final document

Transcription

final document

Global Drought Monitoring Service through the GEOSS
Architecture
Engineering Report
GEOSS Architecture Implementation Pilot (AIP)
Drought and Water Working Group
Version 2.0
Content developed by the GEO Architecture Implementation Pilot
Licensed under a Creative Commons Attribution 3.0 License
Content developed by the GEO Architecture Implementation Pilot
Licensed under a Creative Commons Attribution 3.0 License
Architectural Implementation Pilot, Phase 3
Global Drought Monitoring and European Drought
Observatory-Water SBA Engineering Report
Version:
2.0
Date: 11/Feb/2011
Revision History
Version
Date
Editor and
Content
providers
Comments
1.0
17/Dec/2010 W. Pozzi
1.4
03/Jan/2011
C. Fugazza
Revision to semantics-related sections
1.4
05/Jan/2011
M.J. Brewer
Revision to Global Drought Monitor Portal
1.4
07/Jan/2011
M. Santoro, S.
Nativi
System Architecture for the Discovery
Augmentation Component
1.8
18/Jan/2011
B. Lee
1.9
25/Jan/2011
W. Pozzi
Incorporation of GEO Ontology Registry
2.0
4/Feb/2011
M. Enenkel
Updating GLOWASIS
2.0
10/Feb/2011 M.J. Brewer
Review of NIDIS and GDMP sections
2.0
11/Feb/2011 W.Pozzi
Release
Document Contact Information
If you have questions or comments regarding this document, you can contact:
Name
Organization
Contact Information
W.Pozzi
GEO AIP Water and Drought Working Group &
IGWCO
Will.pozzi@gmail.com
M. Santoro
Italy National Research Council
santoro@imaa.cnr.it
C. Fugazza
Joint Research Center (JRC)
cristiano.fugazza@jrc.ec.euro
pa.eu
J.Vogt
JRC/European Drought Observatory (EDO)
jürgen.vogt@jrc.ec.europa.
eu
S. Nativi
Italy National Research Council
nativi@cnr.it
M. Brewer
US National Integrated Drought Information
System (NIDIS)/NOAA
R. Heim
US National Oceanic and Atmospheric
Administration (NOAA)
J. Sheffield
Princeton University
Page 3
S. Niemeyer
EDO
D. Cripe
GEO Secretariat Scientific Officer for Water
K. Korporal
Environment Canada
A. Howard
Agriculture and Agri-Food Canada
B. LloydHughes
University College London
Version:
2.0
Date: 11/Feb/2011
dcripe@geosec.org
J. Lieberman
W. Wagner
Technical University Wien
Michael
Piasecki
Consortium of Universities for the Advancement
of Hydrologic Science (CUAHSI)/City College of
New York (CCNY)
L. Nunez
Republic of Argentina Servicio Meteorologic
Nacional Drought Monitor
L. Bettio
Australia Bureau of Meteorology
M. Nicholson Australia Bureau of Agricultural and Resource
Economics and Sciences (ABARES)
B. Trewin
Australia Bureau of Meteorology
B. Lee
CSIRO
W. Sonntag
US Environmental Protection Agency
V. Guidetti
European Space Agency
R. Lawford
IGWCO
B. Hofer
EDO/JRC
D. Magni
EDO/JRC
L. Di
George Mason University
Eugene Yu
C. Yang
lawford@umbc.edu
M. Doubkova Technical University of Vienna
M. Enenkel
Technical University of Vienna
Page 4
Version:
2.0
Date: 11/Feb/2011
Table of Contents
A. Global Drought Monitoring Service and the Global Drought Community of Practice
1.1 Scope of this document
1.2 Importance of Global Drought Monitoring as a Critical Earth Concern and a
Prime Activity for GEO
1.3 Identification of Starting Conditions Fostering Drought is not Straightforward
1.3.1 Description of the Water Cycle
1.4 What are the User Requirements for an effective Drought Monitoring and
Forecasting Information System?
8 8 9 9 10 11 2. Drought Monitoring Components and Tools found in Hydrometeorology Drought
Monitoring Services within the Global Drought Community of Practice
13 2.1 European Drought Observatory
13 2.1.1 European Drought Observatory Portal Characteristics: “Drill Down” Capability 13 2.1.2 Importance of Soil Moisture for Monitoring Agricultural Drought
13 2.1.3 EDO-deployed Meteorological Drought Indicator: Standardized Precipitation
Index 15 2.1.4 Hydrologic Drought Indicator
15 2.2 USA National Integrated Drought Information System
16 2.3 Government of Canada Drought Coverage
18 2.4 Commonwealth of Australia Drought Monitoring
18 2.4.1 Commonwealth of Australia Water Availability Project
18 2.5 Africa Continental Drought Monitoring
19 2.5.1 Princeton Experimental African Drought Monitor
19 2.6 New Projects Permitting Further Development of the Global Drought
Monitoring Service
21 2.6.1 European Framework (EF) Drought Early Warning System for Africa-DEWFORA
21 2.6.2 GLOWASIS (Global Water Scarcity Information Service)
21 2.6.3 Satellite Application Facility on Support to Operational Hydrology and Water
Management (H-SAF)
24 2.7 South American Continent
26 2.7.1 Republic of Argentina Servicio Meteorologic Nacional Drought Monitoring
26 3. Capturing User Requirements for the Global Drought Monitor and its
Interoperability with the Global Earth Observation System of Systems (GEOSS)
30 3.1 Assessment of Drought Vulnerability and Susceptibility
30 3.2 Capturing User Requirements and Implementation of Architecture to Design of
the Global Drought Monitor
31 3.2.1 Portal Requirements: Drill-down capability
31 3.2.2 Top-down versus bottom-up Design
31 3.2.3 Soil Moisture and Agricultural Drought Monitoring Requirement
32 3.2.4 Republication of information to help decision makers facilitate drought decision
Page 5
Version:
2.0
Date: 11/Feb/2011
making
3.2.5 Hydrologic Drought Monitoring for Semi-Arid Areas and Meeting Hydrologic
Drought User Requirement through Semantics
3.3 Developing an Architectural Diagram for the GEO Global Drought Monitoring
Service
3.4 Semantic Development Activities within GEO: the Data Integration and
Analysis System (DIAS) Contribution from Japan
3.4.1 Adding Advanced Search and Discovery using Semantics
32 32 33 36 39 4. Global Implementation of the Drought Monitoring Service through GEOSS
40 4.1 Components of the System Architecture of the Global Drought Monitoring
Portal
40 4.2 Actors
41 4.3 Capturing User Requirements for the Global Drought Monitor Portal through
the GDMP Scenario
41 4.3.1 Display of Selection Bar for Drought Indices, Processing to Derive Dehydration
and Drought Severity, and Drought Map Republication
42 4.3.2 Layout and Organization of the GDMP within the NIDIS GIS Server
43 4.3.3 Implementation of Advanced Search and Discovery in the GDMP
44 4.4 Support of Increased Global Coverage within the web-based, real-time GDMP
server
44 4.5 Integration of GDMP with GEOSS Architecture
44 4.6 Remote Sensing Soil Moisture Integration
45 4.7 Adding Water Usage Information Layers, including Agriculture
45 5. Advanced Search and Discovery Capability within the European Drought
Observatory
49 5.1 Components of the European Drought Observatory
49 5.1.1 European Drought Observatory user access
49 5.1.2 Organization and layout of the EDO map server page (scenario step 01—
continued)
50 5.1.3 Selection of Drought Indices
50 5.1.4 Processing Step by Running Drought Indicators over a Selected Spatial Domain51 5.1.5 Automated Email Alerts and Drought Triggers
51 5.1.6 Context and pre-conditions
52 5.2 Implementation of the European Regional Drought Semantic-enhanced
Monitoring and Information System
52 5.2.1 Advanced Semantic Search
54 5.3 EuroGEOSS Deployment of the Foundation Vocabularies
55 5.4 Fine Tuning the Foundation Vocabularies for SBA Application—Specialized
Drought Vocabulary
55 5.4.1 Water Ontology-enablement within the DAC Semantics
56 5.5 How the EuroGEOSS Discover Augmentation Component supports semantic
searches
56 5.6 Operation of the Water Ontology within the EuroGEOSS Discovery
Page 6
Version:
2.0
Date: 11/Feb/2011
5.6.1 Searching for Concepts/Terms
5.6.2 Multilingual Concepts/Terms
5.6.3 European Drought Observatory (Client) Query
5.6.4 WPS Request
5.7 Use of EuroGEOSS Semantic Discovery within the European Drought
Observatory (Returning back to the Scenario)
5.8 Interoperability Arrangements with GEOSS
5.9 Post Deployment Activities
5.9.1 Ontology Engineering
6. Evaluating How the Advanced Semantic EuroGEOSS Search and Discovery System
Works
7. Drought Metadata for fostering interoperability between EDO and EU national
drought monitors
8. Range of Issues Covered by the Water Working Group
9. References
10. EuroGEOSS Drought Vocabulary Keywords
11. EuroGEOSS Water Societal Benefit Area Keywords
12. Acknowledgments
Page 7
58 58 58 59 59 59 59 60 60 61 62 64 65 70 72 72 Architectural Implementation Pilot, Phase 3
Version:
2.0
Date: 11/Feb/2011
Drought Monitoring and Water Activities within the
Group on Earth Observations (GEO)
A. Global Drought Monitoring Service and the
Global Drought Community of Practice
1.1 Scope of this document
This is an overview and documentation of the drought monitoring service as implemented
through the Group on Earth Observations System of Systems (GEOSS) and the European
Drought Observatory implementation of advanced semantic search capability through the
EuroGEOSS Discovery Broker tools. A key deliverable is the specification of a set of tools that
will access information published through a distributed water data infrastructure. The
development of the specification of these tools includes: 1) capturing user requirements through
expressing the GEO Water Societal Benefit Area users within a “scenario,” that is, who might
use the GEO Global Drought Monitoring Service and the types of data and functionality that
these users require or expect; 2)Design of a system architecture and the enabling framework for
this at the component level; 3) integration of this system architecture within the Global Earth
Observation System of System (GEOSS) architecture and its components; and 4)
implementation. The development efforts of the GEO Global Drought Monitoring Portal have
involved multiple parties, including the Architectural Implementation Pilot (AIP) Water and
Drought Working Group, through the GEO Architecture and Data Committee level; the
Scientific Officer for Water of the GEO Secretariat (through the Global Drought Monitoring
Initiative); drought task activities of the Integrated Global Water Cycle Observations (IGWCO)
Community of Practice; Princeton University Land Surface Hydrology Group, USA National
Integrated Drought Information System (NIDIS), the European Drought Observatory, Italian
National Research Council, the Joint Research Centre, the University College of London, the
Technical University of Vienna, Canadian Group on Earth Observations (CGEO), Argentina
Servicio Meteorologico Nacional, Australia Bureau of Agricultural and Resource Economics and
Sciences (ABARES), and the Australia Bureau of Meteorology.
This report is divided into two sections to increase its accessibility. The first section
explains why certain portal Information Technology (IT) capabilities (“user requirements”) were
selected for implementation and deployment within the global drought monitoring service. The
first section deals with development of a web-based, real-time Geographic Information System
GIS server with a distributed database federation, used for hydrologic alerts in drought
conditions, a prototype global drought early warning system. The second section explains why
certain advanced search capabilities (including “semantic” search and discovery)—again, user
requirements—were developed for implementation within the European Drought Observatory
and the EuroGEOSS discovery broker. These technologies can also be eventually migrated for
implementation within the Global Drought Monitoring Portal (GDMP), combining with
concurrent semantic components being built within the Global Earth Observation System of
System through the Data Integration and Analysis System (DIAS), the Japanese Aerospace
Exploration Agency (JAXA) contribution to GEO, and the EuroGEOSS European Union
contribution.
Page 8
Version:
2.0
Date: 11/Feb/2011
1.2 Importance of Global Drought Monitoring as a Critical Earth Concern and a
Prime Activity for GEO
Given current concerns with the increasing frequency and magnitude of droughts in many
regions of the world, especially in the light of expected climate change, drought monitoring and
dissemination of early warning information in a timely fashion is a critical concern. The
European Union experienced intense drought and heat waves in 2003, Argentina in 2008/2009,
southeast Australia in 2009, while, at the same time, the Intergovernmental Panel on Climate
Change (IPCC) climate projections for the 21st century suggests an increased frequency of severe
droughts in continental USA and Mexico, Mediterranean Basin, parts of northern China,
Southern Africa, Australia, and parts of South America. In addition, current agricultural
production is being maintained by multiple crop cycles over the course of a single year in India
and China, for example, and drought is exhausting secondary supplies of groundwater , as the
drought exhausts surface water supplies, creating a dependency upon the groundwater sources
needed to maintain these multiple crop cycles.
Droughts and famine are inseparable from one another: droughts lower agricultural
production. Current agricultural monitoring efforts, such as the European Union (EU)
Monitoring of Agricultural Resources with Remote Sensing (MARS Food-Sec), the USA
Department of Agriculture (USDA) Foreign Agricultural Service, and the Famine Early Warning
System (FEWSNET) have developed methodologies for estimating the impact of drought upon
agricultural production, such as the Food and Agricultural Organization (FAO) Water
Requirements Satisfaction Index (WRSI)(as renamed Global Water Satisfaction Index by MARS
and GeoWRSI by FEWSNET). The MARS convention implies that a WRSI or GWSI of 50
represents a famine condition (actual evapotranspiration of half the plant water requirement).
Advances in Land Surface Modeling, as in more sophisticated representation of soil
water process, including linkage of groundwater with surface water, is just one way in which
new technologies are available to upgrade the more schematic soil water balances incorporated
within WRSI. Additional new technologies are coming online with respect to satellite-based soil
moisture sensors. Standardization of global meteorological datasets has permitted the running
Land Surface Models and distributed hydrological models in near-real-time (NRT). IT
infrastructure and informatics methodologies, combined with all these scientific advances, have
now created the opportunity to develop a more up-to-date, comprehensive, useful-to-decision
making drought monitoring capability. Additional advances in web-based, real-time (RT)
Geographic Information Systems (GIS) with supporting distributed databases (Wangmutitakul,
et. al., 2003; Wang 2005; Chalainanont, et. al. 2007 or web map services in time-critical
applications (Zhang and Li 2005; Ozdilik and Seker).
