Short-term Climate Simulations of African Easterly Waves: A Case Study with a Global Mesoscale Model Dr. Bo-Wen Shen Earth System Science Interdisciplinary Center (ESSIC) University of Maryland, College Park NASA Goddard Space Flight Center [email protected]; [email protected]

The Department of Mathematics and Statistics San Diego State University 28 March 2014 Short-term Climate Simulations of AEWs

1

San Diego State Univ. 28 Mar 2014

Acknowledgements •

NASA/GSFC: Wei-Kuo Tao, William K. Lau, Jiundar Chern, Chung-Li Shie, Zhong Liu

•

NOAA/AOML: Robert Atlas

•

NOAA/NCEP/NHC, Mark DeMaria

•

UCAR: Richard Anthes

•

UC: Roger Pielke Sr.

•

UAH: Yu-Ling Wu

•

NC A&T State U.: Yuh-Lang Lin

•

NASA/JPL: Frank Li, Peggy Li

•

Navy Research Lab: Jin Yi

•

NASA/ARC: Piyush Mehrotra, Samson Cheung, Bron Nelson, Johnny Chang, Chris Henze, Bryan Green, David Ellsworth, control-room Funding sources: NASA ESTO (Earth Science Technology Office) AIST (Advanced Information System Technology) Program; NASA Computational Modeling Algorithms and Cyberinfrastructure (CMAC) program; Modeling, Analysis, and Prediction (MAP) Program

•

Project Title of AIST-08-0049 (PI: Shen) Coupling NASA Advanced Multi-Scale Modeling and Concurrent Visualization Systems for Improving Predictions of High-Impact Tropical Weather (CAMVis), March 2009 – April 2012 Project Title of AIST-11-0012 (PI: Shen): Integration of the NASA CAMVis and Multiscale Analysis Package (CAMVis-MAP) for Tropical Cyclone Climate Study, May 2012 – April 2015 Short-term Climate Simulations of AEWs

2

San Diego State Univ. 28 Mar 2014

Acknowledgment We are grateful for support from the NASA ESTO Advanced Information Systems Technology (AIST) program and NASA Computational Modeling Algorithms and Cyberinfrastructure (CMAC) program. We would also like to thank reviewers for valuable comments, D. Ellsworth for scientific, insightful visualizations, and J. Dunbar for proofreading the related

materials.

Acknowledgment is also made to the NASA HEC Program, the NAS Division, and the NCCS for the computer resources used in this research.

Short-term Climate Simulations of AEWs

3

San Diego State Univ. 28 Mar 2014

Outline

1. Introduction

8 slides

2. Supercomputing, Visualization, and Global Modeling

14 slides

3. Genesis of Twin Tropical Cyclones (TCs) (as model verifications)

8 slides

4. Simulations of AEWs and AEJ

12 slides

5. Sensitivity Experiments

6 slides

6. Predictability and Chaos

2 slides

7. Summary and Future Tasks

5 slides

Short-term Climate Simulations of AEWs

4

San Diego State Univ. 28 Mar 2014

Decadal Survey Missions Two Major Scenarios in Decadal Survey missions are: •

•

Extreme Event Warnings (near-term goal): Discovering predictive relationships between meteorological and climatological events and less obvious precursor conditions from massive data sets  multiscale interactions; modulations and feedbacks between large/long-term scale and small/short-term scale flows Climate Prediction (long-term goal): Robust estimates of primary climate forcings for improving climate forecasts, including local predictions of the effects of climate change. Data fusion will enhance exploitation of the complementary Earth Science data products to improve climate model predictions.

Prediction of Hurricane Tracks and Intensity

Prediction of Hurricane Formation

Hurricane Climate Simulation

Courtesy of the Advanced Data Processing Group, ESTO AIST PI Workshop Feb 8-11, 2010, Cocoa Beach, FL Short-term Climate Simulations of AEWs

5

San Diego State Univ. 28 Mar 2014

Objective Develop a scalable, multiscale analysis tool, based on the Coupled Advanced multiscale Modeling and Visualization system (CAMVis), to improve extendedrange tropical cyclone (TC) prediction and consequently TC climate projection by enabling:

TC01A

Kesiny

TC02B

Errol

• Understanding of the TC genesis processes, accompanying multiscale processes (both downscaling by large-scale events and upscaling by small-scale events), and their subsequent non-linear interactions • Discovery of hidden predictive relationships between meteorological and climatological events. This project targets the ACE, PATH, SMAP, Nextgeneration scatterometer, and 3D-Winds missions. The scientific research cycle consists of Modeling, Observation, Analysis, Synthesizing, Theorizing (MOAST). Short-term Climate Simulations of AEWs

7

San Diego State Univ. 28 Mar 2014

African Easterly Waves (AEWs)

AEW

• • •

During the summer time (from June to early October), African easterly waves (AEWs) appear as one of the dominant synoptic weather systems in West Africa. These waves are characterized by an average westward-propagating speed of 11.6 m/s, an average wavelength of 2200km, and a period of about 2 to 5 days. Nearly 85% of intense hurricanes have their origins as AEWs [e.g., Landsea, 1993]. Contributed by Chris Landsea, http://www.aoml.noaa.gov/hrd/tcfaq/A4.html

Short-term Climate Simulations of AEWs

8

San Diego State Univ. 28 Mar 2014

Formation of Hurricane Helene (2006)

• • • •

http://goo.gl/arWSZ Simulations from Day 20 to Day 30 in a run initialized at 00Z Aug 22, 2006. Upper-level winds in red; middle-level winds in green; low-level winds in blue Low-level CC (cyclonic circulation); Upper-level AC (anticyclonic circulation) Shen, B.-W. W.-K. Tao and M.-L. Wu, 2010b: African Easterly Waves in 30-day High-resolution Global Simulations: A Case Study during the 2006 NAMMA Period. GRL., L18803

Short-term Climate Simulations of AEWs

9

San Diego State Univ. 28 Mar 2014

Multiscale Processes To improve the prediction of TC’s formation, movement and intensification, we need to improve the understanding of nonlinear interactions across a wide range of scales, from the large-scale environment (deterministic), to mesoscale flows, down to convective-scale motions (stochastic). Hierarchical modeling

Zoomed-out

Zoomed-in

Pouch

Shen, B.‐W., B. Nelson, S. Cheung, W.‐K. Tao, 2013b: Scalability Improvement of the NASA Multiscale Modeling Framework for Tropical Cyclone Climate Study. (Sep/Oct issue of IEEE CiSE)

Short-term Climate Simulations of AEWs

11

San Diego State Univ. 28 Mar 2014

Published Articles since 2010 Journal Articles: 1. 2. 3. 4.

