River Basin Information System Open Environmental

International Journal of
Geo-Information
Article
River Basin Information System: Open
Environmental Data Management for Research
and Decision Making
Franziska Zander * and Sven Kralisch
Geographic Information Science, Friedrich Schiller University of Jena, Grietgasse 6, D-07743 Jena, Germany;
[email protected]
* Correspondence: [email protected]; Tel.: +49-3641-9-48897
Academic Editors: Constanze Curdt, Christian Willmes, Georg Bareth and Wolfgang Kainz
Received: 25 March 2016; Accepted: 11 July 2016; Published: 18 July 2016
Abstract: An open, standardized data management and related service infrastructure is a crucial
requirement for a seamless storage and exchange of data and information within research projects, for
the dissemination of project results and for their application in decision making processes. However,
typical project databases often refer to only one research project and are limited to specific purposes.
Once implemented, those systems are often not further maintained and updated, rendering the
stored information useless once the system stops operating. The River Basin Information System
(RBIS) presented here is designed to fit not only the requirements of one research project, but focuses
on generic functions, extensibility and standards compliance typically found in interdisciplinary
environmental research. Developed throughout more than 10 years of research cooperation
worldwide, RBIS is designed to manage different types of environmental data with and without
spatial context together with a rich set of metadata. Beside data management and storage, RBIS
provides functions for the visualization, linking, analysis and processing of different types of data
to support research, decision making, result dissemination and information discovery for all kinds
of users. The focus of this paper is on the description of the technical implementation and the
presentation of functions. This will be complemented by an overview of example applications and
experiences during RBIS development and operation.
Keywords: environmental information system; data management; open science; spatial data
infrastructure; environmental research; decision making
1. Introduction
Integrated environmental resources research in interdisciplinary projects demands for
a comprehensive and holistic system understanding, which in turn relies on powerful and effective data
management and information sharing strategies. At the same time, European and German research
donors increasingly enforce open science and open access policies through ICT (Information and
Communication Technology) tools, networks and media, aiming to make scientific knowledge more
easily accessible and to make research more open, global, collaborative, creative and closer to society [1].
This in turn requires information management and sharing methods based on the best available
knowledge and techniques to support different kinds of users, e.g., researchers, decision makers and
stakeholders. To specifically address the demand for geospatial information sharing, a number of
data format and interface standards have been developed during the last decade, mainly driven by
the Open Geospatial Consortium (OGC) and other standardization organizations. Examples are the
ISO19115 standard for geographic information metadata [2] or the Catalogue Services Specification [3],
both of which are widely applied in geodata infrastructures on national and international levels [4].
ISPRS Int. J. Geo-Inf. 2016, 5, 123; doi:10.3390/ijgi5070123 www.mdpi.com/journal/ijgi

ISPRS Int. J. Geo-Inf. 2016, 5, 123 2 of 20
However, information sharing in environmental sciences also demands for the integration of geospatial
data with other, non-spatial data.
Addressing these demands on the level of specific scientific disciplines, various data repositories
have been developed, such as the Data Observation Network for Earth [5], the CUAHSI Hydrologic
Information System [6] or the PANGAEA information system [7]. Many of these systems are
constrained to specific user groups (e.g., experts) and rather define a framework infrastructure
instead of really offering user and service interfaces or do not support geospatial services at all.
Another problem is the often missing support for the upload and sharing of user-owned datasets.
This is not a problem of missing features in available standards, but simply a lack of support for
these functions on the client’s side. Regarding the data management design, many existing systems
also lack a fine-grained user permission management (e.g., write permission to owned data only).
Summarizing the current situation, it can be stated that, even though standards and software are
readily available, integrating geodata and thematic information together with standardized services
and mapping capabilities in ready-to-use client/server architectures is still challenging.
To address those challenges, the development of the web-based River Basin Information System
(RBIS) started more than 10 years ago, and up to now, the system has been applied in more than
35 research projects of different sizes in South America, Africa, Europe, Australia and Asia (see
examples in Section 7). As most of the study regions were located in developing countries, challenges
related to it were always taken into account during the design and implementation. One of those
is still the accessibility to the Internet (statistics for the current situation can be found under [8]).
However, also the local infrastructure for data management (e.g., for time series data or geodata) from
the authority side is not as developed, such as in Europe, where, e.g., the INSPIRE (Infrastructure for
Spatial Information in the European Community) initiative [4] is driving forward to publish geodata in
a common manner. The existing data often need time-consuming data preprocessing, e.g., if it is only
available as an image, handwritten records, floppy disks, text and Excel files with a changing structure
every year, and so on. It is also not always easy to find out which data exist and how they can be
accessed. Especially at the beginning of a research activity, the data collection can be very challenging
and time consuming. Therefore, it is also important to have an easily-accessible common data base
to avoid every researcher or project member having to do the same work again, which includes not
only the data acquisition, but also preprocessing from different formats to one. Once the data are
prepared, they can be used in further steps of scientific workflows (e.g., hydrological modelling),
which is also one aim that RBIS supports, as well as the dissemination and sharing of research data
among researchers and decision makers in a standardized and service-oriented way. The accessibility
and openness of research data are also increasingly requested by funding agencies [9].
What is RBIS? RBIS is a web-based information system, developed at the Department of
Geography, Geographic Information Science at the Friedrich Schiller University of Jena in Germany,
aiming at researchers and decision makers. Its main objectives are the management of environmental
data from different disciplines focusing on the management of metadata and several data formats
(e.g., spatial, time series and processing data) (Section 2). To find and access the managed data,
there are search and filter functions (Section 3), as well as interfaces and services provided (Section 4).
The administration, especially related to the user and permission management, is described in Section 5.
Its current architecture and used software components are described in Section 6. Examples for the
application and use types of RBIS are part of Section 7. Finally, there is a short discussion and outlook
for further planned developments.
2. RBIS: Data Management, Analysis and Visualization

RBIS is designed to manage different types of environmental data with and without a spatial
context together with its metadata. The data are not limited by the content, spatial location or special
field of research and are primarily grouped by the type (e.g., time series data, geodata (vector and
raster formats), space time data, maps, documents, simulation models, etc.). The current usage is
mainly related to data needed or produced for climate impact assessments from the fields of natural
sciences (e.g., hydro-meteorological time series data, land use, fire, socioeconomic, soil, geology,
elevation, vegetation, indicators, etc.), with a special focus on hydro-meteorological time series data
management, e.g., as is required for hydrological modeling. Beside the management and storage of
data, RBIS provides functions for the visualization, linking, analysis and processing of certain data
types (e.g., time series data) to support researchers, decision makers, research result dissemination and
information discovery and extraction for local stakeholders.
The management of the metadata in general will be described in Section 2.1. The management
and related functions for spatial data, time series data, simulation data and other data types will be
explained in Sections 2.2–2.5.
2.1. Metadata
The main focus of the system is on metadata (“data about data” [10]) to make existing data
findable and to inform about their existence. In this case, the dataset is not required to be uploaded
itself, but is to be described as well as possible. The metadata in RBIS are stored based on a complete
implementation of the ISO 19115 Standard for Geographical Data, which has been extended according
to the specific requirements of the data types supported in RBIS. Accordingly, geodata are described
according to ISO 19115, while all other datasets only use subsets of the standard, e.g., the “Persons and
Organizations” element for the description of responsible parties.
The visualization of metadata datasets, their creation, manipulation and linking is described in
the following sections.
2.1.1. Overview and Detailed View

The overview list may include visible or hidden columns, which can be displayed on demand.
An example for time series data metadata is shown in Figure 1. In the table header, the datasets can be
sorted or filtered using different operators fitting to the data type, e.g., “contains”, “does not contain”
or “ends with” for text or “<”, “ď” or “>” for numbers. In addition, it is possible to export the resulting
lists as a CSV dataset, either limited to the most important dataset attributes or including all available
metadata attributes. Datasets including or linked to spatial coordinates can further be exported as
a vector dataset in Shapefile format. In addition, current search and filter results can be saved for later
use and sharing with others.
The details link in the first columns (see Figure 1) always leads to a more detailed view on single
datasets. Here, it is possible to view, edit or delete the dataset and to access associated or linked data
and functions. This also includes a file upload function that allows one to easily attach arbitrary files
to a dataset, such as pictures, raw data or text documents. As an example, this could be the picture of
a measurement station, making it easier to correctly interpret associated measured data.
2.1.2. Multilingual Support

