Transcript Daily Nowcasts & Forecasts of Sound Speed Profiles

Status of Navy Activities
in support of MISST
28 Nov 2005
NAVOCEANO Activity Status
NAVO Operational Use of Satellite SST
Satellite SST
10km
Analysis
Satellite
Altimetry SSH
MODAS
3D Daily
Analysis
In Situ Profile
• XBT
•Float
•CTD
NCOM or
SWAFS
Daily
72 hr
Forecast
Mission Planning:
Surface Analysis
Tactical Support:
Sound Speed Profiles
And Currents
@ each grid point
@
t0, t1, t2, …72hr
NAVO Satellite SST Data Flow
RSS AMSRE
AVHRR GAC
AVHRR LAC
Operational
Models and
Archive
GOES
Orbital DEF
AATSR
MTSAT
Files
Satellite Processing System
Global
database
Matchup
Database (SSES)
FNMOC
NESDIS
NCEP
NASA-JPL
OcnQC
FNMOC Windsat
MSG
Medspiration
NASA-JPL
L2P
10km Analysis
Fleet Support
World Wide Web
FNMOC
GODAE
Server
NAVO
Operational
Models
Public
NAVO L2P File Status
• Making N-18 GAC and LAC L2P files
since 10/2005
• Sent example files to MeteoFrance and JPL
for feedback. Received and implemented
feedback. New sample files available
• XML metadata file in progress
NAVO L2P File Status
• Contents of NAVO L2P files include….
Latitude
Solar Irradiance
SI Source
Longitude
SI time diff
Sea ice source
File time
Wind speed
AOD source
SST
WS time diff
Sat zenith angle
Time diff
Sea ice fraction
Rejection flag
Bias
Aerosol Optical Depth
Conf flag
StDev
AOD time diff
Proximity conf
Clim diff
WS source
SSES Calculation Status
• Making SSES for N-16, N-17, and N-18 AVHRR
GAC and AVHRR LAC
• Making SSES for GOES-West and GOES-East
• Obtaining AMSR-E, Windsat, MSG, and GOES-9
SST data sets from various sources. Calculating
SSES for evaluation and intercomparison.
• Obtaining AATSR SST data sets.
NAVO L2P File Status (Plans)
• Generate N-17 GAC/LAC L2P
• Partition GAC global L2P orbit files into smaller
regional data files so that XML metafile makes
more sense?
• Evaluate/intercompare satellite SSES values for
operational use
• Generate METOP AVHRR L2P files
• Generate GOES-West and GOES-East L2P files?
NRL Activity Status
Status Navy Activities in Support of MISST
Global Real-time Sea Ice / SST Analyses (IR + in situ)
– 9-12 km resolution, 6 hour update cycle
– available on GODAE server in NRT
http://usgodae2.fnmoc.navy.mil/ftp/outgoing/fnmoc/models/ghrsst
New Observing Systems Added to Ocean Data QC
– orbital AMSR-E microwave SST data from RSS
– MSG SST data from CMS
Blended MW/IR SST Analyses
– measurement error specified by reported SSES
– representation error from background SST gradients
– error consistency and validation studies performed
Navy ocean data quality
control system upgraded to
handle orbital AMSR-E MW
SST data from RSS
QC Procedures:
• background field check (with
appropriate errors)
• climate
• global/regional analysis
• water mass classification
• diurnal warming detection
Examples of sequential orbits
processed for 30 Sep 2005
• rejections (perr>0.99) in red
Global SST analysis backgrounds used in QC of sample AMSR-E orbits
9-12 km resolution, 6 hour update cycle
30 Sep 2005 18Z
AVHRR GAC
in situ
1 Oct 2005 00Z
AVHRR GAC
in situ
Global SST background
error standard deviations
(°C) valid 30 Sep 2005
used to QC AMSR-E orbits
p( x  X )  ( 2 )
1