5. 6.

7. 8.

Shen, B.-W., 2013d: Nonlinear Feedback in a Five-dimensional Lorenz Model. J. of Atmos. Sci. in press. Shen, B.-W., M. DeMaria, J.-L. F. Li and S. Cheung, 2013c: Genesis of Hurricane Sandy (2012) simulated with a global mesoscale model, Geophys. Res. Lett., 40, 4944–4950, doi:10.1002/grl.50934. Shen, B.-W., B. Nelson, S. Cheung, W.-K. Tao, 2013b: Improving NASA’s Multiscale Modeling Framework for Tropical Cyclone Climate Study. IEEE Computing in Science and Engineering, vol. 15, no 5, pp 56-67. Sep/Oct 2013. Shen, B.-W., B. Nelson, W.-K. Tao, and Y.-L. Lin, 2013a: Advanced Visualizations of Scale Interactions of Tropical Cyclone Formation and Tropical Waves. IEEE Computing in Science and Engineering, vol. 15, no. 2, pp. 47-59, March-April 2013, doi:10.1109/MCSE.2012.64. Shen, B.-W., W.-K. Tao, and Y.-L. Lin, and A. Laing, 2012: Genesis of Twin Tropical Cyclones as Revealed by a Global Mesoscale Model: The Role of Mixed Rossby Gravity Waves. J. Geophys. Res. 117, D13114, doi:10.1029/2012JD017450. 28pp Shen, B.-W., W.-K. Tao, and B. Green, 2011: Coupling Advanced Modeling and Visualization to Improve High-Impact Tropical Weather Prediction (CAMVis). IEEE Computing in Science and Engineering (CiSE), vol. 13, no. 5, pp. 56-67, Sep./Oct. 2011, doi:10.1109/MCSE.2010.141. Shen, B.-W., W.-K. Tao, and M.-L. Wu, 2010b: African Easterly Waves in 30-day High resolution Global Simulations: A Case Study during the 2006 NAMMA Period. Geophys. Res. Lett., 37, L18803, doi:10.1029/2010GL044355. Shen, B.-W., W.-K. Tao, W. K. Lau, R. Atlas, 2010a: Predicting Tropical Cyclogenesis with a Global Mesoscale Model: Hierarchical Multiscale Interactions During the Formation of Tropical Cyclone Nargis (2008) . J. Geophys. Res.,115, D14102, doi:10.1029/2009JD013140.

Magazine Articles: 9.

Shen, B.-W., S. Cheung, J.-L. F. Li, and Y.-L. Wu, 2013e: Analyzing Tropical Waves using the Parallel Ensemble Empirical Model Decomposition (PEEMD) Method: Preliminary Results with Hurricane Sandy (2012), NASA ESTO Showcase . IEEE Earthzine. posted December 2, 2013. 10. Shen, B.-W., 2013f: Simulations and Visualizations of Hurricane Sandy (2012) as Revealed by the NASA CAMVis. NASA ESTO Showcase. IEEE Earthzine. posted December 2, 2013. Papers under review/preparation: • Shen, B.-W., 2014a: On the Nonlinear Feedback Loop and Energy Cycle of the Non-dissipative Lorenz Model. (submitted) • Shen, B.-W., 2014b: Nonlinear Feedback in a Six-dimensional Lorenz Model. Impact of an Additional Heating Term. (to be submitted)

Short-term Climate Simulations of AEWs

12

San Diego State Univ. 28 Mar 2014

Tropical Waves and TC Formation Equatorial Rossby (ER) Wave TC Nargis (Shen et al., 2010a)

Mixed Rossby Gravity (MRG) Wave asymmetric Twin TC

symmetric

(Shen et al., 2012)

L

EQ

EQ L An equatorial Rossby wave, appearing in Indian Ocean, is symmetric with respect to to the equator.

MRG waves, asymmetric with respect to to the equator, occasionally appear in Indian Ocean or West Pacific

Hurricane’s Structure

African easterly waves (AEWs) one hemisphere Helene

(Shen et al., 2010b)

LAT:5o Shen, B.-W., B. Nelson, W.-K. Tao, and Y.-L. Lin, 2013a, "Advanced Visualizations of Scale Interactions of Tropical Cyclone Formation and Tropical Waves," Computing in Science and Engineering, vol. 15, no. 2, pp. 47-59, March-April 2013, doi:10.1109/MCSE.2012.64 Short-term Climate Simulations of AEWs

13

AEWs appear as one of the dominant synoptic weather systems. Nearly 85% of intense hurricanes have their origins as AEWs (e.g., Landsea, 1993).