While researchers are typically used to working and communicating in English, local stakeholders
(e.g., in Vietnam) and other interested parties often have limited skills and experience in this
regard. To lower language barriers and thus usage barriers for non-English speakers, RBIS features
a multilingual support for all of its user interfaces. All navigation elements, notices, metadata
field descriptions and, under certain circumstances, the content of selection lists can be translated.
The current state of all translations (stored in the database and files) covers English, Vietnamese,
Spanish, Portuguese and German. In the RBIS web interface, the languages can easily be changed by
using a drop-down list located at the right top corner of the webpage (see Figure 1).
Figure
Figure Typical
1. 1. Typicallayout
layoutforforall
allRBIS
RBISinstances.
instances. The menu on
The menu onthe
theleft
leftside
sideshows
showsallallavailable
available types
types of of
datasets
datasetsthat can
that canbebedescribed
describedininthis
thisRBIS
RBIS instance. The main
instance. The mainwindow
windowshowsshowsthe
theexample
example overview
overview
table of of
table time series
time seriesdata
dataininthe
theOkavango
OkavangoBasin
Basin Information System.Source:
Information System. Source:[11].
[11].
2.1.3.
2.1.3. Creation
Creation andEditing
and Editing
In In order
order totocreate,
create,manipulate
manipulateoror delete
delete datasets
datasets inin RBIS,
RBIS,ititisisrequired
requiredtotohave
havea auser account
user account
with appropriate access permissions (see Section 5). To avoid data inconsistencies and to
with appropriate access permissions (see Section 5). To avoid data inconsistencies and to simplify data simplify
datainformation
input, input, information is entered
is entered using selection
using selection lists whenever
lists whenever possible.possible. Tothe
To ease ease the creation
creation of
of datasets
datasets (e.g., when entering a larger number of similar datasets), RBIS features a template system
(e.g., when entering a larger number of similar datasets), RBIS features a template system that supports
that supports the storage and re-use of user-generated dataset templates. Those templates can also
the storage and re-use of user-generated dataset templates. Those templates can also serve as a base
serve as a base for the automated creation or import of data.
for the automated creation or import of data.
2.1.4. Linking of Datasets
2.1.4. Linking of Datasets
Environmental information is often associated and contextualized with other data. An example
Environmental information is often associated and contextualized with other data. An example
could be a time series dataset that was used within a study, which on the other hand, is described
could be a time series dataset that was used within a study, which on the other hand, is described
within a publication stored in the RBIS “Document” section. To establish the according links, it is
within a publication stored in the RBIS “Document” section. To establish the according links, it is
possible to loosely link every RBIS dataset with any other dataset stored in RBIS. Explicit links are
possible to loosely
also possible link every
and created RBIS metadata
by using dataset with any
fields. other dataset
A common stored
example in RBIS.
for this type ofExplicit linksisare
association
also
thepossible
linkage and
of acreated
person by using metadata
or organization to afields. A expressing
dataset, common example forasthis
their role type of association
a responsible party.
is the linkage of a person or organization to a dataset, expressing their role
Another example is the direct linkage between a study site and a dataset, which is shown as a responsible
in Figureparty.
2
Another
from the study site point of view. This can be used if a dataset itself has no spatial coordinates, but 2
example is the direct linkage between a study site and a dataset, which is shown in Figure
from
stillthe study
refers to asite pointstudy
certain of view. This
site or canunderlining
area, be used if aitsdataset
spatialitself has Another
relation. no spatial coordinates,
way of linking isbut
viastill
refers to a certain study site or area, underlining its spatial relation. Another way of linking is via the
in- and out-put description of a processing step (see Section 2.4). Created linkages are always visible
from each involved dataset.
the in- and out-put description of a processing step (see Section 2.4). Created linkages are always
the in- from
visible and out-put description
each involved of a processing step (see Section 2.4). Created linkages are always
dataset.
visible from each involved dataset.
Figure 2. Screenshot of study sites. The first column shows the study site and the second column the
Figure
Figure 2. Screenshot
2.and
Screenshot of
of study
study sites. The first
first column
sites.datasets.
The column shows
shows the
the study
study site
site and
and the second
theSystem column
second (OBIS) the
column[11].
the
amount type of directly-linked Source: Okavango Basin Information
amount
amount and
and type of directly-linked
type of directly-linked datasets. Source: Okavango
datasets. Source: Okavango Basin
Basin Information
Information System
System (OBIS)
(OBIS) [11].
[11].
2.2. Spatial Data
2.2. Spatial Data
2.2. Spatial Data
Datasets managed in environmental information systems and specifically in RBIS somehow
Datasets managed in environmental information systems and and specifically
specifically in in RBIS somehow
have Datasets
a spatial managed
relation. To in describe
environmental information
this relation, certainsystems
RBIS datatypes contain their RBIS
ownsomehow
location
have a spatial
have a spatial relation.
relation. To describe this relation, certain RBIS datatypes contain their own location
information using point To describe this
coordinates, relation, certain
the bounding box extent RBISordatatypes
a polygon. contain their own
As a second location
option, they
information
information using
using point
point coordinates,
coordinates, the
the bounding
bounding box
box extent
extent or
or aa polygon.
polygon. As aa second
As second option,
option, they
they
can be directly linked to one of those datasets, e.g., each time series dataset is linked to a
can be directly
can bedataset. linked to one of those datasets, e.g., each time series dataset
directly linked to one of those datasets, e.g., each time series dataset is linked to a is linked to a station dataset.
station
Once
station spatial information is available for a specific dataset, it can be displayed in maps.
dataset.
Once spatial information is available for a specific dataset, it can be displayed in maps. Point
PointOncecoordinates
spatial (displayed as
information clustered for
is available points depending
a specific on the
dataset, canzoom level) or inbounding box
coordinates (displayed as clustered points depending on the zoomitlevel) be
or displayed
bounding box maps.
extents Point
are
extents are displayed within a dynamic filled feature layer, which will always change once the listed
displayed within a dynamic filled feature layer, which will always change once the listed datasetsare
coordinates (displayed as clustered points depending on the zoom level) or bounding box extents in
datasets
displayed inwithin
the overview listsfilled
are changed during filtering. Imported shape and raster files can be
the overview lists aare
dynamic
changed feature
during layer,
filtering. which
Imported willshape
always and change
rasteronce
files the
canlisted datasets
be combined in
in
combined
the overview in maps and displayed
lists areinchanged in another separate web client integrated in RBIS. In general, WMS
maps and displayed another during
separate filtering.
web client Imported
integratedshape in and
RBIS.raster files canWMS
In general, be combined
(Web Map in
(Web
maps Map
and Service) orinWMS-T
displayed another (WMS
separate+ Time)
web [12] from
client internal
integrated in (RBIS)
RBIS. and
In external
general, WMSsources
(Web can
Mapbe
Service) or WMS-T (WMS + Time) [12] from internal (RBIS) and external sources can be included in
included
Service) in WMS-T
or every map (WMS(FigureTime)
3). Once[12]an internal WMS is created, it will besources
automatically included in in
every map (Figure 3). Once+an internal from
WMS internal
is created,(RBIS)
it willand
be external
automatically can be included
included in the map
the map
every of the related dataset. External WMS or WMS-T sources must be configured by the application
of the map
related(Figure 3). Once
dataset. an internal
External WMS or WMS WMS-Tis created,
sources it will
mustbe automatically
be configuredincluded in the map
by the application
administrator,
of the related but this also
dataset. allowsWMS
External one toorcreate
WMS-T excludable
sources layer
must lists.
be configured by the application
administrator, but this also allows one to create excludable layer lists.
administrator, but this also allows one to create excludable layer lists.
Figure 3. Screenshot of a map window above the overview list of stations in the in the Vu Gia Thu
Figure
Bon 3. Screenshot
3.
River Screenshot ofaamap
of
Basin Information mapwindow
window
System abovethe
above
(RBIS) the
[13] overview
overview
covering list
list
the of stations
ofregion
stations in the
ofincentral
the in the
in Vietnam.
the VuThe
Vu Gia Giamap
Thu Thu
Bon
Bon
RiverRiver
Basin Basin Information
Information SystemSystem
(RBIS) (RBIS)
[13] [13]
coveringcovering
the the
region region
of of
central central
Vietnam.
shows the location of stations (orange circles) listed in the (not shown) overview table below, but Vietnam.
The mapThe map
shows
shows
the
similar the
location location
foroftime
of stations
as shown stations
series(orange
(orange circles) circles)
data inlisted
Figurein listed
thestudy
1, in shown)
(not the boundaries
site (notoverview
shown)(dark
overview
tableblue) table
below, but
and below,
ansimilar but
enabledas
similar for
shown
external as
landshown
time for time
useseries
Web data
Mapinseries data
Figure
Service 1, in Figure
study
(WMS) site 1,
layer forstudy
boundaries site(dark
the region. boundaries (dark
blue) and an blue) and
enabled an enabled
external land
external
use Web land
Map use Web(WMS)
Service Map Service (WMS)
layer for layer for the region.
the region.
2.3. Time Series Data

