Robert Dumais

1
HIGH RESOLUTION METEOROLOGICAL NOWCASTING TO
MEET U.S. ARMY FUTURE FORCES REQUIREMENTS
Robert Dumais Lou Luces Yansen Wang
Teizi Henmi James Cogan U.S. Army
Research Laboratory Computational
and Information Systems Directorate
Boundary Layer Meteorology Branch
AHPCRC Workshop on Mesoscale and CFD Modeling
for Military Applications Jackson
State University

May 25, 2004
SLIDES CLEARED FOR PUBLIC RELEASE ON MAY 10, 2004
2
WEATHER RUNNING ESTIMATE (WRE) FOR FUTURE FORCES

• The frequent updating of local environmental
  data cubes for use by Army Air Force
  tactical weather systems (IMETS/JET, DCGS-A) and
  decision aids, to improve tactical execution
  strategy and decision making
• A dynamic ability to adjust local environmental
  data cubes to the real and evolving weather

SHADOW TUAV
Centralized Nowcast Use the best models Use a
large database of data Give information through
reachback Local Nowcast Use locally running
models Use locally available data Respond quickly
UAV, UGV, UCV

• Take full advantage of FF/FCS battlefield met
  sensors and obs

3
ARMY WEATHER ROLE

• Fuse higher echelon forecasts with
• local sensor obs
• remote sensing (METSAT)
• Support Brigade ops, UAV, RSTA weather over
  complex terrain and MOUT

• Improve first-in and daily mission support
• - 0-3 hr local nowcasts for WRE Army
• - 0-48 hr regional forecasts (MM5) Air
  Force
• - 3-5 day theater forecasts (NOGAPS/GFS)
• Navy/NCEP

4
Tactical Army Nowcast System

• Design being developed is based on successful
  strategies implemented by other short range
  forecast and analysis systems
• - Cram et al. (2001) , Seaman et al.
  (2002), Zeller et al. (2003), Strahl et al.
  (2003), Daley and Barker (2000), Barker et al.,
  (2003),
• Lazurus et al. (2002), Xue et al. (2003), Case
  et al. (2002), Moninger et
• al. (2003)
• 4DDA chosen over 3DVAR for NWP model
  assimilation (currently MM5), due to anticipation
  of highly asynoptic and sporadic observations,
  and to a lack of well developed and tested
  constraints for 3DVAR systems at mesogamma scales

5
TACTICAL ARMY NOWCAST STRATEGY MULTICOMPONENT
Terrain, land use, morphology
BRIGADE DOMAIN
Metsat EDRs (future)
150 x 150 km mesogamma NWP forecast (with data
assimilation) every hour
Profilers (future)
AFWA MM5 fields
Local surface observations
Lidar (future)
Terrestrial via remote sensing
PLATOON DOMAIN
Multiple platoon domains are possible within
Brigade domain
Radiosondes (with drift)
20 x 20 km objective analysis 3DWF wind
model (every 15 min)
UAVs (future)
Info used to create Weather Running Estimate
NOWCAST CYCLING
Inputs going to objective analysis module
Inputs going to NWP forecast module
6
MM5 with Observation Nudging 4DDA
Hourly Cycling for Brigade Domain (150 x 150 km)

• -Non-hydrostatic full physics
• Target 3 h dynamic assimilation and 3 h forward
  integration period
• Investigate NRL satellite algorithm for enhanced
  moisture fields (for OA backgrounds) using cloud
  imagery
• -Widely used in community and familiar to
  ARL/AFWA
• -Operational model at AFWA and will be used in
  Artillery MMS-P system
• -Has a built in observational nudging/4DDA system
  (and a 3DVAR system for future investigation)
• - Much going into WRF leverages off MM5
• -Serial version, single nest (2.5 km) installed
  on Linux at WSMR