San Diego State Univ. 28 Mar 2014

Supercomputing, Visualization, and Modeling

Short-term Climate Simulations of AEWs

17

San Diego State Univ. 28 Mar 2014

Early Efforts with Global Models (1999~2003) • Unix System and Network Programming: Unix curses, device control (serial I/O), file system control, inter-process communication (pipes, semaphore, shared memory, TCP/IP sockets), process control, signal handling. • System Administration: Unix/Linux/MS windows system installation. • Supercomputing (Parallel/Distributed/Cluster Computing): MPI (Message Passing Interface), MPI-2 remote memory access, MLP (Multi-Level Parallelism), OpenMP, ESMF (Earth Science Modeling Framework), POSIX Threads, and JAVA Threads. Knowledge of Gird computing. • Software: Fortran (F77/F90/F95), OOP (Object Oriented Programming), C/C++, JAVA, Basic, Pascal, UNIX Shells, UNIX m4 script, PERL, Python, PHP, HTML, XML, XHTML, CGI, AWK. CVS (Concurrent Version System), GNU Make, gdb, LaTex, MATLAB, VMWARE, Secure Shell, MSOffice, VIS5D, AVS, GrADS, NCAR Graphics, GEMPAK. • Numerical Models: MM4, MM5-V1/V2, ARPS, WRF, NASA GEOS-4, GEOS-5 (beta), NCAR CAM •

• •

•

Lin, S.-J., S. Nebuda, B.-W. Shen, J.-D. Chern, W. Sawyer, and A. DaSilva, 2001: DAO's suggestions to the software design of CAM. (informal technical note). March 16, 2001. [CAM: NCAR community atmosphere model] Chang, Y., S. D. Schubert, S.-J. Lin, S. Nebuda, B.-W. Shen, 2001: The climate of the FVCCM-3 Model. NASA/GSFC Technical Report Series on Global modeling and Data Assimilation, vol 20, p. 127. Radakovich, J. D., G. Wang, J.-D. Chern, M. G. Bosilovich, S.-J. Lin, S. Nebuda, and B.-W. Shen, 2003: Implementation of the NCAR Community Land Model (CLM) in the NASA/NCAR finite-volume Global Climate Model (fvGCM). 14th Symposium on Global Change and Climate Variations. Lin, S.-J., B.-W. Shen, W. P. Putman, J.-D. Chern, 2003: Application of the high-resolution finite-volume NASA/NCAR Climate Model for Medium-Range Weather Prediction Experiments. EGS - AGU - EUG Joint Assembly, Nice, France, 6 - 11 April 2003.

Short-term Climate Simulations of AEWs

18

San Diego State Univ. 28 Mar 2014

Research News and Highlights (2004~2011) •

2004,

ARC news story: Initial Columbia Results Promising

•

2005,

AGU Highlight, (Atlas et al. 2005)

•

2006,

AGU highlight, featured in ``Science’’ (Shen et al., 2006a,b), cited as a global/mesoscale breakthrough by Prof. R. Pielke Sr.

•

2007,

Genesis simulations of six TCs in May 2002

•

2010,

NASA News story (Shen et al., 2010a). Follow-up stories appeared

in MSNBC, PhysOrg.com, National Geographic--Indonesia, ScienceDaily, EurekAlert, Yahoo News, TechNews Daily, Scientific Computing, HPCwire. •

2011,

featured in the magazine article entitled `` Turning the Tables on

Chaos: Is the atmosphere more predictable than we assume? ‘’ (Anthes, 2011)

Short-term Climate Simulations of AEWs

19

San Diego State Univ. 28 Mar 2014

NASA Supercomputing and Visualization Systems Pleiades Supercomputer (as June 2013) • one of a few petascale supercomputers • Rmax of 1,240 teraflops (LINPACK); Rpeak of 2,880 teraflops • 162,496 cores in total; Intel Xeon processors, X5570 (Nehalem), X5670 (Westmere); ES-2670 (Sandy Bridge, two CPUs/node; 8 cores/CPU), E52680v2 (Ivy Bridge), • 417 TB memory • 3.1 PB disk space • Largest InfiniBand network: • Supercomputer-scale visualization system – 8x16 LCD tiled panel display – 245 million pixels • 128 nodes – Dual-socket quad-core Opterons – 1024 cores, 128 GPUs • InfiniBand (IB) interconnect to Pleiades – 2D torus topology, 32 links to Pleiades – 9x2 switches – High-bandwidth concurrent visualization Short-term Climate Simulations of AEWs

20

San Diego State Univ. 28 Mar 2014

Architecture of the CAMVis v1.0 (the Coupled Advanced Multiscale modeling and concurrent Visualization systems; Shen e al. 2011)

Multi-scale Modeling with “M” nodes

Current Visualization with “N” nodes

Real-time Display

GMM

Large Scale Intra-communication

Parallel I/O Search for the predictability limits

RDMA MMF Coupler •data regridding; Inter-communication •data input and output

Intra-communication

Cloud Scale mgGCE

Simulation Simulation Short-term Climate Simulations of AEWs

Parallel Transfer Extraction

Visualization Visualization 21

comparison Discovery MPEG generation with satellite San Diego State Univ. 28 Mar 2014

Quasi-3D Streamlines

●

Generate multiple 2D slices of streamlines, different colors at different heights

●

Pick an opacity based on velocity magnitude

Control transparency

●

Faster winds are more opaque (less transparent)

●

Stack all of 2D slices vertically

●

Height is currently an estimate with the hydrostatic assumption

The density of streamlines indicate the magnitude of average wind speeds. The appearance of streamlines indicates intensification.

Short-term Climate Simulations of AEWs

23

San Diego State Univ. 28 Mar 2014

Impacts of Opacity High Opacity

Medium Opacity

Low Opacity

Figure: The impact of Opacity on the quasi-3D streamlines generated from exactly the same model data with the only difference in the "alpha" (i.e. opacity) value which is selected for a given wind speed. ‘Z’ in the color bar indicates altitudes (heights). Pink/red, green and blue lines show high-, middle-, and low-altitude winds. Generally we pick increasing alpha as wind velocity increases, so the faster winds are more visible. With alpha set too High (left), we see too much information about winds that are not important. With alpha set too low (right), we see the main feature, but not enough detail about surrounding winds. Short-term Climate Simulations of AEWs

24

San Diego State Univ. 28 Mar 2014

Visualization of Scale Interaction for Katrina

High-resolution animation shows that the intensification is associated with the Interaction of the Katrina’s outflow and an approaching upper-level jet stream. Short-term Climate Simulations of AEWs

25

San Diego State Univ. 28 Mar 2014

Performance of fvMMF 2.0 on Pleiades (the first climate model implemented with MPI inter- and intra- communications)

The benchmark is based on 5-day runs with standard modeling configurations for climate simulations. Key Points: 1. 2. 3. 4.

