A Super-Regional Testbed to Improve Models of Environmental Processes on the U.S. Atlantic and Gulf of Mexico Coasts

yieldingrabbleInternet και Εφαρμογές Web

7 Δεκ 2013 (πριν από 3 χρόνια και 4 μήνες)

272 εμφανίσεις

DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




1



A Super
-
Regional Testbed to Improve Models of Environmental Processes on
the U.S. Atlantic and Gulf of Mexico Coasts

Principal Investigator: L.D. Wright, Director of Coastal and Environmental Research, SURA

wright@sura
.org


Award Number:

NA10NOS0120063


The goal of this
IOOS
-
funded project

is to create a multi
-
disciplinary, community
-
modeling testbed
that will facilitate improved operational coastal ocean prediction. The testbed will allow scientists to
share models, o
bservations, and tools needed to elucidate, prioritize, and resolve issues associated with
interoperable coupling of models. The goals include improving and evaluating models already in
operational use within the Atlantic and Gulf Coast super
-
region, devel
oping and assessing metrics for
model skill and system performance, and facilitating the transition of models from research to
operations. Pilot testbed projects will address three chronic issues of high relevance within the super
region:
(1) coastal in
und
ation (2) estuarine hypoxia

and (3) shelf hypoxia.
An overarching theme will
involve the design and implementation of cyberinfrastructure (CI) to support the three issue
-
motivated
science themes. From the outset, an unstated goal was for this project to b
ecome a showcase for
federal
-
academia
-
industry collaboration. As the team has come together, it is apparent that we are well
on tra
ck to realizing that goal. The t
estbed involves 20 universities, 11 federal centers and programs and
2 private corporations.

This report will highlight the accomplishmen
ts, progress and challenges of each of the t
estbed teams
and the SURA management activity

for the period of August 2010

and is organized as follows:

1.

SURA Management

2.

Cyberinfrastructure Task

3.

Shelf Hypoxia Modeling

in the Gulf of Mexico

4.

Estuarine Hypoxia Modeling in Chesapeake Bay

5.

Coastal Inundation Modeling in the Gulf of Mexico

and the Gulf of Maine


Appendix A: Testbed High Level Deliverables

Appendix B: TAEG Comments on Testbed Leadership telecon on 31 August 20
10



DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




2


1.

SURA Management


SURA is providing overall project leadership, multi
-
institution coordination and management of the
testbed and is insuring that federal funds are expended in a responsible and accountable manner. SURA
has issued subcontract agreemen
ts to our test bed partners, and has communicated the expectation
that invoices are to be submitted on a monthly basis. SURA has been monitoring, approving and paying
the invoices as they are submitted through SURA’s financial and accounting services.

A

dedicated SURA
Linux
server has been set up and is
rapidly
being populated with
examples of
model
output

for the CI team to assess, and to use in their
development of a catalog

of models
. A
collaborative web site has been created by CI team
-
member, Univ
ersity of Alabama Huntsville, to
facilitate communication and information exchange:
http://testbed.sura.org
. SURA continues to work

closely with the website administrators and developers to provide feedback on the usa
bili
t
y of the site.


SURA has been facilitating and participating in all modeling team and CI teleconferences, which have
been occurring regularly, and has recently made available a WebEx

for on
-
line meetings. SURA, with the
team leads and Testbed Adviso
ry and Evaluation Group (TAEG) has developed

a suite of high
-
level
testbed deliverables, or milestones, intended for a broad audience to highlight, in non
-
technical terms,
what will be the major outcomes of the testbed.
These deliverables are found in an
Appendix A.



SURA
is convening
monthly teleconferences among the modeling team and CI leaders

(Friedrichs,
Harding, Howlett, Luettich)
, and the Testbed Advisory and Evaluation Group (TAEG) to discuss progress
and challenges, and to give the TAEG an oppor
tunity to ask questions and provide feedback and
guidance. Subsequently, the TAEG held a teleconference during which they indicated their general
satisfaction wi
th the initial progress of the t
estbed.
Appendix B contains the notes from the TAEG
comments

on the Testbed Leadership conference help on 31 August 2010.



SURA IT Progress Report Input


The
primary

focus area for SURA IT staff this past month
was increasing the pool of high performance
computational resources available to the project teams. The
model teams indicated that their current
resources were insufficient to handle large scale data sets and complete model runs in a timely manner.
In response, SURA p
olled the
testbed groups regarding their need for additional computational
resources.

T
he in
undation modeling group
responded with specific requirements. Members of the
Inundation Team expressed the need for access to computational capacity beyond that available at their
local institutions included;
Harry Wang (VIMS), Chunyan Li (LSU), Lianyuan Z
heng (USF) and William
Perrie (Bedford Institute of Oceanography)
.


To support the needs of the inundation team SURA obtained
startup allocations on
a large shared HPC
system located at LSU (
Queen Bee
) by working directly with staff at LSU and the Louisia
na Optical
Network Initiative (
LO
NI). In addition a TeraGrid startup allocation was obtained on the Ranger
cluster at
Texas Advanced Computing Center (
TACC).
Ranger is
one of
the largest computing system
s

in the world
for open science research.

To date all

four researchers from the inundation team have gained access to
and are using the resources at LONI, including Queen Bee. We are in the process of establishing TeraGrid
user accounts for the same four inundation team members so they will be able to transi
tion to the larger
allocations on TeraGrid when necessary. After only a couple of weeks of access the inundation team has
consumed about half of the initial LONI startup allocation of 50,000 SUs. In anticipation of the
inundation group’s continued need for

HPC resources, SURA
submitted a proposal for a large allocation
DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




3


of 200,000 SUs on
the LONI
Queen Bee

system
. This request
is now in review

by the LONI allocations
committee, and if accepted, this allocation will be available starting October 1
st
.

The req
uest was
coordinated with
Rick Luettich
and Liz Smith to ensure the description of the science, models and
methodology descriptions were accurate.


