NPS Smart Climatology Brief

1
NPS Smart Climatology Brief

• A Smart Climatology of Evaporation Duct Heights
• and Surface Radar Propagation for the 
• Northwest Indian Ocean and Nearby Regions

Lt Katherine L. Twigg, Royal Navy Thesis
Advisors Dr. Tom Murphree and Paul
Frederickson Naval Postgraduate School 07
September 2007
Unclassified
2
Summary

• Author Lt Katherine L Twigg, RN
• Advisors Tom Murphree and Paul Frederickson
• Sponsor Royal Navy Office of Naval Research
• Beneficiaries Users of surface radar (e.g., for
  antisubmarine warfare, ship defense, special ops)
• Methods and Results
• Used smart climatology data and methods to
  improve long term mean climatologies of
  evaporation duct heights (EDH) and radar
  propagation in the Indian Ocean and nearby seas.
• Analyzed impacts of seasonal changes climate
  variations (e.g., ENLN, IOZM) on EDH surface
  radar propagation.
• Results (a) new smart EDH climatology with
  substantial improvements over existing Navy
  climatology (b) identified major spatial and
  temporal changes in EDH, including those caused
  by climate variations (c) determined which
  factors EDH and surface radar propagation are
  most sensitive to for different regions and
  seasons (d) found potential for forecasting EDH
  and surface radar propagation at weekly to
  monthly lead times.
• Products (a) smart climatological environmental
  assessment surfaces for EDH and EDH factors and
  (b) smart climatological performance surfaces for
  surface radar propagation (range, CoF) both for
  varying climate scenarios.
• The methods used in this work are directly
  applicable to developing smart climatologies for
  other regions, and for other EM and acoustic
  propagation phenomena.

Surface radar detection ranges (km) based on our
smart EDH climatology. Values shown are long
term means for Sep, for a C-band radar at 30 ft
and detection threshold of 150 dB.
EDH (m) from our smart EDH climatology. Values
shown are long term means for Aug.
Full thesis available at NPS Smart Climatology
Reports, http//wx.met.nps.navy.mil/smart-climo/re
ports.php
3
Outline

• Introduction to EDH and its importance in EM
  propagation
• Introduction to smart climatology process
• LTM EDH plots and correlation to its factors
• Comparison to existing climatology and WHIO buoy
  data
• Effect of climate variations on EDH
• Operational implications and application
• Conclusions and future work

4
Evaporation Duct and EM Propagation
Image courtesy of Maritime Warfare Center, RN
Unclassified
5
Evaporation Duct

• Evaporation ducts form just above the ocean
  surface due to the rapid decrease in humidity
  with height from saturation conditions at the
  surface.
• The evaporation duct height (EDH) is the height
  in the surface layer where the vertical gradient
  in modified refractivity (M) changes from
  negative to positive.
• EDHs can be highly variable in space and time.
• The trapping layer below the EDH behaves like a
  waveguide and can lead to propagation at
  microwave frequencies over the horizon.

Radar waves bend up
EDH at min M
Radar waves bend down
Height
Evaporation Duct
Modified Refractivity
6
Evaporation Duct

• It is not feasible to routinely measure vertical
  M profiles close to the surface, so EDH is
  determined from an evaporation duct (ED) model
  using bulk measurements.
• NPS ED model is a state of the art model and is
  used in our research.
• Advanced Refractive Effects Prediction System
  (AREPS) computes propagation loss and uses ED
  models to create a low-altitude refractivity
  profile that is blended with refractivity data
  for higher altitudes.
• Higher EDHs often lead to increased signal
  strength and radar detection ranges, depending
  upon the frequency, and the height of the radar
  and target.
• EDH is the best parameter for quantifying
  near-surface microwave propagation.