The role that GEO plays in this process is to provide a rich collaborative environment,
fostering collaboration among the USA, Canada, European Community, Asia, Australia, and
South America.
1.3 Identification of Starting Conditions Fostering Drought is not straightforward
Page 9
Version:
2.0
Date: 11/Feb/2011
The American Meteorological Society Glossary defines “drought” as “a period of
abnormally dry weather sufficiently long enough to cause a serious hydrological imbalance.”
Agricultural drought is defined as “conditions that result in adverse crop responses, usually
because plants cannot meet potential transpiration as a result of high atmospheric demand and/or
limited soil moisture.” Hydrologic drought is defined “prolonged period of below-normal
precipitation, causing deficiencies in water supply, as measured by below-normal streamflow,
lake and reservoir levels, groundwater levels, and depleted soil moisture.” The definition of
agricultural drought stipulates that soil moisture monitoring is the methodology of choice for
monitoring drought afflicting agriculture. The definition of hydrologic drought stipulates that
monitoring of streamflow (including baseflow), groundwater levels, and soil moisture may be
necessary in order to monitor hydrologic droughts. Indeed, the complexities of water cycle
processes found in semiarid terrain, particularly processes in the vadose zone, may be critical in
identifying drought’s early stages. Global drought monitoring capability includes the capability
to monitor drought in many diverse semiarid conditions. The definition of the different types of
droughts, particularly hydrological droughts stipulate that monitoring capability of groundwater,
stream flow, soil moisture, snow storage at the start of spring meltwater season, and river water
level may be prerequisites or user requirements for an effective global drought monitoring
program. These, in turn, establish user requirements for an information system that support
global and regional drought monitoring.
1.3.1 Description of the Water Cycle
The water cycle begins—after evaporation of water over the oceans—as rain out over
land through which precipitation—if temperatures are low enough—which falls as frozen water
which accumulates on top of the surface of land as layers of snow or glacial layers.
Alternatively, precipitation falls—if temperatures are high enough—in its liquid form and
infiltrates into soil (unless the soil has a precondition of already being water saturated. This
infiltration and percolation occurs both as flow through the pores of the soil and flow through
macropores or fractured rock. Drainage may occur from topsoil through thick vadose zones in
semi-arid areas, until the water reaches layers of saturation of pores with water, called
groundwater. The proximity of groundwater to the surface determines whether water is
exchanged between groundwater, with groundwater discharge occurring into streams or rivers or
groundwater recharge occurring through river or streamflow. Furthermore, semiarid areas may
be characterized by ephemeral flashfloods, making the occurrence of such sources of water
difficult to typify statistically. One key difference is that flow of water through pores in soil is a
very slow, diffusive process which occurs over much longer time scales—decades or longer—
than the more rapid, prompt runoff processes occurring at the surface. The point to be made here
is that drought originates as a deficiency of frozen precipitation stored on the surface or liquid
precipitation that slowly works its way through the processes of the hydrologic cycle. Some of
these processes and events, such as decline of soil water, occur rather rapidly and impact the
growth stage of a crop in agriculture (agricultural drought indicator), while other processes,
drawdown of groundwater level and lowered discharge of groundwater to river baseflow can
occur over seasonal time scales or longer (hydrologic drought) (Van Lanen et al., 2004).
Drought can also exhaust municipal water supplies, both as drops in reservoir levels of
Page 10
Version:
2.0
Date: 11/Feb/2011
stored water or declines of water surface elevations within stream and river networks.
Environmental flow requirements are not met, causing environmental impacts as well. Cooling
water for thermal power plants is not available. All of these water cycle processes are often
lumped under the generic term hydrologic drought, but the actual nature of the drought may be
caused by a multiplicity of factors.
Hydrologic droughts can occur through groundwater flow or streamflow. Groundwater
droughts can be the result of long periods with below average precipitation. Van Lanen &
Tallaksen (2007) have compared different terrains having a slow and a fast responding
groundwater system to conclude the effect of the groundwater system on the frequency and
duration of droughts was larger than the effect of different soil types. The groundwater system
has large influence on the propagation of droughts through the hydrological cycle and hence on
drought characterization (Van Lanen & Tallaksen (2008); Wanders, van Lanen, and van Loon
2010). The Total Storage Deficit Index, developed by Yirdaw et al (2008) used NASA Gravity
Recovery and climate Experiment observations to attempt to quantify the groundwater role in
hydrologic drought in the Canadian prairies. Terrestrial water storage changes can also be
adopted for drought monitoring strategies (Rodell). Drought indicators have also been
developed for evapotranspiration (Anderson)
1.3.1.1 Difficulties in Identifying Drought Conditions
Drought lacks a precise and universally accepted definition. The detection of the
threshold beyond which a drought episode begins is difficult to determine out of the statistical
noise that creates random fluctuations (V. Castillo 2009; Moreira et al 2008). Requirements for
drought detection include methodology that can select drought events from the remainder of the
meteorological or hydrological time series, a truncation level or threshold which divides the time
series into “above normal” and “below normal” sections (Dracup et al 1980). The truncation
level can be set to cut the series at several places, and “run length” is the distance between
successive crossings across the threshold; the run intensity is the average deviation from the
threshold (van Lanen et al 2008). Probabilistic prediction tools have also been developed.
1.4 What are the User Requirements for an effective Drought Monitoring and
Forecasting Information System?
The integration of drought information (indices and impact indicators) in a
comprehensive framework (composite index and maps) is the starting point for developing a
drought monitoring system. Several integrating methodologies have been explored in AIP-3.
Drought Monitoring may be summarized as a back-end information system, linked to an
application that, in turn, is at the back-end of a user accessible portal.
Page 11
Version:
2.0
Date: 11/Feb/2011
For example, lack of soil moisture availability is used to define conditions for
agricultural drought, and the shortage of ground-based in-situ soil moisture measurement
stations requires estimation of soil moisture over large land tracts using Land Surface Models
(such as the NCAR Community Land Model, or the ensemble National Land Data Assimilation
System (NLDAS within the USA) or distributed hydrological models (such as the Variable
Infiltration Capacity VIC model or LISFLOOD). Such models are linked systems of partial
difference equations that ingest multidimensional arrays of near-real-time or real-time
meteorological and precipitation data as functions of time. However, despite the fact that such
multi-dimensional data are solved across a lattice of grid cells that emulate spatial locations, the
output arrays for each respective area have to be geographically registered in order to be
imported into a Geographic Information System (GIS). In short, the complex land surface model
and Geographic Information System (GIS) are separate packages (applications). The advantage
of linking together the Land Surface Models or distributed hydrological models with a GIS is
that the soil moisture (as well as other water budget component and drought indicators) can then
be added together or republished as layers within a map, displayed with the drought impact
information, such as crops dependent upon green water. The soil moisture may then be
combined with different layers of information within the GIS, published as maps, and exchanged
using OGC Web Mapping Services (WMS) among individual national hydrometeorological
service drought monitors and the global drought monitor. This is the information system behind
the application (and the front end portal user interface), and this integrates the drought
information (indices and impact indicators) with maps of drought severity rankings and
vulnerability or impact factors. The observing system is comprised of the ground-based or
satellite-based observations used to derive the meteorological and precipitation forcing used in
the Land Surface Models or distributed hydrologic models.
Page 12
Version:
2.0
Date: 11/Feb/2011
2. Drought Monitoring Components and Tools found in Hydrometeorology
Drought Monitoring Services within the Global Drought Community of Practice
This section surveys the different types of Drought Monitoring Systems and why certain
techniques were chosen for a basis of the design of the global drought monitor portal.
2.1 European Drought Observatory
2.1.1 European Drought Observatory Portal Characteristics: “Drill Down”
Capability
Within the European Community, the European Drought Observatory (EDO)’s map
server utilizes a common spatial resolution of 20 km, while the national EU drought monitor
maps have higher spatial resolution. Common registration of datasets through the EuroGEOSS
discovery broker enables the highest resolution maps to be exchanged with the EDO, since the
overall system is utilizing a common set of standards. The EDO map server can exchange map
via web services with the Ministerio de Medio Ambiente (MARM) in Spain, for example, so that
maps of higher spatial resolution can be republished for the benefit of a user query. The design
principles for the European Drought Implementation (the combined EuroGEOSS discovery
broker and EDO and national drought monitors within the EC) were: 1) decentralized data
holdings but direct linkage and exchange using common format and standards; and 2) a set of
products agreed in common among all partners to be made available and exchanged, such as
Standard Precipitation Index and soil moisture anomaly. Common metadata and registration
through the EuroGEOSS discovery broker make the linking of data among river basin, nation,
and regional level possible (as well as interoperable).
2.1.2 Importance of Soil Moisture for Monitoring Agricultural Drought
The EDO currently measures the presence of agricultural drought by estimating soil moisture across the European Union, using the LISFLOOD model. The LISFLOOD model is used for forecasting floods, as part of the European Flood Alert System (EFAS), and the soil moisture outputs of the model are extracted for use in drought monitoring. Continuous
simulations with the LISFLOOD model within the European Flood Alert System produce daily
soil moisture maps of Europe. Having the soil saturated with water is a precondition for flooding,
since any additional liquid precipitation will run off immediately. LISFLOOD is run using near-‐real-‐time meteorological data, including precipitation, derived from measured and
spatially interpolated meteorological point data provided by the MARS-STAT activity of IPSCJRC (so called JRC-MARS data). Due to the reception via the Global Telecommunication
System of WMO and further processing the data are typically one to two days behind the current
date. The LISFLOOD model is run twice daily on a Linux cluster. The spatial resolution of
LISFLOOD on the pan-European scale is currently at 5 km.
Page 13
Version:
2.0
Date: 11/Feb/2011
Daily soil moisture map on is presented in form of soil suction (pF) values of the top soil layer
that commonly range between 1.5 for very wet conditions up to 5.0 for very dry soils.1 The pF
value describes the forces necessary for plants to apply in order to extract water from the soil for
their use.
2.1.2.1 Development of the Soil Moisture Climatology
The “climatology” for soil moisture has been derived as year-to-year outputs from the
LISFLOOD model, as having been generated from the Re-Analysis data of the European Centre
for Medium-Range Weather Forecasts (ERA-40) that comprise the period 1958-2001 (i.e. 44
years), along with updating made available from measured meteorological data from JRC-MARS
from the Global Telecommunication System of WMO covering 1990 to 2006, i.e. a period of 17
years. (Compare this with the Princeton datasets below). Soil moisture anomalies are calculated
from the climatology.2
2.1.2.2 Soil Moisture Anomaly Forecasts prepared from the climatology
Soil moisture and soil moisture anomaly forecasts are derived using the same modeling
approach but, with the exception of using the short term meteorological forecasts rather than
near-real-time meteorology data. In the forecasting mode the European Flood Alert System
produces information on the development of soil moisture in Europe for up to ten days ahead.3
The anomaly forecast is also made.4
The trend map of soil moisture describes qualitatively the change in soil moisture, currently
between today and the seventh day ahead. Orange to red colors indicate drying conditions, while
yellow to green colors predict wetter conditions during the next week.
1
http://desert.jrc.ec.europa.eu/action/php/index.php?action=view&id=19
2
3
4
Page 14
Version:
2.0
Date: 11/Feb/2011
Figure 1 LISFLOOD forecasted normalized top soil moisture suction (pF) for Europe. The pF
values have been normalized by ECMWF ERA-40 statistics.
2.1.3 EDO-deployed Meteorological Drought Indicator: Standardized
Precipitation Index
Precipitation anomalies are expressed the monthly Standardized Precipitation Index (SPI)
of the last month, a well-known meteorological drought index. Monthly SPI values reflect short
term changes in precipitation as compared to the long-term average of the respective month.
Positive SPI values indicate greater than median precipitation, and negative values indicate less
than median precipitation (McKee et al. 1993).
2.1.4 Hydrologic Drought Indicator
A hydrological drought is described usually by the analysis of stream-flow, lake, or
reservoir level data. Opposite to meteorological information, hydrological data are collected
throughout Europe, but are generally stored locally at the national or even regional level, often
with varying formats and qualities less consistent than for meteorological data. Here, after
careful calibration, the hydrological model LISFLOOD might contribute to the forecasting of
Page 15
Version:
2.0
Date: 11/Feb/2011
low flows by predicting discharge as it is already being doing for the prediction of flood events
in major pan-European catchment areas.
2.2 USA National Integrated Drought Information System5
The US National Integrated Drought Information System (NIDIS) is the national drought
early warning system for the US. It employs three key tools: 1) the US Drought Monitor6 (; 2)
the Drought Impact Reporter7; and 3) the US Drought Outlook , and hundreds of supplemental
indicators, services, forecasts, and tools, to provide a snapshot of current drought conditions,
how those conditions are affecting local populations, and whether the drought will continue.
2.2.1.1
USA NIDIS Portal Drill-Down Capability
One interesting feature of the NIDIS map server is that one begins with a national map of
drought conditions within the USA and then “drills down” to the region level and then to the
basin level.
The first drop down tab is “Drought monitor date,” while the second is “Zoom to area.” The
third drop down tab is “Zoom to basin,” which currently includes the Upper Colorado River
Basin and the Lower Colorado River Basin.
Such a “drill down” system—used by both the European Drought Observatory and the
USA National Integrated Drought Information System portal—can integrate the basin scale
drought maps, national scale, continental scale, and global scale and, correspondingly, was
selected for implementation within the Global Drought Monitoring Portal.