SURA IT staff members are also working with the project CI Team and participants from the University of
Alab
ama Huntsville to help ensure that the model data and tools that are being stored on project
testbed servers will be accessible through the Testbed collaboration website.



DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




4


2.

Cyberinfrastructure (CI) Testbed


Progress as of 9/7
/2010

(Submitted by Eoin Howl
ett on behalf of the CI Team)

General

1.

Work has been focused on identifying the data that will be coming from the model teams and
aggregating the model results on the Sura server. The Sura Linux server was made available and we
have been populating the serv
er with data and software.

2.

Some model data was still being sent to the older server, we confirmed that data needs to be
uploaded to the Linux server
(
http://testbed.sura.org/node/100
)

host: testbedapps.sur
a.org

user: sura_ftp

pass: 3ur4_7tp

port: 22 (sftp)

Estuarine Hypoxia

1.

Completed translation of
the
20 year Ch3d hindcast for the Chesapeake Bay in binary format
into a CF compliant NetCDF standard. The data was sent to Hortensia at LSU and has been
posted
on the Sura server and is available via TDS and ncWMS.


Ch3d NetCDF data published with TDS/ncWMS

DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




5


2.

Status of model data on testbed server

Model

Source

Sample on SURA
server

Information on sample output

Notes

Estuarine
Hypoxia





CBOFS2

NOAA

NetCDF





ChesROMS

UMCES

NetCDF





CH3D

USACE

NetCDF





EFDC

VIMS

ASCII

http://testbedapps.sura.org/thredds/fileS
erver/efdc/Readme.txt

Dave Forrest
working on
converter, will be
similar to elcirc.


Inundation



1.

Status of model data on testbed server

Model

Sour
ce

Sample on SURA
server

Information on sample output

Notes

Inundation





WW3 (3
grids)

BIO

WW3 Format

http://testbedapps.sura.org/thredds/fileS
erver/bio_ww3/ww3_extratropical/READ
ME_Noel_Hindcast.doc


WW3 can
produce Grib


can they turn it
on? Howard
to
check.

SWAN also?

SLOSH

UF

None



Don Slinn is
contact, Eoin will
contact.

FVCOM
-

Suituate

UMASS

NetCDF

http://testbedapps.sura.org/thredds/fileS
erver/umass_fvcom/fvcom_extratropical/
read_me_file.txt



ELCIRC

VIMS

NetCDF



Grid for SURA will
be pur
e triangles.
Sample is mixed
tri/quad.

SELFE

VIMS

None

Dave Forrest working on adapting the
ELCIRC => NetCDF translator for use with
SELFE

Check with
Battista’s group,
they have tools.

ADCIRC

UND

ADCIRC Format v49

http://testbedapps.sura.org/thredds/fil
eS
erver/und_adcirc/adcirc_tropical/adcirc_r
eadme.txt

Howard to
contact UND.


Check if running
coupled SWAN
and check that
NetCDF can be
created

FVCOM

USF

Unknown Format

Listing of provided output files here:

Justin to contact,
DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




6


http://testbedapps.sura.org/thredds/catal
og/usf_fvcom/fvcom_tropical/catalog.ht
ml

we need NetCDF.







2.

Resource Issues. The inundation team is planning to run models on several comp
utational grids,
including a Gulf of Mexico grid, which is order millions of nodes. Running the individual models on a
grid this complex requires more powerful cluster resources than many of the team members have
readily available. The inundation team mem
bers, with the assistance of SURA staff, are in the
process of identifying and provisioning additional computational resources. In addition there has
been some discussion of simplifying the computational grids used. The CI team is monitoring these
efforts

and will attempt to lend expertise and experience as possible and needed.

3.

We have completed the initial stage of the model output file format assessment for the inundation
models. As seen from the table above, many of the inundation team models do not sup
port NetCDF
format natively. Based on this analyses, the CI team will develop a set of tasks for converting model
output to a standard format for comparison.

Shelf Hypoxia

1.

Status of model data on testbed server

Model

Source

Sample on SURA
server

Informati
on on sample output

Notes

Shelf Hypoxia





NCOM
-

AMSEAS

NAVOC
EAN

NetCDF

http://edac
-
dap.northerngulfinstitute.org/thredds/ca
talog/ncom/amseas_download/catalog.h
tml



FVCOM

NOAA
/
NOS

None



NOS running
experimentally


Frank Aikman

POM

NOAA
/
NOS

None



Rich Patchen’s
Model



available
on TDS?

HYCOM

NRL

NetCDF
-

One file per
variable

http://tds.hycom.org/thredds/GOMl0.04
/expt_20.1.html



NCOM
-

IASNFS

NRL

NetCDF
-

One file per
variable

http://edac
-
dap.northerngulfinstitute.org/thredds/ca
talog/iasnfs/cat
alog.html



NCOM
-

Gulf of
Mexico

NRL

NetCDF

http://edac
-
dap.northerngulfinstitute.org/thredds/ca
talog/ncom_gomex/catalog.html



ROMS

TAMU

NetCDF

http://megara.tamu.edu:8080/thred
ds/ncss/grid/gcoos/hypoxia/dataset.


DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




7


html

Other

1.

Rich Signell to meet at AS
A on Sept 8
th

to discuss unstructured grid implementation

2.

Eve Devaliere (from Jeff Hanson’s office) will be at ASA Sept 7
-
10 to work with ASA staff on
implementation of skill assessment toolbox with NetCDF/TDS access. The initial plan for the
Matlab
-
based
IMEDS toolbox is:

a.

Select observation data the same way as IMEDS currently does

b.

Select the model we want to validate


this would be linked to the url of the
corresponding opendap server


we will start with CH3D water
-
level

c.

Based on the observation data, I
MEDS would identify the the I,J grid cells from the
model grid and give the user the option to pick the desired time span within the model’s
time

d.

Extract and download the time
-
series from the grid and convert it into wavedat file

e.

Process as usual and produ
ce metrics

3.