ISPRS Int. J. Geo-Inf.
Measured 2016, 5, 123
or simulated time series data usually can be related to a point (e.g., measurement 6 of 20
station), grid cells of different sizes (e.g., climate model data) or a polygon (e.g., spatial modelling
2.3. Time Series Data
entities like hydrological response units). The focus in RBIS lies on the management, processing and
visualizationMeasured or simulated
of station data. time series data usually can be related to a point (e.g., measurement
station), grid cells
Stations are describedof different
by theirsizes (e.g.,type
name, climate modelgauging,
(climate, data) or aborehole,
polygon virtual
(e.g., spatial modelling
station, etc.), spatial
entities like hydrological response units). The focus in RBIS lies on the management, processing and
location (coordinates in the spatial reference system of the original source), year of establishment and,
visualization of station data.
maybe, closing and responsible parties (Figure 4a). Time series metadata can directly be linked to
Stations are described by their name, type (climate, gauging, borehole, virtual station, etc.),
a station
spatial and may have
location one or more
(coordinates parameters
in the (Figure 5b).
spatial reference A time
system of series datasetsource),
the original should be uploaded
year of
as text
establishment and, maybe, closing and responsible parties (Figure 4a). Time series metadata can time
in a certain given format (for more details, see the online tutorial [14]). Depending on the
step,directly
gaps will be analyzed
be linked andand
to a station stored
mayas metadata.
have Once
one or more the dataset
parameters has been
(Figure 5b). Asuccessfully uploaded,
time series dataset
it can be visualized
should be uploadedas as
antext
image
in a or as angiven
certain interactive diagram
format (for (Figuresee
more details, 4b).
theTo further
online analyze
tutorial [14]). data,
it is Depending
possible toon the timeatstep,
explore what gaps will be the
distance analyzed andthe
data of stored
same as parameter
metadata. Once andthe dataset has
temporal been
resolution is
successfully
available, how uploaded, it can be visualized
well it correlates, as an imageand
at which elevation or as
if an
datainteractive
gaps coulddiagram (Figure
be filled by4b). To one
using
further
or more of analyze data, it is possible
the surrounding stations.to With
explore at what
respect todistance
data gaps,the data of the same
a rule-based gapparameter and
filling toolbox is
temporal resolution is available, how well it correlates, at which elevation and if data gaps could be
provided within RBIS [14]. Data values for gaps filled by the implemented interpolation functions (e.g.,
filled by using one or more of the surrounding stations. With respect to data gaps, a rule-based gap
nearest neighbor, linear interpolation, inverse distance weighting) are stored separate from the original
filling toolbox is provided within RBIS [14]. Data values for gaps filled by the implemented
data. The interpolation details (methods, parameters and time series used) are stored as metadata to
interpolation functions (e.g., nearest neighbor, linear interpolation, inverse distance weighting) are
ensure all provenance
stored separate frominformation related
the original data. Thetointerpolation
the gap filling process
details is tracked.
(methods, Finally,
parameters andtime
time series
series data
can beused)accessed andasdownloaded
are stored in different
metadata to ensure ways. The
all provenance followingrelated
information options to are
the available:
gap filling process
‚ isdownload
tracked. Finally,
of thetime series data
originally can be data
imported accessed
file,and downloaded in different ways. The following
options are available:
‚ export of the complete dataset as it is stored in the database (missing date values are filled to ease
further
download of the
use, e.g., fororiginally
modelling imported
tasks), data file,
‚ export
export of the complete dataset
of the time series data with as selected
it is stored in the database
parameters (missingtime
in a selected date interval,
values are filled to
ease further use, e.g., for modelling tasks),
‚ export of the time series data with or without filled gaps,
 export of the time series data with selected parameters in a selected time interval,
‚ export of aggregated
export time data
of the time series series data
with or in different
without filledtime steps,
gaps,
‚ export of data
export from different
of aggregated stations
time series data (one parameter
in different timeonly) as needed in environmental simulation
steps,
models [15].
export of data from different stations (one parameter only) as needed in environmental
simulation models [15].
(a) (b)
Figure 4. Screenshot of a climate station and one of the linked simulated time series datasets.
Figure 4. Screenshot of a climate station and one of the linked simulated time series datasets. (a) Station
(a) Station description with pictures and 22 associated time series; (b) one of the simulated time series
description withseven
datasets with pictures and 22 displayed
parameters associatedintime series; (b)diagram.
the interactive one of the simulated
Source: [13]. time series datasets
with seven parameters displayed in the interactive diagram. Source: [13].
Figure 5.
Figure 5. The
Thescreenshot
screenshotshows
showsininwhich
which data
data sources
sources andand related
related processing
processing stepssteps the land
the land covercover
map
mapGia
“Vu “Vuand
GiaThu
andBon
Thu(VGTB)
Bon (VGTB) LandCover
LandCover 2010
2010 v2” v2”
has hasused.
been been used. Source:
Source: [13]. [13].
A further RBIS module allows one to store and manage space-time-variant data, i.e., a mixture
A further RBIS module allows one to store and manage space-time-variant data, i.e., a mixture of
of time series data and geodata typically given as a series of raster images. The parameters are
time series data and geodata typically given as a series of raster images. The parameters are described
described similarly to time series data, including parameter type, unit and description. The metadata
similarly to time series data, including parameter type, unit and description. The metadata description
description follows the geodata metadata profile, which is extended by additional information, such
follows the geodata metadata profile, which is extended by additional information, such as the value
as the value for missing values and temporal resolution. The dataset itself can be stored as a file as
for missing values and temporal resolution. The dataset itself can be stored as a file as described
described in Section 2.1.1. An example could be climate model results given as netCDF files. Due to
in Section 2.1.1. An example could be climate model results given as netCDF files. Due to their
their hybrid characteristics, space-time datasets can be browsed and searched via the time series data
hybrid characteristics, space-time datasets can be browsed and searched via the time series data and
and geodata views.
geodata views.
2.4. Processing
2.4. Data
Processing Data
Research projects
Research projects using
using RBIS
RBIS maymay alsoalso have
have to to deal
deal with
with the the modelling
modelling of of environmental
environmental
processes (land use, climate, hydrology, etc.), e.g., in the context of scenario-based impact assessment.
processes (land use, climate, hydrology, etc.), e.g., in the context of scenario-based impact assessment.
The description
The description ofof simulation
simulation software,
software, model
model application
application andand model
model chaining
chaining was was therefore
therefore an an
important requirement
important requirement during
during the
the development
development of of RBIS.
RBIS. This
This is is reflected
reflected inin the
the RBIS
RBIS “Processing”
“Processing”
module, which allows the description of lineage processes. It consists
module, which allows the description of lineage processes. It consists of the three parts of the three parts allowing
allowing toto
describe data
describe data source,
source, processing
processing software
software and and processing
processing stepstep (see
(see the
the next
next section).
section). An An RBIS
RBIS dataset
dataset
(e.g., time series data) can directly link to a processing step to build a linkage between
(e.g., time series data) can directly link to a processing step to build a linkage between result data and result data and
how it was created. Conversely, it is possible to see where a certain dataset
how it was created. Conversely, it is possible to see where a certain dataset has been used as input has been used as input
for aa processing
for processing step,
step, as
as shown
shown inin Figure
Figure 5. 5. The
The implementation
implementation aims aims toto still
still allow
allow aa mapping
mapping to to the
the
lineage element of ISO 19115 [2] to provide that information in a standardized
lineage element of ISO 19115 [2] to provide that information in a standardized way while allowing one way while allowing
one
to to store
store much much
moremore detailed
detailed processing
processing information
information in a flexible
in a flexible way.way. The mapping
The mapping to ISOto 19115
ISO 19115
will
will also give the opportunity to use existing external tools like MetaViz
also give the opportunity to use existing external tools like MetaViz [16] for graph-based provenance [16] for graph-based
provenance visualizations,
visualizations, similar to thesimilar to the
schematic schematicinillustration
illustration Figure 6. in Figure 6.
Processing
2.4.1. Data Source, Software and Processing
In order
ordertoto properly
properly describe
describe a dataa processing
data processing workflow,
workflow, RBIS
RBIS allows oneallows one information
to manage to manage
information
about about
input data input
(i.e., datadata (i.e., data
sources), sources),and
the software the method
softwareused
andtomethod
processused to process
the data, the data,
the application
the application of the software to the input data and the
of the software to the input data and the resulting output data. resulting output data.
A data source is described in RBIS by metadata with standard attributes, like name, description
responsible person.
or responsible person. This dataset can then be linked to a list of datasets stored in RBIS or to external
resources, like time series or maps. The second component to be described is the software that has
been used for the processing (e.g.,(e.g., for
for modelling
modelling oror data
data preprocessing).
preprocessing). Special functions, modules
or extensions of software can be described in a new dataset and can be linked to a parent or related
software package. The finally used version of the software or module of a software has to be created
as a separate dataset to avoid having duplicated descriptions. The processing step gives information
about what was done, a linking input data source with the software version, creator, study site, parent
or previous processing steps and output data. A process step can further be described with additional
about what was done, a linking input data source with the software version, creator, study site, parent
ISPRS
or Int. J. Geo-Inf.
previous 2016, 5, steps
processing 123 8 of 20
and output data. A process step can further be described with additional
attributes, like start/end times, type of processing and, in the case of time-variant simulation, the
attributes, like
simulation’s time start/end
period. times, type of processing and, in the case of time-variant simulation, the
simulation’s time period.
Figure6.6. Schematic
Figure Schematic illustration provenance information
illustration of provenance informationbased
basedon
onprocess
processdescriptions
descriptionswithin
withinthe
the
Vu
VuGia
GiaThu
ThuBonBonRBIS
RBIS(changed
(changed after
after [17]).
[17]).
2.4.2. Processing
2.4.2. Processing Examples
Examples
Oneexample
One examplefor foraaprocessing
processingstep stepisisthe
theautomated
automatedimportimportofof the
the Global
Global Surface
Surface Summary
Summary of of
the
the Day (GSOD) data (see Section 4.1). The software used is an automated
Day (GSOD) data (see Section 4.1). The software used is an automated import routine within RBIS; import routine within
RBIS;
the data thesource
data source is an
is an FTP FTP service
service that is accessed;
that is accessed; the processing
the processing step is step is the import
the import or update
or update process
process itself. The date of the last process execution will be updated automatically
itself. The date of the last process execution will be updated automatically on every update. on every update.
Another example
Another example of of aa simplified
simplified process/model
process/model chain chaindescribed
describedwithin
withinthetheVu
VuGia
GiaThuThuBonBon
RBIS [13]
RBIS [13] isis illustrated
illustrated inin Figure
Figure 6. 6. It
It shows
shows thatthat the
the ECHAM5
ECHAM5climate climatemodel
modeldatasets
datasetsofofdifferent
different
climatescenarios
climate scenariosserve serveas as input
input datadata
for afor a regional
regional downscaling
downscaling process
process based onbased on the Weather
the Weather Research
Research and Forecasting (WRF) regional climate model together
and Forecasting (WRF) regional climate model together with historical ERA40 reanalysis with historical ERA40 reanalysis
datasets.
datasets. The results were used as climate data inputs together with soil,
The results were used as climate data inputs together with soil, geology, land cover and elevation geology, land cover and
elevation datasets to assess the impact of changing climate and land cover on
datasets to assess the impact of changing climate and land cover on the water availability and sediment the water availability
and sediment load in the Vu Gia and Thu Bon (VGTB) River Basin in central Vietnam. The simulated
load in the Vu Gia and Thu Bon (VGTB) River Basin in central Vietnam. The simulated discharge at
discharge at several points inside the basin and at its outlet have then been used as input for the
several points inside the basin and at its outlet have then been used as input for the application of
application of MIKE FLOOD, MIKE 11 and MIKE HYDRO models to predict floods, saltwater
MIKE FLOOD, MIKE 11 and MIKE HYDRO models to predict floods, saltwater intrusion and the
intrusion and the impacts of reservoirs [17].
impacts of reservoirs [17].
2.5.Other
2.5. Other Data
Data
Thereare
There arealso
also some
some other
other types
types of
of data
data that
that can
can be
be managed
managedand anddescribed
describedby
bymetadata.
metadata.OneOneofof
them is “documents”, like publications, posters, presentations or others. Datasets that do nothave
them is “documents”, like publications, posters, presentations or others. Datasets that do not have
their own detailed description, like time series data, or that do not fit into the existing structure can
their own detailed description, like time series data, or that do not fit into the existing structure can be
be described under “other data”. This could be just a file (e.g., Excel file with any kind of information)
described under “other data”. This could be just a file (e.g., Excel file with any kind of information)
or the description of an external data repository (see Section 2.4.1). An example of the customized
or the description of an external data repository (see Section 2.4.1). An example of the customized
description of a data type is soil information (sampling profiles and horizons). While the filter
description of a data type is soil information (sampling profiles and horizons). While the filter functions
functions described in Section 2.1.1 can also be applied to search for specific soil datasets (e.g., all
described in Section 2.1.1 can also be applied to search for specific soil datasets (e.g., all samples with
samples with a certain pH value range), the locations of the resulting dataset are easily visualized in
a certain pH value range), the locations of the resulting dataset are easily visualized in overview maps.
overview maps. Making use of these common RBIS functions, datasets with specific properties and
Making use of these common RBIS functions, datasets with specific properties and in locations can
in locations can quickly be identified. A list of soil horizons with restricted view and download
quickly be identified. A list of soil horizons with restricted view and download permissions is shown
permissions is shown in Section 5. For all of these data types, there is a basic set of attributes, but in
in Section 5. For all of these data types, there is a basic set of attributes, but in most cases, it needs to be
most cases, it needs to be extended according to the available information.
extended according to the available information.
Another special type of data is information about future development scenarios and related
Another special type of data is information about future development scenarios and related
indicators. Here, RBIS supports the description of scenarios (e.g., climate change scenarios, story
indicators. Here, RBIS supports the description of scenarios (e.g., climate change scenarios, story
lines) and can be used to link project results to them. Calculated or delineated indicators (e.g., related
to drinking water, population, precipitation, food security) in different formats (e.g., single values,
time series data, diagrams, maps) can be directly linked to a scenario and study area. The scenario
and indicator module is a product of demands raised in a research project, where it was requested to
lines) and can be used to link project results to them. Calculated or delineated indicators (e.g., related
to drinking water, population, precipitation, food security) in different formats (e.g., single values,
time series data, diagrams, maps) can be directly linked to a scenario and study area. The scenario
and indicator module is a product of demands raised in a research project, where it was requested to
describe a great variety of different indicator values in the context of study areas and climate change
scenarios [18].
Within many research projects, not only research data need to be shared, but also information
about events, like, e.g., meetings with stakeholders, project workshops or important deadlines.
To provide such kinds of functionality within RBIS, a calendar was implemented. This calendar
makes use of the existing RBIS permission management system and allows one to easily attach
documents (e.g., the meeting minutes or workshop presentation) to individual events. Moreover,
a central RBIS notification function can be used to inform users about new and upcoming events.
3. Search and Filter