2.2.1.2 Soil Moisture Monitoring for Agricultural Drought
As in the case of EDO, the USA National Integrated Drought Information System
(NIDIS) contains soil moisture and soil moisture anomaly maps.8
5
6
http://www.drought.gov/portal/server.pt/community/drought.gov/202
http://www.drought.gov/portal/server.pt/community/drought_indicators/us_drought_monitor
7
http://www.drought.gov/portal/server.pt/community/impacts/210
8
http://www.drought.gov/portal/server.pt/community/forecasting/209/soil_moisture/338
Page 16
Version:
2.0
Date: 11/Feb/2011
Ensemble soil moisture that is based upon multiple Land Surface Models and distributed
hydrologic models are available from the NASA/GSFC National Land Data Assimilation System
(NLDAS) ensemble Drought Monitor.9
The University of Washington experimental US surface water monitor is based on the
Variable Infiltration Capacity (VIC) distributed hydrologic model.10
The Center for Climate Prediction (CPC) produces “Leaky Bucket Model” soil moisture.11
2.2.1.3 Agricultural Drought Short Term Forecasting
The US Drought Outlook provides an integrated drought forecast, relying heavily on the
NOAA Climate Forecast System (CFS) and is issued for time-scales out to three months12 .
The soil moisture anomaly forecasts are based upon the NOAA Global Forecasting
System (GFS) model; soil moisture anomalies are based upon a 1971-2000 mean climatology.13
2.2.1.4 Indicators for Monitoring Meteorological Drought
A variety of Drought Indicators are made available on the NIDIS site.14 Examples include such
items as Standardized Precipitation Indices and Palmer Drought Indices at short time-scales,
2.2.1.5 Agricultural Impacts Estimation
Agricultural impacts are currently tracked by a system utilizing color coding for pasture and
range land in “poor” and “very poor condition”15
9
http://www.emc.ncep.noaa.gov/mmb/nldas/drought/
10
http://www.hydro.washington.edu/forecast/monitor/
11
http://www.cpc.ncep.noaa.gov/products/Soilmst_Monitoring/
12
http://www.drought.gov/portal/server.pt/community/forecasting
13
http://www.cpc.ncep.noaa.gov/soilmst/forecasts.shtml
14
http://www.drought.gov/portal/server.pt/community/drought_indicators/223
15
http://www.drought.gov/portal/server.pt/community/impacts/210/tracking_agricultural_impacts/
307
Page 17
Version:
2.0
Date: 11/Feb/2011
2.2.1.6 Hydrologic Monitoring
Hydrologic drought monitoring measures and forecasts the amount of water in lakes, rivers, and
aquifers.16 Drought takes longer to show up in hydrological systems than in agriculture,
especially when reservoirs and rivers are managed to balance the extremes of wet and dry years.
Snow is a major component of water supply in the western United States.
2.3 Government of Canada Drought Coverage
Canada in 2004 extended drought mapping coverage from agricultural areas to remainder
of the Canada Provinces. Canada does not carry out drought mapping within the territories
(Yukon, Northwest territories, and Nunavut north-of-tree line and permafrost underlain areas).
Near-Real-Time monitoring is carried out for 508 of 761 ground-based stations by Agriculture
and Agri-Food Canada (AAFC)(Hadwen2008) which runs a national drought model, in which
Standard Precipitation Index is calculated, soil moisture (as percent of average and difference
from normal), and Palmer Drought Severity Index.
2.4 Commonwealth of Australia Drought Monitoring
Water issues are now considered among the most important drivers and constraints on natural
resource management in Australia; from environmental hazards like salinity and drought,
through to security of urban and rural water supplies. At present, Australia has no
comprehensive, consistent source of information on the water balance of its landscapes; that is,
on the relationship between rainfall, evaporation, transpiration, soil moisture, runoff and
drainage to ground and surface water. A better understanding of water availability is needed
across the entire country and is relevant to the implementation of key Australian Government
policies such as Exceptional Circumstances, the National Water Initiative, the Prime Minister’s
National Plan for Water Security and policies in support of improved natural resource
management.
2.4.1 Commonwealth of Australia Water Availability Project
The Australian Water Availability Project is a partnership established in 2004, between the
Bureau of Rural Sciences, CSIRO, the Bureau of Meteorology and the Australian National
University. The project aim is to develop an operational system for estimating soil moisture and
other components of the water balance, at scales ranging from five kilometers (km) to all
Australia, over time-periods ranging from daily to decades. Data from ground-based climate
measurements, remote sensing and models (water, plant and climate) are being combined to
produce maps of historic and current levels of all the main components of the landscape water
balance, including rainfall, evaporation, transpiration, available soil moisture, runoff, stream
flow and deep drainage. The future challenge is to deliver a fully web operational system,
including underpinning procedures for robust real-time product delivery, continuous
16
http://www.drought.gov/portal/server.pt/community/hydrological_monitoring/224
Page 18
Version:
2.0
Date: 11/Feb/2011
improvement and validation, and links to seasonal forecasting of water balance conditions. The
fundamental data derived from this project will help underpin future planning and decisionmaking on a range of issues including drought management and policy, securing urban and rural
water supplies, salinity, biodiversity management, ecosystem services and sustainable farming.
The real-time web operational system to be developed will help agricultural industries
maintain farm profitability before, during, and after drought events and help water and catchment
managers quantify the impact of climate cycles or climate change on surface and groundwater
recharge, vegetation and biodiversity. Additionally, risks to agricultural production may be
assessed by detailed analysis of moisture availability and moisture utilization trends for all
Australia. The Bureau of Rural Science presents Water Balances for Recent Months, Water
Balance Annual Average, Water Maps, Land Use Maps, and Social Data.17
2.5 Africa Continental Drought Monitoring
Regional African drought monitoring networks have been started at AGRHYMET in
West Africa, the Southern Africa Development Center Drought Monitoring Center. The
Princeton Experimental African Drought monitor offers pan-Africa coverage.
2.5.1 Princeton Experimental African Drought Monitor18
The Variable Infiltration Capacity (VIC) Model is used to calculate soil moisture.19 The
Princeton Land Surface Hydrology Group initially developed a meteorological forcing dataset
for global land areas for 1950-2000 to force the Variable Infiltration Capacity (VIC) distributed
hydrologic model. The subsequent task during the first interim period involved updating the
global macro scale modeling to near-real-time over Africa. A particular challenge over the
African continent has been to update the forcing dataset (1950-2000) (and thence the VIC
simulation) to near-real-time (NRT) using available data streams.
2.5.1.1 Climatology
The 1950-2000 meteorology—the climatology—is derived from a blending of reanalysis
(NCEP/NCAR) and gridded observation-based datasets including the Climatic Research Unit's
TS2.0 monthly precipitation and temperature dataset, the NASA Tropical Rainfall Measurement
Mission (TRMM) 3-hourly precipitation products and the NASA Surface Radiation Balance
(SRB) short- and long-wave datasets. In effect, the observation datasets are used to spatially
17
http://adl.brs.gov.au/water2010/water_cycle/index.phtml
18
http://hydrology.princeton.edu/~justin/research/project_global_monitor/index_africa.html
Page 19
Version:
2.0
Date: 11/Feb/2011
downscale the reanalysis, which is available at high temporal resolution, and at the same time
remove biases in the reanalysis. This work is described in detail in Sheffield et al. (2006).
2.5.1.2 Bridging the gap between reanalysis data and real time observing system
data
To bridge the data gap between the beginning of 2001 and near-real-time, these methods
were extended to blend reanalysis with available observations. Although reanalysis data are
available up to real-time, most observation-based datasets are generally only available some
months of even years behind real-time. Therefore for 2001-realtime we have used a number of
different datasets depending on their availability. For 2001-2006, we have used the recently
updated (to 2006) monthly gridded precipitation and temperature dataset of Willmott and
Matsura. This matches well the CRU dataset (used for 1950-2000) over their overlap period at
large scales. From the beginning of 2007, we have used the Global Precipitation Climatology
Project (GPCP) monthly dataset which is available a few months off real-time. Ongoing work is
looking at the differences between these various datasets during their overlap periods and
methods to ensure temporal consistency. For the last few months up to real-time, we are relying
on real-time precipitation products (PERSIANN20 data from University California Irvine,
TRMMM data from NASA) and gauge telemetry (Global Telecommunication System (GTS)
gauge data from NOAA). These products are being downloaded on a daily basis and are
blended into a forcing dataset for VIC over Africa.
Having set the initial meteorological forcing into place, the VIC simulations have been
run, up until near-real-time, in order to establish operational running. Our immediate objectives
are to finalize the data streams for the real-time running of the VIC model. The rapid timing of
real-time operational monitoring creates problems, such as the need to assess whether input data
are available, as well as developing fall-back methods for when data are unavailable or fail
quality control checks. Furthermore, the real-time meteorological data are likely biased, creating
the need to periodically re-run the VIC model up to a few months off real-time when the longterm gridded observation-based products (which are our best estimates of precipitation and
temperature) are updated, to avoid a drift in the land surface states.
The probability distributions of total column soil moisture and runoff for each grid cell
and each month constitute the climatology, against which current conditions can be compared.
The screening tools account for drought areal extent and duration using concepts adapted from
Andreadis et al (2005), which involve a form of spatial cluster analysis to identify drought
patterns from gridded model output. Based on the historic analysis, we will establish a set of
severity-area-duration thresholds that can be used to screen evolving droughts. Within the real
time monitoring framework, we will monitor where drought thresholds are crossed for either soil
moisture or runoff. Once the prescribed drought thresholds have been crossed, we will continue
to track drought evolution in time (i.e., in subsequent forecasts), until the nowcasts indicate that
20
http://chrs.web.uci.edu/research/satellite_precipitation/activities00.html
Page 20
Version:
2.0
Date: 11/Feb/2011
the drought has dissipated. Drought dissipation will be evaluated in comparison with severityarea-duration thresholds estimated using an approach similar to the one used to establish drought
screening thresholds.
2.6 New Projects Permitting Further Development of the Global Drought
Monitoring Service
2.6.1 European Framework (EF) Drought Early Warning System for Africa-DEWFORA
Under the European Framework, the Drought Forecasting and Early Warning System for
Africa (DEWFORA) project has been funded to set up a regional drought monitor for Africa.
DEWFORA also includes local and regional pilot projects. This is the reason why DEWFORA
is listed in parallel with the Princeton African drought monitor.
Figure 2 DEWFORA study regions (Werner et al 2010)
2.6.2 GLOWASIS (Global Water Scarcity Information Service)
GLOWASIS will combine in-situ, satellite derived and statistical data on water supply
and demand and make them available through a public information portal on water scarcity.
Funded under the European FP7 framework, the overall objective is to pre-validate a GMES
(Global Monitoring for Environment and Security) Service for Water Scarcity information, based
on pilot studies in Europe, Africa and on global level. The main objectives are:
•
Assessment of water demand and supply
Page 21
Version:
2.0
Date: 11/Feb/2011
•
Near real-time reporting on disasters (droughts, floods)
•
Medium and long-term forecasting (also with respect to climate change)
•
Promotion of new satellite-capabilities (e.g. Sentinel 1)
•
Matching new satellite-capabilities to specific user requirements
GLOWASIS will be made interoperable with the Water Information System for Europe
(WISE-RTD), linking water demand and supply with existing tools, such as the European
Drought Observatory (EDO) and PCR-GLOBWB, a global hydrological model (the same model
used in DEWFORA), combining complex water cycle variables in a standardized format with
respect to water scarcity information.
Sources of information and data are:
•
Already existing GMES (Global Monitoring for Environment and Security) data, such as
the LMCS (Land Monitoring Core Service) of GEOLAND2,
•
in-situ data from GEWEX (Global Energy and Water Cycle Experiment) and Global
Terrestrial Network on Hydrology (GTN-H) initiatives, such as the International Soil
Moisture Network,
•
statistical databases (e.g. AQUASTAT and SEEAW)
Results of GLOWASIS can be used in research, for practical implementation and
management purposes. Therefore, end-users encompass river basin management organizations
(Rhine, Danube, Elbe, Oder), the European Environment Agency, UN-Water, the Australian
Bureau of Meteorology, etc. GLOWASIS is coordinated by DELTARES, the Netherlands. The
Institute of Photogrammetry and remote Sensing (IPF) leads one work package (user
requirements) and is involved in all others.
As one of IPF’s most successful projects on soil moisture, SHARE will also contribute to
GLOWASIS. The following sub-chapters give an overview about soil moisture products from
ASAR (advanced synthetic aperture radar) and scatterometer sensors.
SHARE is a DUE Tiger Project of the European Space Agency, which offers an
operational soil moisture monitoring service. The synergistic use of ENVISAT's ASAR sensor
and scatterometers (on METOP and ERS) allows for frequent, high resolution monitoring of
regional soil moisture dynamics.
An algorithm was developed at IPF to detect surface soil moisture from active microwave
systems. Active sensors are sensitive to soil moisture mainly due to distinct dielectric properties
of water stored in soil. Microwaves of the Advanced Synthetic Aperture Radar (ASAR) and the
advanced scatterometer (ASCAT) cannot penetrate soil deeper than a few centimetres. In case of
ASCAT an algorithm was developed, which models the soil water content in deeper layers (the
soil water index, SWI). It is obtained by filtering surface moisture time series with an
Page 22
Version:
2.0
Date: 11/Feb/2011
exponential function (WAGNER et al., 1999). Being able to model the profile soil moisture up to
one metre facilitates estimations of infiltration capacities and plant available water (defined as
the difference between field capacity and permanent wilting point). This is the approach used in
the agricultural monitoring and forecasting models cited in section 1.3 above. Flooded soils are
more prone to cause flooding, as noted in section 2.1.2 above.
The two systems to obtain soil moisture data:
•
Medium resolution soil moisture from an imaging Advanced Synthetic Aperture Radar
(ASAR) onboard ENVISAT can be operated in global monitoring or wide swath mode. It
was the first system to deliver global backscatter measurements in C-Band (5.3 GHz) at a
spatial resolution of one kilometre. Spatial resolutions of 150 meters can be achieved by
SCAN SAR wide-swath mode. In the SHARE project, regions on three continents have
been monitored once or twice a week. Soil roughness and vegetation effects of each pixel
are “corrected” by change detection method – the subtraction of a reference image from a
SAR image. This way the inhomogeneous distribution of soil water in the topmost
centimetres of the unsaturated zone, where evapotranspiration takes place, can be
considered. The most recent version of the ASAR data viewer is online at:
http://www.ipf.tuwien.ac.at/radar/dv/ipfdv/index.php?dataviewer=asar2
•
Scatterometers onboard METOP (ASCAT), ERS-1 and ERS-2 (SCAT) are non-imaging
sensors and characterised by higher temporal (1-2 days), but lower spatial resolution.