MM5 2.5 km resolution forecast showing early
morning surface drainage winds over So. California
Brigade MM5 will provide high quality background
fields for platoon analyses
7
MM5 with Observation Nudging 4DDA (2004/early
2005 goals)
Begin using Utah surface mesonet datasets to read
into MM5 (winds only) for observation nudging,
adapting Fortran 90 code provided by Texas Tech
University (Dr. C-Bo Chang) as a guideline for
creating MMOBS_DOMAIN file Begin developing MM5
observational nudging of radiosondes and
PIREP/ACAR observations, also accounting for
radiosonde drift Seeking collaboration with
Texas Tech University (Dr. Chia-Bo Chang), and
using code provided by Texas AM University
(Dr. John Neilson Gammon) and NCEP (Dennis
Keyser) as basic guidelines Begin ingesting
Seaspace-generated GRIB files of high resolution
NOAA-16 seven-day composite land surface
temperature fields into MM5 PREGRID module, and
continue investigating snow cover satellite
products Assistance from New Mexico State
Universitys Physical Science Lab Determine QC
strategies for upper air and surface data
8
Objective Analysis to Fuse Local
Observations to MM5 Background
15 Minute Cycling for Smaller Platoon Sub-Domains
WSMR Mesonet observations
Successive Corrections Approach (Sashegyi and
Madala, 1994)
Univariate (T, TD, U V Winds, Sigma-Z Terrain
Following)
Same or finer horizontal resolution than local MM5
Can make better use of surface T observations
9
Objective Analysis to Fuse Local Observations to
Brigade MM5 Background for Platoon Domains (20km
x 20km) every 15 minutes
Surface objective analysis is now combined with
upper air objective analysis using similarity
theory
Mesogamma resolution MM5 1st guess field
Mesogamma wind field after objective analysis
10
Early Nowcast Proof of Concept Study
Data Set - Three months of Oklahoma Mesonet
Observations 22 Jan- 28 Feb 2000 (38 days) 1
May - 29 Jun 2000 (60 days) Applies simple data
fusion of actual surface sensor observations with
the standard 0-24 hourly outputs from BFM, the
current tactical Battlescale Forecast Model used
in the field on the AN/TMQ-40C Integrated
Meteorological System (IMETS) Reports the
absolute forecast error averaged over the other
96 Oklahoma surface stations at each hour of the
day (0-24) and then averaged over the 38 and 60
day study windows Clearly demonstrates the
progressive improvements to BFM by fusing 1, 3, 5
and 12 on-scene, real-time surface observations
from a few surface observations scattered across
the Oklahoma Mesonet to produce a prototype 0-hr
nowcast for each hour of the day

Reduction in Absolute Error
Using 1 Observation 3
Observations 12 Observations
Temperature 25 - 30 40 - 45 60
- 65 Wind Speed 15
- 25 25 - 35 30 - 45 Wind
Direction 10 - 20 20 - 30
35 - 45 Relative Humidity
0 - 25 20 - 35 40 - 60 Additional
improvements are expected when raw observations
are assimilated directly into the mesoscale
numerical model (as opposed to just the data
fusion post-processing used here)
11
Possible Future Improvements to Fast Objective
Analysis for Platoon

• 3DVAR (beginning investigations in-house with
  MM5 3DVAR, along with at AHPCRC UPOS programs)
• Response Filter being developed by AHPCRC/
  University of North Dakota (Dr. Mark Askelson)

Example Coherent input field looks incoherent
after analysis (upper right) owing to impact of
irregular data distribution on the response
function. Want to improve technique so that
result is coherent (lower right).
Figures from Dr. Mark Askelson, University of
North Dakota
12

Microscale Diagnostic Three-Dimensional Wind
Field Model (3DWF) Platoon Domains

• Given a limited number of observations or
  coarsely modeled wind field in a complex terrain,
  the wind field is physically interpolated in such
  way that the mass conservation is satisfied
  (Sasaki, 1970). Mathematically, to Minimize the
  following functional
• The equation can be cast into a Poisson equation
  for the Lagrangian multiplier
• 
  
• 
  ?(?1/?2)
  and ß11
• with B.C. ?0 on lateral and top ? ?/? n 0 for
  terrain surface
• The wind field is then adjusted as follows
• . A multigrid method to speed up solution, and a
  Cartesian coordinate system.