Scalable with two-level parallelism on Pleiades (distributed memory) supercomputer; A speedup of 79.8 with 3335 cores, which allows to finish a 30-day run within 41 minutes; Bit-by-bit identical results with different CPU configurations; Enabling high-resolutions and higher-dimensions for Goddard Cumulus Ensemble (GCE) model Shen, B.-W., B. Nelson, S. Cheung, W.-K. Tao, 2013b: Scalability Improvement of the NASA Multiscale Modeling Framework for Tropical Cyclone Climate Study. IEEE CiSE. no. 5, pp. 56-67, Sep./Oct. 2013

Short-term Climate Simulations of AEWs

32

San Diego State Univ. 28 Mar 2014

The Global Mesoscale Model

1.

Model Dynamics and Physics: •

The finite-volume dynamical core (Lin 2004);

•

The NCAR physical parameterizations, and NCEP SAS as an alternative cumulus parameterization scheme

•

The NCAR land surface model (CLM2, Dai et al. 2003)

2. Computational design, scalability and performance (suitable for running on clusters or multi-core systems)

Short-term Climate Simulations of AEWs

33

San Diego State Univ. 28 Mar 2014

Physics Parameterizations

•

Moist physics: -Deep convections: Zhang and McFarlane (1995); Pan and Wu (1995, aka NCEP/SAS) -Shallow convection: Hack (1994) -large-scale condensation (Sundqvist 1988) -rain evaporation

•

Boundary Layer - first order closure scheme - local and non-local transport (Holtslag and Boville 1992)

•

Surface Exchange - Bryan et al. (1996)

Pan, H.-L., and W.-S. Wu, 1995: Implementing a mass flux convection parameterization package for the NMC medium-range forecast model. NMC office note, No. 409, 40pp. [Available from NCEP]. Short-term Climate Simulations of AEWs

34

San Diego State Univ. 28 Mar 2014

Global Multiscale Modeling

Limiting factors in traditional global/regional climate models • insufficient resolutions or domain sizes  limited scale interactions; • physics parameterizations such as cumulus parameterizations Challenges • difficult to simulate the life cycle (i.e., evolution) of the “small-scale” processes; • difficult to trace/simulate the “feedback” (impact) of these small-scale flows on the weather events at the system scale

Physics parameterizations: Due to the insufficient resolution in traditional GCMs, some important physical processes such as convection are missing. Therefore, the collective effects of these unresolved physical processes are approximated or ``parameterized ‘’ by the resolved motions, and the statistical relationship between resolved motions and different unresolved processes is determined empirically by a different closure, an extra set of questions with tunable parameters.

Short-term Climate Simulations of AEWs

35

San Diego State Univ. 28 Mar 2014

Grid Cells vs. Grid Spacing

Resolution

x

y

Grid cells

1o (~110km)

288

181

52 K

0.5o (~55km)

576

361

208 K

0.25o (~28km)

1000

721

721 K

0.125o (~14km)

2880

1441

4.15 M

0.08o (~9km)

4500

2251

10.13 M

MMF (2D CRM)

144x64

90

829 K

Y2005

Y2005~2006

The 1/12 degree model with 48 vertical levels has 480 M grid points. In comparison, the hyperwall-2 is able to display 245 M pixels. Short-term Climate Simulations of AEWs

36

San Diego State Univ. 28 Mar 2014

Linear Theory of a 3D flow over an isolated mountain at a wide range of scales

scale buoyancy

hydrostatic

Coriolis

NG

QG

Resolved scale: 10km, 4dx=10km, dx=2.5km.

Shen, B.- W., 1992: The Linear Solution of a Three-Dimensional Flow over an Isolated Mountain, Master Thesis, National Central University, Taiwan, (in Chinese) p. 85 Short-term Climate Simulations of AEWs

37

San Diego State Univ. 28 Mar 2014

Genesis of Twin Tropical Cyclones

•

Shen, B.-W., W.-K. Tao, and Y.-L. Lin, and A. Laing, 2012: Genesis of Twin Tropical Cyclones as Revealed by a Global Mesoscale Model: The Role of Mixed Rossby Gravity Waves. J. Geophys. Res. 117, D13114, doi:10.1029/2012JD017450. 28pp.

•

1st submission: Summer 2007

•

1st presentation: Shen et al, 2007: Forecasts of Tropical Cyclogenesis with a Global Mesoscale Model: Modulation of Six Tropical Cyclones by the MJO in May 2002. CMMAP Team Meeting, Fort Collins, Colorado, August 7-9, 2007.

Short-term Climate Simulations of AEWs

47

San Diego State Univ. 28 Mar 2014

“Twin” TCs

MOAST

EQ

South mirroring north

0900 UTC May 9 2002

Three convective systems are: 1. 2.

ND, (non-developing), Kesiny ( 05/03/06z – 05/11/18z), and TC01A (05/06/18z – 05/10/12z).

3.

EQ

•

A timing lag of 3 days  it is challenging to predict both in a 5-day run!

•

Longitudinal differences  asymmetry

•

Transition from asymmetry to symmetry

Short-term Climate Simulations of AEWs

48

San Diego State Univ. 28 Mar 2014

Visualizations of Twin TCs in May 2002

MOAST

(vortex phasing; init at 00Z May 1) 12Z 05 May 2002

00Z 03 May 2002 (b)

(a) EQ

Kesiny E W

Moving poleward, contracting in scale, and intensifying

14Z 07 May 2002

12Z 06 May 2002 (c)

(d)

TC01A

persistent vs. impulsive forcing Short-term Climate Simulations of AEWs

73

San Diego State Univ. 28 Mar 2014

Evolution of Twin TCs and the MRG Wave

MOAST

The successive formation of multiple TCs may be associated with the appearance of a mixed Rossby gravity (MRG) wave. http://goo.gl/qXH2p

N

•

EQ

S

Shen, Bo-Wen, Bron Nelson, W.-K. Tao, and Y.-L. Lin, 2013a: Advanced Visualizations of Scale Interactions of Tropical Cyclone Formation and Tropical Waves. Computing in Science and Engineering, vol. 15, no. 2, pp. 47-59, March-April 2013, doi:10.1109/MCSE.2012.64