To avoid ending up with a “data cemetery” containing many datasets that are never used again,
it is necessary to provide options to make stored data easily findable and available via their stored
metadata. In order to achieve this, RBIS implements various functions, which focus on different user
groups and application scenarios. The main focus groups are the researchers, project partners and local
stakeholders who already know about the existence of RBIS. A further user group consists of external
researchers or stakeholders interested in the provided data collection within RBIS. While the first
group can make use of RBIS search functions, external users depend on search engines, like Google, to
find RBIS. Once a user has found the system, it is necessary to help him or her in the second step to
find the data and information he or she is looking for. For this purpose, RBIS allows one to view all
metadata even for users not logged in to the system, i.e., including search engines that want to index
RBIS content.
Once a user has found access to RBIS, there are several options to search for data and information.
For a general overview, there is a global search function, which is based on the full text search option
provided by the underlying database management system, PostgreSQL [19]. This also includes a spatial
search option based on a bounding box or a study site representing an area plus an optional buffer
zone (see Figure 7). The spatial search is implemented by using functions of the PostgreSQL extension
PostGIS [20]. As a result, the user will get a list of results grouped by their data type. Without providing
any search criteria, all datasets are listed, which allows a quick overview of RBIS content. Study site,
extent and spatial buffer can also be set as temporary or permanent filters, which are automatically
applied for subsequent searches, regardless of dataset type. This function can be used to create
individual views or to reduce the amount of presented datasets in RBIS installations covering larger
regions, e.g., as found in the Southern African Science Service Centre for Climate Change and Adaptive
Land Management (SASSCAL) Information System (IS) [21], which covers the area of the whole of
southern Africa and belongs to the science service center SASSCAL [22].
Another option to search for datasets is to directly open the overview list of a type of data (e.g.,
“time series data”). Here, the user can search via a map (bounding box) or use the filter options
described in Section 2.1.1. Furthermore, there is a text search over all attributes of a data type as a free
text search or with the help of predefined lists showing only used items.
Finally, datasets in RBIS are also exposed via its integrated Catalog Web Service (CSW) interface
(see Section 4). This open metadata exchange standard also provides search functions, which thus can
be used to find and access data in RBIS.
external researchers or stakeholders interested in the provided data collection within RBIS. While the
first group can make use of RBIS search functions, external users depend on search engines, like
Google, to find RBIS. Once a user has found the system, it is necessary to help him or her in the second
step to find the data and information he or she is looking for. For this purpose, RBIS allows one to
viewISPRS
all metadata
Int. J. Geo-Inf.even
2016, for
5, 123users not logged in to the system, i.e., including search engines 10
that
of 20want
to index RBIS content.
Figure 7. Screenshots of the global search in the Vu Gia Thu Bon RBIS [13], showing the search via the
Figure 7. Screenshots of the global search in the Vu Gia Thu Bon RBIS [13], showing the search via
study site “VGTB Lowlands (watershed)” with a buffer of 10 km.
the study site “VGTB Lowlands (watershed)” with a buffer of 10 km.
4. Interfaces and Services