No Duct
EDH 10m
EDH 20m
EDH 30m
EDH 40m
Height (m)
Range (nm)
AREPS Probability of Detection plots for 10GHz
radar against a small surface target, for varying
EDH
7
Factors Affecting EDH

• Vertical distribution of humidity, which
  determines EDH, is dependant upon SST and
  turbulent mixing near the surface.
• Thus, the main factors which determine EDH are
  SST, wind speed, air temperature, and humidity.
• EDH is sensitive to the sign of the air-sea
  temperature difference (ASTD). Generally
  conditions are unstable (i.e., ASTD lt 0).
  However, if ASTD gt 0 EDH, increases rapidly.
• EDH increases/decreases with increasing wind
  speed when ASTD is negative/positive. Increasing
  wind speeds cause surface conditions to be more
  neutral.
• EDH is extremely sensitivity to the ASTD,
  especially at lower wind speeds and RH.

8
Smart Climatology

• Climate
• Expected state of the environment based on
  scientific observations, analyses, theories, and
  models. Not based just on observational
  analyses. Expected state accounts for long term
  means and variations from these means that occur
  over long periods (e.g., anomalous trends and
  oscillations that occur over weeks, years, or
  longer).
• Smart Climatology
• Climatology applied to military uses that
  employs all relevant tools of modern climatology,
  such as
• Full suite of in situ and remote observational
  data sets
• Reanalysis
• Data access, mining, processing, and
  visualization tools
• Modern statistical and dynamical analysis
  methods
• Long term means and higher order statistics
• Climate variations (e.g., regimes, trends,
  oscillations)
• Climate system modeling
• Statistical and dynamical climate forecasting

Most smart climatology tools are not currently
used by the US or UK military
9
Data Used

• NCEP Global Atmospheric Reanalysis
• Area Indian Ocean, with focus regions in areas
  of operational/academic interest.
• Spatial resolution Guassian grid 210 km x 210
  km (shown below).
• Surface data specific humidity, air temperature,
  skin temperature, u and v components of wind,
  pressure.
• Timescale and temporal resolution 1970-2006,
  every 6 hours.

Blue shaded areas focus regions. Yellow shaded
areas two existing Navy EDH climatology data
squares.
10
Coastal EDH Issues

• Reanalysis data at a given grid point is an
  average for a whole grid box.
• Coastal grid boxes merge land and sea data.
• EDH is dependent on surface factors which can
  change dramatically from land to sea.
• EDH was calculated only for grid boxes lying
  exclusively over sea.
• EDH was estimated for coastal grid boxes that
  include land and sea, using calculated EDH in
  nearest sea grid boxes.
• Thus, EDH values near the coast should be used
  with caution.

Example showing how sea grid points were mapped
to land-sea grid points.
EDH calculated for blue shaded grid boxes.
11
Results Examples Smart EDH Climatology
Long Term Mean EDH for May/Jun/Jul
Long Term Mean EDH for Aug/Sep/Oct
Cautionary note see Coastal EDH Issues slide
EDH unit meters
12
Results Examples Effects of Factors on EDH
LTM SST for Aug/Sep/Oct
LTM EDH for Aug/Sep/Oct
LTM Ta for Aug/Sep/Oct
LTM ASTD for Aug/Sep/Oct
LTM WS for Aug/Sep/Oct
LTM RH for Aug/Sep/Oct
Cautionary note see Coastal EDH Issues slide
Units EDH (m), WS (m/s), RH (), Temperature (C)
13
EDH and Factor Correlations

• Gulf of Aden coastal example
• EDH varies from 10-75 m
• EDH greatest in Jun/Jul/Aug
• Many factors have similar (or mirrored) annual
  patterns.
• EDH vs factor correlations show
• EDH has high correlation to ASTD throughout the
  year.
• RH strongly correlated May-Oct.
• Air Temp strongly correlated Jan-May.
• U-wind strongly correlated Sep-Nov.
• Correlation value of 0.274 equates to a 95
  significance level
• Correlation value of 0.470 equates to a 99.5
  significance level
• EDH highly dependent on wind direction and
  advection of warm, dry air from the land.
• Factor correlations identify key variables to
  monitor, analyze and forecast, and lead to
  improved estimates of uncertainty.

