Assimilating Moisture Information from GPS Dropwindsondes into the NOAA

Global Forecast SystemA NOAA/Joint Hurricane Testbed Project

Jason P. Dunion1 and Sim D. Aberson2 1 University of Miami/RMSAS/CIMAS-NOAA/AOML/HRD

2 NOAA/AOML/HRD

MotivationAssessing atmospheric moisture and predicting its affect

on TC intensity…no easy task.

The NOAA GFS model does not currently assimilate humidity information from GPS dropsondes

Project Objectives

•Begin assessing the impact of this data on GFS forecasts of TC track and intensity

•Perform detailed assessments of the impact of GPS dropsonde humidity data on GFS TC forecasts of track & intensity

•Perform parallel runs of the GFS model that assimilate dropsonde moisture information from NOAA G-IV missions

•Assess how effectively the GFS is able to represent dry layers such as the Saharan Air Layer

Year 1:

Year 2:

•Assess the feasibility of performing targeted observations of humidity to improve GFS forecasts

2005 Atlantic Hurricane Season

**Note to self…pick an El Nino year next time**

G-IV Missions: 10 TCs, 50 missions**

Project Methodology

•In order to be assimilated into the GFS, GPS dropsondes must be transmitted in real - time from the aircraft

•Run GFS parallel runs at lower resolution (T254) for Year 1 assessments

•Compare T254 parallel runs with T254 operational runs for “apples to apples” comparisons

Project Challenges

•“Old dog new trick” syndrome: teaching a satellite guy how to run a global model

•Setting up the code at NCEP to run the parallel GPS dropsonde humidity runs (never been done)•GFS parallel run forecast failures

-recently corrected bug in the radiation code: dry profiles…>…can generate bad profile

temperatures…>…model bombs

-some runs still failing (e.g. Cindy and Dennis)

0

100

200

300

400

500

600

700

800

900

1000

12 24 36 48 60 72 84 96 108 120

Forecast Time (hours)

Tra

ck F

ore

cast

Err

or

(km

)T254 (operational)

T254 (humidity)

26# of cases 25 24 21 19 16 14 11 8 4

0

5

10

15

20

25

30

35

40

45

50

12 24 36 48 60 72 84 96 108 120

Forecast Time (hours)

Inte

nsi

ty F

ore

cast

Err

or

(kt)

T254 (operational)

T254 (humidity)

Current Storm CatalogueArlene; Emily, Irene; Katrina (12 NOAA G-IV missions)

Irene (08-09 September 2005)

Katrina (25-27 August 2005)

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60 70 80 90 100

RH (% )

Pre

ssu

re (

hP

a)

GPS Sonde

Jordan

Drop #21

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60 70 80 90 100

RH (% )

Pre

ssu

re (

hP

a)

GPS Sonde

Jordan

GFS (T254oper)

GFS (T254hum)

Drop #21

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60 70 80 90 100

RH (% )

Pre

ssu

re (

hP

a)

GPS Sonde

Jordan

GFS (T254oper)

GFS (T254hum)

Drop #10

Hurricane Rita22 Sept 2005 00 UTC

G-IV Mission 050921n

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60 70 80 90 100

RH (% )

Pre

ssu

re (

hP

a)

GPS Sonde

Jordan

Drop #10

Impact of GPS Sonde Moisture on the GFSTS Irene 08 Aug 2005 0000 UTC

Saharan Air Layer Experiment (SALEX) G-IV Mission 050807n

SAL 3SAL 2

SAL 1

AEW 1

Irene

Polar 1

551010 1515

20202525

11

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60 70 80 90 100

RH (%)

Pre

ssu

re (

hP

a)

GPS Sonde

Jordan

Drop #19

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60 70 80 90 100

RH (%)

Pre

ssu

re (

hP

a) GPS Sonde

Jordan

GFS (T254oper)

GFS (T254hum)

Drop #19

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60 70 80 90 100

RH (%)

Pre

ssu

re (

hP

a)

GPS Sonde

Jordan

Drop #2

Summary of Findings•Humidity information from G-IV GPS dropsondes can significantly affect the T254 GFS (EMC was smart no to just throw the switch)

•Early results: the GPS dropsonde humidity data doesn’t have a strong negative OR positive impact on GFS forecasts of track/intensity

• For some individual cases (e.g. Irene and Katrina), the GPS dropsonde humidity information had a significant impact on the GFS track forecast

• The GFS analysis of humidity appears to be significantly affected by the GPS dropsonde humidity. This difference appears to grow with successive forecast times.

Future Work

•Continue running the remaining 2005 cases

•Test the impact of the GFS operational vs GPS dropsonde humidity fields on SHIPS (GFDL?) intensity forecasts

•Assess whether targeted observing strategies can optimize GPS dropsonde humidity impacts on the GFS