Web Site/Catalog sub
-
group (Eoin, Charlton, Eric, Kyle) had con
-
call on web site and catalog,
notes below. Focus on re
-
use of Ted Haberman (NOAA) work to manage TDS catalogs and
metadata and portal work using GeoEXT. Preliminary portal is avail
able at
http://testbedwww.sura.org/model/

-

notes attached as appendix.


DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




8


Sample Model Testbed site using SURA testbed catalog

4.

TDS software configured and operational on SURA Server and is live at
http://testbedapps.sura.org/thredds/catalog.html

The following data sets are TDS compatible:

http://testbedapps.sura.
org/thredds/catalog/samples/catalog.html

ERDDAP has been

configured and operational on SURA Server and is live at
http://testbedapps.sura.org/erddap/


Next Month
:

5.

Complete model catalog with closure on
what models will be included. For models that do not
have output in “IOOS standards”, develop plan to implement that.

6.

Skill assessment meeting (Sept 7
-
10) and implementation

7.

Work on observation catalog for comparisons/skill assessment

8.

Work on web site/liv
e catalog

9.

Unstructured Grid work plan (meeting Sept 8)

10.

Evaluate Ramadda from Unidata

11.

Work on tools to make non
-
compliant model data compliant


Challenges:

1.

Lack of clarity on:

a.

Whether all models will be on SURA server, some distributed?

b.

Will all model data
be in gridded NetCDF, some discussion that some models may just
produce time series output at certain points

c.

Where the observation data will come from for comparison and assessment

2.

Finalize how tools and model preview web site will be connected to Drupal
collaboration site.

DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




9



3.

Shelf Hypoxia Modeling



Progress as of 8/26
/2010

(Submitted by John Harding on behalf of the Shelf Hypoxia Team)

Overall Project Management



John Harding (Lead)

Tasks from work plan
:

-

Oversee sub
-
project task progress & milestone c
ompletion

-

Provide monthly status report for SURA management


Progress/ Accomplishments

(August 2010):

August Shelf Hypoxia Team telcon conducted with review of sub
-
projects and identification of
necessary action items by sub
-
project team leads

Planned Wo
rk (September 2010)

based on work plan

Work with sub
-
project team leads to monitor September action items

Conduct September Shelf Hypoxia Team telcon


Issues/ Challenges

See Hetland post
-
doc issue identified under sub
-
project 1 & 2

See Wiggert student issu
e identified under sub
-
project 2


Sub
-
project 1



Assess hydrodynamic skill


Rob Hetland (Lead)

Tasks from work plan
:

-

Identify / assemble physical evaluation data (Due 31 Aug)

-

Compile/organize/store data (Due monthly beginning July 31, ending Februar
y 28)

-

Provide NGOM POM hindcast output (Due 31 July)


Delayed to September

-

Provide IASNFS NCOM hindcast output (Due 31 July)


Delayed to September

-

Provide HYCOM GOM hindcast output (Due 31 July)


Task Complete

-

Couple Shelf ROMS to NGOM POM (Due
30 Sep)


See Challenges/Issues

-

Couple Shelf ROMS to NGOM POM (Due 30 Sep)


See Challenges/Issues

-

Couple Shelf ROMS to IASNFS (Due 30 Sep)


See Challenges/Issues

-

Couple Shelf ROMS to GOM HYCOM (Due 30 Sep)


See Challenges/Issues

-

Run uncoupled Sh
elf ROMS (Due 30 Sep)


See Challenges/Issues

-

Evaluate/ develop physical skill assessment tools (Due August 31, November 30, December 31)


Progress/ Accomplishments

(August 2010):

Identify/compile/organize/store of physical evaluation data continues.

Ha
rding & Ko met to discuss 2004
-
2009 IASNFS transfer as soon as new EDAC / OceanNOMADS
server operational.

New EDAC/ OceanNOMADS servers delivered to MSU week of August 23.

DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




10


Hetland postdoc identified and hired to conduct coupling and hindcast runs. Exp
ected start date
mid
-
October


Planned Work (September 2010)

based on work plan

Matt Howard will contact John Lehrter to gain access to EPA data

Matt Howard will provide Bruce Lipphardt pointers to model output and data access

Rob Hetland will provide sampl
e shelf hypoxia output to Matt Howard for copy to SURA
THREDDS server

Eugene Wei contacting Rich Patchen for September delivery of 2004
-
2009 NGOM POM to Matt
Howard for to conversion NetCDF/OPeNDAP and storage on SURA server.

Harding/ Ko will undertake tr
ansfer of 2004
-
2009 IASNFS from NRL to OceanNOMADS server.

Rob Hetland & Katja Fennel to confer and determine work plan modifications based on Hetland
post
-
doc issue identified below.

Matt Howard to contact Rich Patchen to arrange for transition of CSDL mo
del evaluation tools to
SURA server.

Issues/ Challenges

Original Hetland post

doc hire fell through. New post
-
doc identified and hired but won’t start
until mid
-
October. This will likely delay model coupling and hindcast runs originally scheduled for
com
pletion by end of November and may impact schedules of Subprojects 1 & 2. Task 1 & 2 leads will
confer on work plan modifications to accommodate this delay.

High Priority Items Identified for Year 2


Use of larger domain shelf hypoxia grid for potential y
ear 2 hindcasts


Sub
-
project 2



Assess ecosystem model skill


Katja Fennel (Lead)

Tasks from work plan
:

-

Identify / assemble hypoxia
-
specific evaluation data (Due 31 Aug, monthly thereafter until February 28)

-

Compile/organize/store data (Due monthly
beginning July 31, ending February 28)

-

Evaluate/ develop biogeochemical skill assessment tools (Due August 31, November 30, December 31)


Progress/ Accomplishments

(August 2010):

We performed two simulations for the Texas
-
Louisiana shelf domain with our
biogeochemical
model. The simulations use climatological boundary conditions, and we refer to the first of these as the
baseline. The only difference to the second simulation is in the wind forcing. While the baseline uses
high
-
frequency winds form a buoy
in the Gulf, the second uses NARR winds. Chlorophyll
-
based metrics
were calculated for both (see below).