13
Vertical wind (z9m) from 3DWF simulation and
selected model wind profiles (JUT 03)
(T is the 90 m tower location)
Selected wind profiles at points shown in left
panel
T
4
2
5
3
2
1
1
4
3
5
14

3DWF Run Compared to JUT03 Data
15
Current and Future 3DWF Work and Other
Considerations

• Uses a Cartesian coordinate system to achieve
  better vertical resolution near the surface
• Has used sonic anemometer, tower, and Lidar VAD
  winds as input from JUT03
• Applies a multigrid numerical method (Briggs et
  al., 2000) for speed-up in solving the elliptical
  equation and for improved real time simulation
  (20 to 30 times faster than traditional method)
• The mass consistent model only attempts to get a
  first order approximation of the mean flow, and
  is frequently used in situations of limited
  computing power.
• The shortcoming to a mass-consistent approach is
  that there is no turbulent parameterization, no
  building wakes, and no effects of vegetation
  canopy
• - However, we are in the
  process of putting these things into the 3DWF
  model by using observed data or large-eddy
  simulation data as bases for parameterization.
•  In very calm wind conditions the mean flow is
  not well defined, so this model is not as useful
  since turbulence becomes a dominant factor.
• This model does not deal with stability although
  some people try using an alpha term to fudge
  this effect
• -The stability can only be
  handled properly when coupled with momentum and
  energy eqautions
• - The best hope is that we
  initialize with several wind profiles which
  already have stability information info contained
  within (e.g. well mixed mixed layer in convective
  condition, very stratified profile in stable
  condition),
• - In this situation,
  the mass consistent interpolation will include
  some of the stability information.
• Merge higher resolution 3DWF wind fields ( 100
  m) with objective analysis wind fields in
  production of final WRE cube

16
Postprocessing to complete WRE Cube
Extend basic model prognostic parameter set using
diagnostic methods / identify weather hazards
Cross-Sections
17
Final Merge to Complete WRE Data Cube
Most recent operational AFWA 15 km MM5 fields
Locally Run MM5 at 2.5 km resolution for Brigade
(every hour)
obs
0h 1h 2h 3h
WRE DATA (extrapolate/fuse) static analysis
forward
Example of a conceptual 90-minute WRE cube,
updated every 15 minutes
Static Objective analysis 3DWF for Platoon
domains (every 15 minutes)
-Creating the final WRE cube will require smart
interpolation/fusion time extrapolation of
fields produced from different nowcast modeling
components (MM5, objective analysis, 3DWF,
postprocessing algorithm, most recent WRE ),
spatial resolutions coordinate systems
18