EDH (m)
Wind (m/s)
Temp (C)
unstable
unstable
stable
ASTD (C)
SWly Wind
ENEly Wind
RH
Correlation
14
EDH and Factor Correlations

• Diego Garcia equatorial example
• EDH varies from 8-11 m
• Biannual cycle, peaks in Feb Jul.
• Unstable conditions throughout
• EDH vs factor correlations show
• EDH has high correlation to wind speed for the
  entire year, wind speed dominates EDH cycle.
• RH strongly correlated Jul - Nov.
• Air Temp strongly correlated Mar-May and Oct
  -Jan.
• U-wind strongly correlated Jun Nov.
• Correlation value of 0.274 equates to a 95
  significance level
• Correlation value of 0.470 equates to a 99.5
  significance level.
• EDH highly dependent on wind speed.
• Factor correlations identify key variables to
  monitor, analyze and forecast, and lead to
  improved estimates of uncertainty.

EDH (m)
Wind (m/s)
Temp (C)
ASTD (C)
NEly Wind
RH
SWly Wind
Correlation
15
Existing Climatology vs Smart Climatology

• Existing climatology plotted in map view
  normally only displayed in bar graph form.
• Resolution
• Plots highlight the benefit of higher resolution
  data and improve understanding of EDH
  variability.
• Data
• Existing climo data collected primarily along
  shipping routes. Some 10x10 grid boxes had 100
  observations, and some gt3000.
• Comparison
• Some basic similarities, but clearly large
  differences in magnitude and spatial patterns.
• EDH in Gulfs dissimilar due to open ocean
  adjustment of Paulus-Jeske model.
• Requirement for ground truthing

Smart EDH LTM Sep
Existing EDH LTM Sep
EDH units meters
Cautionary note see Coastal EDH Issues slide
16
Existing Navy Climatology vs Smart Climatology

• Existing EDH climatology
• Constructed in early 1980s from 1970-1979 COADS
  data.
• COADs data primarily from volunteer ships of
  opportunity
• LTM monthly values
• Resolution 10x10 boxes (1000 km x 1000 km)
• EDH computed using Paulus-Jeske ED model with
  open ocean adjustment.
• Smart EDH climatology
• Constructed in this study from NCEP reanalysis
  data.
• Monthly mean of 6 hrly data.
• Time periods 1970-1979 1970-2006.
• Resolution 210 km x 210 km
• EDH computed using NPS ED model.
• Comparison
• Smart climo comparison used average of 36 grid
  points covering the same area as existing climo.
• Both climatologies show similar seasonal cycle.
• Large magnitude differences due to differences in
  data and differences in model.

Existing climo grid boxes 10x 10 degree Marsden
squares
Existing Climo Smart Climo (1970-2006)
Reanalysis Means (1970-1979)
17
EDH Climatologies vs EDH From In Situ Buoy

• Buoy, existing climo, smart climo, and 1994-1995
  reanalysis monthly means should not necessarily
  be the same, as they do not represent the same
  areas and time periods.
• So there is no one absolute truth.
• However, the values should be similar.
• Buoy one point in space for one year, located
  where the Somali Jet position, structure, and
  strength will have a distinct effect on
  observations in the vicinity.
• All EDH data sets show similar temporal
  variations.
• Existing climo consistently has highest values.
• Smart climo and 1994-1995 means generally have
  better agreement with the buoy than existing
  climatology.
• Worst agreement with buoy is Jul/Aug when Somali
  Jet is strongest.

18
Correlation of Diego Garcia EDH with Large Scale
EDH Factors

• Local EDH significantly correlated with remote
  factors.
• Fluctuations in remote factors are indications of
  global scale climate variations. E.g.
• Correlation with SST and 850 hPa geopotential
  height indicate Diego Garcia EDH is strongly
  correlated with El Nino-La Nina (ENLN) and Indian
  Ocean Zonal Mode (IOZM).
• Thus, we hypothesized that climate variations
  cause significant deviations from the LTM EDH
  values.

