Nearcasting Severe Convection
Using the GOES Sounder
Robert M. Aune
Scott S. Lindstrom
Center for Satellite Applications and Research (STAR) Review
09 – 11 March 2010
Image:
MODIS Land Group,
NASA GSFC
March 2000
Requirement, Science, and Benefit
Requirement/Objective
•
Mission Goal: Weather and water
– Increase lead time and accuracy for weather and water warnings and forecasts
– Improve predictability of the onset, duration, and impact of hazardous and severe
weather and water events
– Increase development, application, and transition of advanced science and
technology to operations and services
Science
• Can observations from a geostationary IR sounder be used
to predict severe weather outbreaks 1 to 6 hours in
advance, filling the gap between radar nowcasts and NWP
models?
Benefits
•
•
•
Reduce loss of life, injury and damage to the economy
Better, quicker, and more valuable weather and water information to support
improved decisions
Increased customer satisfaction with weather and water information and services
VISITview Teletraining
2
Nearcasting uses GOES Sounder Data
• The GOES Sounder includes three separate
water vapor channels
• The water vapor channels have weighting
functions that peak in different parts of the
troposphere (longer wavelengths see farther
down into the atmosphere)
• Therefore have a three-dimensional look at
atmospheric moisture
VISITview Teletraining
3
VISITview Teletraining
4
VISITview Teletraining
5
VISITview Teletraining
6
VISITview Teletraining
7
VISITview Teletraining
8
VISITview Teletraining
9
VISITview Teletraining
10
VISITview Teletraining
11
VISITview Teletraining
12
Premise: Sounder gives information on distinct layers in atmosphere at observation time
Winds from a numerical model can move those slabs of moisture around
Question: Where does Convective Instability develop because of the moving slabs?
Very Dry Layer
Somewhat Moist Layer
Very Moist Layer
VISITview Teletraining
13
How is nearcasting done?
fcst time increasing
Data
1
obs time increasing
Data
Start at an initial time.
Use a Lagrangian model.
Step forward 6 hours.
Output hourly forecasts
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
Data
Data
Data
Data include winds
and sounder observations
of qe and qe that has
moved to a point at time=0
and geopotential heights at
t=0, 3 and 6h
Data
VISITview Teletraining
6
14
Nearcasting Severe Convection
Using the GOES Sounder
• Research description
– The GOES sounder can provide hourly snapshots of layer-averaged stability parameters.
These observations can be assimilated at multiple levels using a simple approach to
provide fast, short-term projections of atmospheric stability.
• Recent science accomplishments (~FY08 to present)
– In collaboration with CIMSS, a Lagrangian approach was selected that moves the GOES
observations along forward trajectories. Observation error growth remains small to 4 hours
and beyond.
– GOES sounder retrieved parameters such as equivalent potential temperature (Theta-E) at
750hPa and 500hPa are projected forward 6 hours. Destabilization is indicated when
Theta-E500 (5800m) minus Theta-E750 (2500m) becomes negative.
– The nearcasting model has been tested in real time at CIMSS using the GOES-12 sounder.
Products are displayed on the internet (http://cimss.ssec.wisc.edu/model/nrc).
– Hourly nearcasts are currently being transmitted to NWS Central Region AWIPS for
evaluation.
– Product will be evaluated at the NWS Storm Prediction Center’s Spring Experiment in May
2010.
VISITview Teletraining
15
Filling the Guidance Gap and
Atmospheric Stability Basics
The GOES sounder can detect water vapor at 2-3 layers
in a clear atmosphere. Gradients of water vapor can be
tracked using multiple GOES sounder scans. Upper
level drying over lower level moistening conditions lead
to autoconvection.
The Guidance Gap
Very-short-range NWP
precipitation forecasts often
either:
When the layer is lifted
the inversion bottom cools
less than top and it
becomes absolutely
unstable
3hr Model Forecast
Valid 1700 UTC 7/2/08
1) miss significant moisture
features
2) have difficulty with exact
position and timing of
events / phenomena
If moisture is
present in the
stable layer and
the entire layer is
lifted, it can
become unstable.
