Convective Systems in the 2006 West African Monsoon: A Radar Study

1
Convective Systems in the 2006 West African
MonsoonA Radar Study

• Presented to the SJSU Meteorology Dept
• by Nick Guy
• 1 July 2008

2
Outline

• Section 1 General Overview Introduction
• Observational Projects
• West African Monsoon (WAM)
• African Easterly Waves (AEWs)
• Mesoscale Convective Systems (MCSs)
• Section 2 Radar Review Analysis
• Radar Background
• Radar Data
• Processing
• Classification
• Dataset Quality Control (QC)

3
Analysis Outline

• Section 3 Rainfall Estimation Analysis
• Methodology
• Spatial and Temporal coverage
• Conclusion

4
Section 1

• General Overview Introduction

5
AMMA Project

• African Monsoon Multidisciplinary Analyses
• Cooperative international project
• Objectives as stated at
• http//amma-international.org/index

6
AMMA Observational Network
7
Previous Studies

• Synoptic study of atmospheric structure of WAM
• Global Atmospheric Research Project (GARP)
  Atlantic Tropical Experiment (GATE )
• West African Monsoon Experiment (WAMEX)
• Regional study of ground-recorded rainfall
  (continental Africa)
• Etudes des Precipitations par Satellite (EPSAT)

8
Tropical Convection

• Continuous surface heating
• Wide range of spatial scales
• Isolated convective cores (O1 km)
• Organized systems (O1000 km) Mesoscale
  Convective Complex (MCC)
• Mesoscale Convective Systems (MCSs)
• Group of contiguous t-storms (O100 km) with
  associated stratiform rain
• Many types, squall line MCS (SLMCS) common in
  Africa

9
MCS Schematic
Houze (2004)
10
MCS

• Definition for this study
• Organized t-storms with contiguous precipitation
  region with horizontal scale gt100 km

11
SLMCS

• Linear organization and propagation
• Large trailing stratiform region
• Large impact of thermodynamic and dynamic
  structure of environment
• High prevalence of this type throughout season
• Large contributor of precipitation totals in some
  areas (more later)

12
African Precipitation

• Northward progression of rainfall
• Banded structure

Observational Data
GCM
GCM
Lawrence (Climate and Global Dynamics Division
NCAR)
13
West African Monsoon

• Seasonally dependent thermally-induced low over
  African continent
• Migration northward during boreal summer

Ferreira (2007)
14
WAM Characteristics I

• Two distinct phases (Sultan and Janicot 2003b)
• Preonset migration of southwesterly winds and
  ITF past 15N
• Onset - abrupt northward shift of the ITCZ from
  5N to 10N
• Precipitation time-frame April October
• Results in 99 of annual rainfall (Shinoda et al.
  1999)
• Mid-June September generally defines WAM period

15
African Monsoons
First part of season April - June
Second part of season July - October
Simmon, Gu, and Adler (2004)
http//www.nasa.gov/centers/goddard/news/topstory/
2004/0510africanwaves.html
16
WAM Characteristics II

• MCSs account for estimated 80-90 of annual
  rainfall in Sahel (Mathon et al. 2002)
• Convective portion of total rainfall average of
  65 in tropics (Schumacher and Houze 2006)
• Convective portion of total area average of 10
  in tropics (Houze 1993)
• Formation of MCSs (largest contributor of
  rainfall) is highly correlated to AEWs
• Parallel investigation (Ferreira et al. 2008)

17
Section 2
Radar Review Analysis
18
Radar Location
19
Radar Review
Echo top height
Brightband melting level
Houze (2004)
20
Radar Review

• Common standards
• Transmit/receive microwave pulse
• Doppler measure reflectivity, wind velocity,
  spectral width
• Reflectivity given by the simplified radar
  equation
• Rainfall estimates provided by a Z-R relationship
• Z aRb

21
Radar Review

• Scan techniques
• Survey
• 360, 3-D field of view
• Adjust PRF to increase unambiguous scan range
• Range Height Indicator (RHI)
• Vertical scan steps along a constant radial
• Produces vertical profile
• Volume
• 360, 3-D field of view
• Composed of 15 elevation angle sweeps

22
Radar Review

• Background noise ground clutter, equipment
  noise
• Anomalous Propagation (AP) Normal beam
  propagation altered by change in refractivity
  with height
• Range aliasing (2nd trip echo)

Rinehart (1997)
23
Data

• MIT C-band Doppler radar reflectivity field
• Collected June September 2006
• Full volume scans from 5 July 27 September 2006
• gt 11,250 scans analyzed
• Rmax 150 km

24
Classification System

• Classification structure based on scope of study
• Simplified version of that used by Rickenbach and
  Rutledge (1998)
• Sub-MCS and MCS-scale events
• Visual inspection - subjective

25
Software Flowchart
26
Radar Data Quality Control

• Removal of non-meteorological data
• Maximize meteorological echo retained
• The most difficult false echo to remove is that
  which is embedded in meteorological echo
• Not necessarily removed
• In some cases it requires suppression of some
  real echo
• Generally favorable results from the QC operation

