IEEE Geoscience and Remote Sensing Magazine - September 2016 - 8

Data from Space event organized by the ESA in Frascati,
Italy, which was followed by two equally successful conferences co-organized by the ESA, the Joint Research Center of
the European Commission,
and the European Union
Satellite Center in 2014 and
THESE FIVE ARTICLES
2016, respectively. The goal
PROVIDE A SAMPLING THAT
of Big Data from Space is to
HIGHLIGHTS THE DIVERSITY
foster research, science, inOF THE BIG DATA ISSUES
novation, and technological
EMERGING IN EO.
developments that tackle
the big data challenges in
EO to increase and promote
our capacity to extract meaningful information.
For this special issue on big data from space for geoscience and remote sensing, we have collected five articles that
present a snapshot of the recent developments and advances. These articles were selected through a two-step submission/review process. The resulting five contributions cover
some of the key challenges and leading-edge developments
in the field of big data from space for the field of remote
sensing and geoscience.
The first article, "Big Data Challenges in Climate Science,"
distinguishes itself by striking a balance between data analytics and data storage within a cyberinfrastructure for climate science. The article proposes a new service model, i.e.,
Climate Analytics-as-a-Service, where adaptive analytics services coexist with a climate data service that is based on an
evolution of the Open Archive Information System model
while securing interoperability with the Earth System Grid
Federation (ESGF). This type of new model is urgently needed in view of the research community's collective work that
is expected to generate up to 100 PB of data in the framework
of the next Intergovernmental Panel on Climate Change Assessment Report, which is scheduled for 2020.
The second article, "Managing Big, Linked, and Open
Earth-Observation Data," addresses the challenge posed by
the many applications that require access to data originating from different data silos. Indeed, linked data is a new
paradigm that studies how data following the Resource
Description Framework can be interconnected to increase
their value. The developed stack of tools is demonstrated
in the context of wildfire-monitoring services. Particular
attention is given to exploration and visualization as well
as discovering services that allow the selection of spatial
and temporal relations.
The third article, "Supporting Earth-Observation Calibration and Validation," introduces an emerging field
exploiting the vast amounts of data provided by citizens
in the form of in-situ data as well as visually interpreted
and manually digitized EO data sets. The article presents a review of the existing crowdsourcing initiatives of
relevance to EO and details a series of case studies before expanding on the challenges and barriers limiting
the adoption of citizen science and crowdsourcing in
support of EO.
8

The fourth article, "Big Data Management in Earth Observation," emphasizes that big data management in EO
is not just an information technology issue and that close
collaboration and interaction between operations engineers, data managers, system and software engineers, and
scientists are the key to the successful exploitation of satellite data archives.
The fifth and last article, "A Multi-GPU Program for
Uncertainty-Aware Drainage Basin Delineation," addresses
the timely processing of massive digital elevation models
(DEMs) for river network delineation thanks to the use of
graphics processing units (GPUs) combined with Message
Passing Interface via the host central processing units. The
proposed solution is scalable and will cope with the data
volumes originating from future countrywide DEMs at a
spatial resolution down to 2 m.
These five articles provide a sampling that highlights the
diversity of the big data issues emerging in EO. The velocity at which the data volumes and variety of data sources
is increasing will call for further advances and innovative
approaches. Further effort will be needed, particularly in
metadata processing, to ensure that it will be possible to
select, in an efficient and effective way, the smallest subset
of the available (linked) data needed to provide the desired
information while bringing the uncertainty to an acceptable level. In addition, big data in geosciences and remote
sensing, unlike in other fields faced with big data issues,
cannot be reduced to data-driven science by simply looking
for correlations and patterns in the input data. Indeed, expert knowledge in geosciences needs to be combined with
advanced data science to maximize the extraction of valuable information in an effective and efficient way. At a technical level, it will become unsustainable to keep bringing
the data to geoscientists, for the evolution of the bandwith
of network connections is not keeping up with the pace of
data volume increase. Therefore, it will become more and
more important to develop seamless ways to bring the
processing close to the data, including such innovative approaches as the ESGF developed in the first article of this
special issue.
AUTHOR INFORMATION
Pier Giorgio Marchetti received the Laurea (M.Sc.) degree
in electronic engineering from the University of Rome "La
Sapienza," Italy, in 1977. From 1979 to 1985, he served as
researcher with the Fondazione Ugo Bordoni, Rome, Italy,
where he carried out research projects on microwave atmospheric propagation, digital signal processing, digital
radio communication, and cochannel interference for radio relay and satellite links. He participated in the cochannel interference experiment between the Italian SIRIO
and the ESA's Orbital Test Satellite with small receiving
stations in Italy and China. In 1978, he served as a junior
engineer in the Radio Laboratory of Autovox Spa (a subsidiary of Motorola Inc.), Rome, Italy. From 1986 to 2000,
he served as a staff member of the ESA in various areas,
ieee Geoscience and remote sensing magazine

SEPTEMBER 2016
http://www.M.Sc
Table of Contents for the Digital Edition of IEEE Geoscience and Remote Sensing Magazine - September 2016
