The MODIS Science Data Support Team (SDST) is
dedicated to the generation of high quality
Level 2 (L2) through L4 MODIS science
products for distribution to the Distributed
Active Archive Centers (DAACs) for archival
and to the MODIS science team for quality
control. SDST reports to the MODIS Team
Leader, Dr V. Salomonson, and functions
within a Principal Investigator (PI)
processing paradigm under the terms of which
SDST works closely with the MODIS PIs at
every phase of the work performed. SDST has
the following major responsibilities:
Development and Testing of Product
Generation Executives
MODIS supports the land, oceans and
atmospheres disciplines and the MODIS science
team generates in excess of 140 science
algorithms for L2 through L4 products for
each MODIS instrument. These algorithms are
combined by SDST into groups called Product
Generation Executives (PGEs). SDST is
responsible for:
-
Generating the PGEs for processing the L0
data to L1A data;
-
Generating the geolocation algorithm and
performing trend analysis to increase the
accuracy of the algorithm;
-
Testing these algorithms and, cloud mask
algorithm and the L1B PGE received from the
MODIS Characterization Support Team (MCST)
and delivering these PGEs to the GDAAC for
operational processing;
-
Developing and maintaining utility code to
support the SDST environment;
-
Interacting with the EOSDIS Core System
(ECS) staff in the generation of ESDTs to
allow the science products to be inserted
into the DAACs;
-
Receiving science algorithms from the MODIS
science team and integrating them with
metadata and production rules to generate
PGEs for L2 through L4 processing;
-
Documenting the PGEs, production rules, and
procedures generated at SDST;
-
Operating a Configuration Management group
for controlling the evolution of the PGEs
and the MODIS processing system described
below, and;
-
Operating a science test environment with
capability to process up to 32 days of data
in support of the science team's efforts to
validate their science algorithms.
System Development and Maintenance
The MODIS Adaptive Processing System (MODAPS)
receives Geolocation, L1B, Cloud Mask, and
ancillary data from the GDAAC and uses
approximately 100 PGEs per MODIS instrument
to generate L2 through L4 products. SDST is
responsible for:
-
Developing and maintaining the MODAPS
software for both IRIX and distributed
linux environments, and;
-
Exploring the potential applications of
MODAPS to other missions and data sets such
as OMI, NPP, Landsat, and AVHRR.
Data Production
The generation of the L2 through L4 science
products is accomplished using two super
computers: an Origin 3000 with 64 cpus, a
cluster of 136 Pentiun processors, and 50TB
of disk and an Origin 2000 with 80 cpus, a
cluster of 80 Pentiun processors, and 60TB
of disk. The former is used for data
reprocessing and the latter is used for
forward processing that is normally 2-5 days
behind real time. These systems have a
combined production capacity for all three
discipline products of 6x (6 datadays per
production day). SDST is responsible for:
-
Reporting plans and accomplishments and
interacting with the MODIS PIs to generate
long range processing plans;
-
Generating daily production plans for
operations staff and reconciling production
with the science products archived at the
DAACs;
-
Executing the production plans for the
forward and reprocessing machines on a
7x24h basis;
-
Documenting operational and production
procedures;
-
Conducting performance testing and
processing special requests from the
science disciplines;
-
Operating a rapid response data system for
generating special data products within
hours of real time to support such
activities as fire fighting, and;
-
Performing sustaining engineering to
support the on-going evolution of the
hardware and system software associated
with the operations computers.
Quality Control
As part of routine production, subsets of the
science products are distributed directly to
the MODIS science team members for QA
purposes. Each discipline is responsible for
maintaining a website identifying product
quality issues. SDST is responsible for:
-
Conducting daily QA reviews of atmospheres
and oceans science products and interacting
with the PIs to develop plans for
corrective action if necessary, and;
-
Operating the Land Data Operational Product
Evaluation (LDOPE) facility for land data
assessment, interacting with the land
science team to improve science algorithm
performance, and maintaining a land QA
database.
|