5 05 DA Liu TAMDAR Presentation

Document Sample
5 05 DA Liu TAMDAR Presentation Powered By Docstoc
					Mesoscale Data Assimilation and Prediction with
  Commercial Aircraft (TAMDAR) Observations
                               Yubao Liu
      Collaborators: Scott Swerdlin, Mark Anderson1, Tom Warner
       Laurie Carson, Ming Ge, Wei Yu and Francois Vendenberghe
         (NCAR/Research Applications Lab, USA; 1AirDat llc.)


        TAMDAR and GLFE
        The NCAR/ATEC RTFDDA
        Data-denial experiment results
        Comparison with NAM and RUC
        Summary

       BACIMO 2005                         yliu@ucar.edu
                  TAMDAR
Tropospheric Airborne Meteorological DAta Reporting


                             Temperature
                             Moisture (two RHs)
                             Winds
                             Pressure
                             GPS height
                             Icing
                             Turbulence
                             Others
           Motivations for TAMDAR

  Upper-air data are sparse and limited
     Radiosondes: only observe twice daily
     Satellite winds: single layer, clustered
     ACARS: mostly at upper-troposphere
     Wind profilers: low-spatial density
     Indirect satellite and radar obs: exploratory
     Insufficient lower-level moisture observation
TAMDAR is to fill the upper data gaps with
 high-frequency and high-density lower
 level soundings and plane observations
    GLFE – Great Lakes Field Experiments



Mesaba SAAB 340:       15 Jan -15 Jul 2005

63 equipped
~400 flights a day
between 75 airports
~20k obs a day
Masaba Airline Routes Map
The NCAR/ATEC RTFDDA System
•   PSU/NCAR MM5 / WRF based,
•   Multi-scale: meso-g  meso-a (dx = 0.5 – 45 km),
•   Rapid-cycling: flexibly at intervals of 1 – 12 hours,
•   FDDA: 4-D continuous data assimilation,
•   Forecast ( 0 – 48 hours), and
•   Real-time, retrospective and relocatable.
Main objective: To produce best-possible real-
time local-scale analyses and nowcasts/forecasts
by effectively combining a full-physics mesoscale
model with all available observations
Related BACIMO presentations
 First introduced on: BACIMO-2001
 Enhancements and applications: BACIMO-2003



 WRF-transition: 2.08 (Knieval)
 Feature-based verification: P2.04 (Rife)

 Athens Olympics: 5.06 (Hahmann)

 Probabilistic forecasts: 5.08 (Hacker)

 Application modeling: 5.09 (Sharman)

 4DWX-on-MOVE (GMOD): 5.10 (Betancourt)

 Global Climate Analysis Tool: P5.10 (Vendenberghe)
4-D Continuous Data Assimilation and Forecast



 All WMO/GTS       GOES          Radars
                                            New 12 - 48 h forecast every 3 hrs,
                                                using all obs up to “now”

                  RTFDDA
 Wind Prof
               Regional-scale                                                     t
                                                         FDDA
               model, based on      Cold
                                    start
                PSU/NCAR
                 MM5 /WRF
                                                        Forecast

MESONETs



                Etc.

   ACARS                   TAMDAR
RTDDA: Advantages of Continuous Relaxation


     Cold                                              t
                              FDDA
     start

                               Forecasts

 Allows to use all synoptic and asynoptic observations. In
  particular, it allows to fully weight time-space irregularly
  distributed observations, such as TAMDAR data,
  according to the observation time, location and quality;
 Mitigates dynamics and cloud/precipitation “spin-up”
  problem that exists in all cold-start operational models.
 Both properties are critical for mesoscale analyses and
   short-term (0-12 hour) forecasts.
D1                    DX=36km    Two real-time
                                 RTFDDA systems

                                 AIRDAT:
     D2        DX=12km            with TAMDAR

                                 AIRNOT:
                                  w/o TAMDAR



          D3




               D3: 400x400 km2
          RTFDDA Observations: A Snapshot
Sndgs                                             Sat
Prof                         AirDat               Aircraf
                                                  t




850 hPa                                           > 600 hPa




                                      SFC

                             00Z
                             June
Sat                          24                   Sat
Aircraf      600 - 350 hPa   2005     < 350 hPa   Aircraf
t                                                 t
RTFDDA NO-TAMDAR
04Z, 20041111, 3h-fcst
                                                    Frontal rainband (2)
                                                    Nov. 11, 2004
                                                    Model 3-h forecasts

                  RTFDDA with TAMDAR
                  04Z, 20041111, 3h-fcst




                                   WSR-88D Reflectivity
                                   0419Z, 20041111
Without TAMDAR                           Weak snowbands (1)
                                         00Z, Dec. 09, 2004
                                         Radar reflectivity


                 With TAMDAR




RTFDDA                         WSR-88D
1h forecasts
Without TAMDAR   Snowbands
                 18Z, Feb. 02, 2005
                 Radar reflectivity
                 RTFDDA Analyses




  With TAMDAR                 WSR-88D
                         Daily Evolution of
                          Forecast errors
T    DIR                    At 850 hPa
                             Between
                           Jan. 29, 2005
                                And
                           Feb. 20, 2005


Q    SPD



           With TAMDAR
           W/O TAMDAR
RH
             Temperature

12h fcst
6h fcst                   Domain 2
  Analysis
                       Jan. 28 – Feb. 6
                             2005

             Water Vapor
             Mixing Ratio




             Vector Wind
             Difference

              Key:
              W/O TAMDAR
              WITH TAMDAR
Rainbands:15Z, March 12, 2005,1-h accu. rain (mm)




RTFDDA
4h Forecast                  Stage IV




RUC-13km
3h forecast                  WSR-88D
                    Rainbands
                    21Z, March 12, 2005
                    3-h accu. rain (mm)

                RTFDDA presents better
RTFDDA
10h Forecasts
                Distribution and structures




NAM-218
9h forecast     Stage II
Comparison of 3 hr Rain, valid at 20050815, 00Z
                          7h fcst




                StageIV                RTFDDA


6h fcst                   6h fcst




                RUC13                  NAM218
               Summary
• TAMDAR data are evaluated with the NCAR
  RTFDDA system through data-denial experiments
  for mesoscale analyses and forecasts.
• A general positive impact of TAMDAR was found
  in the analyses and forecasts of upper-air
  variables and surface precipitation.
• RTFDDA analyses and 1 – 12h forecasts of
  surface precipitation with TAMDAR data
  outperform NOAA RUC and NAM running at
  similar resolutions.
• The benefit of TAMDAR appears to fluctuate
  significantly with weather situations.
• On-going work:
   study TAMDAR impact on cloud-scale models;
   optimize TAMDAR data assimilation algorithms;
   and provide modeling guidance for the next-
  phase, broad-scale TAMDAR implementation.

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:5
posted:4/25/2012
language:
pages:20