We acquired 4
-
km monthly mean SeaWiFS and MODIS
-
Aqua chlorophyll products from the
NASA Goddard Server and interpolated these to our model grid. Based
on these data we calculated a
time series of metrics: the ratio of standard deviations of model and satellite, bias, root
-
mean
-
square
error, and model efficiency (or skill) for the two simulations describe above (see example in Figure 1).

DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




11


We began compilin
g other available data, specifically dissolved oxygen (DO), and participated in
a discussion via e
-
mail and telecon, which involved Russ Beard at NOAA NODC. We will continue to
interact with this group in our efforts to compile relevant data. An overview o
f the stations with DO
currently available to us is given in Figure 2.


Figure 1: Metrics based in satellite chlorophyll. 050 refers to the baseline simulation. 051 uses NARR
winds instead.

DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




12



Figure 2: Number of stations with dissolved oxygen (DO) measur
ements available to us thus far.
LUMCON refers to data from Nancy Rabalais’ group. MCH refers to data from Steve DiMarco’s projects.


Planned Work (September 2010)

based on work plan

Matt Howard will contact John Lehrter to gain access to EPA data (See als
o sub
-
project 1)

Katja Fennel, with team, will focus on determining additional metrics that are more directly
related to hypoxia, including hypoxic area, based on the data currently available and will continue in our
data compilation efforts.

Eugene Wei wi
ll contact Rich Patchen about getting the hindcast COAMPS fields to Matt for
translation to THREDDS and inclusion on the SURA server

Katja Fennel and DOng Shan Ko will correspond on what he has available and how to best
conduct their hypoxia model compari
son.

Jerry will work with Scott Cross (NCDDC) on test case getting Alan Shiller's DO and other data
onto NODC site.

Katja Fennel will contact Ruoying He about using his biogeochemical model output for our
boundary conditions (biogeochemical variables only)
.

DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




13



Issues/ Challenges

The delay in hiring a post
-
doc for Hetland’s group will affect the ecosystem skill component of
this project if it delays the provision of physical boundary conditions.


High Priority Items Identified for Year 2


None at this time.



Sub
-
project 3



Evaluate/transition pre
-
operational AMSEAS NCOM



Jerry Wiggert (Lead)


Tasks from work plan
:

-

Identify / assemble real
-
time ocean & atmospheric evaluation data (Due 31 Aug, monthly thereafter
through May 31)

-

Provide real
-
time AMSEAS

NCOM output (Due June 30, monthly thereafter through May 31)

-

Store NAVO tools and accumulate AMSEAS on NGI/NCDDC OceanNOMADS (Due July 31, monthly
thereafter through 31 May)

-

Evaluate/develop additional real time skill assessment tools (Due 30 November
)

-

Evaluate AMSEAS (Due 31 March)

-

Evaluate AMSEAS COAMPS forcing (Due 31 March)


Progress/ Accomplishments

(August 2010):

New EDAC/ OceanNOMADS servers delivered to MSU week of August 23.

Continue
in situ

data compilations from NAVO (glider, ctd, xbt)
. In hand are 17
-
Feb
-
2010
through 12
-
Aug
-
2010

Continue accumulation of pre
-
operational 3 hrly AMSEAS NCOM 4 day forecasts on NGI/NCDDC
EDAC in OPenDAP format beginning 25 May.

Past OPTEST Reports and preliminary analysis of AMSEAS
-
NCOM provided by Frank
Bub include
a number of metrics examples that are the starting point for tools to apply.

These include:


T and S comparisons at depths (10, 50, 100, 200, 500 and 1000 meters).



T and ∆S (i.e., model
-
obs) at these depths.


T and S scatter plots of model
vs. obs .


Acoustic properties



Sonic Layer Depth (SLD)



In
-
layer sound speed gradient (ILG)



Below
-
layer sound speed gradient (BLG)



SLD Cut
-
off Frequency (COF)


Additional metrics under consideration



Taylor diagrams



Target diagrams


Assessment
metrics being employed by Estuarine Group

DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




14


Planned Work (September 2010)

based on work plan


NGI staff & Harding will set up new EDAC/ OceanNOMADS servers

Accumulation of Assessment Data


Wiggert will coordinate ongoing acquisition of NAVO compilations w/ F
rank Bub.

Wiggert will continue accumulation of remote sensing data streams (SST and SSH)

Wiggert will work with Howden (USM) to obtain CODAR and CenGOOS mooring data

Develop Metrics Toolkit for AMSEAS evaluation


Wiggert will continue effort to identify u
seful metrics for model skill evaluation

Wiggert will connect with Bruce Lipphardt for his input on useful assessment metrics to include

Develop Facility with MAVE/PAVE tools

Wiggert will continue to familiarize with this Matlab
-
based toolkit and bring pro
grammer (J.
Davis) up to speed. As needed, Frank Bub will be tapped for assistance

Coordinate w/ Scott Cross on test case for moving MAGMIX (Alan Shiller) cruise data on to NODC
repository

Wiggert/Shiller will evaluate how this process works and suggest a
ny ways process may be
simplified. Metadata aspect of transferring data to NODC/NCDDC is a particular issue to assess.

Issues/ Challenges

Initially, we expected that Wiggert's student (B. Denton) would be primary Matlab programmer.
However, she has obtaine
d separate funding for her PhD studies so an alternative was needed. Need
thus developed to identify individual to take on programming tasks. Role of research associate, Jamie
Davis (USM) will be expanded to be primary programmer for USM's effort. He had

already been slated
for serving in a more advisor/assistance role. Bringing Jamie in as lead is probably a more appropriate
use for the model assessment activity assigned to USM. It will promote more timely generation of
results and he works closely with
Howden on the CODAR array and CenGOOS mooring. Availability of his
time has been confirmed.