27
QC approaches

• Hardware setup
• Setup before data collection
• Data processing algorithm
• Employed during data collection or in archived
  data
• Network approach
• Employed during data collection or in archived
  data
• Method dependent upon processing time required
  and equipment
• Employed data processing method
• Performed at 1 km CAPPI
• Algorithm based on a modified approach developed
  by Rosenfeld et al. (1995)

28
QC Algorithm Theory

• Eight adjustable variables to determine removal
  of spurious echo
• 3 echo height thresholds (H1, H2, and H3)
• 5 radar reflectivity thresholds (Z0, Z1, Z2, Z3,
  and dBZnoise)
• Parameter values are customized to particular
  region being studied
• Topographical Characteristics
• Climatological Characteristics
• Default based on those used at radar site in
  Senegal, Africa

29
QC Algorithm
Vertical development
Min echo top
Precipitable echo
Stronger clutter
Ztop Echo top height Zmax(h) - Maximum
reflectivity at height, h

• Conditions of filter application
• Once any condition put forth in Eq. 1-4 is
  satisfied, the reflectivity value is rejected.
• Application of 1 x 1 km mask over raw data to
  suppress echo

30
QC Algorithm Application - Comparison
No QC filter
GSFC default parameters
Region-specific special case parameters (SJSU)
31
QC Algorithm Decision Tree
Widespread convection, MCSs
Local Convection
Least Stringent
Default Parameters
AP Parameters
Complete Removal
Most Stringent
32
QC Application Weak Echo
Before
After
33
QC Application Strong Echo
Before
After
34
QC Analysis

• Compared rainrate values of raw data and QC data
• bias QC / Raw data
• (removing bias from data)

35
QC Analysis Rainfall Retention
MCS bimodal structure
Sub-MCS broad distribution
36
QC Analysis Echo Cover Retention
MCS broad distribution
Sub-MCS total removal
37
Section 3

• Rainfall Estimation Analysis

38
Rainfall Estimation

• Convective-Stratiform partitioning algorithm
  applied to data
• Assess spatial radar reflectivity gradients
• Convective intense, horizontally variable
• Stratiform weaker, homogeneous
• Intensity criterion Z gt 40 dBZ Convective
• Peakedness criterion Z lt 40 dBZ but must
• exceed the mean reflectivity in an 11 km circular
  
• radius about the point by 4.5 dBZ
• Once designated, a circular cell of 1 5 km
• radius around point designated as convective
• Cells not designated convective, but with echo
• are designated as stratiform
• Z-R relationship applied
• Analysis of disdrometer data (Sauvageot and
  Lacaux 1995)

39
Convective-Stratiform Map
SLMCS event plotted in terms of
convective-stratiform components
SLMCS event plotted in terms of reflectivity
40
Rainrate Timeseries

• July and August core monsoon season months
• R gt 0.5 mmhr-1 ? MCS-scale systems
• Spectral analysis revealed strong diurnal cycle
  harmonics and 2-4 day frequency

41
MCS Contributions
Seasonal mean of 59 convective component of
rainfall
Seasonal mean of 12 convective component of
area coverage
42
Diurnal Composites
MCS-scale systems
Sub- MCS-scale systems
Note the difference in vertical scales ? MCS
component dominates total
43
Conclusion

• AMMA 2006 West African Monsoon data processed
• Quality controlled data via GVS (1C51) software
• Parameter settings verified via visual inspection
  and analysis
• Reflectivity retained highly correlated to SLMCS
  events
• Some stratiform precipitation sacrificed to
  optimize algorithm performance
• July possessed the most vigorous convection,
  though convective occurrences peaked in August
• Slightly more convective contribution to total
  rainfall than by stratiform
• Strong 1 (diurnal harmonic) and 2-4 day (AEWs)
  frequencies were observed in rainrate data
• SLMCS propagates preferentially at 07Z
• Vertical development of convection deepest in
  July
• Data is disseminated by Medias France

44
Acknowledgements

• Dr. Tom Rickenbach
• Bart Kelley, Jason Pippitt, Dave Wolfe at GSFC
• Drs. Craig Clements and Eugene Cordero
• The SJSU faculty
• SJSU students

45
Questions?
46
Vertical Structure
47
Data Processing I

• RSL
• Only full volume scans are retained
• GVS Level 1 1C51
• Only processes full volume scans with more than 4
  tilt levels

48
Data Processing II

• REORDER
• Interpolation to Cartesian coordinate system
• 1 km horizontal vertical spacing
• IDL routines
• Visual output and calculations
• Use range only 130 km, discard outer 20 km

49
Radar Review

• Error sources
• Instantaneous measurements
• Conversion of reflectivity to precipitation
• DSD affects Z-R relationship
• Presence of hail and snow
• Scan volume increases as beam propagates, beam
  width too large
• Ground clutter from nearby topography and
  buildings
• Anomalous propagation, range and velocity
  aliasing, etc.
• Radar beam always AGL does not produce ground
  precipitation measurements
• Except Radar beam affected by temp inversions

50
(No Transcript)
51
QC Application Sub-MCS Systems
Before
After
52
QC Application SLMCS Systems
Before
After
53
QC Application Stratiform
Before
After
54
MCS Event Duration

• Mean duration of 12.1 hr (in radar view field)
• Monthly mean duration decreases throughout season

55
SLMCS Frequency
Early morning (LT) SLMCS passage over the Niamey
radar site