Short-term Climate Simulations of AEWs

74

San Diego State Univ. 28 Mar 2014

Precipitations of the Three Convective Systems

MOAST

Moistening NASA TRMM

CNTL

Precipitation (mm/day) averaged over 000 UTC May 03-06, 2002 from (a) NASA TRMM, (b) the control run which is initialized at 0000 UTC May 1, 2002. Relative errors (Ei) are 0.81, 2.35, 4.6, 24.67 for the large domain, sub-domains 1, 2, 3 (D1, D2, D3), respectively. Each of the sub-domains contains a 10ox10o box. Ei = (Pi-PTRMM)/PTRMM, where Pi (mm/d) indicates the domain average precipitation, and PTRMM is the corresponding domain average precipitation from TRMM. Short-term Climate Simulations of AEWs

80

San Diego State Univ. 28 Mar 2014

Sensitivity Experiments Case id

ICs

GSC

CPs

SST

Remarks

CNTL

05/1

x

none

weekly

EXP-A

“

x

NCAR

weekly

EXP-B

“

x

NCEP

weekly

EXP-C

“

x

Channel

weekly

no CPs in 0-40oN

EXP-D

“

x

Channel

weekly

no CPs in 0-40oS

EXP-E

05/02

x

none

weekly

Different ICs

EXP-F

05/01

x

none

weekly

Aqua planet

EXP-G

“

x

none

Climate

EXP-H

“

x

none

November

EXP-I

“

x

none

February

EXP-J

“

x

regional

weekly

Regional changes In the model configurations

GSC: grid-scale Condensation CPs: Cumulus parameterizations Short-term Climate Simulations of AEWs

83

San Diego State Univ. 28 Mar 2014

Sensitivity Experiments CNTL

NCAR CP

05/02

EXP-C

False alarms

NCEP CP

TC01A

Kesiny

EXP-D

No CPs

No CPs

• • •

The performance of a specific CP may be case dependent, (dependence of “largescale conditions”?) The spatial distribution of minimum sea level pressure (MSLP) overwith the 10-day The regional improvement (or change) in the moist processes a different CP integration, which is initialized at 0000 UTC 1 May 2002, qualitatively shows the initial may not be sufficient for improving the formation of a specific TC. location and subsequent movement of the simulated TCs. A specific CP may also affect the simulation of environmental conditions (such as the mixed Rossby gravity wave) and thus TC formation

Short-term Climate Simulations of AEWs

87

San Diego State Univ. 28 Mar 2014

Interactions of Twin TCs and MRG wave Wavelength reduction of mixed Rossby gravity (MRG) wave

MRG Wave Development

Analytical Solution

Time/longitude diagram of meridian winds from NCEP analyses (a) and the 10-day control run initialized at 00Z 1 May 2002 (b). Northerly (southerly) winds are indicated in red (blue). The westward-propagating disturbances with the sloping northerly to southerly flow couplets that are nearly asymmetric about the equator are likely associated with an MRG wave.(c) Time/longitude diagram of 850mb vertical velocity (shaded) and meridian winds (contour). 10-day Fcst of Twin TCs’ Track and Intensity

Simulations of 850-hPa zonal winds at the initial time and the integration of 96 hours.

Evolution of Low-level Cyclonic Circulation

W

E

N

S

120h simulations of zonal (left) and meridian winds (right) Scale Interactions between MRG gyres

Track (a) and Intensity simulations of TCs 01A (b) and 23S (c) from the 10-day run initialized at 00Z 1 May 2002, as compared to the observations. The first record for TC 23S (01A) was issued at 06Z 3May (18Z 5 May). Short-term Climate Simulations of AEWs

850-hPa winds (vectors) and geopotential heights (shaded) at 00Z 5 May 2002 from NCEP analyses (left) and the control run (right). 96

San Diego State Univ. 28 Mar 2014

Short-term Climate Simulations of AEWs and AEJ:

Shen, B.-W., W.-K. Tao, and M.-L. Wu, 2010b: African Easterly Waves and African Easterly Jet in 30-day High-resolution Global Simulations. A Case Study during the 2006 NAMMA period. Geophys. Res. Lett., L18803, doi:10.1029/2010GL044355. (4 figures) Shen, B.-W., W.-K. Tao, and M.-L. Wu, 2010b: Auxiliary Materials for Paper 2010GL044355. (8 figures)

Short-term Climate Simulations of AEWs

98

San Diego State Univ. 28 Mar 2014

African Easterly Waves (AEWs)

AEW

• • •

During the summer time (from June to early October), African easterly waves (AEWs) appear as one of the dominant synoptic weather systems in West Africa. These waves are characterized by an average westward-propagating speed of 11.6 m/s, an average wavelength of 2200km, and a period of about 2 to 5 days. Nearly 85% of intense hurricanes have their origins as AEWs [e.g., Landsea, 1993]. Contributed by Chris Landsea, http://www.aoml.noaa.gov/hrd/tcfaq/A4.html

Short-term Climate Simulations of AEWs

101

San Diego State Univ. 28 Mar 2014

African Easterly Jet (AEJ) (Thorncroft and Blackburn, 1999) • • • • • • •

• •

The initiation of an AEW is found to be related to the release of barotropic and/or baroclinic instability TEJ associated with an African Easterly Jet (AEJ). AEJ is a midtropospheric jet located over tropical north Africa during the northern hemisphere summer AEJ is seen as a prominent feature in the zonal wind, with AEJ a maximum of around 12.5 m/s at 600-700 hPa and 15oN. Its vertical shears are crucial in organizing moist convection and generation of squall lines; Its horizontal and vertical shears are important for the growth of easterly waves; Below the AEJ, the easterly wind shear is in thermal wind balance with the surface temperature gradient (BWS) One may view the AEJ as resulting from the combination of diabatically forced meridional circulations which maintain it and easterly waves which weaken it. As the nature of diabatic forcing (e.g., moist or dry convection) differs between models, simulated AEJ is likely to be different (e.g., different heights and/or strengths, which will subsequently affect the easterly waves - model climate The rate at which the AEJ is maintained is likely to be particularly sensitive to the boundary layer parameterization Other factors should be included: radiation; the establishment of the surface temperature and humidity gradients themselves; etc

Short-term Climate Simulations of AEWs

103

San Diego State Univ. 28 Mar 2014

An Integrated View for Climate Studies Multiscale Interaction during Hurricane Formation

From a climate perspective, it is important to understand to what extent these multiscale processes can be simulated realistically by a numerical model.