High Priority Items Identified for Year 2


Identification and archiving of NAVO, NOAA, and other Deepwater Horizon data for use in any
Year 2 proposal


DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




15


4.

Estuarine

Hypoxia



Progress as of 9
/3/2010

(Submitted by Carl Friedrichs on behalf of the Estuarine Hypoxia Team)

Estuarine Hypoxia Update,
draft version September

3
, 2010

Progress

in August
:


--

In collaboration with the CI team, the full
CH3D

output for
Chesapea
ke Bay hydrodynamics for 1985
-
2005
has been
converted

to NetCDF
.


--

The
CBOFS2

synoptic hindcast for 2004
-
2005 Chesapeake Bay hydrodynamics has been submitted to
the SURA server in NetCDF format, including the model grid, initialization fields and forcing
s.

--

ChesROMS

forcing and model output (3D fields daily and 128 stations hourly) for 1991
-
2005 has been
submitted to the SURA server in NetCDF format.

--

EFDC

model simulations for Chesapeake Bay hydrodynamics for 2004 have been submitted to the
SURA ser
ver in ASCII format.

--

A
model intercomparison

of base runs for CH3D, CBOF2, ChesROMS and EFDC for salinity at 28
stations in the Chesapeake Bay for all of 2004 has been completed and presented to the Estuarine
Hypoxia Team in the form of an REU (NSF Rese
arch Experience for Undergraduates) report. Systematic
model
-
data mismatch was found in all models as a function of station latitude and/or salinity suggesting
oceanic and/or riverine forcing can be improved.

--

UMCES ROMS

has completed hydrodynamic simula
tions for 2004
-
2005 (output is NetCDF).

--

A 100K+
unstructured grid

of the Chesapeake Bay has been completed for use with SELFE and/or
FVCOM.

-
--

Initial 2004
-
2005 runs of
ChesROMS

for
hypoxia

with realistic physical forcing have been completed
with a 1
-
t
erm temperature
-
dependent oxygen respiration formulation.

-

Observational

data

of for various Chesapeake Bay
parameters

(e.g., S, T, TSS, O2, nutrients


~20
parameters) at EPA monitoring stations
from 1991 to 2006 have been processed and submitted to
the
SURA server.

--

A software engineer at University of Boulder
(Jisamma Kallumadikal)
has been identified to support
“plug
-
and
-
play”
interoperability

of ROMs
-
based models with multiple grids, forcings and material flux
and fate choices.

--

Finalized date fo
r all
-
hands
Team Meeting

at VIMS


October 1, 2010.

DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




16





Figure 1. Unstructured Grid of the Chesapeake Bay for SELFE, FVCOM



DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




17


Challenges
:


--

Delays in transfer of funds to subcontracts has slowed initiation of work at Rutgers
.

--

Time required to hire new
staff at VIMS (post
-
doc beginning in October) and CSDMS (software
engineer beginning in September) has slowed progress somewhat.


Look ahead to next month

--

Post team
-
reviewed copy of base
-
line intercomparison report to SURA website.

--

In collaboration w
ith the CI team, the NetCDF version of
CH3D output for
Chesapeake Bay
hydrodynamics will be transferred to LSU for SURA project on
-
line posting.


--

A new
skill assessment metric

for vertical salinity stratification is being developed based on
Average
Disc
rete Frechet Distance (ADFD)
.

--

We will begin work advancing interoperability
of multiple ROMS grids of Chesapeake Bay (e.g.,
CBOFS2,

ChesROMS, UMCES ROMS, CHIMP) linked with multiple physical forcing data sets (e.g., winds,
riverine input) and multiple c
hoices for material flux and fate subroutines (e.g., simplistic to more
complex oxygen models, Lagrangian tracking


c.f. oil spills).

--

Calibration of the unstructured SELFE model for Chesapeake Bay hydrodynamics will begin.

--

Initial model
-
data compari
son will begin for MAB hydrodynamic models (e.g., ESPreSSO, HyCOM,
NCOM).

--

Continue to m
ove forward with improved/standardized forci
ng (to be led by L. Lanerolle) and use
standardized forcing to run next generation of intercomparison.

--

Continue to move

forward with initial model
-
data comparison for CB oxygen from hypoxia models
(ChesROMS, ICM, EFDC).





DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




18



5.

Coastal Inun
dation Modeling Progress as of 9/
7/2010

(Submitted by Rick Luettich on behalf of the Testbed Inundation Team)

General

1. UNC


SURA cont
ract in place

2. Subcontracts between UNC and UND, USF, UMassD are complete. Subcontracts between UNC and
LSU, UF nearly complete (hopefully done this week). Substantial disagreement remains in contract
language in subcontract between UNC and BIO. UNC con
siders BIO to be a subcontractor. Gov of Canada
considers BIO to be a collaborator. BIO is looking for a path forward.

3. SURA has secured computer time on LONI and more recently TeraGrid computers. Several PIs have
attempted to use LONI machines with par
tial success (Zheng


USF, Wang


VIMS, Toulany


BIO, Li


LSU). Gary Crane indicated that only Zheng’s usage has shown up on SURA account. Requested that
others email him their login ids so he can check on their accounts.

4. Provided information on vari
ous model outputs to CI group to assist with model file conversion
efforts.

5. Inundation group has been using the testbed archive at sura.org. The CI team has set up a new
archive and will provide access information for the new archive and migration of t
he data on the existing
archive site to the new archive.

6. Mid
-
year inundation group meeting being planned in Chapel Hill, NC. Preliminary discussion (9/7/10
call) identified week of Dec 5 as good.


Skill Assessment

USACE


Progress to date

• Presente
d capabilities of IMEDS as a skill assessment software package


important to have well
defined methods, since often subtitles often exist in typical metrics.

• Provided a draft white paper on IMEDS capabilities as a contribution to an overall skill asses
sment
requirements white paper for our group.