Change detection works similar to the SAR system. ASCAT is a collaboration of
EUMETSAT and IPF. It was declared operational in December 2008 and is now
produced in near real-time by EUMETSAT, using the WARP-NRT software. This
software had been prototyped by EUMETSAT and developed by IPF. ASCAT soil
moisture is a Level 2 product delivered in orbit geometry at two different grid spacings:
25 km and 12.5 km. The two products are derived directly and on the same grid as the
equivalent ASCAT Level 1b products (normalized backscatter).Consequently, the
resolution of the soil moisture values is approximately 50km and 35 km.
Thorough validation of ERS scatterometer and ASAR demonstrated a good
correspondence of satellite and in-situ data (DORIGO, 2010). The correlation of ASAR results to
in-situ measurements is slightly weaker than the ones of scatterometers on board ERS, mainly
due to its lower radiometric resolution. However, the correlation of ASAR and in-situ data
improves significantly when averaged over larger areas (PATHE et al., 20009, MLADENOVA
et al., 2010). ASCAT products are spatially variable with high quality over grassland and
agricultural areas and lower quality in more densely vegetated areas and deserts. The
investigation of soil moisture at medium scale is a critical assess for IPF’s efforts for
downscaling of active and passive sensors. Field studies showed that, despite high spatiotemporal variability of soil moisture, its correlation to the mean soil moisture over a larger area is
significant in the temporal domain.
Recent flood events (January 2011) in Eastern Australia affected more than 200 000
people and an area as big as the size of France and Germany combined. ASAR observations can
Page 23
Version:
2.0
Date: 11/Feb/2011
now be used to increase the reliability of information that is fed into models for monitoring and
forecasting. The Australian Commonwealth Scientific and Research Organization (CSIRO)
currently rely on optical data in combination with passive microwave technologies and digital
elevation models. Incorporating ASAR in the system would result in several advantages: on one
hand, reliability and accuracy increases through higher resolution, while, on the other, cloudindependent continuous monitoring is possible. Figure 3 illustrates relative soil moisture
saturation on Australia’s Eastern coast during the flood events of December 2010.
Figure 3 Relative soil moisture from ASAR onboard ENVISAT during Floods in Australia (25th
of December). Blue colours represent highly saturated soils, while brown stands for extremely
dry soil conditions (IPF, 2010)
2.6.3 Satellite Application Facility on Support to Operational Hydrology and
Water Management (H-SAF)
H-SAF was established by the EUMETSAT Council in July 2005. The Development phase
started in September 2005. Within the H-SAF framework the focus for new satellite products lies
on:
Page 24
•
•
•
Version:
2.0
Date: 11/Feb/2011
Precipitation rate and cumulate precipitation, including liquid/solid discrimination,
Soil moisture in the surface layer and possibly in the roots region and
Snow parameters such as effective cover, wet/dry discrimination and water equivalent.
H-SAF membership includes 11 EUMETSAT member or cooperating States (Austria,
Belgium, Finland, France, Germany, Hungary, Italy, Poland, Romania, Slovakia and Turkey)
and ECMWF. Host of H-SAF is the Italian Met Service. The algorithms for satellite rainfall
estimation used in H-SAF will be considered and tested with respect to requirements of
GLOWASIS.
IPF is again contributing to H-SAF with expertise on soil moisture. The basis for all soil
moisture products in H-SAF is the radar scatterometer ASCAT on Metop. The three soil
moisture products that emerged from the development phase were:
• Large-scale surface soil moisture derived from ASCAT for the H-SAF area (SM-OBS- 1),
• Small-scale surface soil moisture resulting from disaggregation of the EUMETSAT CAF
global soil moisture from ASCAT (SM-OBS-2);
• Volumetric soil moisture (SM-ASS-1) for the H-SAF area (four soil layers up to a depth of
three metres). SM-ASS-1 is now part of ECMWF’s operational service and not
considered an H-SAF product anymore.
The development of two other products is planned in the current phase:
• A Soil Wetness Index in the root zone resulting from assimilation of CAF global ASCATSoil Moisture product in a NWP model (SM-ASS-2) and
• Global large-scale surface soil moisture derived from ASCAT for the H-SAF area (SMOBS-3) as the successor of the EUMETSAT CAF global soil moisture product.
Page 25
Version:
2.0
Date: 11/Feb/2011
2.7 South American Continent
2.7.1 Republic of Argentina Servicio Meteorologic Nacional Drought Monitoring
Figure 4 (a) Hydrological Balance and (b) Precipitation is deducted from potential
evapotranspiration (left) to estimate hydrologic balance difference with the previous decade
(Nunez 2010).
The Republic of Argentina SMN assembles maps of the hydrologic balance, which are
included within daily, monthly, and decadal bulletins which also include maps of number of
consecutive dry days, and NDVI-based vegetation health. Precipitation exhibits a non-normal
statistical distribution. Argentina uses 1960-1990 as its base “climatology” in determining
“average” conditions out of which drought episodes is detected.
Page 26
Version:
2.0
Date: 11/Feb/2011
2.7.1.1 Integration of Republic of Argentina SMN Drought Coverage into the
Global Drought Monitor
Users should be able to access the Global Drought Monitor from anywhere on the World
Wide Web and see the drought coverage for their respective countries in the form that they see it
and utilize it within their native countries. Yet, at the same time, the methodology that is
identifying droughts over Brasil or Paraguay should be able to identify droughts, if they are
present, over Argentina, as well. Hence, some standardization of drought indicators is required,
since one of the objectives of Global Drought Monitoring is to improve the accuracy with which
drought is being recorded all over the planet.
One approach is to find a proportional relationship between the range of the
drought indicator used in the native country and the drought severity range, either used by the
US National Drought Mitigation Center or by EDO, as illustrated in Figures 5 and 6. The
drought severity ranking system employed by the North American Drought Monitor was
developed by the USA University of Nebraska Lincoln. This system has a colorized code which
is linked to Standardized Precipitation Index (Figure 5). Figure 5 (a) and (b) compare the
National Drought Mitigation Center drought severity ranking system with the drought severity
ranking system employed by the European Drought Observatory. Eventually, the two systems
will have to be linked (or made interoperable).
Page 27
Version:
2.0
Date: 11/Feb/2011
Figure 5 (a) Drought Severity Ranking System of the USA National Drought Mitigation Center
(UNL) and (b) Drought Severity Ranking System of EDO (Iglesias and Schlickenrieder 2010)
Page 28
Version:
2.0
Date: 11/Feb/2011
Figure 6 (a) Drought Severity Status (Vargus 2008a) and (b) Drought Severity Indicator in
practice (Vargus 2008b)
Page 29
Version:
2.0
Date: 11/Feb/2011
B Informatics Section
3. Capturing User Requirements for the Global Drought Monitor and its
Interoperability with the Global Earth Observation System of Systems (GEOSS)
A key deliverable is the specification of a set of tools that will access information
published through a distributed water data infrastructure. The tools in this case are represented
by the applications which constitute the GEO Global Drought Monitoring Service. The tools are
specified through completion of three phases:
1). Capture of User Requirements—who might use the GEO Global Drought Monitoring
Service and the types of data and the types of functionality these users might require or expect
2) Design of a System Architecture—and associated enabling framework at the
component level
3) Implementation Plan
The GEOSS Architecture Implementation Pilot (AIP) task develops infrastructure
components for the GEOSS Common Infrastructure (GCI) and the broader GEOSS architecture
as a means of coordinating and deploying cross-disciplinary interoperability, such as the display
on top of drought map layers, combined with layers of different water usage and agricultural
water needs. The architectural implementation (AIP) task is envisioned as a way of developing
the GEOSS informatics capability and architecture through pilot projects. The process includes
user interactions; component deployment and interoperability testing; and SBA-focused
demonstrations.
3.1 Assessment of Drought Vulnerability and Susceptibility
The first section of this report dealt with drought indicators currently utilized by the
drought community of practice. Indicators do not correlate well with historic drought impacts,
and they need to be correlated with vulnerability. A direct linear proportionality between the
severity of the drought, as expressed by a drought indicator and the observed and recorded
impacts of a drought should not be expected. That is the role of the drought vulnerability factor
(Iglesias and Schlickenrieder 2010). Values of indicators change with the region and social
conditions.
Page 30
Version:
2.0
Date: 11/Feb/2011
The same level of drought severity can cause a wide variety of drought impacts due to
different underlying vulnerability of different regions. The multiple disciplinary information
sources that assist decision makers in evaluating drought impacts include information on regional
infrastructures, land use, residential water use, etc, which either are impacted by drought or may
mitigate drought severity (such as groundwater availability). Land use information (forage for
pasture animals in agricultural lands), crop type information with crop growing seasons, power
plant locations (for identifying cooling water requirements), groundwater springs (to identify
area of groundwater export) are all different types of data that can be combined together as
“layers” within a Geographical Information System. The display of layers, one type of
information on top of other layers, is the basis for the integration of multi-disciplinary
information. Several types of multi-disciplinary data integration exist, and several tools were
explored through testing for deployment for regional and global drought monitoring.
3.2 Capturing User Requirements and Implementation of Architecture to Design
of the Global Drought Monitor
3.2.1 Portal Requirements: Drill-down capability
Both the European Drought Observatory and the US NIDIS drought monitoring system
portals support “drill down” capability from continental to national scale and from national scale
to river basin scale. The spatial resolution of the drought maps are progressively higher, moving
from global scale to continental scale to national scale and finally to river basin scale. This is not
simply a matter of display preference, since a drought early warning system should be developed
for local scales, particularly in the case of small-scale agricultural plots. Although existing
national drought monitoring coverage (at its existing resolution) is incorporated into the GDMP,
the GDMP is not simply the assembly of a collection of web page graphics into one location.
3.2.2 Top-down versus bottom-up Design
There are several possible candidates for designing a global drought monitoring service:
1) a single, top-down system at coarse resolution; or 2) a single, top-down system at fine
resolution; 3) a bottom-up system, or 4) a bottom-up system complemented with some top-down
coverage where coverage is lacking.
One example of a top-down global drought monitor is the University of College London
Global Drought Monitor.21 Another is the Beijing Climate Center Drought Monitor.22
21
http://drought.mssl.ucl.ac.uk/drought.html?map=%2Fwww%2Fdrought%2Fweb_pag
es%2Fdrought.map&program=%2Fcgibin%2Fmapserv&root=%2Fwww%2Fdrought2%2F&map_web_imagepath=%2Ftmp
%2F&map_web_imageurl=%2Ftmp%2F&map_web_template=%2Fdrought.html
22
http://bcc.cma.gov.cn/Website/index.php?ChannelID=82&show_product=1
Page 31
Version:
2.0
Date: 11/Feb/2011
However, there are good reasons for embarking upon a bottom-up system. A global top
down system can imply that drought is present within a local area, although the local area might
be drought-free due to availability of secondary sources of water such as groundwater. Having
participation of members who are familiar with local conditions on the ground is invaluable in
setting up a global drought monitoring system. A global network of national
hydrometeorological service and ministry-based drought experts can provide the expertise to
carry out retrospective validations of drought forecasts, along with fine tuning of the drought
forecasting system, part of the life cycle by which “experimental” (research stage product)
becomes “operational.”
Drought monitoring and forecasting is intended for applications. For example soil
moisture monitoring and forecasting can support farmers’ activities. Since the size of farms may
vary, a drought early warning system is best applied at local scales. This means that a coarsescale system may not be very valuable for providing decision support. A drill down system is
built upon a combined bottom up-top down system, in which the highest resolution drought maps
of the system, i.e., those at river basin scale or national scale can be used for drought early
warning applications or used for agricultural support.
3.2.3 Soil Moisture and Agricultural Drought Monitoring Requirement
Given the importance of soil moisture in agricultural drought monitoring, and the
importance of agriculture in the world food problem, remote sensing-based and modeled-based
soil moisture should be utilized within the system.
3.2.4 Republication of information to help decision makers facilitate drought
decision making
Integrating together multiple disciplinary and cross-disciplinary information, such as
drought severity information and agricultural production data, require different informatics
strategies to carry out such integration. While layers can be added together and removed within
a Geographic Information System (GIS), more sophisticated tools are required in order to
assemble all of the information in a form that can be immediately used for decision making. As
noted by Lemon et. al (2010): “The ‘Discover, Display, and Download’ Use Case has misled us.
No one simply wants to find, look at, and collect data. To the contrary, they all want to do
something with the data: subject it to some analysis, make a map, or prepare a basis for making
rapid decisions.” Search and discovery alone will not make GEOSS a viable system, valuable to
end users: its information has to be repackaged into a user-friendly form that provides
application knowledge and accelerates decision making.
3.2.5 Hydrologic Drought Monitoring for Semi-Arid Areas and Meeting
Hydrologic Drought User Requirement through Semantics
The hydrologic drought indicator user requirement creates a need for assembling
information on water budget components, such as groundwater, streamflow, precipitation, soil
water, snow cover, etc. The sheer volume of information, particularly if assembled over large
Page 32
Version:
2.0
Date: 11/Feb/2011
segments of the globe, will require some integrative technology in order to accommodate the
utter complexity of multiple languages, multiple scientific terms within different languages,
differences in place names to describe geographic entities, and multiple variable names within
database schema. These are the requirements for a semantic-based information system: datasets
and records have to be registered at the level of water budget components, i.e., stores of
groundwater, river water elevation, precipitation, etc, to meet the requirements for hydrologic
drought monitoring. This also means, conversely, that a semantic ontology has to include these
concepts, as well, within the water ontology, for the purposes of organizing information. This
level of detail is a critical requirement.
