NASAs Shortterm Prediction and Research Transition SPoRT Center

1
NASAs Short-term Prediction and Research
Transition (SPoRT) Center
Presentation at Rapid Prototyping Workshop April
19-20, 2006 Dr. Gary Jedlovec, NASA/MSFC Earth
Science Office Mission of the SPoRT Center
Apply NASA measurement systems and unique Earth
science research to improve the accuracy of
short-term (0-24 hr) weather prediction at the
regional and local scale
(http//weather.msfc.nasa.gov/sport/)
The SPoRT Center serves as one of the
research-to-operations (R2O) testbeds for the
SMDs Weather focus area
2
NASA Relevance and Partners
NASA SPoRT Center
3
(No Transcript)
4
Matching Data to Customer

• Develop a collaborative systems approach
• link data to forecast problems
• development and testing of new
• products for operations
• integrate capabilities into in NWS / WFOs
• decision support system and verify /
• validate performance
• develop / conduct training
• user feedback / interaction
• benchmark products

5
Unique NASA Data to Operations

• EOS satellite data provided in near real-time to
  6 WFOs (Huntsville since 2/2003) to address
  specific forecast issues
• convective initiation, morning minimum
  temperatures, fog and low cloud detection,
  sea/land breeze convection/temperatures,
  off-coast precipitation mapping, and coastal
  processes
• MODIS high resolution visible/infrared imagery
  derived products
• 4 times / day 15-30 minutes latency, full
  resolution
• 8-10 channels available simulating NPOESS and
  GOES-R capabilities
• TPW, LST/SST, cloud and fog products, composite
  imagery for smoke/haze detection, surface
  vegetation patterns and snow cover
• benchmarked night-time cloud detection capability
  and snow cover map
• AMSR-E products to coastal offices (Miami and
  Mobile)
• 15-20 minute latency, rain rate off-shore
  precipitation mapping
• Nowcasting
• convective initiation products for thunderstorm
  development
• flash density of total lightning (LMA) relation
  to severe weather

6
Unique Modeling / Assimilation to Operations

• Transitioned model products to WFOs
• twice daily WRF runs provide supplemental
  information to NWS at 12km with focus over
  Southeastern US 0-48hr suite of products
• Initialized with MODIS 1km composite SSTs since
  Fall 2005
• benchmarked performance in several WFOs
  positive impact of QPF as supplemental product
• transitioned WRF capabilities to SERVIR model
  run by CATHALAC
• Future transitional activities
• testing assimilation of AIRS profiles of
  temperature and humidity within ADAS/WRF impact
  of data in void regions
• conducting high resolution (2 km) simulations to
  assess impact of MODIS SST composites Project
  Columbia
• coastal circulations
• oceanic clouds and precipitation
• hurricane forecasts near real-time as part of
  MAP06

7
SPoRT What it is not!

• SPoRT provides end-to-end transition capabilities
  to move products from the research environment to
  operations. SPoRT IS
• an end-to-end transition facility
  identification of products, rapid prototyping,
  VV, training, benchmarking
• discipline / focus area specific -- narrow focus
  (short-term, regional draws on expertise)
• customer-centric with specific (identified) users
  / problems
• dynamic
• SPoRT is NOT
• a RPC clearing house (one-stop shop) for EOS data
  and products

8
Thoughts on What a RPC should be!

• Discipline specific cant do it all
• Knowledgeable of the EOS science and data
  community
• know what is out there and what can be done
• have good ties to EOS science data (IT) teams
• A resource (source) for customers
• use of EOS products
• data, tools, and formats
• training
• Flexible one size doesnt / cant fit all
• Forward looking beyond EOS to GEOSS

9
SPoRT Lessons Learned

• Work the transition / prototyping at all levels
  (political managerial)
• Establish good end user relationship
• match product / capability to end user needs
• gain buy-in from customer
• Products need to be timely and in customers DSS
• Training is important for customer usage (do not
  just throw products over the fence)
• Cant do it all ---- look for low hanging fruit
  and pick your opportunities
• Stay flexible - need different paradigms for
  different customers

10
SERVIR Ecological / Disaster Forecasting
NASA/MSFC Huntsville, Alabama
RadarSat data indicating flooded regions (in red)
from Hurricane Stan in Guatemala supports
relief efforts
Programmatic buy-in at all levels ensures
success of rapid prototyping activities
Disaster agencies integrate relief information
with GIS / flooding data
CATHALAC, City of Knowledge, Panama
11
Examples of Observation, Modeling, and
Nowcasting Products at the WFOs
12
MODIS/AMSR-E Data Access in AWIPS

• Data provided in D2D
• access like GOES
• satellite data
• correspond to WFO
• coverage areas
• at highest resolution
• Examples

• color composites

• fog product

• morning min. temp.