Images created at NOAA/ESRL Physical Sciences
Division web site at using NCEP reanalysis data.
19
Climate Variations Indian Ocean Zonal Mode
(IOZM)

• Coupled ocean-atmosphere oscillation in the
  Indian Ocean (IO).
• Positive Phase
• Anomalous surface Ely winds over equatorial IO.
• Anomalous cooling of SE IO and anomalous warming
  of W IO.
• Enhanced (reduced) convection over the western
  (eastern) IO.
• Negative phase often follows/ precedes positive
  phase.
• Negative Phase opposite configuration to
  positive phase.
• Dipole Mode Index (DMI) is a measure of IOZM.
• Positive phase often coincides with El Nino event

Images from http//www.jamstec.go.jp/frsgc/resear
ch/d1/iod/
20
ENLN and IOZM Indices and Event Selection
--- IOZM index (DMI, 5 month smoothed) --- ENLN
index (MEI, 5 month smoothed)
IOZM Aug-Oct is period of high amplitude and
large anomalies in IO. ENLN Oct-Dec is period of
high amplitude and large anomalies in
IO. Developed composites based on top 5 IOZM and
ENLN events.
MEI from http//www.cdc.noaa.gov/people/klaus.wolt
er/MEI/index.html
21
Composite Results Positive IOZM Anomalies
Vector Wind IOZM Anomaly ASO
SST IOZM Anomaly ASO
RH IOZM Anomaly ASO
Wind Speed IOZM Anomaly ASO
Units EDH (m), WS (m/s), RH () Temperature (C)
22
Composite Results Positive IOZM Anomalies
EDH IOZM Anomaly ASO
ASTD IOZM Anomaly ASO
RH IOZM Anomaly ASO
Wind Speed IOZM Anomaly ASO
Cautionary note see Coastal EDH Issues slide
Units EDH (m), WS (m/s), RH () Temperature (C)
23
EDH Anomalies Seasonal Composite vs Single Month
EDH IOZM Composite Anomaly ASO
EDH Anomaly Sep 1997
-10
10
-3
3

• Comparison of 5-event composite seasonal anomaly
  to a single month anomaly during a strong IOZM
  event.
• Similar patterns, but magnitudes 3x larger in
  single month anomaly.
• Opposite anomalies in Gulf of Oman and other
  areas. This highlights complex nature of the
  impacts of climate variations on EDH.

Cautionary note see Coastal EDH Issues slide
Units EDH (m)
24
Correlations of EDH with Climate Variations -
Examples
ENLN
IOZM

• Colors show significant correlations
• warm colors positive correlation.
• cool colors negative correlation.
• dark colors 95 significance.
• EDH lags climate variations by 0,1, and 2 months.
• Correlation varies with the climate variation,
  location, and season.
• EDH often lags the climate variation by 1-2
  months.
• Thus, EDH variations may be predictable at medium
  to long range lead times.

25
Climatological Sensor Performance Impacts

• Figure to left shows how radar propagation
  strength decreases with range for different EDHs.
• Radar at 20m, target at 5m.
• APM propagation model used.
• Propagation loss at a given range decreases as
  EDH increases
• Greater EDH less propagation loss, greater
  detection range.
• Thus, detection range will vary as EDH varies.
• Example C-band radar detection ranges for Sep
  LTM Sep 1997, shown in climatological sensor
  performance surfaces, or maps, below

Cautionary note see Coastal EDH Issues slide
26
Key Results

• Produced improved LTM EDH
• Increased data time period and temporal
  resolution.
• Increased spatial resolution from 1000 x 1000 km
  to 210 km x 210 km.
• Generated map displays
• EDH correlated to different factors in different
  seasons and locations
• Factor correlations identify key variables to
  monitor, analyze and forecast, and lead to
  improved estimates of uncertainty.
• Where upwelling occurs stable conditions do not
  necessarily result in increased EDH.
• EDH varies with climate variations (e.g. IOZM,
  ENLN)
• Spatial and temporal variations, with varying
  significance levels.
• Climate variations impact EM detection ranges.
• EDH often lags the climate variation by 1-2
  months, thus EDH and surface ranges may be
  predictable at medium to long range lead times .
• Proof of concept use of smart climatology for
  environmental and performance assessments.
• Demonstrated value of smart climatology for
  improving military climatologies
• Caveats and limitations
• Coastal issues.
• Detection ranges are based on evaporation duct
  only and does not account for surface-based ducts.