Several possible methodologies for achieving multidisciplinary interoperability take
advantage of the possible integrative power of Semantic Web technologies, developed by Tim
Berners-Lee (Berners-Lee, Hendler, Lassila 2001; Yu(2007).
What, simply put, does the semantic web do? It tries to lift the burden off the user of
having to process huge amounts of information by automating (and making machine readable)
the collection and processing of information, so that the processing burden may be shifted from
the user to the machine. Semantic web techniques improve irretrievability of the correct
document or resource or dataset by providing semantic annotation through Resource Description
Framework (RDF) or RDFS, perhaps combined with an ontology which provides the structural
arrangement of the resources in context with one another, along with possibly including some
simplified artificial intelligence application for sorting or selection.
Semantics can be directly employed within the decision support services developed by
GEO, i.e., within the software applications and processing of data. For example, SEAMLESS
links together application modules (such as used in Delft- Flooding Early Warning System or
FEWS) and component-based applications that can be orchestrated into a workflow run over a
framework, in this case, OpenMI (Rizolli, et. al 2007).
Another use of semantics is the more traditional search and discovery role. This use
case of semantics is what has been explored within this session of AIP-3, as a test case project
within the European Union among the architects of the EuroGEOSS discovery broker, the AIP-3
Semantics Working Group, the European Drought Observatory, and the AIP-3 Water and
Drought Working Group.
3.3 Developing an Architectural Diagram for the GEO Global Drought
Monitoring Service
Figure 6, derived from the Australia Water Resources Information System, illustrates
some of the components that are prerequisites for the Global Drought Monitor Portal (GDMP).
The “system architecture” is a diagram of the applications and the tools, combined with the
enabling framework at the component level. Figure 6 shows the bottom rung of data entering
through the observing system, as, respectively, “Numeric data” (as in soil moisture generated by
the VIC and LISFlood models), or satellite source “Sensor output” originating from space-based
scatterometer soil moisture data. The upper tier illustrates in schematic boxes some additional
components, the Open Geospatial Consortium (OGC) geospatial Web Mapping Services (WMS)
exchange of drought maps. The exchange of drought maps among the European Drought
Page 33
Version:
2.0
Date: 11/Feb/2011
Observatory, the Princeton African Drought Monitor, the NIDIS Server, the University College
London drought monitor, and the Argentina SMN drought monitor make the functionality of the
GDMP possible.
In between the bottom rung (the sensor and model data sources of the observing system)
and the upper tier (and the OGC-supported web service data exchange over the WWW) are
additional layers which include controlled vocabularies, dictionaries, ontologies, semantic
mappings, and transfer format and protocols, along with common data models, for data
integration and processing this information at multiple levels, republishing the information in a
format immediately available for decision making. This functionality is depicted as schematic
boxes in Figure 8(a), but the actual operations are set out below.
This overall strategy is a scalable system that permits integration multiple data stores
(information hubs), along with information of multiple disciplines.
Figure 7 Australia Water Resources Information System (Boston 2010)
Page 34
Version:
2.0
Date: 11/Feb/2011
Figures 8 (a) and (b) Commonwealth of Australia Water Resources Information System
(Boston 2010)
Page 35
Version:
2.0
Date: 11/Feb/2011
Figure 9 EuroGEOSS Broker Discovery Augmentation Component Expansion of
Drought (diagram kindly provided by Mattia Santoro)
3.4 Semantic Development Activities within GEO: the Data Integration and
Analysis System (DIAS) Contribution from Japan
The DIAS approach is illustrated within Figures 10 and 11. Each GEO Societal Benefit
Area, i.e., Disasters, Water, Sustainable Agriculture, Biodiversity, Health, Energy is represented
by a domain within the “application layer.” The overall user requirement is to support crossdisciplinary or multidisciplinary sharing of information and data. The DIAS system supports the
organization of information within each of these areas.
Figure 9 illustrates a SBA area, in this case drought, using EuroGEOSS (see below). At
the center of the graph is “drought.” “Drought” is linked to “Species impoverishment” (within
the Biodiversity cluster) in one node and to “famine” (within the Sustainable Agriculture cluster)
as another node. The arrangement of concepts for each GEO SBA is the ontology for each GEO
SBA. DIAS also separates this conceptual scientific terminology from geographic locations and
place names (Figure 12). The drought lexicon, for example, comprises the lexicographic
content. Each ontology or collection of ontologies for each of these areas can be loaded into the
semantic network dictionary. A semantic network illustrates the relationship among concepts.
Page 36
Version:
2.0
Date: 11/Feb/2011
Figure 11 Integration of DIAS into Web-based Information System (Koike 2010)
Page 37
Version:
2.0
Date: 11/Feb/2011
Figures 12 (a) DIAS cross disciplinary areas (Koike 2010) and (b) DIAS Ontology Development
Page 38
Version:
2.0
Date: 11/Feb/2011
Architecture (Nagai 2010)
Figure 13 (Nagai 2010)
3.4.1 Adding Advanced Search and Discovery using Semantics
The AIP-3 video23, “Drought—European” includes a walkthrough demonstrations, in
which users select scientific terms (or “concepts”) and, secondly, the geographic region or spatial
domain which is defined within a bounding box. This reflects the division into lexicographic and
geographic content which has been cited in Section 3.5. The purpose of the semantic enrichment
was to supplement keyword searches, such as used on Google, by adding search capability that
could search through concepts; this is tantamount to adding “Semantics.” In other words, one is
not searching for “drought” as a keyword; one is searching for “drought” as a concept, combined
with search functionality that allows the user to select broader or narrower searches within the
drought field or within allied fields. For example, “water” is a concept which can be broken up
into underlying processes of “evapotranspiration,” “streamflow,” “precipitation,” “soil
moisture,” “snow cover,” and “groundwater.” The stores of water are a subset or part of water,
and this class structure is affected in the arrangement of the terms within the semantic network.
Datasets can be registered to each of these terms, so that queries of “hydrologic drought
indicators” reveal “groundwater” data, “streamflow” data, such as baseflow, and other water
budget information for a selected area. The EuroGEOSS discovery broker has the capability to
access the GI-Cat registered datasets on groundwater, river discharge, water usage, and other
data. Search-by-concept is intended to not only improve the “hit-or-miss” success rate of recall
of datasets through keyword searches alone but also reduce the high amounts of irrelevant
returned results in keyword searches.
Figure 8 is a screen capture of the search interface in the AIP-3 “Drought—European”
23
http://www.ogcnetwork.net/pub/ogcnetwork/GEOSS/AIP3/pages/Demo.html
Page 39
Version:
2.0
Date: 11/Feb/2011
video. Note that the interface and tool has not only has a list of scientific terms but also has
these terms “arranged,” so that the terms are linked to one another. This tool enables a user to go
from the general term—“drought”—to an associated term “drought indicator” to specific drought
indicators, such as “meteorological drought indicator,” then to “precipitation” and then to
“Standard Precipitation Index.” Alternative branches are “drought” to “drought indicator” to
“agricultural drought indicator” to “soil moisture” or “drought” to “drought indicator” to
“hydrologic drought indicator” to “groundwater” to “terrestrial water storage change.” Datasets
also have tags to the appropriate geographic area, such as “England.”
4. Global Implementation of the Drought Monitoring Service through GEOSS
4.1 Components of the System Architecture of the Global Drought Monitoring
Portal
The Global Drought Monitor utilizes and is designed to have a “drill-down” capability.
One begins at the global (and coarsest spatial resolution) and then is directed toward higher
spatial resolution regional maps. The user can follow the sequence of events by accompanying
the steps while watching the “Drought—Global” video.24
One begins with a Global World Map (a Global Drought Map, as well), accessible on the
Graphical User Interface (GUI). The Global Drought Monitoring Portal is accessible from the
World Wide Web.25
The layout of the entry web page (index or home page) includes the title bar “”Beyond
Drought: Global Participation for Better Planning and Response,” underneath of which are four
underlying header tabs, arranged from left to right: “Current Conditions,” “Interactive Maps and
Data,” “Regional Drought monitoring,” and “About.”
The “Current Conditions” tab displays the “Global Drought Monitor” of the University
College London global drought monitor.26 As noted in the Disclaimer to the University College
London Global Drought Monitor, the (UCL) Global Drought Monitor provides the 'overall
drought picture' on a ~100km spatial scale. The maps are not designed to depict local conditions.
As a consequence, there could be water shortages or crop failures within an area not designated
as drought, just as there could be locations with adequate water supplies in an area designated as
'extreme' or 'exceptional' drought.”
24
25
26
http://www.drought.gov/portal/server.pt/community/global_drought
http://drought.mssl.ucl.ac.uk/drought.html?map=%2Fwww%2Fdrought%2Fweb_pages%2Fdrou
ght.map&program=%2Fcgibin%2Fmapserv&root=%2Fwww%2Fdrought2%2F&map_web_imagepath=%2Ftmp%2F&map
_web_imageurl=%2Ftmp%2F&map_web_template=%2Fdrought.html
Page 40
Version:
2.0
Date: 11/Feb/2011
The differences between the second “Interactive Maps and Data” tab and the third
“Regional Drought Monitoring” tab are the display of the drought zones on the Interactive Maps
and Data” tab, while the Regional Drought Monitoring” tab displays highlighted continental
areas. The drought zones that are displayed on the “Interactive Maps and Data) map viewer are
not necessarily the same drought zones that are displayed on the “Current Conditions” first tab.
This is because the “Current Conditions” Global Drought map is largely based upon Standard
Precipitation Index (SPI), while the “Interactive Maps and Data” second tab displays drought
coverage for members of the Global Drought Monitor and the Global Drought Monitoring
Community of Practice. The “Current Conditions” calculates drought globally, based upon API
and is a top-down system. The “Interactive Maps and Data” drought displays are integrated
together from national coverage in North America, derived from the regional European
Community LISFLOOD model application with real time data, or derived from continental scale
coverage for the African continent.
The third “Regional Drought Monitoring” tab27 displays the North American,
European and western Asian, and African continents highlighted in different colors with the
indent being to allow users to access regional drought portals for additional, higher resolution
drought information. The remainder of the terrestrial globe is designated a common color. If
one points the mouse and clicks anywhere within the Canada, USA, or Mexican spatial domain,
i.e., anywhere within North America, one is redirected to the North American Drought Monitor.28
If one points the mouse and clicks anywhere within the European Community, one is redirected
to the European Drought Observatory home page.29 If one points the mouse and clicks
anywhere within the African continent, one is redirected to the Princeton University African
Drought Monitor. 30 The remainders of the terrestrial continental areas share a common color,
because these areas have yet to be integrated and made interoperable within the Global Drought
Monitor Portal (GDMP).
4.2 Actors
While the administrative user has been identified above, the main actors will be officials
working in the national hydrometeorology drought monitoring services, as well as officials
working in national and private relief agencies, such as the case for famine relief, countrysponsored agricultural agencies, and agricultural commodities insurers.
4.3 Capturing User Requirements for the Global Drought Monitor Portal through
27
http://www.drought.gov/portal/server.pt/community/global_drought/314/regional_drought_moni
toring/1097 .
28
http://www.drought.gov/portal/server.pt/community/nadm/303
29
http://edo.jrc.ec.europa.eu/php/index.php?action=view&id=2
30
http://hydrology.princeton.edu/~justin/research/project_global_monitor/
Page 41
Version:
2.0
Date: 11/Feb/2011
the GDMP Scenario
As noted above, a scenario is the listing, step-by-step, of the user requirements (for
drought monitoring), showing which GEOSS resources and components are utilized within the
retrieval of drought maps and information (such as soil moisture) for users.
Table 4
Objective: Obtain a Drought Overview of a Given Area on the Terrestrial Earth, along with
detailed information on affected regions
Step 01: Obtain Drought Indices from the Global Drought Monitor
Step 01.1: Obtain Drought Indices through standard services
Step 01.2: A dedicated WPS processes the drought index and calculates the drought hazard
Step 02: A dedicated WPS processes the drought index and calculates the drought hazard
Step 02.1: The WPS retrieves the Drought Index through the WCS
Step 02.2: The WPS executes the hazard detection model and, where detected, sends an alert to
the decision support tool
Step 03: Drought Hazard Related Information Discovery
Step 03.1: The decision maker uses the decision support tool to submit a query to the augmented
search component in order to discover drought hazard related information (datasets)
Step 03.2: The EuroGEOSS Broker mediates the query request, distributing it to its federated
services
Step 03.3: The Decision maker uses the Decision Support System to select one or more drought
hazard related information datasets, among the ones returned by the query
Presentation of Reachable Services and Alerts
Step 03.4: The Decision Support Tool submits an access request to the EuroGEOSS Broker in
order to retrieve the user-selected drought hazard information datasets
Interact with Services
Step 04: Visualization and Assessment of Information
Step 04.1: The Decision Support Tool displays the accessed drought hazard information datasets,
combining them with the potential hazard layer
Step 04.2: The decision maker assesses the drought hazard impact
4.3.1 Display of Selection Bar for Drought Indices, Processing to Derive
Dehydration and Drought Severity, and Drought Map Republication
The Global Drought Scenario steps, by themselves, are relatively abstract, and are best
understood by following the actual presentation given within the videos, “Drought—Global.”31
31
Page 42
Version:
2.0
Date: 11/Feb/2011
Step 01 within the European Drought Observatory is equivalent to providing with users
the ability to select one of multiple drought indicators. EDO makes available to users the choice
of Standard Precipitation Index (SPI), Soil Moisture, Soil Moisture Anomaly, and several remote
sensing-based measures of vegetation health, such as fraction of Absorbed Photosynthetic Active
Radiation (fAPAR) in the case of EDO and VegDRI in the case of NIDIS (over the USA). The
GDMP does not yet have the extensive development to support independent display of multiple
drought indicators and indices. As noted above, SPI is already being displayed on the “Current
Conditions” map.