Once a user has found access to RBIS, there are several options to search for data and
As the
information. Forsystem aims overview,
a general to integratethere
data from different
is a global sources,
search to provide
function, dataisfor
which further
based onuse
the(e.g.,
full text
for modelling) and to inform about existing data in a standardized way, several
search option provided by the underlying database management system, PostgreSQL [19]. This service interfaces are also
implemented. Other interfaces support the automated import and preprocessing of time series data
includes a spatial search option based on a bounding box or a study site representing an area plus an
from different sources or the export of data in specific formats. These import and export interfaces
optional buffer zone (see Figure 7). The spatial search is implemented by using functions of the
and services are briefly described in Section 4.1. The third option is the exposure of metadata via the
PostgreSQL extension PostGIS [20]. As a result, the user will get a list of results grouped by their data
standardized CSW. Its integration and usage is shown in Section 4.2.
type. Without providing any search criteria, all datasets are listed, which allows a quick overview of
RBIS4.1. Integration
content. Studywith Data
site, Providers
extent andand Usersbuffer can also be set as temporary or permanent filters,
spatial
which areBeside
automatically
the manualapplied
import for subsequent
of data searches,
and metadata regardless
via its of dataset
web frontend, type. This
RBIS provides function
various
functions for the automated import and preprocessing for selected data types. One example is the
interfaces for the automated import and update of time series data. The data can be located locally
or in online repositories, as well as from web services. One of those online repositories is the Global
Surface Summary of the Day (GSOD) product. Provided by the National Climatic Data Center (NCDC)
in Asheville, USA, GSOD is a global, freely-available climate dataset based on World Meteorological
Organization (WMO) weather stations and the Integrated Surface Database (ISD) (DSI-3505) [23,24].
When importing these data into RBIS using automated functions, units are converted from English to
SI-units [15]. GSOD data do not substitute the need for further time series data acquisitions necessary
for in-depth analyses in a region, but help greatly in the early stages of research or if no other data are
available. Similarly, climate data measured at weather stations all over Germany and provided by the
Climate Data Center of the German Weather Service [25] can be easily imported into RBIS.
A further use case for automated data import is the continuous integration of online measurements
done within individual research projects (e.g., soil moisture data). Once data can be requested or
accessed by the RBIS server, import processes can be easily configured by a user or customized data
reader implemented by the administrator. For the configuration by a user, the data must be in a certain
format (date and parameter in columns). Furthermore, it is necessary to create a preprocessing rule
(column delete list, allow deletion of duplicate dates) and to combine it with an import rule describing
the data source location, file name rule (prefix), time series data import template and link to the
destination dataset. All kinds of import or update processes implemented in RBIS can be described
and linked with the process description introduced in Section 2.4.2. Not just data imports, but also the
creation of metadata datasets can be automated based on a set of equally-structured base data and
predefined templates.
For further use and analyses of data or metadata in external tools, it is important to have functions
for mass export and download. One of them is a function for metadata export in the CSV format.
Datasets exported like this can be easily compared, checked or analyzed in external tools, like MS Excel,
and allow further processing and analysis of data. RBIS also includes specific functions for exporting
data to the Jena Adaptable Modelling System (JAMS) [26], allowing one to easily provide time series
time series data inputs for a large variety of quantitative and qualitative hydrological simulation
data inputs for a large variety of quantitative and qualitative hydrological simulation models.
models.
4.2. Services
4.2. Services
To expose metadata about raster, vector, soil and time series datasets on the Internet, an OGC
To expose metadata
standard-compliant about
CSW [3] raster,
based vector,
on the pycswsoil and time series[27]
implementation datasets on set
has been theupInternet, an OGC
and integrated.
standard-compliant CSW [3] based on the pycsw implementation [27] has been set up
While exposing only datasets that do not have any read permission restrictions, the RBIS CSW can be and integrated.
While exposing
accessed from anyonly datasets
CSW client.that do not are
Examples have any read permission
a web-based restrictions,
search interface, the RBIS
integrating CSW can
a number of
RBIS instances [28] and the GLUES Geodata infrastructure [29] (see Figure 8), a joint data andnumber
be accessed from any CSW client. Examples are a web-based search interface, integrating a service
of RBIS instances
platform [28]projects
for research and theworldwide
GLUES Geodatafundedinfrastructure [29] (seeprogram
within the research Figure 8), a joint dataLand
“Sustainable and
service platform for research projects worldwide funded within
Management” of the German Ministry of Education and Research (BMBF) [30]. the research program “Sustainable
LandTheManagement” of the Observation
setup of a Sensor German Ministry of (SOS)
Service Education andtime
[31] for Research
series (BMBF) [30].on istSOS [32] is
data based
The setup
currently underofdevelopment.
a Sensor Observation Service (SOS) [31] for time series data based on istSOS [32] is
currently under development.
(a) (b)
Figure 8. Two examples for RBIS Catalog Web Service (CSW) applications. (a) Screenshot of a
Figure 8. Two examples for RBIS Catalog Web Service (CSW) applications. (a) Screenshot of
aCSW-based
CSW-basedsearch
searchfunction
functionon
onall
allRBIS
RBISinstances
instances[28].
[28].The
Thered
red boxes
boxes indicate
indicate the
the overall extent of
overall extent of
available data
available data within
within individual
individual RBIS
RBIS instances.
instances. (b) Screenshot of
(b) Screenshot of aa GLUES
GLUES geodata
geodata infrastructure
infrastructure
tool [29] showing bounding boxes of datasets available in Vietnam from different research
tool [29] showing bounding boxes of datasets available in Vietnam from different research projects projects
within the research program “Sustainable Land Management”.
within the research program “Sustainable Land Management”.
5. Administration, User and Permission Management

5. Administration, User and Permission Management
For the administration and maintenance of RBIS instances, it is necessary to care about the
For the administration and maintenance of RBIS instances, it is necessary to care about the general
general configuration (e.g., overall spatial extent, UTM zone for distance calculations, activation or
configuration (e.g., overall spatial extent, UTM zone for distance calculations, activation or deactivation
deactivation of modules), user and permission management, user support, error detection and
of modules), user and permission management, user support, error detection and reporting and the
reporting and the collection of user demands. All write accesses (user name, type of action, date, error
collection of user demands. All write accesses (user name, type of action, date, error message and
message and snapshot of the previous state of the dataset) are logged, and recent changes are shown
snapshot of the previous state of the dataset) are logged, and recent changes are shown under the
under the menu “report”. Furthermore, notification mails can be sent automatically, to keep project
members informed about new or changed datasets.
The user and permission management functions in RBIS provide fine-grained options to protect
or publish metadata and data. For this reason, any access to RBIS requires the use of a user account.
The creation of user accounts follows a two-step procedure. First, a user can apply for a user account
via an online registration form. In a second step, the system’s administrators will decide if a user
menu “report”. Furthermore, notification mails can be sent automatically, to keep project members
informed about new or changed datasets.
The user and permission management functions in RBIS provide fine-grained options to protect
or publish metadata and data. For this reason, any access to RBIS requires the use of a user account.
The creation of user accounts follows a two-step procedure. First, a user can apply for a user account
via an online registration form. In a second step, the system’s administrators will decide if a user
account is created and which general permissions will be granted (e.g., permission to read, download
or modify information). Once the account is created, the user will be informed about his or her
user account and password. Accounts may further have expiration dates. The assignment of view,
export or manipulation permissions to a user account is done by the assignment of permission groups.
Those freely-designable groups indicate which type of permission is granted within each of the
different RBIS modules. The permission management via groups is based on the assumption that
everybody within a project is allowed to have access to all datasets within a module, and user accounts
are only
ISPRS provided
Int. J. to project
Geo-Inf. 2016, 5, 123 members. As this assumption does not fit always with the requirements, 12 of 20
it is completed by a dataset-based permission management with an assigned owner. Such requirements
provided
can originate to project members.
from dataset As this
owners assumption
(protection does not fit
of intellectual always with
property), the requirements,
use restrictions it is
of datasets
completed
(license, useby a dataset-based
limitation), as wellpermission
as from themanagement with anuser
demand to provide assigned owner.
accounts for Such requirements
external users, like
can originate
local stakeholders fromordataset owners (protection
other researchers. of intellectual
In addition property),
to the group-based use restrictions
permissions, of datasets
each dataset can
(license, use limitation),
have individual permissions, i.e., as well as from the demand to provide user accounts for external users, like
local stakeholders or other researchers. In addition to the group-based permissions, each dataset can
‚havePublic metadata
individual and datai.e.:
permissions,
‚ Public metadata, data download for registered users (default)
 Public metadata and data
‚ Public metadata, data download on request
 Public metadata, data download for registered users (default)
‚ Metadata visible for registered users, download for registered users
 Public metadata, data download on request
‚ Metadata
Metadata visible
visible for
for registered
registered users,
users, data download
download on requestusers
for registered
‚ Metadata
Metadata and data
visible fornot accessible
registered users, data download on request
 Metadata and data not accessible
Data in this case refer to uploaded data, like documents or time series data (Figure 9b), but can
Data
also be in this
related to case refer
single to uploaded
fields, as shown data, like documents
in Figure 9a. or time series data (Figure 9b), but can
also be related to single fields, as shown in Figure 9a.
(a) (b)
Figure 9.
Figure 9. Two
Two examples
examples for
forpermission
permissionmanagement.
management.(a)(a)Screenshot
Screenshot of of
soil horizon
soil descriptions
horizon in
descriptions
OBIS
in [11]
OBIS with
[11] withprotected
protectedvalue
valuecolumns;
columns;(b)
(b)screenshot
screenshotofofaatime
time series
series dataset
dataset in
in OBIS without
OBIS without
download permission for the currently logged in
download permission for the currently logged in user.user.
All RBIS instances have per default a guest account (guest/guest), which is an account with view
All RBIS instances have per default a guest account (guest/guest), which is an account with view
permissions. With the option for an automatic guest login, it is not necessary to always actively login,
permissions. With the option for an automatic guest login, it is not necessary to always actively login,
which is a very useful and successful option to allow search engines to index metadata datasets.
which is a very useful and successful option to allow search engines to index metadata datasets.
6. Software Components and Architecture
The software framework RBIS is built on open source software in a modular way to ensure an
easy reuse and adaptation of the system for different research projects and their demands. The system
as a whole can be used and modified without any restrictions or license fees. It is possible to run it in
different environments, e.g., in a native Linux installation on a server or within VirtualBox [33] in a
6. Software Components and Architecture

The software framework RBIS is built on open source software in a modular way to ensure an easy
reuse and adaptation of the system for different research projects and their demands. The system as
a whole can be used and modified without any restrictions or license fees. It is possible to run it in
different environments, e.g., in a native Linux installation on a server or within VirtualBox [33] in
a virtual environment as a copy on a local computer with or even without Internet access. This is not
only useful for presentations for locations without or with insufficient Internet connectivity, but it
is also an easy way to distribute the whole system for further use at the end of the research project
to local stakeholders or organizations. The software components used and the internal application
structure of RBIS will be explained in the following sections.
6.1. Software Components

RBIS is built for a Linux environment and was successfully deployed on the Debian and Ubuntu
operating systems. The major software packages used and their role within RBIS are shown in
ISPRS
FigureInt.10.
J. Geo-Inf. 2016, 5, 123 13 of 20
Figure 10. RBIS

