Navy Altimetry

1
Navy Altimetry NOWCAST/FORECAST Sensitivity in
Undersea Warfare Systems

• Guillermo Amezaga LT, USNR
• Advisor Prof. Peter Chu
• Second Readers Eric Gottshall, CDR USN
• Past Thesis Michael Perry ENS, USNR
• Steve Mancini LCDR, USN
• Collaborator David Cwalina
• Todd Drury

2
Past Thesis

• Michael Perry, June 2003
• GDEM vs MODAS with 3 altimeters
• March 15, 2001
• 117 vs 1633 profiles
• 35.0-40.0N
• 70.0-75.0W
• Area coverage is an effective metric for
  comparing weapon presets

3
Past Thesis Process Flowchart (Mancini,2004)
Relative Difference
4
Thesis Process Flowchartfor NOWCAST/FORECAST
POM
MODAS
SCSMEX
Relative Difference
OBS
5
Thesis Process FlowchartNavy Altimetry
Relative Difference
6
(No Transcript)
7
Area of Interest (AOI)

• Taiwan
• High Oceanographic variability
• NE/SW Monsoon
• Meso-scale (Eddies)
• Typhoons
• Tactical Significance
• SCSMEX data availability
• MODAS validation
• POM validation
• Hydrographic data availability

8
Taiwan (AOI)
(From Laing et. al., 2001)
9
Snapshotof Variability
(From Laing et. al., 2002)
100m depth
30m depth
10
Satellite orbitology

• GEOSAT Follow On (GFO)
• TOPEX (TPX)
• JAN 2001

11
TPX launched AUG 1992 Exact overhead repeat 10
days Orbit 1336 km 66 degree inclination Circular
GFO launched FEB 1998 Exact overhead repeat 17
days Orbit 800 km 108 degree inclination 0.001
eccentricity
12
TPX
GFO
13
Satellite orbitology

• GEOSAT Follow On (GFO)
• TOPEX (TPX)
• JUL 2001

14
TPX
GFO
15
NOWCAST/FORECASTVALIDATION

• South China Sea Experiment (SCSMEX) validation of
  MODAS and POM

16
SCSMEX

• SCSMEX was large scale multi-national experiment
  in the SCS
• Main goal of gaining insight into the water and
  energy cycle of the Asian monsoon cycle (Chu et
  al, 2001)
• SCSMEX provided a unique opportunity to evaluate
  both the POM and MODAS.
• SCSMEX was conducted in the SCS from April
  through June 1998.
• During SCSMEX, the oceanographic data set
  included 1742 CTD and mooring stations (Chu et
  al, 2001).

17
SCSMEX Observation

• SCSMEX provided a unique opportunity to evaluate
  both the POM and MODAS
• SCSMEX was conducted in the SCS from April
  through June 1998
• Oceanographic data set included 1742 CTD and
  mooring stations (Chu et al, 2001).

(From Chu et. al., 2001)
18
Methodology of SCSMEX evaluation of MODAS

• Both observational and climatology where used in
  the verification of the value added of MODAS (Chu
  et al, 2004)
• The observational data was used as the benchmark
  to determine the error statistics for MODAS and
  climatology data.
• MODAS has added value if the difference between
  MODAS and observational data is smaller than the
  difference between climatological and
  observational data. (Chu et al, 2004).

19
Methodology of SCSMEX evaluation of MODAS (POM)

• MODAS, climatology , and observational data are
  represented by (temperature, salinity). The
  difference in between MODAS and observational
  data is
• And, the difference in between climatology and
  observational data is

.
20
Methodology of SCSMEX evaluation of MODAS (POM)

• The bias , mean-squareerror (MSE), and
  root-mean-square-error (RMSE) for MODAS ,

Where, N is the total number of horizontal points
(Chu et al, 2004)
21
Results of SCSMEX evaluation of MODAS

• MODAS provides reasonably good temperature and
  salinity NOWCAST. (Chu et. al., 2004)
• Gaussian-type distribution of errors
• Mean Temperature error nearly zero (deg C)
• Mean Salinity error of -0.2 ppt
• MODAS temperature NOWCAST better than salinity
  NOWCAST

22
Results of SCSMEX evaluation of MODAS
(From Chu et. al., 2004)
23
Results of SCSMEX evaluation of POM

• POM without data assimilation has capability to
  predict circulation patterns, temperature fields
  reasonably well, but has no capability to predict
  salinity fields. (Chu et. al., 2001)
• POM errors for temperature have a Gaussian-type
  distribution
• POM errors for salinity have a non-Gaussian type
  distribution
• Data assimilation enhances POM HINDCAST
  performance
• POM errors for BOTH temperature and salinity have
  Gaussian type distribution

24
MODAS vs POM

• MODAS is the US Navys premier dynamic
  climatology tool.
• MODAS provides the capability of modifying the
  historical climatology with remotely sensed SSH
  and SST
• Dynamic MODAS assimilates in situ measurements of
  the temperature and salinity by method known as
  Optimum Interpolation techniques (Fox et al
  2002).
• The surface structure projected downward using
  empirical relationships of the historical data
  which relates both SST and SSH to the subsurface
  temperature (Fox et. al., 2002).