Step 02 is the processing loop of taking the selected drought indicator, running the indicator over
a selected spatial domain or region, and then returning a republished map which displays the
level of drought intensity or severity within this area (if drought is present at all). The current
GDMP configuration displays the integrated drought severity ranking system of Figure 5 for
North America and the integrated drought severity ranking system deployed by EDO for the
European Community. (A drought severity ranking system is being developed for Africa and is
not yet deployed. The Figure 5 system can be adapted to soil moisture percentiles, which is a
drought indicator within the Princeton African Drought Monitor system). Under the current
system, users would have to drill down to the North American Drought Monitor and from there
to the USA, Canada, or Mexico, in order to select individual drought indicators, such as those on
display on the NIDIS portal. Drought Indicators are made available on the NIDIS site.32
SPI may not be an adequate drought indicator for semiarid areas, such as areas where
sources of water may be water crossing a national boundary from a snowmelt runoff mountain
zone (like Central Asia), or areas where complicated moisture fluxes within the vadose zone may
reverse direction and return back to the surface. The North American map currently displays
drought zones using the National Drought Mitigation Center drought severity ranking system
shown in Figure 5(a), while EDO deploys the “indicator” alert system shown in Figure 5(b).
4.3.2 Layout and Organization of the GDMP within the NIDIS GIS Server
The home page (index page) of the NIDIS portal is accessible from the World Wide
Web.33 Underneath the NIDIS banner, “US Drought Portal” is a list of subcategories: “Home,”
“What is NIDIS,” “Current Drought,” “Forecasting,” etc. The URL for “current drought,”34 the
URL for “forecasting”35 show the navigation within the portal: the category item is located in the
32
http://www.drought.gov/portal/server.pt/community/drought_indicators/223
http://www.drought.gov/portal/server.pt/community/drought.gov/202
http://www.drought.gov/portal/server.pt/community/current_drought/208
35
http://www.drought.gov/portal/server.pt/community/forecasting/209
33
34
Page 43
Version:
2.0
Date: 11/Feb/2011
URL path after community, so that the path to the global drought monitor is analogous.36
4.3.3 Implementation of Advanced Search and Discovery in the GDMP
Advanced semantic-enriched search and discovery is discussed below for the European
Drought Observatory. Eventual semantic deployment within GDMP would likely be limited to
national drought monitors that are part of GDMP. In addition, the information resources built
up by DEWFORA can be incorporated into the system, including water cycle component
datasets for Africa. The water usage datasets, such as for the Water Information System for
Europe (WISE), and other areas, from GLOWASIS can also be integrated into the system over
time, funding permitting. This type of implementation would require registration of the datasets
within GI-Cat, tantamount to the addition of another drought catalogue (Figure 13). The datasets
would also be registered with the concepts of the water ontology. More information on
establishing interoperability of EuroGEOSS with GEOSS is contained in Section 5.8.
4.4 Support of Increased Global Coverage within the web-based, real-time GDMP
server
The GDMP is a web-based, real time (RT) Geographical Information System (GIS) server,
which is built on top of a distributed database federation and ingests meteorological information
and hydrologic information in real-time, in order to provide alerts, a prototype Drought Early
Warning System. An overview of the alert system is provided for the European Drought
Observatory in Section 5.1.5.
As mentioned in the “Drought—Global” video, the global drought server is integrated
and interoperable with continental drought servers, while the national hydrometeorology drought
monitors within the continental areas are integrated with and made interoperable with the
continental drought servers. The European Framework project Drought Early Warning System
for Africa (DEWFORA) would be expected to possibly serve as an African continental (pan
Africa) continental drought monitor with intercomparisons being prepared between the
DEWFORA and Princeton African drought monitors. The meteorological forcing data sets are
being assembled for South America to integrate with real time meteorological observing system
data to create a continental scale system there. The GEO Community of Practice partner, the
Asian Water Cycle Initiative Drought Working Group (Ichiro Kaihotsu, Hiroshima University) is
developing a regional drought network (with ground-based stations) that may be made
interoperable with the GDMP. An expected, possible configuration for the continental servers is
depicted in Figures 14 through 16.
4.5 Integration of GDMP with GEOSS Architecture
The GDMP is integrated into GEOSS via: 1) metadata creation and addition; 2) catalogue
addition (via EuroGEOSS and GI-Cat; and 3) utilization of OGC Web Mapping Service, via
36
http://www.drought.gov/portal/server.pt/community/global_drought/
Page 44
Version:
2.0
Date: 11/Feb/2011
installed Minnesota MapServer (University College London, Princeton, and EDO) and ESRI GIS
Server (NIDIS). Additional web services are slated for installation. See section for more details
in interoperability of EuroGEOSS with GEOSS.
4.6 Remote Sensing Soil Moisture Integration
Figure 7’s lowest tier displayed both numeric and sensor sources of data originating from
the observing system. The numerical models are forced by real-time meteorological data from
the observing system. Sections 2.6.2 and 2.6.3 showed how space-based scatterometers could
provide soil profile and root zone soil moisture. These results can be displayed alongside
modeled soil moisture data; direct data assimilation into a common product is also a possibility
or even the latter with the display of separate inputs. The US National Aeronautics and Space
Administration (NASA) Soil Moisture Active and Passive (SMAP) results can also be added
when they go live and come on line.
4.7
Adding Water Usage Information Layers, including Agriculture
GLOWASIS may provide a basis for incorporating water usage information and data into
the GDMP. This would also be a prerequisite for assessing drought vulnerability.
Page 45
Figures 14 (a) and (b)
Page 46
Version:
2.0
Date: 11/Feb/2011
Page 47
Version:
2.0
Date: 11/Feb/2011
Page 48
Version:
2.0
Date: 11/Feb/2011
Version:
2.0
Date: 11/Feb/2011
5. Advanced Search and Discovery Capability within the European Drought
Observatory
A set of tools have been developed for deployment within the European Drought
Observatory, also serving as a contribution to GEO. These technologies provide capability to
integrate water information for the Water (and drought) Societal Benefit Area, including possible
deployment within the Global Drought Monitoring Service. This report captures the user
requirements for who would be using this system (EDO), what data types would be required, and
the type of functionality users would expect. These user requirements are embodied within a
“scenario,” with the development of a system architecture providing the associated enabling
framework. The GEO Architectural Implementation Pilot (AIP) develops components for the
GEOSS Architecture through component deployment and subsequent testing, interoperability
testing, followed by Societal Benefit Area (SBA) demonstrations, i.e., demonstrations of a
decision support service, such as the global drought monitoring service for the Water SBA. GEO
AIP projects are run by framing a “scenario” which expresses and embodies the user
requirements, such as GEO tasks.
5.1 Components of the European Drought Observatory
The European Drought Observatory is based upon a loosely-coupled system having as
components: 1) drought indicators; 2) drought climatologies; 3) drought observing systems; 4)
water usage observing system; 5) internet-based and web-based services which make
interoperability and exchange of data and maps possible; 6) common formats among the system;
7) user network to verify nowcasts and forecasts; and 8) hardware infrastructure and technical
support staff.
The European Community has identified drought, biodiversity, and forestry as targets for
GEOSS activity. This AIP effort has included development of the EuroGEOSS search
capability. As has been mentioned above (Section 2.1.1), common registration of datasets permit
maps and data to be shared and exchanged among the EDO, the national drought monitors, such
as MARM and SIA, and the river basin authorities, such as Confidercion Hidrografica del Ebro.
In short, joint registration supports interoperability. At the same time, some of the data retrieval
capabilities of EDO were time series of soil moisture, soil moisture anomalies, and Standardized
Precipitation Index over different time periods. The longer the period of observation for
precipitation falling upon a landscape, the longer the period of time over which water works its
way through the soil and drains down into groundwater.
5.1.1 European Drought Observatory user access37
The EDO map server is separated from the index page.38
37
38
Page 49
Version:
2.0
Date: 11/Feb/2011
5.1.2 Organization and layout of the EDO map server page (scenario step 01—
continued)
The upper left hand column has “European Drought Products,” underneath of which are
listed:
1. Soil Moisture;
2. Precipitation;
3. Precipitation from Archive;
4. Remote Sensing Indicators;
5. Drought-related products; and
6. Generate Graphs and Time Series
Underneath this list are three tabs:
1. Information from EuroGEOSS Drought Catalogue;
2. National/International Drought Information; and
3. Regional/Local Drought Information
Selecting Soil Moisture, number 1 from the top panel (button), causes a list to fall down, having
the selection:
1. Daily Soil Moisture;
2. Daily Soil Moisture Anomaly;
3. Forecasted Soil Moisture Anomaly;
4. Forecasted Soil Moisture Trend. Etc.
The EuroGEOSS drought catalogue box is accessible.39
5.1.3 Selection of Drought Indices
The Drought Indices of Step 01 are Standardized Precipitation Index and Soil Moisture
Anomaly (and Soil Moisture). No hydrologic drought indicator is included, although some
remote sensing drought indicators are given. We have not included remote sensing drought
indicators here, in order to reduce the length of this report.
Step 01.1 entails “obtaining drought indices through standard services”: this step is tantamount to
the process of selecting one of the drought indicators above, such as daily soil moisture anomaly,
and sending a query to the EDO server from a local machine browser window, in order to request
a returned map showing daily conditions calculated for western Euro Asia for that day.
Returning to the EuroGEOSS drought catalog box above, the returned web page contains a given
drought vocabulary (Section 9), along with a thesaurus button underneath. Pressing the
thesaurus button prompts a new button to appear “Add term” with a dialog box popping up
“Select thesaurus from list ‘SBA_EuroGEOSS.” The GEO Societal Benefit Area groupings are
listed, including water.
NOTE: This is an important step which is not included within the original scenario. The
implications of this EuroGEOSS interface—with respect to semantics—are outlined below in
Section 6.4.
39
http://eurogeoss.unizar.es/Search/Search.html?opener=EDOMapServer
Page 50
Version:
2.0
Date: 11/Feb/2011
A returned map is loaded and returned.
5.1.4 Processing Step by Running Drought Indicators over a Selected Spatial
Domain
Step 02 is the process of processing current daily conditions using the drought indicator (“daily
soil moisture anomaly” and the display of the map within the browser on the local machine.
Step 02 determines whether there is a drought over a given spatial domain, as well as the
severity (ranking) of the drought. EDO utilizes a drought severity ranking system,
corresponding to drought of increasing severity: 1) 0-green; 2) 1-yellow; 3) 2-orange; 4) 3-red;
5) 4-brown. This is the drought severity ranking system that differs from the North American
Drought Monitor ranking system, as presented within Figure 5.
The currently displayed drought severity color coding system on the EDO map server
system does not yet implement this drought severity ranking system. It current system is
simpler, exhibiting wetter or drier conditions (than average) only: green (indicating wetter
conditions), yellow (indicating normal or 0), and orange for progressively drier, until red.40
5.1.5 Automated Email Alerts and Drought Triggers
Scenario step 02.2
If any area within the European Union (including adjacent areas, such as Turkey) is
designated as having drought of a particular severity, an automated email alert can be sent to
decision makers, if they have already signed up to be a recipient for such an alert service. This is
the type of automated email alert system used by the USA National Oceanic and Atmospheric
Administration (NOAA) Integrated Coral reef Observing Network (ICON), which dispatches
automated emails to alert the oceanographic community of possible coral reef bleaching, when
pre-assigned bleaching thresholds are passed. The EuroGEOSS broker offers support for the
GeoRSS alert mechanism.
If, after being notified by email alert of a designated drought ranking, a decision maker
wants to retrieve more information about drought conditions, the semantics-supported advanced
search and discovery is set up to make it easier to retrieve the information. This is in keeping
with the philosophy of tailoring a decision support system to make it easier to utilize
information.
40
Page 51
Version:
2.0
Date: 11/Feb/2011
5.1.6 Context and pre-conditions
Some of the existing Global Earth Observation System of System functionality has been
described above. This section will provide further documentation of EuroGEOSS in providing
components supporting advanced search and discovery. The following datasets and services are
assumed to be available before the scenario begins:
• GEO Portal, through this portal the end user will be able to search, find and access
the services which are needed for the Scenario execution;
• EuroGEOSS/GENESIS Client Application is registered on the Components and
Services Registry (CSR) and accessible through the GEO Portal;
• EuroGEOSS Discovery Augmentation Component (DAC) Service. This service
federates both semantics (e.g. SKOS repositories) and ISO-compliant geospatial
catalog services. The DAC can be queried using common geospatial constraints (i.e.
what, where, when, etc.). The service exposes a semantics-extended OpenSearch
interface.
• EuroGEOSS Discovery Broker Service. This is a distributed catalogue which
federates several services (exposing them through the CSW-ISO interface). Federated
services publish the following datasets:
o Environmental datasets (WCS);
o Climate Change datasets (WCS);
• GENESIS Vocabulary Service. This repository publishes a SPARQL interface for
navigating the aforementioned SKOS-based thesauri.
• WPS Client. This is web client for configuring and running data retrieval for graph
construction through the European Drought Observatory
• GEOSS Ontology Registry
• GEOSS Geographic Gazetteer
• Application (WPS): the search interface is designed to be accessible through the
European Drought Observatory (EDO) portal interface
• A workflow engine. This component manages all phases of the scenario (browse
semantic repository, retrieve concepts of interest, search for resources related to such
concepts, execute WPS)
5.2 Implementation of the European Regional Drought Semantic-enhanced Monitoring and
Information System
The Drought Scenario that has been used for AIP-3 was originally introduced within
presentations of Barbara Hofer and Stefan Niemeyer (European Drought Observatory),
“EuroGEOSS for Drought--Linking the EDO to Local and Global Scales,” at the INSPIRE
conference in June 2010 and “Drought Data, Metadata, and Interoperability,” at the Training
Workshop on Drought Risk Assessment for the Agricultural Sector in September 2010. These
Page 52
Version:
2.0
Date: 11/Feb/2011
identical drought scenarios are also given within Report D.2.1.1 Report on Requirements for
Interdisciplinary Interoperability (L. Vacarri, S. Nativi, and M. Santoro), 2 Nov 2010.
The actual scenario which was used is presented in Table 1 below. This scenario is pretty
abstract, and the reader may find viewing this scenario more useful by accompanying the reading
with a viewing of the video “Drought—European.” 41along with reading this scenario.