Nowcast References of Interest
Askelson, M., and H. Lin, 2003 The Response
Filter An Approach for Adapting to Irregular
Data Distributions, AHPCRC, Presented to US Army
Research Laboratory, WSMR, NM, 2 Oct 2003.
Barker, D., W. Huang, Y.R. Gao, and A.
Bourgeois, 2003 A three-dimensional variational
(3DVAR) data assimilation system for use with
MM5, NCAR, Mesoscale and Microscale Division,
TN-453STR, Feb 2003, 71 pp. Briggs, W.L., V.E.
Henson, and S.F. McCormick, 2000 A Multigrid
Tutorial SIAM Publishing. 193 pp. Case,
Jonathan L., Manobianco, John, Oram, Timothy D.,
Garner, Tim, Blottman, Peter F., Spratt, Scott M.
2002 Local Data Integration over East-Central
Florida Using the ARPS Data Analysis System.
Weather and Forecasting Vol. 17, No. 1, pp.
326. Cram, J.M., Y. Liu, S. Lo-Nam, R.S. Sheu,
L. Carson, C.A. Davis, T. Warner, and J.F.
Bowers, 2001 An operational mesoscale RT-FDDA
analysis and forecasting system, Preprints of the
18th Weather Analysis and Forecasting, American
Meteorological Society, Ft. Lauderdale, FL.,
29-31 Jul 2001. Daley, R., and E. Barker, 2000
NAVDAS Source Book 2000, NRL/PU/7530-00-418, US
Naval Research Laboratory, Monterey, CA, Aug
2000, 151 pp. Dumais Jr., R.E., L. Luces, Y.
Wang, T. Henmi, and J. Cogan, 2004 High
Resolution Meteorological Nowcasting to meet U.S.
Army Future Forces Requirements, to be presented
at the AHPCRC Workshop on Mesoscale and CFD
Modeling, Jackson State University, Jackson, MS,
May 2004 Grell, G.A., J. Dudhia, and D.R.
Stauffer, 1995 A description of the
Fifth-Generation Penn State/ NCAR Mesoscale Model
(MM5), NCAR Technical Note, NCAR/TN-398STR, 122
pp. Henmi, T., 2003 Development of a nowcasting
method for three-dimensional meteorological data
preliminary report, US Army Research Laboratory
Technical Report, ARL-TR-3120, WSMR, NM, Nov
2003, 30 pp. Lazarus, Steven M., Ciliberti, Carol
M., Horel, John D., Brewster, Keith A. 2002
Near-Real-Time Applications of a Mesoscale
Analysis System to Complex Terrain. Weather and
Forecasting Vol. 17, No. 5, pp. 971-1000.
19
Nowcast References of Interest
Moninger, W.R., R.D. Mamrosh, and P.M. Pauley,
2003 Automated meteorological reports from
commercial aircraft, Bulletin of the American
Meteorological Society, Vol. 84, No. 2, pp
203-216. Passner, J.E., 2003 Post-processing
for the Battlescale Forecast Model and Mesoscale
Model Version V, U.S. Army Research Laboratory
Technical Report, ARLTR-2988, June 2003, 43
pp. Rachele, H., A. Tunick, and F.V. Hansen,
1995 MARIAH- A similarity-based method for
determining wind, temperature, and humidity
profile structure in the atmospheric surface
layer, Journal of Applied Meteorology, 1995, 34,
pp. 1000-1008. Sashegyi, K.D., and R.D. Madala,
1994 Initial conditions and boundary
conditions, Mesoscale Modeling of the
Atmosphere, American Meteorological Society
Meteorological Monograph, 1994. Eds. R. A. Pielke
and R.P. Pearce, 25 (47), pp. 1-12. Seaman,
N.L., D.R. Stauffer, A. Deng, A.M. Gibbs, A.J.
Schroeder, and G.K. Hunter, 2002 Evaluation of
a rapidly relocatable high-resolution numerical
model for meteorological nowcasting based on
MM5, Preprints of the 19th Conference on Weather
Analysis and Forecasting, San Antonio, TX, 12-16
Aug 2002. Strahl, J., D. Geiszler, J. Cook, P.
Harasti, G. Love, L. Phegley, Q. Zhao, F. Franco,
M. Frost, R. Mantri, D. Martinez, and S. Wells,
2003 Nowcast for the next generation Navy
Recent progress in Naval Nowcast Technology,
Proceedings of the 2003 BACIMO, Monterey, CA, Sep
2003. Wang, Y., J.J. Mercurio, C.C. Williamson,
D.M. Garvey, and S. Chang, 2003 A high
resolution, three-dimensional, computationally
efficient, diagnostic wind model, US Army
Research Laboratory Technical Report,
ARL-TR-3094, Oct 2003, 26 pp. Xue, M., D.-H.
Wang, J.-D. Gao, K. Brewster, and K. K.
Droegemeier, 2003 The Advanced Regional
Prediction System (ARPS), storm-scale numerical
weather prediction and data assimilation. Meteor.
Atmos. Physics, 82, 139-170. Zeller, K., J.
McGinley, N. Nikolov, P. Schultz, B. Shaw, S.
Albers, and J. Snook, 2003 A consortium for
comprehensive mesoscale weather analysis and
forecasting to monitor fire threat and support
fire management operations, 5th Symposium on Fire
and Forest Meteorology Joint with 2nd
International Wildland Fire Ecology and Fire
Management Congress, Orlando, FL, 16-20 Nov
2003.