27
Future Work

• Use higher spatial resolution data.
• Improved dynamical understanding and coastal
  values.
• Investigate finer temporal effects.
• Accounting for diurnal effects should increase
  correlation significance.
• Construct conditional climatologies based on
  different climate regimes (e.g., strong wind vs
  light wind regimes).
• Investigate smarter correlations (e.g.,
  correlations with offshore/onshore winds).
• Further analyses of climate variation impacts.
• Include North Atlantic Oscillation, Madden-Julian
  Oscillation, monsoon effects, different ENLN
  magnitudes, combined ENLN IOZM years.
• Develop correlation plotting tools for
  user-selected data (e.g., EDH vs IOZM, ENLN).
• Potential to predict EDH and surface radar ranges
  based on climate variations (ref Hanson Moss
  theses).
• Develop forecaster rules of thumb.
• More local analysis, including backward smart
  climatology methodology.
• Ground truth results.
• Extend smart climatological process to full
  atmospheric M profiles.
• Thus enabling detection range prediction during
  seasons of subsidence and prediction at any
  height.
• Use smart climatological approach to develop an
  improved global EDH climatology.
• Extend methodology to optical and acoustic
  climatologies and sensor performances.

28
Any Questions?
A Smart Climatology of Evaporation Duct Heights
and Surface Radar Propagation for the Indian
Ocean and Nearby Seas
Acknowledgements Bob Creasey, Mary Jordon, Arlene
Guest, Bill Little
29
Contact Information
Katherine Twigg, Lt, Royal Navy Fleet Numerical
Meteorology and Oceanography Center Stop 1, 7
Grace Hopper Avenue Monterey, CA
93943-5598 Voice 831-656-4096 katherine.twigg.uk
_at_navy.mil Tom Murphree, Ph.D. Department of
Meteorology Naval Postgraduate School 254 Root
Hall, 589 Dyer Road Monterey, CA
93943-5114 Voice 831-656-2723 Fax
831-656-3061 murphree_at_nps.edu Paul
Frederickson Department of Meteorology Naval
Postgraduate School 254 Root Hall, 589 Dyer
Road Monterey, CA 93943-5114 Voice 831-595
5212 Fax [phone] pafreder_at_nps.edu
The full thesis is available at NPS Smart
Climatology Reports http//wx.met.nps.navy.mil/sma
rt-climo/reports.php
30
Backup Slides
30
31
Motivation

• Investigate seasonal and anomalous variations in
  the air / sea boundary layers of the Indian Ocean
  (IO) and their effects on electromagnetic
  propagation via EDH variations.
• Aim to provide tactical advice to naval command
  and operators on radar/ electronic support
  measure /communication propagation. 
• Our study is designed to
• enhance understanding of long term mean
  climatological EDH variations primarily in the
  northwest Indian Ocean,
• hopes to provide the METOC/HM community with
  tools to predict how the EDH varies on medium to
  long timescales.
• Primary analysis conducted using atmospheric and
  oceanographic reanalysis data composited for the
  anomalies of interest, and then EDH height
  calculated using the NPS EDH model (used in
  AREPS).  
• Individual case studies to highlight operational
  impacts and ground truth conclusions. 

32
Thesis Approach

• Goal
• Determine sensitivity of EDH and radar
  propagation in NWIO to climate variations.
• Data sets
• NCEP/NCAR reanalysis,
• SODA reanalysis, etc.
• Analysis methods and models
• Smart climatology composite analysis and
  correlation analysis to identify how the seasonal
  cycles of the major atmospheric-oceanic factors
  that determine EDH are altered by ENLN, IOZM,
  MJO.
• Use of EDH model and AREPS to assess sensitivity
  of EDH and radar propagation to climate
  variations.
• Decision analysis models to develop tools for
  aiding military decision makers etc.
• Anticipated results / products
• Analyses of EDH factors, EDH, and radar
  propagation by season and by climate variation
  type prototype operational products for use in
  military planning.