• TPW

• SSTs

composite SST

• rain rates

Previews available http//weather.msfc.nasa.gov/sp
ort/sport_observations.html
13
Total Lightning Impacts Decision Making.

• Has directly contributed to several correct
  severe warning decisions at HUN and BMX
• the LMA density map gives you a great overall
  view of where storms with intensifying updrafts
  are located. So it gives you a good map of where
  to concentrate attention.
• I believe the flash density rates were the
  primary factor in holding off on a warning.
• Used in Warning Event Simulator (WES) for office
  training

14
Advanced Modeling and Data Assimilation

• WRF - Operational (since 2/04)
• 12-km resolution 28 vert. levels
• 12Z 00Z cycles ? 48 h Fcsts
• Standard dynamics and physics
• Initialized with AWIP212 NCEP EDAS analyses, NAM
  forecasts for LBCs
• Uses MODIS composite SSTs

• WRF - research mode
• Down to 2km resolution (Project Columbia)
• Up to 51 vertical levels
• ADAS for assimilation of AIRS profiles
• LIS

15
Real-time MODIS Composite SST
Lower boundary layer forcing important in
mesoscale models Current real time products used
by modelers (e.g., the RTG SST analysis) lacks
sufficient spatial resolution for regional
models Single pass of MODIS provides limited
coverage/impact on forecast

• Generate a simple composite generated from
  multiple passes of MODIS data (previous few days)
  
• SST does not change significantly from
• one day to the next
• high resolution (1 km), full coverage
• small scale features in SST gradients
• retained
• Expanding coverage area for SERVER and MAP06
  applications

16
Validation of WRF forecasts (with MODIS SSTs)
GOES imagery provides validation for convective
rolls
Small but significant differences due to MODIS
SSTs produce better forecasts of surface fluxes/
convergence patterns over ocean
17
WRF Hurricane Forecasts w/SSTs
Hurricane Katrina 06 UTC August 29, 2005

• Evaluate WRF simulations of tropical system
  sensitivity to MODIS SSTs
• collaboration with GSFC
• 24 48 h forecasts at 2km resolution
• Initialized with 40 km NAM analyses
• NAM 3h forecasts used for LBCs
• parallel forecasts with either RTG SSTs
• or MODIS SST composite

42h forecast of 3h accumulated precip (in)
Initial cases indicate that WRF hurricane
intensity forecasts are sensitive to high
resolution changes in SSTs.
18
SPoRT Research WRF with AIRS Data

• WRF Configuration
• 36km domain with 37 vertical levels
• state-of-the-art dynamics and physics
• initialized with NCEP 1 GFS grids, with 6-h
  forecasts used as LBC
• ADAS Configuration
• vertical resolution to match WRF
• Bratseth SCM 3-D weights and AIRS error
  characteristics
• selective use of AIRS T,q profiles based on
  quality indicators
• Assimilation / Forecast
• GFS interpolated to WRF
• domain, 1h forecast
• WRF analysis at 07 UTC used as background for ADAS

WRF Domain for November 2005 Case Study
1h forecast
19
AIRS Data Overview

• Prototype v5.0 AIRS retrievals much improved
• Quality indicators by levels
• pressure for each sounding indicating level
• of valid data
• error estimates for each profile at each
• level for T and q
• More data are assimilated
• number of assimilated profiles reduced in
• v5.0, however
• higher data volume more data are used
• in the mid-troposphere (previously
• ambiguous QI here)
• assimilating a larger volume of higher
• quality data produces an analysis that
• provides better initial conditions

4.0 - black 5.0 - red
20
29h Forecasts - Validation against RAOBs
Cumulative Precipitation

• Inclusion of v5.0 AIRS data
• reduces bias in T and q forecasts at most levels
• impact of data is greater more data, greater
  coverage
• spatial distribution of precipitation is improved

AIRS v5.0 No AIRS Stage IV
Prudent use of quality indicators with AIRS
profiles can have a positive impact on short-term
weather forecasts