• POM is a general three dimensional gridded model
  that is time-dependent and utilizes primitive
  equations to model general circulation with
  realistic topography and a free surface (Chu et
  al, 2001).
• POM was specifically developed to model nonlinear
  processes and mesocale eddy phenomena.

25
Navy Altimetry requirements for MODAS and
numerical models

• A minimum of one instrument is required. With
  only one instrument, this data must be used in
  conjunction with systems such as MODAS and NRL
  Layered Ocean Model (NLOM). (NRL/FR/7320
  99-9696)

26
Process Flowchart
POM
MODAS
SCSMEX
Relative Difference
OBS
27
Weapons Acoustic Preset Program (WAPP) Objectives

• To Provide the Fleet with an On-Board Automated
  Interactive Means for Generating Mk 48 Mk 48
  ADCAP Acoustic Presets and Visualizing Torpedo
  Performance
• Base Computations on In Situ Environmental,
  Tactical, Target, and Weapon Parameters
• Track the Evolution of Weapon, Tactical, Target,
  and Environmental Models
• Provide Interfaces to
• Support Fleet Exercises, Training, and Program
  Deliveries

28
(No Transcript)
29
Data SourcesProjected Environment (NUWC, NPT,
RI, 2005)
Latitude/Longitude Date-Time Group
DBDB-V v4.2 Level 2
Bottom Depth
Shallow/Deep Water
GDEM-V v3.0
Sound Speed Profile
Sound Speed Profile
HIE (SN v5.3)
Open Water/MIZ/Ice Cover
Under Ice
SMGC v2.0
Historic Wind Speed
Sea State
BST v1.0
Bottom Sediment Type
Bottom Type
VSS v6.3
Volume Scattering Strength Profile
30
Generated Output

• Percentage Area Coverage
• ASUW and ASW Scenarios
• Shallow, Mid, and Deep Search Bands
• High and Low Doppler Targets
• Values Normalized Over Acoustic Modes

31
Acoustic Preset ModuleRanked Listset

• List Set of Search Depth/Pitch Angle/Laminar
  Distance/Effectiveness Values
• List Set Ranked Based on Acoustic Coverage
  Effectiveness and Recommendation Made Accounting
  for Cavitations and Depth Separation
• Laminar Distance Utilized in Weapon Order
  Generation for Gyro/RTE

32
Acoustic Preset Module Signal Excess Display

• Signal Excess Selectable from Pull-Down Menu
• Provide a Visual Interpretation of Mk 48 Acoustic
  Performance Over Depth Band of Target

33
5 Tactical Scenarios Evaluated

• HD Deep ASW
• LD Deep ASW
• LD Shallow ASW
• HD ASUW
• LD ASUW

34
WAPP output for MODAS and POM

• Here, the subscripts m denotes MODAS and p
  denotes. (Manicini, 2004)

The relative difference was calculated using
statistical package, which produced absolute
values of the relative differences (RD) in area
coverage (AC) for the identical SD/SA combination
generated by WAPP,
35
WAPP output for MODAS and POM
MEAN 18.04 STD 7.76
MEAN 12.08 STD 5.51
Mean
36
WAPP output for MODAS and POM
37
WAPP output for MODAS and POM
38
Scenario Prob (RDgt0.1) () Prob (RDgt0.15)() Mean Std Dev
MODAS HD Deep ASW 43.75 3.25 11.3 4.88
POM HD Deep ASW 6 0.25 8.98 2.95
MODAS LD Deep ASW 23.75 3 9.66 4.41
POM LD Deep ASW 3 0.74 7.59 3.56
MODAS LD Shallow ASW 25.75 3 10.04 4.76
POM LD Shallow ASW 3.25 1 7.58 3.62
MODAS HD ASUW 81 71 19.83 7.89
POM HD ASUW 54 21.21 12.73 5.79
MODAS LD ASUW 73.5 65.25 18.04 7.76
POM LD ASUW 55 13.25 12.08 5.51
POM out performs MODAS
39
Navy Altimetry Sensitivity

• GFO vs TPX
• JAN 2001

40
ECS
SCS
GFO
TPX
In January 2001, the orbital coverage by TPX for
the ECS and SCS did not vary. The same orbital
track was updated every ten-days. GFO orbital
change every 17 days. GFO orbital track
overlapped the TPX orbital track during the
ten-day period of 21-31 January 2001.
41
Analysis of Input Data

• ECS
• JAN 05, 2001

42
MODAS SSP delta in ECS JAN 05, 2001
43
STATS in ECSJAN 05, 2001
yx
44
STATS in ECS JAN 05, 2001
45
SSP in ECS JAN 05, 2001
46
Analysis of Input Data

• SCS
• JAN 05, 2001

47
MODAS SSP delta in SCSJAN 05, 2001
48
STATS in SCSJAN 05, 2001
yx
49
STATS in SCS JAN 05, 2001
50
SSP in SCS JAN 05, 2001
51
Process Flowchart
Relative Difference
Key assumption GFO is more accurate than TPX
52
WAPP output for MODAS and POM