33
Evaporation Duct Height

• Figures show how EDH varies with changing RH
  (left) and changing (wind speed).
• The lower figure shows error on EDH for one set
  of conditions

EDH determined by Modified Refractivity M T (K)
atmos temper, p (hPa) total atmos pressure e
(hPa) water vapor pressure.
Ref Fredrickson et al,01
34
Evaporation Height

• Error Graphs for differing wind speed situations

35
Correlation of Diego Garcia EDH with Large Scale
EDH Factors
H
Weaker RSJ
Stronger westerlies
Decreased HOA IO Precipitation
850hPa Geopotential Height for Anomalous
Convection in Maritime Continent (La Nina) LaJoie
Thesis, Mar06.
Image provided by the NOAA/ESRL Physical Sciences
Division, Boulder Colorado from their Web site at
using NCEP reanalysis data
36
Climate Variations Indian Ocean Zonal Mode
(IOZM)

• Coupled ocean-atmosphere oscillation in the
  Indian Ocean (IO).
• Positive Phase
• Anomalous surface Ely winds over equatorial IO.
• Anomalous cooling of SE IO and anomalous warming
  of W IO.
• Enhanced (reduced) convection over the western
  (eastern) IO.
• Negative phase often follows/ precedes positive
  phase.
• Negative Phase opposite configuration to
  positive phase.
• Dipole Mode Index (DMI) is a measure of IOZM.
• Positive phase often coincides with El Nino event

Ocean-atmos feedbacks in the E IO controlling
the growth of the IOZM
Positive Dipole Mode
http//www.jamstec.go.jp/frsgc/research/d1/iod/
Meyers et al, 2007 J Climate,20, p 2872
37
Climate Variations IOZM 1997/1998 event

• Diagram of the sequence of events in 97-98.
• a. The climatological alongshore winds off
  Sumatra (E) and the east African coast (F). The
  winds observed in the late summer and early
  autumn are denoted by G and H, respectively. The
  right-hand panel shows the effect at the Equator
  on the upper ocean induced by increased upwelling
  in the east and decreased upwelling in the west.
• b. Distribution of the winds resulting from the
  anomalous SST gradient along the Equator and the
  changes in the SSH distribution.
• Formation of the Ekman ridge in the central
  Indian Ocean and the forcing of
  westward-propagating downwelling equatorial
  Rossby waves to the west. The right-hand panel
  shows the effect on the upper ocean near 58 S.
• d. Subsequent cooling of the western Indian Ocean
  through enhanced mixing and coastal Ekman
  transports from stronger than average monsoon
  winds and through circulation changes associated
  with the weakening of the 199798 El Nino.
• Hypothesis adds a coupled dynamical component
  that has a longer timescale than the
  thermodynamics of the mixed layer, and which may
  form a link from one monsoon season to the next.
  Suggests that the Indian Ocean may not be a
  passive player in climate variability on seasonal
  to interannual timescales, but may enact a very
  active and independent role.

Webster et al.,1999 NATURE VOL 401 23
http//webster.eas.gatech.edu/Papers/Webster1999a.
pdf
38
Climate Variations El NinoLa Nina (ENLN)

• Magnitude of the global fluctuations during El
  Nino and La Nina are second only to those of the
  seasonal cycle.
• Occur about every 2 to 7 years.
• Usually last about one year, starting in May-Jun
  of one year and lasting until the following
  May-Jun.
• For EN events, the basic chain is 
• Pacific subtropical highs become anomalously low
  and/or tropical southeast Asian Pacific lows
  become anomalously high.
• Pacific trade winds become anomalously weak.
• Sea surface temperature (SST) becomes anomalously
  cool (warm) in western (eastern) tropical
  Pacific.
• Tropical convection become anomalously weak
  (strong) in western (eastern) tropical Pacific.
• Energy, moisture, and momentum transports into
  and out of the tropical Pacific becomes anomalous
  in many ways.
• Phase magnitude measured by variety of climate
  indices eg Multivariate ENSO Index (MEI)
• La Nina sequence approx opposite

Dr Murphree, Modern Climatology Module16
39
Compositing Results DMI Anomaly
Air Temp DMI Anomaly ASO
EDH DMI Anomaly ASO
RH DMI Anomaly ASO
SST DMI Anomaly ASO
Cautionary note see Coastal EDH Issues slide
Units EDH (m), WS (m/s), RH () Temperature (C)
40
Sensor Performance Impacts
Cautionary note see Coastal EDH Issues slide