African Easterly Wave (AEW) Hurricane

African Easterly Jet (AEJ)

Surface mechanic and thermodynamic effects

Dry/moist processes

?

Horizontal temperature gradient • • •

One way interaction; downscale cascading; no or limited feedbacks from smaller-scale flows

Short-term Climate Simulations of AEWs

104

Radiation; land surface processes

San Diego State Univ. 28 Mar 2014

IVP vs. BVP

BVP

IVP

Scale

? Short-term climate/ Extended-range / Sub-seasonal Seasonal/climate simulations

Short-term forecasts

Lead Time

IVP: initial value problem BVP: boundary value problem / forced problem Short-term Climate Simulations of AEWs

105

San Diego State Univ. 28 Mar 2014

AEWs in late August 2006 •

The NASA African Monsoon Multidisciplinary Analyses (NAMMA) field campaign was launched in August 2006, providing a great opportunity to characterize the frequency of AEWs, their evolution over continental western Africa.

•

During the 30-day observation Time period between late August ↓ and late September, there were six AEWs documented that appeared over Africa, propagated westward, and then passed by the Cape Verde Islands. In early September, an observed AEW developed into a Cape Verde storm-Hurricane Helene (Brown, 2006).

•

Short-term Climate Simulations of AEWs

106

1 2

3

4 5 6

San Diego State Univ. 28 Mar 2014

Five AEWs in 30-day Simulations V winds averaged over 5o-20oN

30-day averaged U winds (10oE)

30-day averaged U winds (20oE)

22 Aug

NCEP Analysis

TEJ

AEJ

21 Sep 22 Aug

GMM

TEJ

AEJ

H 21 Sep 50oW

Longitude

40oE

30oN

Latitude

EQ

30oN

Latitude

EQ

(init at 00zz Aug 22, 2006) Short-term Climate Simulations of AEWs

107

San Diego State Univ. 28 Mar 2014

30-day Averaged U Winds and Temp NCEP Reanalysis

Model Simulations

U g (init at 00zz Aug 22, 2006) Short-term Climate Simulations of AEWs

z 108



 R T fH y

108 San Diego State Univ. 28 Mar 2014

Verification with NAMMA Observations Praia, CV @ (23.5oW, 14.9oN) NAMMA

NAMMA OBS

NCEP

NCEP Reanalysis

Model

Model simulation

false one?

Aug 22 Aug 22

Aug 22

Left panel: Schmidlin, F. J., B. Morrison, T. Baldwin, E. T. Northam, 2007: High Resolution Radiosonde Measurements from Cape Verde: Details of Easterly Wave Passage. AGU 2007 Fall Meeting.

In general, the timing and location of the simulated maximum southerly winds (indicated by a yellow color) are quite close to those from the NAMMA observations and NCEP analysis except that an additional signal (with a circle) appears on about 31 Aug. This event, which can be also identified by Zipser et al. [2009, Figure 2b shown in the next slide], is stronger than the one in the NCEP analysis and has a time lag of about 1 day. Shen, B.-W. et al., 2010b: African Easterly Waves and African Easterly Jet in 30-day Highresolution Global Simulations. A Case Study during the 2006 NAMMA period. Geophys. Res. Lett., L18803, doi:10.1029/2010GL044355. NAMMA: NASA African Monsoon Multidisciplinary Analyses Short-term Climate Simulations of AEWs

109

San Diego State Univ. 28 Mar 2014

AEW

TC

TC

Short-term Climate Simulations of AEWs

AEW

Upper-level winds in red

•

Low-level winds in blue

Figure 7: Formation of Hurricane Helene (2006) and its association with the intensification of an African Easterly Wave (AEW) in a 30-day run initialized at 0000 UTC August 22, 2006. Upperlevel winds are in pink, middle-level winds in green and low-level winds in blue. (a) Initial formation of Helene as the AEW moves into the ocean, validated at 0000 UTC Sep. 13 (day 22); (b) initial intensification associated with intensified low-level inflow with counter clockwise circulation, validated at 2100 UTC Sep. 14; (c) further intensification with an enhanced outflow with clockwise circulation (indicated in pink), validated at 2200 UTC Sep. 16. An animation can be found: http://tiny.cc/j9ul9

AEW

AEW

•

AEW

113

San Diego State Univ. 28 Mar 2014

Formation of Hurricane Helene (2006)

• • • •

http://goo.gl/arWSZ Simulations from Day 20 to Day 30 in a run initialized at 00Z Aug 22, 2006. Upper-level winds in red; middle-level winds in green; low-level winds in blue Low-level CC (cyclonic circulation); Upper-level AC (anticyclonic circulation) Shen, B.-W. W.-K. Tao and M.-L. Wu, 2010b: African Easterly Waves in 30-day High-resolution Global Simulations: A Case Study during the 2006 NAMMA Period. GRL., L18803

Short-term Climate Simulations of AEWs

114

San Diego State Univ. 28 Mar 2014

Simulations of Helene (2006) between Day 22-30 (Helene: 12-24 September, 2006)

Track Forecast

Intensity Forecast

OBS

model

OBS model

Future work: to study multiscale interactions among TEJ, AEJ, AEWs, hurricanes and surface mechanic and thermodynamic processes Shen, B.-W., W.-K. Tao, and M.-L. Wu, 2010b: African Easterly Waves and African Easterly Jet in 30-day High-resolution Global Simulations. A Case Study during the 2006 NAMMA period. Geophys. Res. Lett., L18803, doi:10.1029/2010GL044355.