• Here, the subscripts mg denotes MODAS-GFO and mt
  denotes MODAS-TPX

The relative difference was calculated using
statistical package, which produced absolute
values of the relative differences (RD) in area
coverage (AC) for the identical SD/SA combination
generated by WAPP,
53
(No Transcript)
54
MODAS-GFO vs MODAS-TPXASW Scenario in ECSJAN
2001
55
MODAS-GFO vs MODAS-TPXASUW Scenario in ECSJAN
2001
56
MODAS SSP delta in ECSJAN 05, 2001
57
MODAS SSP delta in ECSJAN 10, 2001
58
MODAS SSP delta in ECSJAN 15, 2001
59
MODAS SSP delta in ECSJAN 20, 2001
60
MODAS SSP delta in ECSJAN 25, 2001
61
MODAS SSP delta in ECSJAN 30, 2001
62
(No Transcript)
63
MODAS-GFO vs MODAS-TPXASUW Scenario in SCSJAN
2001
64
MODAS-GFO vs MODAS-TPXASUW Scenario in SCSJAN
2001
65
MODAS SSP delta in SCSJAN 05, 2001
66
MODAS SSP delta in SCSJAN 10, 2001
67
MODAS SSP delta in SCSJAN 15, 2001
68
MODAS SSP delta in SCSJAN 20, 2001
69
MODAS SSP delta in SCSJAN 25, 2001
70
MODAS SSP delta in SCSJAN 30, 2001
71
Mean RD decreasing
72
Conclusions (Mancini, 2004)

• WAPP output is sensitive to satellite altimetry
  data assimilation
• Especially when MODAS fields differ significantly
  in the depth zone of interest (due to better
  depiction of mesoscale features by the field with
  altimetry)
• Satellite altimeter data contributed as much as
  an 80-90 chance of having a different engagement
  outcome
• Assuming RD of 0.1-0.2 in AC is enough

73
Coverage Score 47.7
Coverage Score 33.8
In the first case on the left the percentage of
tracks that wind up in the lobe is 94.2. Of
these 46.7 will enter homing. When the
likelihood is taken into consideration an overall
coverage score is derived which is 47.7. For the
second case 89.6 of the tracks wind up in lobe,
but only 16.3 enter homing, for an overall
coverage score of 33.8. As you can see the delta
in the overall coverage score is significant.
( Cwalina, 2005)
74
Analysis of Input Data

• ECS (Summer)

75
(No Transcript)
76
(No Transcript)
77
uillermo,     Enclosed is a file with two
depictions of a torpedo sweeping through an area
of uncertainty. Each dot represents a plausible
contact position with an associated speed and
course along with a likelihood. The targeting
algorithms determine the appropriate gyro and
run-to-enable to best cover the area. This plays
out as an animation in MABLAB and generates some
statistics. I took a snapshot during the middle
of the run. It is interpreted as follows   A dot
is red until the acoustic cone of the torpedo
passes over it. When this happens the torpedo has
had a detection opportunity on the  dot and it
turns yellow. It takes some time (based on the
waveform/beamset and signal processing  logic)
for the torpedo to go through the detection,
acquisition, and verification phase. If the dot
remains in the lobe for this time then it is
likely that the torpedo would be able to enter
homing on the dot target and the dot would turn
green.   For these two cases the vertical
acoustic coverage differs by 20 and affects the
size of the cone in the horizontal plane. Each
dot is considered a track. In the first case on
the left the percentage of tracks that wind up in
the lobe is 94.2. Of these 46.7 will enter
homing. When the likelihood is taken into
consideration an overall coverage score is
derived which is 47.7. For the second case 89.6
of the tracks wind up in lobe, but only 16.3
enter homing, for an overall coverage score of
33.8. As you can see the delta in the overall
coverage score is significant. I hope this gives
you a feeling for the impact of some of the
acoustic coverage differences on the targeting
problem in general.
78
(No Transcript)
79
(No Transcript)
80
(No Transcript)
81
(No Transcript)
82
(No Transcript)
83
MODAS SSP delta in ECS (North)JUL 10, 2001
84
STATS in ECSJUL 10, 2001
yx
85
STATS in ECS JUL 10, 2001
86
SSP in ECS JUL 10, 2001
87
Analysis of Input Data

• LZ (Summer)

88
MODAS SSP delta in LZ (South)JUL 10, 2001
89
STATS in LZJUL 10, 2001
yx
90
STATS in LZ JUL 10, 2001
91
SSP in LZ JUL 10, 2001
92
Process Flowchart
Relative Difference
93
Acoustic Preset Module Ray Trace Display

• Ray Trace Selectable from Pull-Down Menu
• Provide a Visual Interpretation of Mk 48 Acoustic
  Performance
• Impact of Boundary Interactions and Refraction
  Shown
• Variable Target Depth Bands(Near-Surface, Depth
  Zone of Interest, Target Max Depth)
• Effects of Reverberation Apparent for Low Doppler
  Targets

94
Acoustic Preset Module Signal Excess Display

• Signal Excess Selectable from Pull-Down Menu
• Provide a Visual Interpretation of Mk 48 Acoustic
  Performance Over Depth Band of Target