RBIS architecture
architecture and main software components used.
The
The database
database management
management systemsystem PostgreSQL
PostgreSQL [34] [34] together
together with
with the
the PostGIS
PostGIS extension
extension [20]
[20] is
is
used to store application data (e.g., user accounts), metadata and parts of the data
used to store application data (e.g., user accounts), metadata and parts of the data within RBIS. within RBIS.
The
The main
main part
part of
of the
the RBIS logic is
RBIS logic is written
written in
in PHP
PHP [35].
[35]. The
The Apache
Apache HTTP
HTTP Server
Server [36]
[36] is
is used
used asas
aa webserver.
webserver. The template engine Smarty for PHP is used to separate application
The template engine Smarty for PHP is used to separate application logic and content logic and content
from its presentation
from its presentation [37].
[37]. Furthermore,
Furthermore, for for the
the processing
processing andand creation
creation of maps, MapServer
of maps, MapServer [38]
[38] is
is
used. To expose metadata of spatially-related datasets, the CSW server pycsw [27]
used. To expose metadata of spatially-related datasets, the CSW server pycsw [27] written in Python written in Python
is
is integrated.
integrated.
For theweb-based
For the web-based frontend,
frontend, the Bootstrap
the Bootstrap [39]framework
[39] web web framework
together together with JavaScript
with different different
JavaScript libraries
libraries (e.g., jQuery (e.g., jQuery
1.11 1.11jQuery
[40] and [40] andUIjQuery UI used
[41]), are [41]), to
arecreate
used atouser-friendly
create a user-friendly and
and intuitive
intuitive graphical user interface. For the display of map data in web browsers,
graphical user interface. For the display of map data in web browsers, OpenLayers 2 [42] is used. OpenLayers 2 [42] is
used. The use of an older version of both jQuery and OpenLayers ensures
The use of an older version of both jQuery and OpenLayers ensures the compatibility of older the compatibility of older
web-browsers still
web-browsers still in
in use,
use, especially
especially inin developing
developing countries.
countries.
6.2.
6.2. Application
Application Architecture
Architecture
The
The structure
structureandandmain
mainfunctional
functionalcomponents
components of RBIS are shown
of RBIS in Figure
are shown 11. The
in Figure data
11. storage
The data
is organized in three separate PostgreSQL databases. The main RBIS database is
storage is organized in three separate PostgreSQL databases. The main RBIS database is used forused for storing
application data, metadata
storing application and other
data, metadata data. data.
and other The storage of the
The storage metadata
of the metadatais isbased
basedononaa full
full
implementation of ISO 19115 as the relational database schema extended by data type-specific
implementation of ISO 19115 as the relational database schema extended by data type-specific tables. tables.
Geodata and time series data are managed in separate databases. Time series data are organized in
generic tables, each associated with one time series metadata dataset. The date value is stored as a
UNIX timestamp for easier processing during gap detection and filling. Values are stored as all equal,
independent of their parameter type, each parameter one column. All changed values (e.g., from gap
filling) are stored in separate tables to ensure that the original dataset will never be changed. Prepared
Geodata and time series data are managed in separate databases. Time series data are organized
in generic tables, each associated with one time series metadata dataset. The date value is stored
as a UNIX timestamp for easier processing during gap detection and filling. Values are stored as
all equal, independent of their parameter type, each parameter one column. All changed values
(e.g., from gap filling) are stored in separate tables to ensure that the original dataset will never
be changed. Prepared database views simplify the access of gap filled time series data, e.g., from
external applications, like simulation models. The geodata database manages imported map data (e.g.,
Shapefiles) and the Mapfile configuration for the MapServer. All uploaded files (e.g., time series data,
documents, pictures) are stored in the file system of the server.
The RBIS framework is written on the server-side mainly in PHP following object-oriented
programming paradigms. The common access to stored metadata datasets (e.g., persons and
organizations, stations, geodata, time series data, etc.) is realized with a description layer [43] using
XML documents to hold all information needed for access, manipulation, visualization and linking of
database contents; e.g., in case of the creation or change of a dataset, the data will first pass several
checks against the expected type, while a second step will generate all required SQL statements in
a delineated order based on a dependency analyses. The implemented object-relational mapping
allows one to easily adapt the application logic and frontend to changes in the underlying database
ISPRS Int. e.g.,
schema, J. Geo-Inf. 2016, 5,
in order to123 14 of
adapt to the specific requirements of a research project. RBIS does not use one20
of the existing object-relational mapping tools for PHP, as they have not been shown to provide the
one of thethat
flexibility existing object-relational
is required for RBIS. mapping tools for PHP, as they have not been shown to provide
the flexibility that is required for RBIS.
Figure 11. RBIS architecture and components.

Figure 11.
XML documents
XML documents alsoalso hold
hold information
information on on how
how datasets
datasets oror other
other components
components of of RBIS
RBIS are
are
interlinked. An important one is the “Time Series Management” providing
interlinked. An important one is the “Time Series Management” providing import and export import and export
functionalities, as
functionalities, as well
well as
as processing
processingsteps,
steps,like
likegap
gapdetection
detectionororfilling.
filling.Depending
Depending onon thethe
amount of
amount
possible data values (e.g., more than 2,000,000 values) and access permissions (e.g.,
of possible data values (e.g., more than 2,000,000 values) and access permissions (e.g., no download no download
permission), time
permission), time series
series diagrams
diagrams areare either
either visualized
visualized as
as an
an image
image withwith the
the help
help of
of the
the open
open source
source
graph library
graph library JpGraph
JpGraph [44]
[44] or
or as
as an
an interactive
interactive diagram
diagram atat the
the client
client side
side using
using the
the open
open source
source
JavaScript charting library dygraphs
JavaScript charting library dygraphs [45]. [45].
The graphical user interface of RBIS for the management of different types of data (e.g., geodata,
time series data) is created based on templates using the Smarty template engine [37]. The jQuery
Grid Plugin jqGrid [46] is used for the overview list of datasets. In comparison to other similar
JavaScript libraries for visualizing data tables, jqGrid performed best under the given conditions and
requirements. To reduce requests to the server, the full list of datasets is loaded only once and
rendered on demand. This means once the list is loaded, no further requests to the server are
The graphical user interface of RBIS for the management of different types of data (e.g., geodata,
time series data) is created based on templates using the Smarty template engine [37]. The jQuery
Grid Plugin jqGrid [46] is used for the overview list of datasets. In comparison to other similar
JavaScript libraries for visualizing data tables, jqGrid performed best under the given conditions and
requirements. To reduce requests to the server, the full list of datasets is loaded only once and rendered
on demand. This means once the list is loaded, no further requests to the server are necessary for
sorting and filtering, which is a big advantage if the available Internet connectivity is low.
Once there is spatial location information related to a dataset, the RBIS mapping functions are
activated. The location or bounding box of datasets is displayed and directly linked for a map-based
search in a map visualized with the help of OpenLayers [42]. Background WMS layers from RBIS
(based on MapServer WMS [38]) or external sources can be visualized. Besides WMS, other services
like, e.g., the CSW catalogue service are provided by RBIS and can be accessed from outside.
Because of the demand to easily adapt RBIS to upcoming project requirements, the internal
structure is modularized to ensure high flexibility and extensibility. Functions or sets of functions
related to RBIS modules can be activated or deactivated in the configuration file. A selection of
important modules with related special features of RBIS are:
‚ Management of metadata datasets and responsible parties.

‚ Management, analysis and visualization of measured and simulated time series data.
‚ Management and visualization of spatial data.
‚ Management of documents and any kind of files.
‚ Management of study sites and observation data.
‚ Management of simulation models and linking in- and out-put data.
7. Applications
RBIS has been successfully applied in more than 35 research projects mainly related to integrated
land and water resources management and climate change impact assessment in South America,
Africa, Europe, Australia and Asia. They differ in the size of the covered study region (from a few
km2 to several 100,000 km2 ), the amount of registered and active users, the amount and type of stored
data and the addressed target groups (e.g., students, researchers, stakeholders). This section will give
an overview of RBIS applications and will highlight their major differences.
7.1. Application Types