Short-term Climate Simulations of AEWs

115

San Diego State Univ. 28 Mar 2014

Sensitivity Experiments Case id

Dynamic IC

Clm and Physics IC

SST

cntl

08/22

08/22

weekly

A

08/23

08/23

weekly

B

08/24

08/24

weekly

C

08/25

08/25

weekly

D

08/22

Climate clm

weekly

E

08/22

06/22

weekly

F

08/22

08/22

climate

G

04/22

08/22

weekly

H

06/22

08/22

weekly

I

08/22

08/22

weekly

Short-term Climate Simulations of AEWs

118

Guinea Highlands

Remarks

Impact of initial “perturbations”

Impact of initial land surface conditions Impact of SSTs

Impact of “physics” A factor of 0.6 in heights

Changed date to be 08/22/2006 Changed date to be 08/22/2006 Impact of terrains

San Diego State Univ. 28 Mar 2014

Sensitivity Experiments

•

Sensitivity to initial perturbations (e.g., AEJ)

Case D

•

Sensitivity to initial land surface conditions  e.g., dissipation of an initial AEJ  impact of soil moisture

•

Sensitivity to surface sea temperatures (SSTs)  oceanic feedbacks on AEW simulations; impact on large-scale flows in the upstream, subsequent atmosphereland interactions, initiation of multiple AEWs, intensification of the 4th AEW and formation of the model ‘Helene’

Case G

•

Sensitivity to physics (with realistic land surface conditions)  e.g., initial development of an AEJ

•

Impact of a reduced mountain height on the simulations of upstream flows Examining other factors (forcing) that control the evolution of the AEJ, AEW and thus hurricane formation!

Short-term Climate Simulations of AEWs

119

San Diego State Univ. 28 Mar 2014

Initial Zonal Winds

AEJ

Short-term Climate Simulations of AEWs

120

San Diego State Univ. 28 Mar 2014

Experiment D: AEJ Dissipation

Day 2

Day 3

30oN

Latitude

Day 4

EQ

Dynamic IC: 08/22 Land Surface IC: cold start run

Short-term Climate Simulations of AEWs

121

San Diego State Univ. 28 Mar 2014

Experiment G: AEJ development

Day 15

Day 20

Day 25

Day 30

AEJ

Dynamic IC: 04/22  08/22 Land Surface IC: 08/22

Short-term Climate Simulations of AEWs

122

San Diego State Univ. 28 Mar 2014

Sensitivity Experiments

Climate SST

Climate CLM

Different terrain heights

Different dynamic ICs (04/22 and 06/22)

Short-term Climate Simulations of AEWs

123

San Diego State Univ. 28 Mar 2014

Multiscale Interactions of AEJ, AEWs, Hurricane and Surface Processes in a 30-day run

Altitude

(a)

Time-averaged African Easterly Jet (AEJ)

AEJ

(c) African Easterly Waves (AEWs) Day 0

AEJ Day 30

Altitude-latitude cross section of 30-day averaged zonal winds along longitude 20oE, showing African Easterly Jet (AEJ) from GFS analysis (left) and a 30‐day simulation (right).

Time-longitude diagram of meridional winds, showing six AEWs in a 30‐day simulation. The black circle roughly indicates the timing and location of Helene’s formation.

Detection of Multiple AEWs

(d) Track and Intensity in a 30-day run

Altitude-time cross sections of meridional winds at (23.5oW, 14.9oN) from NAMMA (NASA African Monsoon Multidisciplinary Analyses) observations (left) and model simulation (right).

Track (left) and intensity (right) forecasts for Hurricane Helene from Day 22 to 30. Red and blue lines indicate model predictions and best track, respectively.

Altitude

(b)

Short-term Climate Simulations of AEWs

125

San Diego State Univ. 28 Mar 2014

The 3D and 5D Lorenz Models 1. Are the simulations of TC genesis consistent with Chaos theory? 2. Why can the high-resolution global model have skills? → For generalized LMs, under which conditions could the increased degree of the nonlinearity improve solution stability? 3DLM with r=25

5DLM with r=25

Negative Nonlinear Feedback

stable critical points

strange attractors

Anthes, R., 2011: the tables on chaos: the atmosphere more predictable The studies byTurning Lorenz laid the foundation forIschaos theory, which was viewed asthan the we assume? NCAR/UCAR third scientific revolutionAtmos of theNews. 20th century after relativity and quantum mechanics, Shen, 2013d: Negative Feedback in a as Five-dimensional Lorenz Model. J. of Atmos. andB.-W., is being applied in various fields such Earth science, mathematics, Sci. 10.1175/JAS-D-13-0223.1 (in press, DecAnthes 16, 2013). philosophy, physics, etc (e.g. Gleick, 1987; 2011). Short-term Climate Simulations of AEWs

135

San Diego State Univ. 28 Mar 2014

Lorenz Models D, H, and N refer to as the dissipative terms, the heating term, and nonlinear terms associated with the primary modes (low wavenumber modes), respectively. Ds, Hs, and Ns refer to as the dissipative terms, the heating term, and nonlinear terms associated with the secondary modes (high wavenumber modes), respectively. NLM refers to the non-dissipative Lorenz mode.