X

e
 12 ( x   ) 2 /  2
x = observed SST
 = background SST
 = background error SD
p = probability error
reject if p>0.99, ~3 
Background Error Variances in Navy Analysis System
• vary by position, depth, and analysis variable
• evolve with time, updated continuously using analyzed increment fields
• error growth parameterization in data void areas
– function of age of data on grid and temporal autocorrelations
– errors asymptote at climate variability in long term absence of observations
dx
AMSR-E SST data very consistent with
AVHRR/in situ analysis backgrounds analysis differences ~ +/- 2°C
climate background differences ~ +/- 4°C
20 UTC
21 UTC
Frequency distributions of AMSR-E SST minus
background SST for 30 Sep 05 sample orbits
Red - global 9-12 km analysis
Green - regional 9 km north pole analysis
Blue - GDEM SST climatology
23 UTC
Locations AMSR-E Rejections
1 July - 1 October 2005
# retrievals = 512,939,677
# rejections = 1,384,503 (0.26%)
Frequency distribution rejected AMSR-E
anomalies from global analysis
Binning interval = 0.5°C
Max frequency = 339,191at -1.5°C
17,940 retrievals with anomalies > 4°C
Bias rejected observations = -1.95°C
Temperature °C
Water Mass Classification AMSR-E SST
Frequency distribution of AVHRR GAC
SST in Brazil Malvinas Current region for
month of July (2001-2004) with mixture of
normal probability distribution functions
that best describes observed frequency
distribution
p (i | x)  p ( x | i ) p (i ) / p ( x)
p ( x | i )  N ( x | i, i )
Water mass classified AMSR-E SST
retrievals for 16-17 July 2005
• clearly shows fronts and eddies
• water mass classification categories
used in analysis data thinning
procedures
• maintains SST gradients in analysis
Snapshot Blended
MW-IR-In situ SST
Analysis
Satellite SST
AVHRR/GAC AMSR-E
In Situ SST
Ship, Buoy (fixed, drifting)
27 Aug 2005 00Z
9-12 km resolution
12-hr update cycle
Analysis Increments
Analyzed SST
Blended MW/IR Analysis Verification - August 2005
12-hr update cycle, 9 km resolution North Atlantic Basin
• cold start from climatology
on 1 Aug 2005
• systematic reduction in
RMS error from climate to
innovation (previous analysis)
to residual (current analysis)
• analysis residuals unbiased
• innovation biases show
diurnal variability
Analysis Observation Error Standard Deviations
• sum of a measurement error (em) and a representation error (er)
eo = e m + e r
• measurement errors specified by SSES from data providers
- computed from buoy match-ups, varies with retrieval type and time
• representation errors function of resolution model and resolution of
observing system
- satellite retrieval error computed as function background horizontal
gradient (b) when retrieval resolution (rs) exceeds grid resolution (rg)
er = b · (rs / rg)
rs > rg
- primarily affects AMSR-E 25-km retrievals at analysis grid
resolutions of 9-12 km (AVHRR 8-km GAC retrievals not affected)
- AMSR-E errors increased 2-fold in western boundary currents
Jmin Diagnostic MW/IR Analysis - August 2005
measures consistency of specified
error variances (observation and
background) with innovation vector
expected value is one if consistent
< 1 error variances too large
> 1 error variances too small or
erroneous data assimilated
AMSR-E SSES plus gradient
representation error very consistent
Plans Navy Activities in Support of MISST
Impact Studies Blended MW/IR SST Analyses
– NOGAPS forecasts (T239, ~55 km resolution), focus on
2005 hurricane season (Jun-Oct)
– add AMSR-E MW data to real-time GHRSST analyses
Ocean Data QC
– work with RSS to understand AMSR-E rejections
– improve background error statistics in QC
– evaluate AMSR-E diurnal warming estimates
Jmin Diagnostic
– continue evaluation of gradient representation error
– feedback to data providers on inconsistencies in SSES
measurement errors