In general, three cases for the usage of RBIS can be identified based on the different scales of
application. The first one belongs to single research topics (small scale), which includes RBIS instances
related to small projects, like PhD theses with a small amount of registered users. The main focus lies
on climate and hydrological data management, pre-processing, analysis, access and preservation of
time series data collected for the related study [47]. Examples are:
‚ KosiRBIS [48] with data collected in the Kosi River Basin within a PhD project evaluating
upstream-downstream linkages of hydrological dynamics in the Himalayan region [49].
‚ JakartaRBIS [50] as part of a PhD research assessing the impacts of land use and climate changes
on hydrology in Western Java, Indonesia using hydrological models [51].
‚ OTI RBIS [52] with a special focus on the Kara River Basin in Togo and Benin as a part of a PhD
research project about river basin assessment and hydrologic process modelling for integrated
land and water resources management in West Africa [53].
Once the advantages have been discovered there is a high motivation of the author to incorporate
data and to make use of a broad range of system functions.
The second is groups of researchers working together on a single field of research (medium scale).
RBIS instances are set up for project cooperation at an institutional and bilateral level for the exchange
and management of jointly-used datasets, like geodata or time series data. Examples are:
‚ DrysalRBIS: Geodata and environmental data in the BMBF-funded project DRYSAL (Multi-scale
analyses of dryland salinity impacts and adaptive Integrated Water Resources Management
(IWRM) strategies for salinity mitigation in the Berg River basin, Western Cape Province,
South Africa (2005–2008)).
‚ SeyhanRBIS: Time series data collected and measured in the Seyhan River Basin during the project
GIRMASEM (GIS supported Integrated Water Resources Management System (IWRMS) for the
Eastern Mediterranean: A regional clean water action plan for the Seyhan River (GIRMASEM)
(2006–2009)).
The last group of application cases refers to interdisciplinary research projects (large scale),
involving a large number of researchers, and local stakeholders beyond the project consortium are also
involved and addressed. Those RBIS installations usually have many registered users from different
disciplines, institutions, organizations and authorities. Examples are:
‚ BrahmaRBIS [54] and DanubeRBIS: environmental data from the Brahmaputra and Upper Danube
River Basins collected and produced within the EC project BRAHMATWINN (Twinning European
and South Asian river basins to enhance capacity and implement adaptive integrated water
resources management approaches [55]) (2006–2010).
‚ Okavango Basin Information System (OBIS) [11] stores environmental information from the
Okavango River Basin for the BMBF-funded project “The Future Okavango” (2010–2015) [56,57].
‚ Vu Gia Thu Bon RBIS [13] manages environmental data from the Vu Gia and Thu Bon River Basin
from the BMBF-funded LUCCi project (Land Use and Climate Change interactions in the Vu Gia
Thu Bon River Basin/Central Vietnam; BMBF; 2010–2015) [58].
Due to its flexible structure and ease of integration, existing RBIS instances have been further
used for projects covering bigger or smaller study regions. Examples for further usage of existing data
collections are:
‚ Assam RBIS [59] is a subset view of the BrahmaRBIS where new datasets have been uploaded
within the Guwahati Field Trial Project financed by the state government of Assam (2013–2016).
‚ The SASSCAL Information System (IS) is the project database for water-related tasks within
Phase 1 of the ongoing SASSCAL project (Southern African Science Service Centre for Climate
Change and Adaptive Land Management (SASSCAL)—a Regional Science Service Centre (RSSC)
in Southern Africa; BMBF; 2014–2017). SASSCAL IS is a newly-created RBIS instance that
integrates all datasets of OBIS, LetabaRBIS and DrysalRBIS.
7.2. RBIS System Performance

The dedicated SQL-based storage of RBIS information allows one to efficiently manage large
amounts of table data. For example, the metadata and application data representation in the RBIS
database involves more than 300 tables with 1500 attributes overall. This, on the other hand, requires
a careful design of the overall relational structure and SQL statements, the latter being safeguarded
by the object-relational mapping interface implemented in RBIS (see Section 6.2). As a result, RBIS
allows one to store a large amount of individual datasets within a single database management server.
The overall amount of over 8000 time series datasets, 900 geodata datasets and some hundreds of other
datasets on a single PostgreSQL server proved to create only minor SQL execution times and related
server load. For example, the SASSCAL IS instance [21] includes more than 230 geodata datasets
and 680 time series datasets, which refer to more than 500 associated files. The fact that this system
performs very well even in a virtual environment on a notebook further emphasizes the efficient
overall system design.
7.3. RBIS Deployment and Maintenance

All RBIS instances mentioned above are hosted and maintained on dedicated server hardware
at the Department of Geography of the University of Jena in Germany. With the increasing visibility
and robustness during the last 10 years, project partners decided to use RBIS for their own research
activities. One example is the Institute for Technology and Resources Management in the Tropics and
Subtropics (ITT) in Cologne, Germany. One of their coordinated projects is the CNRD project (Center
for Natural Resources and Development). This worldwide university network promotes academic
exchange and cooperation in the area of natural resources management, in particular related to water,
land, ecosystem and renewable energy [60]. One part of the activities is MSc student exchanges and
joint student projects on case studies, which use RBIS as the data infrastructure. For this reason, ITT
hosts its own RBIS server with currently 12 RBIS instances [61].
Furthermore, research activities initiated by a research group of the ITT and numerous partners
working in the area of river basin assessment and management are cooperating within the “basin-info”
network [62]. One of the basins is the Limari catchment, where geodata and measured time series data
have been incorporated during several research projects into the “Centro-Norte-Chile RBIS:Limari and
Petorca River Basin Information System” [63,64]. The Centro-Norte-Chile RBIS is currently hosted at
the Center for Advanced Research in Arid Zones (CEAZA), a scientific research and technological
center in the Region of Coquimbo in Chile [65].
8. Discussion and Outlook

The development of RBIS during the last 10 years was driven by needs and requirements raised
in research projects and in cooperation with different disciplines. Due to the application in many
different research projects, it was possible to continuously extend and adapt the system according
to user needs. As shown in the previous section, not only the source code, but sometimes even the
data collection could be reused during that period. Throughout the many years of RBIS’s evolution,
a close cooperation between the software development team, domain experts and RBIS users proved
to be of utter importance for a successful implementation of the system. This close cooperation
does not only support fast and informal error reporting, but it also eases the selection of features for
implementation. It should not be forgotten that comprehensive user support plays a very important
role to keep such a system alive and used. Building trust, raising acceptance and motivating user are
more essential and time consuming than the provision of technical solutions. This includes training
courses for data providers and users, taking care of multilingual web interfaces, providing online
tutorials and a reliable and fast response to user requests. Experiences during various research projects
showed that provisioning of added values through the use of centralized data management systems
like RBIS is the best option to convince data owners to share data and metadata. The availability
of effective and easily-usable access control mechanisms can further increase acceptance for such
systems. Providing the information system as a virtual appliance (e.g., based on VirtualBox) can be
a further incentive for collaboration and data sharing. Even though this typically limits its use (e.g., no
collaborative access/read access only), this option proved to be especially beneficial in regions with
limited Internet availability, as can still often be found in developing parts of the world.
Future improvements and the addition of new features will be mainly driven by research project
demands. One will be the provision of more data integration services of data available in other
repositories (e.g., derived from satellite imagery). Another focus will be on the support for processing
services and to create a closer link between web-based model applications and process descriptions.
Acknowledgments: The authors acknowledge the support of the German Ministry of Education and Research,
which has funded the RBIS development within various research programs (Grant Numbers 03IP514, 01LL0912,
01LB0801, 01LL0908D, 01LG1201E).
Author Contributions: Sven Kralisch and Franziska Zander designed the presented system. Franziska Zander
implemented most parts of the presented work and is technically maintaining and administrating the core system.
Franziska Zander wrote the paper, and Sven Kralisch contributed to it.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. European Commission. Digital Science in Horizon 2020; European Commission: Brussles, Belgium, 2013.
2. ISO. ISO Geographic Information—Geographic Information—Metadata—Part 1: Fundamentals (ISO 19115-1:2014);
ISO: Geneva, Switzerland, 2014.
3. Nebert, D.; Whiteside, A.; Vretanos, P. OpenGIS Catalogue Services Specification 2.0.2 2007; Open Geospatial
Consortium: Wayland, MA, USA, 2007.
4. EU. Directive 2007/2/EC of the European Parliament and of the Council of 14 March 2007 establishing
an infrastructure for spatial information in the European Community (INSPIRE). Off. J. Eur. Union L 2007,
108, 50.
5. Michener, W.; Vieglais, D.; Vision, T.J.; Kunze, J.; Cruse, P.; Janée, G. DataONE: Data observation network
for earth—Preserving data and enabling innovation in the biological and environmental sciences. Lib Mag.
2011, 17. [CrossRef]
6. Tarboton, D.G.; Maidment, D.; Zaslavsky, I.; Ames, D.; Goodall, J.; Hooper, R.P.; Horsburgh, J.;
Valentine, D.; Whiteaker, T.; Schreuders, K. Data interoperability in the hydrologic sciences—The CUAHSI
hydrologic information system. In Proceedings of the Environmental Information Management Conference,
Santa Barbara, CA, USA, 2011; pp. 132–137.
7. Grobe, H.; Diepenbroek, M.; Dittert, N.; Reinke, M.; Sieger, R. Archiving and distributing earth-science
data with the PANGAEA information system. In Antarctica; Fütterer, D.K., Damaske, D., Kleinschmidt, G.,
Miller, H., Tessensohn, F., Eds.; Springer-Verlag: Berlin, Germany, 2006; pp. 403–406.
8. Telecommunication Development Sector (ITU-D)—Statistics. Available online: http://www.itu.int/en/ITU-
D/Statistics/Pages/stat/default.aspx (accessed on 18 March 2016).
9. Corti, L.; Van den Eynden, V.; Bishop, L.; Woollard, M. Managing and Sharing Research Data: A Guide to Good
Practice; Sage: London, UK, 2014.
10. ISO. ISO Geographic Information—Geographic Information—Metadata—Part 1: Fundamentals (ISO 19115-1:2003);
ISO: Geneva, Switzerland, 2003.
11. Okavango Basin Information System (OBIS). Available online: http://leutra.geogr.uni-jena.de/obis
(accessed on 18 March 2016).
12. De La Beaujardiere, J. OpenGIS Web Map Server Implementation Specification; Open Geospatial Consortium:
Wayland, MA, USA, 2006.
13. LUCCi Consortium Vu Gia Thu Bon RBIS. Available online: http://leutra.geogr.uni-jena.de/vgtbRBIS/
14. Using the Okavango Basin Information System (OBIS)—ILMS-Wiki. Available online: http://ilms.
uni-jena.de/ilmswiki/index.php/Using_the_Okavango_Basin_Information_System_(OBIS) (accessed on
23 March 2016).
15. Zander, F.; Kralisch, S.; Flügel, W. Data and information management for integrated research–requirements,
experiences and solutions. In Proceedings of the 20th International Congress on Modelling and Simulation,
Adelaide, Australia, 1–6 December 2013; pp. 1–6.
16. Henzen, C.; Mäs, S.; Bernard, L. Provenance information in geodata infrastructures. In Geographic Information
Science at the Heart of Europe; Springer: Berlin, Germany, 2013; pp. 133–151.
17. Henzen, C.; Mäs, S.; Zander, F.; Schroeder, M.; Bernard, L. Representing Research Collaborations and Linking
Scientific Project Results in Spatial Data Infrastructures by Provenance Information. In Proceedings of the
19th AGILE Conference on Geographic Information Science, Helsinki, Finland, 14–16 June 2016.
18. Flügel, W.-A. Development of adaptive IWRM options for climate change mitigation and adaptation.
Adv. Sci. Res. 2011, 7, 91–100. [CrossRef]
19. PostgreSQL: Documentation: 9.3: Full Text Search. Available online: http://www.postgresql.org/docs/9.3/
static/textsearch.html (accessed on 10 March 2016).
20. PostGIS Developers PostGIS—Spatial and Geographic Objects for PostgreSQL. Available online: http:
//postgis.net/ (accessed on 10 March 2016).
21. SASSCAL Consortium SASSCAL Information System. Available online: http://leutra.geogr.uni-jena.de/
sasscalRBIS/ (accessed on 10 March 2016).
22. SASSCAL. Available online: http://www.sasscal.org/ (accessed on 10 March 2016).

