Im pact of AIRS Retrievals on Im pact of AIRS Retrievals on - - PowerPoint PPT Presentation

Im pact of AIRS Retrievals on Im pact of AIRS Retrievals on Forecast Skill using the Forecast Skill using the GEOS-5 DAS GEOS-5 DAS

Oreste Oreste Reale Reale, Bob Rosenberg , Eug enia , Bob Rosenberg , Eug enia Brin Brin, Joe , Joe Terry , Lars Peter Terry , Lars Peter Riishojg aard Riishojg aard, Juan Carlos , Juan Carlos Jusem Jusem , Joel Susskind, Max Suarez, Larry , Joel Susskind, Max Suarez, Larry Takacs Takacs, Ricardo , Ricardo Todling Todling

Special Thanks Special Thanks

  Michele

Michele Rienecker Rienecker and the GMAO for providing and the GMAO for providing the GEOS-5 DAS and the technical support the GEOS-5 DAS and the technical support needed to set up and run these experim ents. needed to set up and run these experim ents.

  AIRS team at JPL and the Sounder Research

AIRS team at JPL and the Sounder Research Team at NASA Goddard for providing the Team at NASA Goddard for providing the retrievals used in this experim ent. retrievals used in this experim ent.

Assim ilation Runs Assim ilation Runs

  W e generated two GEOS-5 assim ilation runs:

W e generated two GEOS-5 assim ilation runs:

  Control assim ilation is the GEOS-5 DAS Version beta7 P2 at

Control assim ilation is the GEOS-5 DAS Version beta7 P2 at 1 x1 deg ree resolution, run 1 2 /1 7 /0 2 throug h 1 /3 1 /0 3 1 x1 deg ree resolution, run 1 2 /1 7 /0 2 throug h 1 /3 1 /0 3

  AIRS assim ilation is the 1 x1 -deg . GEOS-5 DAS Version

AIRS assim ilation is the 1 x1 -deg . GEOS-5 DAS Version beta7 P2 with sam e data as control plus AIRS retrievals beta7 P2 with sam e data as control plus AIRS retrievals added as added as rawinsonde rawinsonde tem perature prof i

• les. It was run from

tem perature prof i

• les. It was run from

1 /1 /0 3 throug h 1 /3 1 /0 3 . 1 /1 /0 3 throug h 1 /3 1 /0 3 .

  Retrievals used for the AIRS assim ilation were produced by

Retrievals used for the AIRS assim ilation were produced by the Version 4 AIRS alg orithm , and thinned for input to the the Version 4 AIRS alg orithm , and thinned for input to the assim ilation by using every other data point. assim ilation by using every other data point.

Forecast Runs Forecast Runs

  W e ran two sets of 2 6 5 -day forecasts, initialized

W e ran two sets of 2 6 5 -day forecasts, initialized at 0 0 Z each day, from 1 /6 /0 3 through 1 /3 1 /0 3 : at 0 0 Z each day, from 1 /6 /0 3 through 1 /3 1 /0 3 :

  “

“Control Control” ” set initialized from the control set initialized from the control assim ilation assim ilation

  “

“AIRS AIRS” ” set initialized from the AIRS assim ilation set initialized from the AIRS assim ilation

  W e skipped f i

rst 5 days to allow for spin-up of W e skipped f i rst 5 days to allow for spin-up of the AIRS assim ilation. the AIRS assim ilation.

  W e verif i

ed both the control and AIRS W e verif i ed both the control and AIRS forecasts against the NCEP analysis forecasts against the NCEP analysis

Result Result ‒ ‒ average of 2 6 GEOS-5 AIRS forecasts vs. average of 2 6 GEOS-5 AIRS forecasts vs. 2 6 GEOS-5 Control forecasts 2 6 GEOS-5 Control forecasts

Individual Forecast Anom aly Individual Forecast Anom aly Correlations Correlations

Case Selection Case Selection

  W e then looked in the Northern Hem isphere

W e then looked in the Northern Hem isphere Extra-tropics for a link between the data at Extra-tropics for a link between the data at the initial condition and a synoptic event the initial condition and a synoptic event that was better forecast when initialized that was better forecast when initialized using the AIRS assim ilation using the AIRS assim ilation

  W e wanted to look at a case where the

W e wanted to look at a case where the control did reasonably well and AIRS showed control did reasonably well and AIRS showed notable im provem ent. The forecast notable im provem ent. The forecast initialized at Jan. 2 5 (case 2 0 ) m et this initialized at Jan. 2 5 (case 2 0 ) m et this criterion. criterion.

Averaged 3 0 N to 6 0 N Averaged 3 0 N to 6 0 N

AIRS data over N. Pacif i c used by assim ilation AIRS data over N. Pacif i c used by assim ilation

• n 0 0 GMT Jan. 2 5 , 2 0 0 3
• n 0 0 GMT Jan. 2 5 , 2 0 0 3

  (will show plot of global coverage of AIRS data

(will show plot of global coverage of AIRS data

• n January 25, 2003)
• n January 25, 2003)

Top panel shading is AIRS m inus Control Top panel shading is AIRS m inus Control Bottom panel shading is NCEP m inus control Bottom panel shading is NCEP m inus control

Top panel shading is AIRS m inus Control Top panel shading is AIRS m inus Control Bottom panel shading is NCEP m inus control Bottom panel shading is NCEP m inus control

Sum m ary Sum m ary

  The addition of AIRS tem perature retrievals in the

The addition of AIRS tem perature retrievals in the GEOS-5 DAS produced a m eaningful im provem ent in GEOS-5 DAS produced a m eaningful im provem ent in 5 -day forecast skill in the Northern Hem isphere 5 -day forecast skill in the Northern Hem isphere extra tropics, averaged over a sam ple of 2 6 extra tropics, averaged over a sam ple of 2 6 forecasts in January 2 0 0 3 . forecasts in January 2 0 0 3 .

  A selected case showed a link between a

A selected case showed a link between a perturbation in the initial height f i eld over the perturbation in the initial height f i eld over the northeastern Pacif i c Ocean by the insertion of AIRS northeastern Pacif i c Ocean by the insertion of AIRS data, and an im proved 5 -day forecast over Canada data, and an im proved 5 -day forecast over Canada and the N Atlantic. and the N Atlantic.

Future W ork Future W ork

  This was a very prelim inary study. W e plan to evaluate

This was a very prelim inary study. W e plan to evaluate all cases of m eaningful im pact, both positive and all cases of m eaningful im pact, both positive and negative, and docum ent the im pact and its dynam ic negative, and docum ent the im pact and its dynam ic response. response.

  W e plan to extend the forecasts to 1 0 days, to be

W e plan to extend the forecasts to 1 0 days, to be consistent with other NW P studies. consistent with other NW P studies.

  W e want to collaborate with the GMAO to docum ent the

W e want to collaborate with the GMAO to docum ent the im pact of using AIRS radiance assim ilation in GEOS-5 , im pact of using AIRS radiance assim ilation in GEOS-5 , and exam ine cases of signif i cant im pact. and exam ine cases of signif i cant im pact.

  W e plan to re-run these experim ents using Version 5

W e plan to re-run these experim ents using Version 5 retrievals. retrievals.