D

H

Linearzied 3DLM

V

V

3DLM

V

V

V

X c  Yc   b(r  1)

V

V

( X c , Yc )  (  2r ,0)

3D-NLM

N

Ds

Hs

Ns

Critical points for (X,Y)

rc

remarks Unstable as r>1

“1”

5DLM

V

V

V

V

6DLM

V

V

V

V

V V

V

X c  Yc ~  2b(r  1)

24.74 conservative

42.9

X c  Yc   b( Z c  2Z1c )

41.1

Shen, B.-W., 2014a: Nonlinear Feedback in a Five-dimensional Lorenz Model. J. of Atmos. Sci. in press. Shen, B.-W., 2014b: On the Nonlinear Feedback Loop and Energy Cycle of the Non-dissipative Lorenz Model. (submitted to NPGD) Shen, B.-W., 2014c: Nonlinear Feedback in a Six-dimensional Lorenz Model. Impact of an Additional Heating Term. (submitted to JAS) Short-term Climate Simulations of AEWs

136

San Diego State Univ. 28 Mar 2014

Mathematics Courses • Ordinary and partial differential equations • Complex variables • Linear Algebra • Vector, tensors • Perturbation methods • Multiscale Analysis (Bender and Orszag) • Advanced Calculus, II, III, IV, V • Numerical Methods; • Numerical Modeling • Signal processing Bender and Orszag, Advanced Mathematic methods for Scientisits Short-term Climate Simulations of AEWs

137

San Diego State Univ. 28 Mar 2014

Summary 1. We first discuss the model’s performance and the enabling roles of supercomputing technology with the simulations and visualizations of three convective systems, two of which turn into a twin TC. Sensitivities of simulations to model configurations are also examined. 2. The statistical characteristics of multiple AEWs (including initiation and propagation) are realistically simulated in short-term climate simulations. Remarkable simulations of a mean African Easterly Jet (AEJ) are also obtained. 3. While land surface processes may contribute to the predictability of the AEJ and AEWs (as a boundary value problem), the initiation and detailed evolution of AEWs still depend on the accurate representation of dynamic and land surface initial conditions and their time-varying nonlinear interactions (as an initial value problem). 4. Of interest is the potential to extend the lead time for predicting hurricane formation (e.g., a lead time of up to 22 days) as the 4th AEW is realistically simulated. 5. In the experiment with climate SSTs, differences appear in the 5th and 6th AEWs, implying that the effects of using climatological SSTs on the simulation of AEW initiation begin to occur after 15-20 days of integration. 6. The reduced height of Guinea highlands causes significant differences in the simulations of AEWs since Day 15. For example, the initiation of the 4th, 5th and 6th AEWs are influenced by this change, and the downstream development of AEWs (e.g., the 2nd and 4th AEWs) becomes weaker. Short-term Climate Simulations of AEWs

138

San Diego State Univ. 28 Mar 2014

Future Tasks

Current Project NASA AIST CAMVis: 2012-2015 MAP: Multiscale Analysis Package HHT: Hilbert Huang Transform SAT: Stability Analysis Tool

•

to what extent can large-scale flows determine the timing and location of TC genesis? MAP/HHT

•

to what extent can resolved small-scale processes impact solutions stability (or predictability)? MAP/SAT

Tropical Waves

Short-term Climate Simulations of AEWs

MAP with HHT

Tropical Cyclone Formation

139

MAP with SAT

smallscale processes

San Diego State Univ. 28 Mar 2014

EMD as Bank Filters EMD performs like filter banks (e.g., a dyadic filter) and generates IMFs each of which has features with comparable scales (Wu and Huang 2004), which indicates its potential for hierarchical multiscale analysis. The right figure displays the first 9 IMFs for the Gaussian White Noises with 220 (1 million) points, showing the characteristics of the bank filters (i.e., a dyadic filter). Here, f and T represent frequency and period, respectively.  1/ T

Small Scale

log 2( )   log 2(T )

Medium Scale

log 2( )  log 2( )  7 log 2(Tn  1)  log 2(Tn) 1 2

9

Tn  1 / Tn  2

Large Scale

Doubling of the mean period Reproduced with a different presentation by Shen Short-term Climate Simulations of AEWs



9 140



2

San Diego State Univ. 28 Mar 2014

II: Three Level Parallelism: The 3-Level parallelism is achieved with the fine-grain OpenMP inside all the N members in each M process. M Groups

. . .

mp_eemd()

1D domain decomposition

mp_eemd()

ensemble decomposition

eemd()

eemd()

N

N

…

…

…

…

…

OpenMP OpenMP Processes Processes

…

Loop decomposition

OpenMP OpenMP Processes Processes

Decomp in Ensemble

Short-term Climate Simulations of AEWs

Speedup M 2 2 4 4 25 100 100

N 1 2 2 4 4 4 16

OMP1 OMP2 OMP4 1.99 3.66 6.28 3.79 6.33 10.92 7.46 12.52 21.57 13.72 21.65 33.99 80.40 127.79 200.50 286.35 459.04 721.30 449.16 100 nodes

Multiple runs for the MRG case with 1001x1001 points and en=1000 were performed on Pleiades. Sandy processors were used; each CPU has 8 cores, and each node has 16 cores. Using 100 nodes, the MPI-OMP hybrid parallelism produces the best performance.

Decomp in Ensemble

141

San Diego State Univ. 28 Mar 2014

Decompositions of MRG wave with the PEEMD U’

WWB

Total Analytical Solutions

IMFs

Differences

September 2013

Short-term Climate Simulations of AEWs

142

San Diego State Univ. 28 Mar 2014

Summary 1. We first discuss the model’s performance and the enabling roles of supercomputing technology with the simulations and visualizations of three convective systems, two of which turn into a twin TC. Sensitivities of simulations to model configurations are also examined. 2. The statistical characteristics of multiple AEWs (including initiation and propagation) are realistically simulated in short-term climate simulations. Remarkable simulations of a mean African Easterly Jet (AEJ) are also obtained. 3. While land surface processes may contribute to the predictability of the AEJ and AEWs (as a boundary value problem), the initiation and detailed evolution of AEWs still depend on the accurate representation of dynamic and land surface initial conditions and their time-varying nonlinear interactions (as an initial value problem). 4. Of interest is the potential to extend the lead time for predicting hurricane formation (e.g., a lead time of up to 22 days) as the 4th AEW is realistically simulated. 5. In the experiment with climate SSTs, differences appear in the 5th and 6th AEWs, implying that the effects of using climatological SSTs on the simulation of AEW initiation begin to occur after 15-20 days of integration. 6. The reduced height of Guinea highlands causes significant differences in the simulations of AEWs since Day 15. For example, the initiation of the 4th, 5th and 6th AEWs are influenced by this change, and the downstream development of AEWs (e.g., the 2nd and 4th AEWs) becomes weaker. Short-term Climate Simulations of AEWs

143

San Diego State Univ. 28 Mar 2014