Database Reference
In-Depth Information
10.4.1 Environmental Modeling Task—Evapotranspiration
We elaborate on a particular use case involving the integration of real-time
evapotranspiration observations and its comparison with estimations from
a weather model for accuracy assessment. Evapotranspiration (or ET) is a
term used to describe the sum of evaporation and plant transpiration from
the Earth's land surface to the atmosphere, which is an important part of
the water cycle. ET estimations are used in irrigation scheduling, watershed
management, weather forecasting, and the study of long-term effects of land
use change and global climate change.
78
A standard reference evapotranspira-
tion, denoted ETo, can be determined by using meteorological measurements,
which can be obtained from multiple sources for the same region. Compared
with station-observed ETo (point data), the weather model-based ETo (raster
data) has a continuous spatial coverage. In general, the model output accuracy
needs to be verified against the station observations. Assuming the observed
and the predicted data are published as WFS and WCS services, respectively,
data integration is needed to retrieve the model predicted data at the sta-
tion locations and compare the values, possibly after normalization and unit
conversions.
The scientific goal in this scenario is the visualization, monitoring, and
validation of model-based evapotranspiration for different eco-regions and se-
lected locations in California. This requires the overlay of ETo from the vari-
ous sources on a single display for visual analysis. We show next a geospatial
integration approach that allows the realization of this scenario.
10.4.2 Integration Platform
The overall conceptual architecture for geospatial data integration and in-
teroperability, which consists of data sources, structured repositories, service
middleware systems, and client applications, is depicted in Figure 10.6. Data
sources include remotely sensed imagery, model outputs, GIS stores, sensor
network systems, as well as other service-enabled data providers. The struc-
tured data repositories in general refer to data management systems including
databases and data stream engines. Service middleware refers to service en-
abling infrastructures that make the data available to clients by means of
Web service standards. Client applications allow users to search, query, and
retrieve metadata and data by using the provided Web services (CSW, WFS,
SOS, etc.), possibly in combination with more traditional access mechanisms
(HTTP, FTP, etc.) with the back-end data repositories.
In general, a service-oriented scientific data integration framework consists
of a set of interconnected service-enabled computation nodes. A service en-
abled computation node consists of both structured data repositories and ser-
vice middleware. Structured data repositories, such as DBMS, GIS data stores,
and data stream engines, store relational data, vector/raster geospatial data,
and real-time data streams. They can be connected to physical devices to