The video is actually a “walkthrough,” showing step-by-step how a user interested in
drought will use the drought information system, showing the implementation of the scenario.
The scenario itself in Table 1 is actually the user requirements before construction of the system,
while the video displays the components that have been assembled and implemented to meet
these requirements.
The use cases Semantics Enabled Search and Ontology Engine Search have been
developed in conjunction with the Semantics WG; further details are contained in the
EuroGEOSS Broker documentation and the Semantics Working Group Report.
Table 3 European Drought Observatory Scenario
European Drought Observatory Scenario
Step 01: Obtain Drought Indices from European Drought Observatory
Step 01.1: Obtain Drought Indices through Standard Services
Step 02: A dedicated WPS processes the drought index and calculates the drought hazard
Step 02.1: The WPS retrieves the Drought Index through the WCS
Step 02.2: The WPS executes the hazard detection model and, where detected, sends an alert to
the decision support tool
Step 03: Drought Hazard Related Information Discovery
Step 03.1: The decision maker uses the decision support tool to submit a query to the augmented
search component in order to discover drought hazard related information (datasets)
Use Case: Semantics Enabled Search
Step 03.1.1: The augmented search component submits a query to the ontology query engine and
extracts 0,…N terms to be inserted into the geospatial query
Specialized Use Case: Ontology Enabled Search
Step 03.1.2: The augmented search component generates one or more geospatial queries based
on the user selection as geospatial constraints and/or as keywords from the previous step and
41
Page 53
Version:
2.0
Date: 11/Feb/2011
submits the queries to the EuroGEOSS Broker
Use Case: Discovery: Client Search of Metadata
Step 03.2: The EuroGEOSS Broker mediates the query request, distributing it to its federated
services
Step 03.3: The Decision maker uses the Decision Support System to select one or more drought
hazard related information datasets, among the ones returned by the query
Presentation of Reachable Services and Alerts
Step 03.4: The Decision Support Tool submits an access request to the EuroGEOSS Broker in
order to retrieve the user-selected drought hazard information datasets
Interact with Services
Step 04: Visualization and Assessment of Information
Step 04.1: The Decision Support Tool displays the accessed drought hazard information datasets,
combining them with the potential hazard layer
Step 04.2: The decision maker assesses the drought hazard impact
5.2.1 Advanced Semantic Search
Scenario Step 03 embodies the advanced semantics incorporated into the European
Drought Implementation.
Upon being notified of a drought alert in an area of interest, the decision maker (drought
expert) can go online to consult the common EuroGEOSS broker-EDO interface. For example,
the EuroGEOSS broker incorporates both: 1) Discovery Augmentation Component (DAC) and
2) the workflow engine, which increases the power of search by integrating together catalogue
and semantic components. In other words, the workflow engine browses the semantic
repositories to retrieve concepts. Once the expert drought user has identified a concept of
interest, the resources (datasets and services) linked to each of these concepts can be retrieved.
5.2.1.1 Ontology Registration
An ontology is a technology for organizing information which includes the organization
of the information together into relationships that are reminiscent of the class structure found in
programming languages. The ontologies are stored within the Semantic Network within DIAS,
which preserves this class structure.
5.2.1.2 Geographic Registration
As can be seen in Figure 11, two types of ontological information are developed and
expanded: 1) lexicographic ontologies, i.e., ontologies of scientific disciplines and remote
sensing; and 2) geographic ontologies, as represented by gazetteers. A gazetteer is defined as a
reference for information about places and place names used in conjunction with an atlas (hill et
Page 54
Version:
2.0
Date: 11/Feb/2011
al 1999). The gazetteer assembles a correspondence between place names and spatial
information. Each concept (i.e., place) can be uniquely identified by Resource Description
framework via a Universal Resource Identifier (URI). Gazetteers can record a triple (place
names; geographic footprints (locations); and class of described feature or representation of the
real world geographic entity. The place name is a “handle” to support communication.
5.3 EuroGEOSS Deployment of the Foundation Vocabularies
Ontologies are created out of scientific vocabularies, controlled vocabularies, where
terminology has accepted meaning. So the starting point for the vocabularies in Figure 10 would
be general vocabularies, such as:
The GEO SBA ontologies were constructed out of:
•
•
•
•
The General Multilingual Environmental Thesaurus (GEMET): 28 of the 29 languages
currently provided by the EIONET portal.
The INSPIRE Feature Concept Dictionary and Glossary: 21 of the 23 EU official
languages for INSPIRE Themes, monolingual the other terms.
The ISO 19119 categorisation of spatial data services: 21 of the 23 EU official languages.
The GEOSS Societal Benefit Areas: 5 languages.
5.4 Fine Tuning the Foundation Vocabularies for SBA Application—Specialized
Drought Vocabulary
At the start of AIP-3, Pozzi parsed the GEMET water thesaurus, in order to identify the
extent of drought terminology and concepts contained within it. An extensive, developed
network of drought concepts would be necessary to support the presentation graphs in the user
interface on the EDO portal (Figure 9) and also provide concepts linked to the drought data and
information. However, parsing and browsing the GEMET thesaurus for water shows it lacks any
specialized drought vocabulary.42 Even the precipitation it lists only includes chemical
precipitation.43 Hence, GEMET, by itself, is not adequate to express meteorological drought,
agricultural drought, and hydrologic drought indicators or their associated water budget
components (groundwater, streamflow, baseflow, snow cover); the base thesaurus must be
supplemented by a water ontology. Pozzi (of the Water Working Group), C. Fugazza of the
AIP-3 Semantics Working Group, and M. Santoro and S. Nativi, the EuroGEOSS architects,
concurred in developing a water ontology that would include a drought specialization that could
be used for both EuroGEOSS (and EDO) and the DIAS ontology registration within the semantic
42
http://www.eionet.europa.eu/gemet/theme_concepts?letter=0&start=390&th=40&langcode=en&
ns=4
43
http://www.eionet.europa.eu/gemet/theme_concepts?letter=0&start=270&th=40&langcode=en&
ns=4
Page 55
Version:
2.0
Date: 11/Feb/2011
network registry. Such a specialized drought vocabulary or a water ontology would link to
drought indicators and drought and water datasets used in common with the European Drought
Observatory. The Simple Knowledge Organizing System (SKOS) had been used to express this
GEMET data structure, so the water ontology, developed in AIP-3, would also be translated into
SKOS data structures and linked to relevant terms in the reference thesauri as an AIP-4 activity.
5.4.1 Water Ontology-enablement within the DAC Semantics
One possible candidate as a foundation water ontology was the USA Consortium of
Universities for the Advancement of Hydrologic Sciences (CUAHSI) water ontology, version 1.
The CUAHSI water ontology does list water stores of surface and subsurface (soil) water,
thereby meeting some of the requirements needed in a hydrologic drought indicator. Fugazza
has converted the CUAHSI Ontology Web Language (OWL) into SKOS, as an AIP-3
contribution (See AIP-3 Semantics Engineering Report).
5.5 How the EuroGEOSS Discover Augmentation Component supports semantic
searches
To conclude, the Discovery Augmentation Component enables semantics-aware
discovery by matching the search patterns entered by the end user against a collection of
multilingual, SDI-related thesauri: these are controlled vocabularies providing multiple textual
representations for terms and organizing them according to specificity and relatedness. As a
consequence the user’s query is first related to a set of language-neutral identifiers (URIs) (like
the URIs used for geographic spatial entities, noted above in section 5.2.1.2). These URIs
represent entities in a concept graph that the user may navigate for identifying related terms that
are relevant to her search. These data structures are hosted by the GENESIS Vocabulary Service.
These thesauri are provided in the Simple Knowledge Organizing System (SKOS)
format, a lightweight ontology for expressing knowledge organization systems (such as
taxonomies, classification schemes etc.), and have been harmonized in the context of the
EuroGEOSS project by relating terms from distinct thesauri, thus allowing the user to move from
one categorization to the other, i.e., one scientific discipline to another within a GEO SBA or
from one SBA (water) to another SBA (agriculture). Once the user has identified an exhaustive
set of terms that are relevant to her query, the broker translates the corresponding URIs back to a
customizable set of languages and executes multiple queries against the catalogs it is federating
(recalling the desired datasets).
The EuroGEOSS Discovery Augmentation Component (DAC) implements a query
expansion strategy deriving multiple traditional geospatial queries from a single semantic query.
The DAC is able to accept a semantic query and, accessing a configurable set of external
semantic services (e.g. controlled vocabularies, gazetteers, etc.), split it into several geospatial
queries directed to a set of federated traditional/standard services for geospatial resources
discovery. Results are finally combined in a meaningful way and sent back to the client.
This framework realizes the Separation-of-Concerns pattern assigning specific tasks to
Page 56
Version:
2.0
Date: 11/Feb/2011
different components, making the architecture flexible and scalable. Moreover, it does not affect
existing geospatial service interfaces implementing a loosely-coupled solution in compliance
with the GEOSS architectural principles.
The system design for the DAC applies the well-known principle of Layered Architecture
(ISO, 1994), as depicted in Figure 11. Functionalities are grouped and layered according to their
abstraction level. Figure 11 shows the three layers of the proposed architecture, implementing
each layer on a different distribution tier:
• in the Presentation Layer we find components implementing graphic user interfaces
(GUIs);
• the Integrated Semantic Layer is composed of components which implement the business
logic necessary to integrate semantic and geospatial services;
• The Single Semantic and Geospatial Query Layer provides query functionalities towards
a set of different services (geospatial, semantic, etc.).
Figure 17 – System Architecture for the DAC System
The DAC clearly falls into the integrated semantic layer and makes use of the services in
the single semantic and query layer in order to implement the query expansion strategy.
The choice of service interfaces was mainly driven by the need of being as compliant as
possible with widely adopted catalog service specifications to be interoperable with existing
systems. Thus, for the interaction between the DAC and the catalog service, the OGC CSW/ISO
AP (Application Profile) interface is used. Among the present application profiles of the OGC
CSW core specification; this is presently one of the most widely implemented. Moreover, this is
the INSPIRE compliant catalog service interface. The access to the semantic service takes place
through SPARQL (the Query Language for Resource Description Format (RDF) semantic
Page 57
Version:
2.0
Date: 11/Feb/2011
documents, a W3C standard) syntax for queries. However, the DAC was conceived to be flexible
and federate also semantic services that use different interfaces.
The DAC shall also provide an interface (towards the presentation layer) for being
queried with any combination of semantic, geospatial and free text constraints. At the time being
there is no well-recognized standard interface or syntax allowing such combined queries. Hence,
the choice was to use the lightweight OpenSearch44 interface. The OpenSearch is a basic
interface, allowing querying a catalogue with a simple free text search. There exist several
extensions of the basic OpenSearch syntax; two widely used extensions to submit geospatial
queries are:
• Geo extension: allows to specify a spatial extent/location as constraint in a query;
• Time extension: allows building queries based on time and time spans constraints.
In addition to the above extensions, we defined a “Concept-driven” extension to allow the
discovery of well-defined concepts and relations between concepts form semantic services.
These three extensions form the DAC query interface.
A detailed documentation describing the “Concept-driven” extension will soon be
published on the OpenSearch Web Site. The AIP-3 Engineering Report Best Practices Wiki45
will be updated as soon as the detailed documentation will be available.
5.6 Operation of the Water Ontology within the EuroGEOSS Discovery
5.6.1 Searching for Concepts/Terms
EuroGEOSS DAC communicates with the GENESIS Vocabulary Service using SPARQL
interface. According to user’s request the EuroGEOSS DAC performs different actions:
1. When the user has searched for concepts/terms related to a keyword of interest (e.g.
“drought”), the EuroGEOSS DAC performs a “GetConcepts” request; that is,
EuroGEOSS DAC builds a SPARQL query to retrieve from the GENESIS
Vocabulary Service all concepts/terms containing the searched keyword in the label
and/or in the description. The matching concepts/terms are returned to the
EuroGEOSS/GENESIS Client Application.
2. When the user is extending a set of concepts/terms, the EuroGEOSS DAC
transforms the selected relation type (e.g. “more specific concepts”) into formal
SKOS relations (e.g. skos: narrower and skos: narrowMatch). Using these relations
a set of SPARQL queries is executed, and matching concepts/terms are returned to
the EuroGEOSS/GENESIS Client Application.
5.6.2 Multilingual Concepts/Terms
Each of the selected concepts/terms is identified by a URI. The EuroGEOSS DAC submits
a SPARQL query to the GENESIS Vocabulary Service in order to retrieve all available
44
45
http://www.opensearch.org/Home
http://wiki.ieee-earth.org/Best_Practices/GEOSS_Transverse_Areas/Data_and_Architecture
Page 58
Version:
2.0
Date: 11/Feb/2011
translations for each of the selected concepts/terms.
5.6.3 European Drought Observatory (Client) Query
EuroGEOSS DAC communicates with the EuroGEOSS Discovery Broker through the
OGC CSW ISO AP 2.0.2 interface. For each of the selected concepts/terms, EuroGEOSS
DAC creates a query that contains geographic (i.e. the envelope characterizing the specific
AOI), and multilingual Keywords constraints (i.e. the concepts/terms selected through the
EuroGEOSS/GENESIS Client Application). This set of queries is submitted to the
EuroGEOSS Discovery Broker. As shown in the “Drought—Europe” video, the user has
identified a geographic area on the map interface, while at the same time, highlighted and
clicked on a concept. Before sending back the results to the client, the EuroGEOSS DAC
groups them according to the matched concept/term.
5.6.4 WPS Request
The European Drought Observatory (EDO) WPS Client sends an Execute request to the
WPS Server, including references to the input thematic layers selected by the user (WCS
endpoint and coverage name).
5.7 Use of EuroGEOSS Semantic Discovery within the European Drought
Observatory (Returning back to the Scenario)
The User Interface to DAC includes two tabs: 1) “Search” and 2) “configuration.” The
process begins with a “Simple Search” text box, in which a user types in the overall query item
of interest “drought.” The “Advanced Search” panel becomes active, containing a text string box
in which the user types the keyword, underneath of which are buttons “Get concepts,” “relation,”
“extend node,” “clean selection,” and “search.” The “Get concepts” button obviously retrieves
the concepts, which are then displayed in graph form as nodes on a tree or graph.