USACE


Next Steps

• Eve Devalier will visit with Eoin Howlett at ASA this week to initiate the development of a common
NetCDF interface for IMEDS. It is anticipated that the role out of IMEDS for use by the

group will wait
until after the development of the common NetCDF interface plus conversion routines from each model
output format to NetCDF.

• Eve (eve.devaliere@unc.edu) will provide IMEDS to any team member who is ready to use it. Currently
it supports

ADCIRC & SWAN formatted ascii files.

• Bob Beardsley (WHOI) has agreed to take a first stab at drafting an overall skill assessment document.


USACE


Challenges

• Lack of standard model file formats

• Better definition of skill metrics required for p
roject

DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




19


Extratropical storms

Group


Progress to date

• Scituate Harbor was chosen as the primary response focus area. This area responds strongly to
regional extratropical storms, has been identified by the area NWS Weather Forecast Offices as a high
pr
iority area for storm surge/inundation modeling and has a high resolution grid that was developed for
the companion IOOS NECOFS project.

• April 17, 2007 “Patriot’s Day” and May 24, 2005 Nor’easter storms were selected for study because
they significantly

impacted the region and they are recent enough that our current topography and
bathymetry may be valid during these times.

• July 10


August 10, 2010 chosen as a model validation period because a water level gauge has only
recently been installed in Sci
tuate Harbor and does not date back to 2005 or 2007.


Group


Next steps

• Agree on skill metrics for model


data and model


model inter
-
comparisons. Bob B. has agreed to
initiate a white paper on this for distribution to the group and discussion


Gro
up


Challenges

• File interchangeability including structure of grid files and use of NetCDF.


UMassD/WHOI


Progress to date


Scituate FVCOM inundation grid
and basic bathy/elevation data uploaded to testbed archive


Regional Gulf of Maine FVCOM
-
SW
AVE grid
and basic bathymetry data uploaded to testbed archive


FVCOM v3.1 source code


code uploaded to testbed archive


FVCOM v3.
2
-

FVCOM
-
SWAVE coupling problem found and corrected, code is currently being tested
and will be provide to the group in

September.


July 10
-
August 10, 2010 validation run


Forcing (tides and winds), boundary, and initial conditions for Scituate grid and FVCOM results uploaded
to testbed website together with a pdf format report for the explanation and instruction. Based

on a
visual time series plot, the response appears to be primarily tidal.


April 2007 storm


Regional Gulf of Maine grid
-

forcing and boundary conditions, uncoupled FVCOM and SWAVE results for
month of April 2007, pdf report including model


data com
parisons uploaded to testbed website.

Scituate grid
-

forcing and boundary conditions, FVCOM results, pdf report uploaded to testbed website.


May 2005 storm


Regional Gulf of Maine grid


uncoupled FVCOM and SWAVE runs completed, forcing and boundary
c
onditions, results will be uploaded to test bed archive this week

Scituate grid


FVCOM runs completed, forcing and boundary conditions, results will be uploaded to test
bed archive this week


Scituate ground truth data
-

High water level at Ball’s hous
e near Scituate lighthouse on a high hill at
entrance to harbor


2005 storm: 23 ft; 2007 storm: 25.25 ft above ground level. Need method to
convert to height above local datum.

DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




20


UMassD/WHOI


Next steps

• Try to identify a means for obtaining a datum cor
rection for high water marks

• Continue to look for any additional data

• Re
-
run 2005 & 2007 storms for coupled FVCOM


SWAVE using FVCOM 3.2.

• Distribute FVCOM 3.2 to group and continue to assist LSU and USF with implementation


UMassD/WHOI


Challen
ges

• Observational data for skill assessment in Scituate


BIO


Progress to date

• Installing latest available version of unstructured SWAN. Can compile on LONI machine (Louie) but will
not run


MPI error. Will switch to version of unstructured SWAN t
hat UND is running. UND has online
instructions and will help to compile/run.

• Evaluating how to convert provided FVCOM grid files (with bathy) and wind files into formats used by
WaveWatch 3 and unSWAN

• TS Earl passage provided opportunity to evaluate

WW3 performance for a tropical cyclone.


BIO


Next steps

• Convert regional Gulf of Maine grid to SWAN (ADCIRC) & WW3 formats

• Convert 2007 storm wind forcing to SWAN & WW3 formats

• Run 2007 storm


BIO


Challenges

• Getting unstructured SWAN to
run on multi
-
processor machine.

• Negotiating subcontract between BIO and UNC.


UF SLOSH


Progress to date

• Received extratropical files/code from NOAA MDL


UF SLOSH


Next steps

• Convert FVCOM wind fields (open BCs?) into format for use in SLOSH e
xtratrop.

• Run 2005, 2007 storms

• Communicate with MDL to determine whether archived SLOSH results exist for the 2005 & 2007
storms.


UF SLOSH


Challenges

• Key person at MDL, Arthur Taylor, is on detail during the month of September and is not able

to assist
with testbed activities.


UND


Progress to date

• Scituate FVCOM grid converted to ADCIRC grid. Is not stable for tidal runs using M2 forcing only from
ADCIRC tidal database. Questions about irregular bathymetry in grid and run parameters.


DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




21


UND


Next steps

• Continue to work on getting stable tide runs on Scituate grid.

• Convert FVCOM Gulf of Maine Grids to ADCIRC format

• Convert FVCOM wind fields (open BCs) into format for use in ADCIRC.

• Run 2010 validation, 2005, 2007 storms


UND


Challenges

• Stability of Scituate grid.


VIMS


Progress to date

• Converted Scituate grid into SELFE grid.

• Performed test tidal run for Scituate for April 2007 and compared to tidal time series for that period
reconstructed from observationally b
ased tidal harmonics.


VIMS


Next steps

• Convert FVCOM wind fields (open BCs) into format for use in SELFE.

• Run 2010 validation, 2005, 2007 storms


VIMS


Challenges

• Ability to include waves?


DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




22


Tropical storms

Group


Progress to date

• Wester
n Louisiana, eastern Texas region selected for test bed due to extensive surge and inundation
caused by Hurricanes Rita 2005 and Ike 2008 in this region and the availability of substantial
observational data sets for model validation

• Selected Rita 2005
and Ike 2008 as study storms

• Agree to try to work with a high resolution western LA/eastern TX grid containing ~2million nodes/~4
million elements

• All groups have successfully run this grid for tides. Compute intensive, large files especially from
FV
COM which only has global output.


Group


Next steps

• Fall back to smaller grid covering only the Gulf of Mexico, ~300,000 nodes.

• Need to agree on meaningful tidal run and comparison in 2D and 3D.

• Agree on skill metrics for model


data and model



model inter
-
comparisons. Bob B. has agreed to
initiate a white paper on this for distribution to the group and discussion


Group


Challenges

• File interchangeability including structure of grid files. In particular, ADCIRC and FVCOM have a
different

way of representing sub
-
grid scale structures such as levees, roadbeds, etc.


UND will convert
ADCIRC sub
-
grid scale structures to a form FVCOM can use.

• Size of grid makes it difficult for most participants to modify using standard grid tools (e.g., S
MS) and
makes model runs infeasible on in house clusters, forcing the use of shared (SURA Grid, TeraGrid)
resources.

• Lingering difficulty getting models to run / achieving desired turnaround time on LONI shared
machines.


UND


Progress to date

• High

resolution LA/TX 2 million node grid uploaded to testbed archive

• 2008 tidal boundary condition (for Hurricane Ike time period) uploaded to testbed archive

• OWI gridded 10m wind fields for Hurricane Ike uploaded to testbed archive


note, these are 30
-
min
averaged, marine roughness winds on a set of moving, nested grids.

• Performed initial tide and Ike ADCIRC and ADCIRC+SWAN runs, studied proper treatment of refraction
terms in SWAN


UND


Next steps

• Upload full set of observational points and da
ta to testbed archive

• Assist other groups translate ADCIRC grid (and particularly sub
-
grid scale features) into FVCOM and
SELFE grids

• Provide Hurricane Ike wind fields that have been adjusted to a 10
-
min averaging period and modified
with ADCIRC’s la
nd cover routine to reduce overland windspeed on 2 km regular grid

• Provide ADCIRC + SWAN results for Hurricane Ike

• Provide 300,000 node grid along with tidal BCs

DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




23



• Provide tidal boundary condition and wind forcing for Hurricane Rita.

• Put LSU hig
h resolution bathymetry on LA/TX 2 million node grid.


UND


Challenges

• Refraction terms in unstructured SWAN model.


UF SLOSH


Progress to date

• Ran Hurricane Ike using “out of the box” SLOSH wind model with NHC best track on two SLOSH basins
(Gal
veston and Sabine Pass). Provided MEOW of results.


UF SLOSH


Next steps

• Upload basin, forcing, model, results uploaded to testbed archive.

• Provide Ike time series results / maximum elevations at locations of observational data

• Developing a sequ
ence (3
-
4) of nested grids to allow running of SWAN for Hurricane Ike. Requires
output and manipulation of windfields from SLOSH for use in SWAN.


UF SLOSH


Challenges

• Limited scope of SLOSH means lots of recoding/extensions to compare with other mode
ls.


LSU


Progress to date

• Uploaded recently acquired high resolution bathymetry in several important passes to testbed archive

• Successfully able to get FVCOM to compile and run on LONI machines for tides during Ike period.


LSU


Next steps

• Co
mparison of tidal results with observation

• Run Hurricane Ike


waiting for UND to convert ADCIRC grid(s) to FVCOM format including sub
-
grid
scale features and also to provide 2km Ike gridded wind fields


LSU


Challenges

• Working with 2 million node
tropical grid

• Computational resources


USF


Progress to date

• Converted ADCIRC grid into FVCOM grid for tidal runs and provided revised grid to the other team
members. Did not include sub
-
grid scale features conversion.

• Initial evaluation of mode
l performance (2D, 1.5s external, 15s internal time step; 96 processors = 4
hrs run time/day simulation time).

• Successfully ran tides (boundary forcing and internally forced) using FVCOM on 288 processors on
LONI and compared results to NOAA tidal gauge
s from FL to TX. Results uploaded to testbed archive?


USF


Next steps

DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




24



• Additional tidal runs


3d? May wait for 300,000 node grid

• Run Hurricane Ike


waiting for UND to convert ADCIRC grid(s) to FVCOM format including sub
-
grid
scale features and a
lso to provide 2km Ike gridded wind fields


USF


Challenges

• Working with 2 million node tropical grid

• Computational resources


VIMS


Progress to date

• Converted ADCIRC grid into SELFE grid. Did not include sub
-
grid scale features in conversion

• Provided performance benchmarks for model runs on differing #s of processors for tidal runs (2D, 90 s
time step, 64 processors = 1 hrs run time / day simulation time).

• Ran SELFE for tides and compared results to 22 NOAA tidal gauges along North Gulf
and FL shelves,
better skill found using a Manning n coefficient = 0.025 than =0.02 or 0.03. Identified two worst
comparison stations and question the need for spatially varying friction coefficients. (Does SELFE include
the internal tidal loading?)

• Ran

Hurricane Ike using NCEP reanalysis wind forcing.


VIMS


Next steps

• Upload tidal results?

• Run Hurricane Ike


waiting for UND to convert ADCIRC grid(s) to FVCOM format including sub
-
grid
scale features and also to provide 2km Ike gridded wind fiel
ds


VIMS


Challenges

• ?

DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




25



Appendix A
-

Super Regional Modeling Testbed Deliverables




Develop protocols for how to transition research models and skill assessment tools, from the research
environment into operational centers.