23. Lott, N. Global Surface Summary of Day; National Climatic Data Center: Asheville, NC, USA, 1998.
24. Smith, A.; Lott, N.; Vose, R. The Integrated Surface Database: Recent Developments and Partnerships.
Bull. Am. Meteorol. Soc. 2011, 92, 704–708. [CrossRef]
25. DWD Wetter und Klima—Deutscher Wetterdienst—CDC (Climate Data Center). Available online: http:
//www.dwd.de/DE/klimaumwelt/cdc/cdc_node.html (accessed on 20 March 2016).
26. Kralisch, S.; Fischer, C. Model representation, parameter calibration and parallel computing—The JAMS
approach. In Proceedings of the International Congress on Environmental Modelling and Software, Sixth
Biennial Meeting, Leipzig, Germany, 1–5 July 2012.
27. Kralidis, T.; Tzotsos, A. Pycsw—Metadata Publishing Just Got Easier. Available online: http://pycsw.org/
28. Zander, F. RBIS CSW Global Search. Available online: http://leutra.geogr.uni-jena.de/RBISsearch/
metadata/global_search.php (accessed on 20 March 2016).
29. GLUES GLUES Geoportal. Available online: http://geoportal-glues.ufz.de/index.php (accessed on
20 March 2016).
30. Bernard, L.; Mäs, S.; Müller, M.; Henzen, C.; Brauner, J. Scientific geodata infrastructures: Challenges,
approaches and directions. Int. J. Digit. Earth 2013, 7, 1–21. [CrossRef]
31. Sensor Observation Service | OGC. Available online: http://www.opengeospatial.org/standards/sos
32. istSOS Open Software for Standard Sensor Data. Available online: http://istsos.org/ (accessed on
20 March 2016).
33. Orcale Oracle VM VirtualBox. Available online: https://www.virtualbox.org/ (accessed on 12 March 2016).
34. PostgreSQL Global Development Group PostgreSQL: The World’s Most Advanced Open Source Database.
Available online: http://www.postgresql.org/ (accessed on 10 March 2016).
35. PHP Group PHP: Hypertext Preprocessor. Available online: http://php.net/ (accessed on 12 March 2016).
36. Apache Software Foundation Apache HTTP Server Project. Available online: https://httpd.apache.org/
37. PHP Template Engine | Smarty. Available online: http://www.smarty.net/ (accessed on 12 March 2016).
38. Open Source Geospatial Foundation Welcome to MapServer—MapServer 7.0.1 Documentation. Available
online: http://mapserver.org/ (accessed on 12 March 2016).
39. Bootstrap the World’s Most Popular Mobile-First and Responsive Front-End Framework. Available online:
http://getbootstrap.com/ (accessed on 12 March 2016).
40. jQuery Foundation jQuery—Write Less, Do More. Available online: https://jquery.com/ (accessed on
12 March 2016).
41. jQuery Foundation jQuery UI. Available online: http://jqueryui.com/ (accessed on 12 March 2016).
42. OpenLayers Contributors OpenLayers 2. Available online: http://openlayers.org/two/ (accessed on
12 March 2016).
43. Kralisch, S.; Zander, F.; Krause, P. Coupling the RBIS Environmental Information System and the JAMS
Modelling Framework. In Proceedings of the 18th World IMACS/and MODSIM09 International Congress
on Modelling and Simulation, Cairns, Australia, 13–17 July 2009.
44. JpGraph—Most Powerful PHP-Driven Charts. Available online: http://jpgraph.net/ (accessed on
23 March 2016).
45. Dygraphs. Available online: http://dygraphs.com/ (accessed on 23 March 2016).
46. Trirand jqGrid—jQuery Based Grid HTML5 Component for Javascript. Available online: http://www.
guriddo.net/demo/guriddojs/ (accessed on 12 March 2016).
47. Zander, F.; Kralisch, S.; Busch, C.; Flügel, W.-A. Data management in multidisciplinary research projects
with the river basin information system. In Open Data and Industrial Ecological Management; Hans-Knud, A.,
Knetsch, G., Eds.; Shaker Verlag: Herzogenrath, Germany, 2012; pp. 137–143.
48. Kosi RBIS. Available online: http://leutra.geogr.uni-jena.de/kosiRBIS/ (accessed 27 January 2013).
49. Nepal, S. Evaluating Upstream Downstream Linkages of Hydrological Dynamics in the Himalayan Region.
Ph.D. Thesis, Friedrich-Schiller-University, Jena, Germany, 2012.
50. Jakarta RBIS. Available online: http://leutra.geogr.uni-jena.de/jakartaRBIS/ (accessed on 17 March 2016).
51. Julian, M.M.; Fink, M.; Fischer, C.; Krause, P.; Flügel, W.-A. Implementaion of J2000 hydrological model
in the western part of Java Island, Indonesia. In Proceedings of the MacroTrend Conference on Applied
Science, Paris, France, 20–21 December 2013.
52. OTI River Basin Information System. Available online: http://leutra.geogr.uni-jena.de/otiRBIS/ (accessed
on 17 March 2016).
53. Badjana, H.M.; Zander, F.; Kralisch, S.; Helmschrot, J.; Flügel, W.-A. An information system for integrated
land and water resources management in the Kara River Basin (Togo and Benin). Int. J. Database
Manag. Syst. 2015. [CrossRef]
54. BrahmaRBIS. Available online: http://leutra.geogr.uni-jena.de/brahmaRBIS/ (accessed on 10 March 2016).
55. Brahmatwinn: Projekt. Available online: http://www.brahmatwinn.uni-jena.de/ (accessed on 24
March 2016).
56. Kralisch, S.; Zander, F.; Flügel, W. OBIS—A data and information management system for the Okavango
Basin. Environ. Assess. Okavango Reg. Biodivers. Ecol. 2013, 5, 213–220. [CrossRef]
57. The Future Okavango (TFO). Available online: http://www.future-okavango.org/ (accessed on 18
March 2016).
58. LUCCi. Land Use and Climate Change Interaction in Central Vietnam. Available online: http://www.lucci-
vietnam.info/ (accessed on 18 March 2016).
59. Assam RBIS. Available online: http://leutra.geogr.uni-jena.de/assamRBIS (accessed on 10 March 2016).
60. CNRD—Centers for Natural Resources and Development. Available online: http://www.cnrd.info/
61. TH-Köln ITT RBIS Server. Available online: http://rbis.itt.fh-koeln.de/ (accessed on 17 March 2016).
62. BASIN INFO—Web Based River Basin Information System. Available online: http://www.basin-info.net/
63. Centro-Norte-Chile RBIS. Available online: http://rbis.ceaza.cl/limariRBIS (accessed on 17 March 2016).
64. Zander, F.; Kretschmer, N.; Nauditt, A.; Astudillo, O.; Orrego, R.; Godoy, R.; Alvarez, P. Centro-Norte-Chile
River Basin Information System (RBIS). In Proceedings of the International Expert Symposium “Coping with
Droughts”, Santiago, Chile, 19–21 November 2014.
65. CEAZA—Centro de Estudios Avanzados en Zonas Áridas. Available online: http://www.ceaza.cl/en
© 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC-BY) license (http://creativecommons.org/licenses/by/4.0/).

River Basin Information System Open Environmental

Uploaded by

Copyright:

Available Formats

River Basin Information System Open Environmental

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

River Basin Information System Open Environmental

Uploaded by

Copyright:

Available Formats

International Journal of

ISPRS Int. J. Geo-Inf. 2016, 5, 123; doi:10.3390/ijgi5070123 www.mdpi.com/journal/ijgi

2. RBIS: Data Management, Analysis and Visualization

2.1.1. Overview and Detailed View

2.1.2. Multilingual Support

2.3. Time Series Data

3. Search and Filter

4. Interfaces and Services

5. Administration, User and Permission Management

6. Software Components and Architecture

6.1. Software Components

Figure 10. RBIS

Figure 11. RBIS architecture and components.

‚ Management of metadata datasets and responsible parties.

7.1. Application Types

7.2. RBIS System Performance

7.3. RBIS Deployment and Maintenance

8. Discussion and Outlook

Conflicts of Interest: The authors declare no conflict of interest.

22. SASSCAL. Available online: http://www.sasscal.org/ (accessed on 10 March 2016).

You might also like