see project progress note for details. PDF
March
1) Update the WRFNL to work with the lastest WRF-Var;
2) Test the computation efficiency of the latest WRF-4Dvar system;
3) Rework the three concurrent executables of the 4DVAR system into into a single executable code;
Please see project progress note for details. PDF
April
1) Generate a BE over IKE 3Dvar experiment domain at 54 Km resolution;
2) Carry out the 3Dvar and 4Dvar experiments for IKE case on NCAR IBM machine;
3) Continued reworking the three concurrent executables of the 4DVAR system into a single executable sequential code;
Please see project progress note for details. PDF
May
1) Carried out cycling 4Dvar experiments for Ike case on NCAR IBM machine;
2) Continued with project to create a single executable version of WRF-4Dvar using ESMF;
3) Prepareing the extend abstract for 10th WRF working shop;
Please see project progress note for details. PDF