1

Impact of Infrared, Microwave and Radio Occultation Satellite

Observations on Operational Numerical Weather Prediction

Lidia Cucurull (1) and Richard A. Anthes(2)

(1 ) Office of the Director, Global Systems DivisionEarth System Research Laboratory (ESRL) / CIRESNOAA Office of Oceanic and Atmospheric Research

(2) University Corporation for Atmospheric Research

12th JCDSA Workshop, 21-23 May 2014, College Park, MD..

2

JCSDA-ESRL related activities Close collaboration with NCEP

– Lead Radio Occultation (RO) assimilation activities with GSI – Algorithm development done in collaboration with Jim Purser– Currently working on improving the use of bending angles in the lower

troposphere under large vertical gradients of refractivity OSE/OSSE Sandy Supplemental Project

– New Nature Run, ESRL leading the RO Component NOAA SHOUT Program (UAS Global Hawk) Collaborative activities between ESRL and UCAR to investigate the

complementarity between RO and satellite radiances (microwave and infrared).– RO improves the bias correction of the satellite radiances– RO is a valuable complement to NPP and JPSS– A balanced system of satellite radiances and RO is most efficient and cost

effective NOAA RO Chief Scientist for COSMIC-2 and Jason-CS Programs

3

Motivation

RO limb soundings and passive MW & IR nadir-viewing observations are together the most effective observational systems in reducing forecast error

The limb-viewing and nadir-viewing systems are highly complementary The assimilation of satellite radiances in operational weather forecasting

benefits from the assimilation of unbiased observations (i.e. RO) that reduce the drift of a weather model towards its own climatology

The goal of the study is to investigate the differences and similarities between the assimilation of RO, MW, and IR observations in the NCEP’s global data assimilation system (GSI/GFS)

Results of the study are under current review (Cucurull and Anthes 2014a, MWR)

A follow-up study to analyze the impact of loss of MW and RO observations in operational NWP has been conducted in support of the US Data Gap Mitigation Activities (Cucurull and Anthes 2014b, WAF, in preparation) – a couple of main results will be shown here

4

SYNOPAIREPDRIBUTEMPDROPPILOT

GOES-AMVMeteosat-AMV

MODIS-AMVSCATHIRS

AMSU-AAIRSIASI

GPS-ROAMSR-E

SSMISTMI-1MERIS

MHSAMSU-B

Meteosat-RadMTSAT-Rad

GOES-RadO3

0 5 10 15 20 25

Contribution in Forecast Error Reduction (%)

ECMWF System (June 2011)

RO bending angles ~2-3% of assimilated data

Radio Occultation ranks very high

Cardinally et al. 2011

C

Radio Occultation Concept

Raw measurement is traceable to atomic clock time standard

5Courtesy of UCAR

6

Experiment Design

Six experiments– CTL, operational configuration with all the observations– BASE, CTL without IR, MW and RO– IR, BASE with IR added– MW, BASE with MW added– RO, BASE with RO added

All experimental forecasts begin 00 GMT and ran for 8 days Time period: 21 February – 31 March 2013 (first seven days used for

model spin-up) NCEP’s global configuration (hybrid GSI, T574, 64 levels in the vertical) We looked at fit to radiosondes; horizontal maps of the analysis

differences & RMS differences; vertical profiles of global and temporal averages of mean differences, RMS differences and correlations; and anomaly correlation score.

7

Fit to Radiosondes

• IR, MW, CTL are cold in stratosphere

• Warm bias in the upper troposphere

• MW produces the largest coldest bias in stratosphere and the warmest bias in the troposphere

Moist troposphere

Dry stratosphere

Temperature Moisture

8

850 mb horizontal maps (mean diff)• Impacts within 0.5K in the

monthly mean• Greatest differences are in the

Southern Hemisphere• MW produces a much warmer

analysis over Arctic region, northern Africa, North Atlantic, the Southern Ocean and the western Antarctica.

• IR and RO show cooler patterns over the Southern Ocean

• All three analyses show a cooler region off the western coast of South America

• IR and MW show similar patterns – wetter over Africa, off the west coasts of North and South America, and over the tropical Atlantic; and drier over Australia and off the eastern coast of South America

• RO shows less impacts

Temperature

Moisture

9

850 mb horizontal maps (RMS diff)• Maximum impacts generally in

SH, with maxima over Antarctica, Southern Ocean and the South Pacific

• Region of very large impact off the coast of South America

• Large RMS differences over the tropical North Pacific

• Similar patterns in IR, MW and RO

• Large impact across the globe north and south of the Equator to about 40N and 40S

• Strong maxima over Equatorial Africa, the Indian Ocean, tropical North Pacific, and Caribbean and off both the east and wests coasts of South America

Temperature

Moisture

10

TemperatureGlobal mean impacts of IR and RO are similar, MW quite different

Largest variability in LT, and betweenMW and RO

In general, high correlation except in LT

11

Moisture Overall, drier analyses

Overall, small rms differences; smallest between IR and MW

High correlations except in LT

12

500 mb AC geopotential heights

13

Impact of loss of MW and RO Time period: March-April 2013 NCEP’s operational configuration Verification done against consensus analysis (average of NCEP,

ECMWF and UK Met Office analyses) Experiments:

prctl: control, operational configuration with all the observations prnogps: prctl without RO observations prnoatms: prctl without ATMS observations

A potential gap in RO is a serious problem (see next slides)

14Cucurull and Anthes 2014b, in preparation

15

Evolution ofbias at 300 mb

Fit to radiosondes

Anal-O (K)

RO reduces the bias of the NCEP’s model

Cucurull and Anthes 2014b, in preparation

Overall, the patterns of the impact of the different satellite systems are similar, with the MW observations producing the largest impact on the analyses and RO the smallest – due to the lower number of RO observations being assimilated

However, without RO observations, satellite radiances are over or under bias corrected, producing less accurate analysis and forecasts

The largest biases are found for the experiment that only assimilates MW observations

Positive correlation coefficients of temperature impacts are generally found between the different satellite observation analysis, indicating that the three satellite systems are affecting the global temperatures in a similar way

Correlation for temperature and moisture in the lower troposphere among all three systems is surprisingly small

A gap in RO could potentially be a serious problem 16

Conclusions