Assimilating Moisture Information from GPS Dropwindsondes into the NOAA

Global Forecast SystemA NOAA/Joint Hurricane Testbed Project

Jason P. Dunion1 and Sim D. Aberson1 1 NOAA/AOML/Hurricane Research Division

MotivationAssessing atmospheric moisture and predicting its affect

on TC intensity…no easy task.

…GPS dropsonde humidity was not being assimilatedinto the GFS model

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

“995. DATACARDS - IBM Jobs GFS_PREP, GDAS_PREP. (Keyser, NP22). This program PREPOBS_PREPDATA prepares observational data for subsequent quality control programs and for subsequent analysis in all forecast networks, using data card switches in the input parm cards to control processing based on the forecast network. The input parm cards for the GFS and GDAS networks, prepobs_prepdata.gfs.parm and prepobs_prepdata.gdas.parm, respectively, are being modified to no longer flag Gulf Stream dropwindsonde moisture data. These data, on all levels, will now be assimilated by the Global SSI analysis. USAF dropwindsonde moisture will continue to be flagged and not assimilated in all networks, as will Gulf Stream dropwindsonde moisture in the NAM, NDAS and RUC networks.”

•Issues with the conversion to “haze” at NCEP:  -HPSS and mirroring with other computer systems

was not available for more than a month after “haze” was installed.  These data are necessary for the parallel runs.

-The GFS was not available on haze for weeks after the above difficulties were overcome

-The NCEP cyclone tracker (and GFDL model) were only installed on “haze” within the last two weeks.

-The parallel cycle now works and runs will be completed soon.

Project Challenges

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

Barbados

•Harsh working conditions

200

300

400

500

600

700

800

900

1000

0 5 10 15 20 25 30 35

WindSpeed (kt)

Pre

ssu

re (

hP

a)

GPS Sonde

GFS

ENE

SSW

NE

NE

NE

NEDrop #24

200

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

GFS

Jordan

Drop #24

200

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

GFS

Jordan

Drop #23

200

300

400

500

600

700

800

900

1000

-40 -30 -20 -10 0 10 20 30

Temperature (C)

Pre

ssu

re (

hP

a)

GPS Sonde

GFS

Drop #23

3oC Temp Inv

G-IV SALEX Mission 060915n (15 September 2006)

200

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

GFS

Jordan

Drop #25

200

300

400

500

600

700

800

900

1000

0 5 10 15 20 25 30 35 40

Wind Speed (kt)

Pre

ssu

re (

hP

a)

GPS Sonde

GFS

SE

ESE

E

NE

ENE

NEDrop #25

200-850 mb shear:GPS Sonde: 25 kt

GFS: 5 ktSHIPS: 02 kt

200-750 mb shear: GPS Sonde: 36 kt

GFS: 7 kt

P-3/G-IV SALEX Mission 060918n18 September 2006

200

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

GFS

Jordan

Drop #16

200

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

GFS

Jordan

Drop #17

200

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

GFS

Jordan

Drop #22

200

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

GFS

Jordan

Drop #15

Hurricane Helene 16 September 2006

SAL 3SAL 2

SAL 1

AEW 1

Irene

Polar 1

551010 1515

20202525

11

200

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

200

300

400

500

600

700

800

900

1000

0 5 10 15 20 25 30

Wind Speed (kt)

Pre

ssu

re (

hP

a)

GPS Sonde

E

NE

ENE

E

NNW200

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

200

300

400

500

600

700

800

900

1000

0 5 10 15 20 25 30

Wind Speed (kt)

Pre

ssu

re (

hP

a)

GPS Sonde

E

NE

E

S

NW200

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

200

300

400

500

600

700

800

900

1000

0 5 10 15 20 25 30

Wind Speed (kt)

Pre

ssu

re (

hP

a)

GPS Sonde

ESE

ENE

ENE

S

WNW200

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

200

300

400

500

600

700

800

900

1000

0 5 10 15 20 25 30

Wind Speed (kt)

Pre

ssu

re (

hP

a)

GPS Sonde

SE

NE

NE

S

200

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

200

300

400

500

600

700

800

900

1000

0 5 10 15 20 25 30

Wind Speed (kt)

Pre

ssu

re (

hP

a)

GPS Sonde

E

NE

ENE

E

SSE

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

400 nm360

320

280

240200

160

120

80

40

665 km600

530

465

400

330

265

200

130

65

StormMotion

25N

20N

15N

10N

5N

35W50W 45W 40W55W60W65W

Saharan Air Layer

200

300

400

500

600

700

800

900

1000

0 5 10 15 20 25 30

Wind Speed (kt)

Pre

ssu

re (

hP

a)GPS Sonde

E

NE

ENE

E

NNW

Saharan Air Layer

Getting Dry Air in to the TC Circulation

200

300

400

500

600

700

800

900

1000

0 2 4 6 8 10 12 14 16 18

Mixing Ratio (g kg-1)

Pre

ssu

re (

hP

a)GPS sonde (SAL)

GPS sonde (non-SAL)

GFS (SAL)

GFS (non-SAL)

Jordan

200

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 (SAL)GPS sonde (non-SAL)GFS (SAL)GFS (non-SAL)Jordan

200

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 (SAL)

GPS sonde (non-SAL)

Jordan

GFS Analyses vs GPS Dropsondes (2006 SALEX)SAL drops: 79; non-SAL drops 27

200

300

400

500

600

700

800

900

1000

0 5 10 15 20 25 30 35 40 45 50

RH RMS Error (%)

Pre

ssu

re (

hP

a)

GFS

Jordan

SAL

GFS Analyses vs GPS Dropsondes (2006 SALEX)SAL drops: 79

Targeted Humidity Observations

Summary of Findings

•As of August 2006, G-IV GPS dropsonde humidity data is being assimilated into the GFS model (not from P-3s or C-130s)

•GFS analysis fields appear to overestimate the SAL’s mid-level moisture and underestimate its mid-level easterly jet

•Targeting humidity observations (GPS dropsondes) shows promise

Future Work

•Continue parallel runs (w/o GPS sondes) for 2006 G-IV TC cases

•Continue assessing targeted observing strategies for optimizing GPS dropsonde humidity impacts on the GFS

•Assess feasibility of operationally assimilating humidity data from the NOAA P-3s and AF C-130s GPS dropsondes

•Begin similar moisture studies with higher resolution models (e.g. HWRF; GFDL)