Advance a cyberinfrastruc
ture capability that integrates tools, methodologies, and best practices to
support a modeling testbed, and includes:

o

Conventions for representation of unstructured grid data in NetCDF;

o

Implementation of an unstructured grid model standard in NetCDF
-
Java.I
;

o

A standard model toolkit in Matlab to perform standard operations on
climate and forecast

(CF)
-
compliant OpenDAP data;

o


A Web Map Service (WMS) that works with all testbed structured and unstructured grid models;

o

Tools for model grid subsetting and re
-
g
ridding.




Provide a testbed that:

o

Allows for rigorous, transparent and replicable testing of models;

o

Uses the aforementioned cyberinfrastructure capability;

o

Demonstrates the skill and performance of models in the following regions:



Shelf hypoxia in the Nor
thern Gulf of Mexico



Estuarine hypoxia in Chesapeake Bay



Storm surge and land inundation for tropical and extra
-
tropical storms in the

Gulf of Mexico and Gulf of Maine.




Develop and provide meaningful, quantitative and comparable analyses including sensit
ivity analyses, skill
scores and other skill metrics for evaluating numerical model performance. These analyses will include
model
-
to
-
model comparisons, and model
-
to
-
data comparisons. A byproduct of these analyses will be the
compilation and integration,

through the cyberinfrastructure initiative, of relevant observational data (in
situ and satellite) that will heretofore be available for future model validation efforts.




Begin the transition of the most promising research model(s) and skill assessment to
ols from the research
environment into operational centers.




Provide skill assessment guidance and advice to the operational centers for the models that are currently
in operational use in particular with respect to model resolution, atmospheric and latera
l boundary
forcing, specific physics, and model efficiency in the context of computational resources.




Develop a demonstration website to facilitate sharing of observations and model results, model testing,
and visualization of model
-
to
-
model and model
-
to
-
observations comparisons incorporating the analysis
and skill assessment tools.

DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




26



Appendix B
-

TAEG
comments on Testbed Team Leads
meeting:
Aug 31
, 2010

Attendees: Denny Kirwan, Rich Signell, Frank Aikman, Yi Chao, Bruce Ebersole, Chris Mooers

Overall, the

TAEG is satisfied the progress during month 2 of the IOOS Testbed, with evidence of Team
Leads overseeing continual progress toward the Year 1 high level goals. Some issues, however, need to
be addressed.

The high
-
level goals should reflect that establi
shed metrics for estuarine and shelf hypoxia do not
currently exist and may involve significant scientific effort to develop. This project will probably not
solve the problem, but is expected to make significant progress by coming up with some prototype
m
etrics and documenting the manner in which they were selected.

The TAEG is concerned that the operational side of the house is not being sufficiently engaged. We
suggest adding Gregg Jacobs (NRL), Frank Bub (NAVO), Hendrik Tolman(NOAA/NWS), and AJ Zhang
(
NOAA/CO
-
OPS) to the TAEG to complement Bruce Ebersole (USACE) and Frank Aikman (NOAA/NOS) so
that they can better track and provide input to the program. Suggest inviting Jonathan Westcott from
FEMA to the Inundation mailing list so he can track the inun
dation group discussion and meetings.

Team leads are encouraged to communicate any roadblocks or problems, not just successes in their
reports.

A joint GCOOS/ SURA/OceanLeadership team for model testbed activities for the Deepwater Horizon
modeling seems l
ike a perfect follow
-
on to the existing testbed.

How can we facilitate the longer term storage and distribution of data (sites like NCDDC?)

The teams should analyze preliminary findings to see if particular models or scenarios should be
dropped from f
urther computation or analysis.

Some administrative suggestions:

-

Group reports should be available the day before (not day of) the telecon either by email or
from a web site

-

A standard 1
-
pager PPT slide should also be provided the day before the telecon to

summarize
the key points and issues

-

Use webex ony if really necessary, given time constraints.

Cyberinfrastructure:

The codes and procedures used to convert model output to CF
-
compliant NetCDF (e.g. the code used to
convert Ch3D) should be documented and

made available as a community resource.

IMEDS should be enhanced to access CF
-
compliant files or datasets via OPeNDAP. An easy way to do
this is by utilizing a NetCDF
-
Java based Matlab Toolkit, such as http://njtbx.sourceforge.net.

DRAFT

--

SURA Monthly Progress


Super
-
Regional Modeling Testbed

September 10, 2010


Compiled by Elizabeth Smith (
lsmith@sur
a.org
)




27


The UAH Drupal site
(
http://testbed.sura.org
) should feel like the one site that people need to visit to
find information about the IOOS Modeling Testbed. It should have descriptions of and links to the
services provided by the Linux s
erver at LSU. At some point in the future it might also be a portal for
actually configuring and running models, but development of those capabilities should not take place
during this project, as this would take resources away from the other CI efforts
.

Estuarine Hypoxia
:

Having a software engineer working for Scott Peckham on a GUI for integrating different ROMS versions
as part of the estuarine hypoxia group seems inconsistent with the high level goals and with the
management structure. Shouldn’t t
his software engineer should be part of the CI group and tasked to
support the high
-
level goals?

We didn’t hear anything about the shelf model comparison (Wilkin’s effort to compare RTOFS,
ESPRESSO, HOPS, NYHOPS). Is that work on track?

Inundation
:

The

codes and procedures used to convert model grids from one model to another should be
documented and made available as a community resource. Perhaps CI can help with this.

It seems that the 2,000,000 node grid has been a burden to some groups. On the con
call it was
mentioned that the 2M node grid was actually smaller than the 5M node grid that the ADCIRC team
would prefer. Could we see the assessment from ADCIRC that shows the cost/benefit of grid
refinement from several resolutions (e.g. 0.5M, 2M and
5M) to help build our intuition for how
important these grid refinement is?

Shelf Hypoxia
:

What is the potential impact on the first year deliverables by Hetland not getting a postdoc?

Wiggert’s Matlab programmer should interface with CI on MAVE and PAV
E development, since ASA did
a substantial part of the original development work.