2hr Model Forecast
Valid 1700 UTC 7/2/08
Fill the Gap
Between Nowcasting & NWP
0
2
4
10
12 hours
Verification
7/2/08 1700 UTC Radar
To detect the development of areas becoming
convectively unstable, we need to monitor not only
the increase of low level moisture, but areas where
low-level moistening and upper-level drying overlap
VISITview Teletraining
16
Lagrangian Nearcasting Approach
GOES-12 900-700 hPa precipitable water analysis valid
21 UTC 13 April 2006
GOES-12 900-700 hPa precipitable water retrievals valid
00 UTC 14 April 2006
Dry
0-hour Nearcast
3-hour Nearcast
Dry
O
O
Moist
Moist
Starting location
New location
How it works
NWP models use randomly spaced moisture
observations interpolated on to a fixed grid, and use
gridded wind data to advect the moisture information
forward in time at fixed grid points. This process
smooths horizontal gradients.
The Lagrangian approach interpolates wind data to
each observation location (~10km spacing) which is
then projected forward to a new location forced by
dynamically changing wind forecasts. A relatively long
time step (10 min) can be used. The new data
locations are then transferred back to a regular grid.
VISITview Teletraining
17
Nearcasts of Severe Weather
6-hour NearCast for 2100 UTC
Low level Theta-E
6-hour NearCast for 2100 UTC
Mid - Low level Theta-E Differences
Low-level Theta-E
nearcasts shows
warm moist air band
moving into far NW
Iowa by 2100 UTC.
Oklahoma City tornado
Vertical Theta-E
Differences predict
complete convective
instability by 2100
UTC.
De-stabilization predicted
by nearcast
Radar indicates rapid
development of
convection over NW
Iowa between
2000 and 2100 UTC, 9
July 2009. Correct
shape is indicated.
4-hour nearcast of precipitable water lapse rate (mm
differences) near Oklahoma City valid 22UTC Feb 10,
2009. De-stabilization potential is indicated.
VISITview Teletraining
18
Evaluation by NWS Forecast Office,
Sullivan, Wisconsin
NWS Milwaukee is evaluating the CIMSS
precipitable water nearcasting product. The
example below shows a good relationship
between amount of convective clouds (or lack
of them) and strong vertical gradients of PW.
1-hour nearcast of
vertical precipitable
water differences (mm)
valid 19 UTC, July 2,
2009.
Visible image for
19 UTC, July 2, 2009.
Red indicates where
convection is likely.
Black indicates areas
where convection is
not likely.
CIMSS nearcast products are currently
being inserted into the operational AWIPS
data stream for NWS evaluation.
CIMSS precipitable
water lapse rate
product valid
15UTC April 13,
2009.
Same CIMSS
nearcast product
displayed on an
AWIPS workstation.
VISITview Teletraining
19
Challenges and Path Forward
• Continuing science challenges
– Current GOES sounders cannot see low-level temperature inversions
and dryness which leads to false alarms
– Accurate mesoscale wind information is needed to initialize trajectories
• Next steps
– Assimilate additional products from the GOES sounder to determine
which is the best indicator of de-stabilization
– Perform Observing System Simulation Experiments (OSSEs) to
determine the impact of using a hyperspectral sounder
• Path into applications/operations
– Evaluation at NWS forecast office has commenced (AWIPS, webpage)
– Evaluation at Storm Prediction Center Severe Weather Test Bed to
commence May, 2010 (supported by GOES-R Risk Reduction)
VISITview Teletraining
20
Nearcasting Severe Weather using a Hyperspectral
Environmental Sounder (HES)
Weak gradients of low-level Theta-E
are indicated by ABI which has only
two water vapor channels.
Simulated ABI
A WRF model simulation of the June 12, 2002
IHOP case was used to generate simulated
radiances from an Advanced Baseline Imager
(ABI), a geostationary Hyper-spectral
Environmental Sounder (HES), and simulated
radar reflectivity.
Strong low-level Theta-E gradients
are indicated by HES which has the
ability to detect low-level moisture.
Simulated HES
Temperature and moisture profiles were
retrieved from the radiance datasets and
assimilated by the CIMSS Nearcasting Model
and compared. Detailed Theta-E gradients
were resolved by HES.
5-hour NearCast for 2000 UTC
Low level Theta-E
Simulated ABI
5-hour NearCast for 2000 UTC
Low to Mid level Theta-E Differences
Simulated composite reflectivity from
nature run indication the formation of
convection.
Rapid Development of Convection over Texas and
Nebraska between 2000 and 2100 UTC 12 June 2002
VISITview Teletraining
5-hour NearCast for 2000 UTC
Low level Theta-E
Simulated HES
5-hour NearCast for 2000 UTC
Low to Mid level Theta-E Differences
21