The “relation” button, when pressed, offers the selection of “more general terms,” “more
specific terms,” “corresponding terms,” and “related terms.” The color coding used in the graphs
are: 1) “orange” for more specific, 2) “yellow” for more general, and 3) “green” for
corresponding. For example, “drought indicator” is a “more specific” example of “drought.”
As shown in the demonstration video, a user can draw a box around an area of interest on
a map, while simultaneously having entered the concept of interest to the user. Then the search
results will be retrieved. The returning data matches the requested concepts. Then the selected
datasets can be displayed upon a map server (Scenario Step 04) (see section 6.4)
5.8 Interoperability Arrangements with GEOSS
We use the following service interface:
Page 59
Version:
2.0
Date: 11/Feb/2011
OGC CSW ISO AP, published by the EuroGEOSS Discovery Broker
W3C SPARQL, published by the GENESIS SKOS repository
OpenSearch interface (with geo, temporal and semantic extensions), published by
the EuroGEOSS DAC
Use of the GEOSS Common Infrastructure (GCI):
The EuroGEOSS discovery broker is registered in the GCI. It is accessible through the GEO
Portal using the following standard interfaces:
CSW/ISO 2.0.2
CSW/ebRIM-EO 2.0.2
OpenSearch with Geo and Time extensions
5.9 Post Deployment Activities
5.9.1 Ontology Engineering
As noted in Section 5.4.1, the CUAHSI version 146 water ontology contains the
subclasses of surface hydrology, subsurface hydrology, atmospheric hydrology, land, water
quality, aquatic biology, and infrastructure subclasses, but it lacks an extensively developed
subsurface water subclass and surface subclassification. CUAHSI is preparing to revise and
update a version 2 release of the water ontology, but not in time for AIP-3. Correspondingly,
some development work was undertaken, along with more expected for AIP-4, in preparing a
specialized drought module for the CUAHSI water ontology.
A concise overview of the documentation for these modules follows. However, a standalone drought vocabulary was prepared to use within the EuroGEOSS search tools built into the
European Drought Observatory. This stand-alone drought vocabulary is already operational and
provides the tool to test whether concept-oriented drought searches improve retrieval of drought
information for users—structuring a tool to facilitate the user, as in section 3.2.4. The
operational tool and its search results are reviewed in section 6.
The ontology documentation is provided here. Every water domain specialist, drought
specialist, hydraulic engineer, land surface modeler, hydrologist, and ecologist will recognize
two basic equations: the surface water equation and the surface energy equation. The
“atmospheric hydrology” concept contains subclasses “precipitation” and “radiation” and “wind”
(which creates mechanical turbulence and affects turbulent transfer of water vapor fluxes of
evaporated water back to the atmosphere.
A starting point for more comprehensive water ontology is to build upon the ALMA
convention.47 The ALMA convention has also been used in the distributed hydrologic model
intercomparison experiments undertaken by the EU Water and Climate Change (EU-WATCH).48
46
http://water.sdsc.edu/hiscentral/startree.html
47
http://web.lmd.jussieu.fr/~polcher/ALMA/
48
http://www.eu-watch.org/templates/dispatcher.asp?page_id=25222765
Page 60
Version:
2.0
Date: 11/Feb/2011
FLUXNet controlled vocabulary, the surface water budget equation, the surface energy budget
equation, and the land use classification system of Boston University and the University of
Maryland (as modified by GlobeCover Product Specification of MERIS. These provide a
framework in which the drought vocabulary module can be included and expanded.
An example of the expansion of the land module of the water ontology is presented in
Figure 18.
Figure 18 Land Surface Module of the Water Ontology
6. Evaluating How the Advanced Semantic EuroGEOSS Search and Discovery
System Works
By the end of AIP-3, the EuroGEOSS search interface had been incorporated into the
operational European Drought Observatory web site, even though the water ontology was not
fully functional within the EuroGEOSS Discovery Augmentation Component. Be that as it may,
a specialized drought vocabulary (section 10) was available and linked to the EuroGEOSS
broker. How effectively does this system retrieve specialized drought datasets, its stated user
objective?
The EuroGEOSS Drought Catalog page pops up, containing a list of drought terms
(Section 10), followed by a list of terms representing the GEO Societal Benefit Areas (Section
11). As has been noted, the list of drought terms contains the water budget components, i.e.,
groundwater, discharge, evapotranspiration, low flow, piezometric level, precipitation, snow, soil
moisture, soil moisture deficit, and snow pack.
Clicking the groundwater term and hitting the search button brings up two pages of
references.
Page 61
Version:
2.0
Date: 11/Feb/2011
Highlighting the soil moisture term and clicking on the search button retrieves: daily soil
moisture per region; daily soil moisture anomaly per region (EDO): humedad del suelo en
Espana, composite drought indicator, forecasted soil moisture trend (EDO); forecasted soil
moisture anomaly (EDO); daily soil moisture anomaly EDO; and daily soil moisture.
Although this is not a direct test utilizing the full ontology, the results are encouraging in
supporting the use of a specialized concept-oriented vocabulary, such as that of drought, in
supporting more effective searches.
7. Drought Metadata for fostering interoperability between EDO and EU national
drought monitors
European efforts, independent of AIP, have constructed drought metadata and the
registration of drought datasets (for Spain),part of the metadata creation and registration upper
tier “Catalog” box of the Australia Water Resources Information System diagram (Figure 6).
Page 62
Version:
2.0
Date: 11/Feb/2011
Figure 19
Drought metadata were defined, as documented in D.5.2, Metadata Catalogue for
Drought Information (J Nogueras, et. Al.49
At the start of the project only two WP5 partners (CNIG and CHE) had metadata
catalogues describing drought related resources together with other types of resources, while
some of the other partners had no metadata available for their drought related resources they use.
In addition to this, CNIG and CHE catalogues have been included in the WP2 broker (GI-Cat),
together with this WP5 drought catalogue, so all metadata resources available from WP5 partners
are accessible in a distributed way.
The technology of this catalogue has been developed by the Universidad de Zaragoza.
The EuroGEOSS drought catalogue has been registered as a service accessible through the
EuroGEOSS discovery broker component.50 Thanks to the connection to the IOC brokering
framework, EuroGEOSS users can discover drought related resources in a distributed way.
49
www.eurogeoss.eu
50
http://217.108.210.73/broker/
Page 63
Version:
2.0
Date: 11/Feb/2011
The EuroGEOSS discovery broker component is based on GI-Cat software51.
8. Range of Issues Covered by the Water Working Group
The AIP-3 Call for Participation originally included Water Quality and Drought
(including Agricultural Drought). W. Sonntag and C. Spooner (USA Environmental Protection
Agency), V. Guidetti of the European Space Agency (ESA), and J. Lieberman joined to raise
water quality issues with regards to an EO2Heaven project to be based in Africa and beach
closures in the Gulf of Maine.
The CSIRO Tasmanian ICT Centre has developed the Hydrological Sensor Web (HSW)
based on OGC-SWE standards. Near real-time hydrologic observations and flow forecasting are
published and accessed through the OGC Sensor Observation Service (SOS).
Fig. 18(a) shows the generation process of flow forecasting. Firstly, rainfall observations
are collected from different sensor sites, owned by different agencies, and stored in databases.
The observations are published on the HWS via SOS. A Kepler workflow obtains rainfall
observations from SOS and generates the gridded rainfall surface. A forecast model then
consumes the gridded rainfall data and produces flow forecasts. Finally, the forecasting results
are published onto the HSW through SOS. It can be seen that different agencies are involved in
producing flow forecasting results.
For this use case, a provenance information model has been developed which is demonstrated in
Fig. 18(b). Three sets of ontologies have been adopted, which are the Sensor Ontology, the
WaterML2 Ontology and the Process Ontology to describe information/ knowledge in the sensor
domain, the water domain and the data processing domain, respectively. Then, the Proof Markup
Language (PML) is used to describe the generation processes of information products and link
multi-domain ontologies together. This allows tracking the lifecycle of hydrologic data products,
as well as record-related factors that may impact on data qualities, e.g., sensor setting, model
calibration.
51
http://zeus.pin.uinfi.it/cgi-bin/twiki/view/GIcat
Page 64
Version:
2.0
Date: 11/Feb/2011
Figure 10 (a) and (b)
9. References
Agboma, C.O., S. Z. Yirdaw, & K. R. Snelgrove 2009. Intercomparison of the Total Storage
Deficit Index (TSDI) over two Canadian Prairie catchments. Journal of Hydrology, 374,
351 – 359
American Meteorological Society Glossary of Meteorology
Anderson, M.C., W. P. Kustas, J.R. Mecikalski, and C. R. Hain, 2009 “A GOES-based drought
product using thermal remote sensing of evapotranspiration, 23rd Conference on Hydrology,
Session 2, Drought Prediction, Monitoring, and Mitigation, 17 Jan 2009, American
Meteorological Society
Anderson, M.C., and W. P. Kustas, “Mapping Evapotranspiration and Drought at Continental
and Local Scales with a Thermal-based Surface Energy Balance Model”
Page 65
Version:
2.0
Date: 11/Feb/2011
Andreadis, K.M., E.A. Clark, A.W. Wood, A.F. Hamlet, and D.P. Lettenmaier, 2005. 20th
Century drought in the conterminous United States, J. Hydrometeorology. 6, 985-1001
Berners-Lee, T., J. Hendler, and O. Lassila 2001 “The Semantic Web,” Scientific American,
May, 2001
Boston, T 2010 Australian Water and Climate Networks: the Changing Environment of Data
Sharing,” Workshop on Standards-based Frameworks underpinning Linked Information Sharing
Networks, Canberra, 5 Nov 2010
V. Castillo 2009 Brief note on the Inter-‐Regional Workshop on Indices and Early Warning Systems for Drought, 8-‐11 December 2009, Lincoln, Nebraska U.S.A. Dai, A., K. E. Trenberth, & T. Qian, 2004. A global dataset of Palmer Drought Severity Index
for 1870 - 2002: Relationship with soil moisture and effects of surface warming. Journal of
Hydrometeorology, 5, 1117 – 1130
Doubkova, M. A. Bartsch, C. Pathe, D. Sabel, and W. Wagner The Medium Resolution Soil
Moisture Dataset: Overview of the SHARE ESA TIGER Project
Dracup, J. A., K. Seong Lee, and E. G. Paulson 1980 “On the Definition of Droughts,” Water
Resources Research, 16, 2, 297-302
EuroGEOSS 2010, Specification of the EuroGEOSS Initial Operating Capacity
Fleig, A.K., L. M. Tallaksen, H. Hisdal, & S. Demuth 2006. A global evaluation of streamflow
drought characteristics. Hydrology and Earth System Sciences, 10, 535 – 552
Gibbs, W.J., & J. V. Maher 1967. Rainfall deciles as drought indicators. Australian Bureau of
Meteorology
Hadwen, T. 2008 “The North American Drought Monitor—The Canadian Perspective,”
presentation, March 16-18, Canmore, Alberta.
Iglesias, A. and J. Schlickenrieder 2010 “Overview of Indicators,” Mediterranean Joint Process
Water Scarcity and Drought Working Group Meeting, Madrid, 17 Feb 2010
Institute of Hydrology. 1980. Low flow studies. Tech. rept. Institute of Hydrology, Wallingford,
UK
International Organization for Standardization 1994, “Information technology—Open systems
interconnection—Basic Reference Model: The Basic Model,” ISO/IEC 7498-1
Jones, D.A., W. Wang, and R. Fawcett 2009 “High quality spatial climate datasets for Australia,”
Australia Meteorologic and Oceanographic Journal, 58, 233-248
Page 66
Version:
2.0
Date: 11/Feb/2011
Keyantash, J.A., & J. A. Dracup 2004. An aggregate drought index: Assessing drought severity
based on fluctuations in the hydrologic cycle and surface water storage. Water Resources
Research, 40, W09304
Koike, T 2010 “GEOSS/AWCI Summary Report, including updates of the demonstration
projects,” presentation at the 6th GEOSS/AWCI International Coordination Group (ICG)
Meeting, Bali, Indonesia, 13 March 2010
Lawrence, B. 2010 “British Experience with Building Standards-based Networks for Climate
and Environmental Research,” presentation, Standards Based Information Sharing Networks
workshop, Canberra, Australia, 5 Nov 2010
Lemon, D. P. Box, and R. Atkinson 2010 “Towards a National Environmental Information
Repurposing System, presentation, Standards Based Information Sharing Networks workshop,
Canberra, Australia, 5 Nov 2010
Lloyd-Hughes, B. and M. A. Saunders 2002 “A drought climatology for Europe,” International
Journal of Climatology, 22, 1571-1592
Luo, L., J. Sheffield, and E. F. Wood “Towards a global drought monitoring and forecasting
capability,” 33rd NOAA annual Climate Diagnostics and Prediction Workshop, 20-24 October
2008, Lincoln, NE
Mendicino, G., A Senatore, & P. Versace, 2008. A Groundwater Resource Index for drought
monitoring and forecasting in a Mediterranean climate. Journal of Hydrology, 357, 282 – 302
McKee, T.B., N. J. Doesken, & J. Kleist 1993. The relationship of drought frequency and
duration to time scales. In: Eighth Conference on Applied Climatology. 17-22 January, Anaheim,
California
Moreira,E., C. Coelho, A. Paolo, L. Pereira, and J. Mexia 2008 “SPI-based drought category
prediction using log linear models, Journal of Hydrology 354, 116-130.
Nagai, M. 2009 “Interoperability Arrangements for Geospatial Data,” presentation delivered at
Workshop on Geospatial Information for Developing Countries, December 16, 2009, India
Naresh Kumar, M., C. S. Murthy, M. V. R. Sesha Sai, & P. S. Roy 2009. On the use of
Standardized Precipitation Index (SPI) for drought intensity assessment. Meteorological
applications, 16, 381 – 389
Nunez, L. N. “National Meteorological Service Drought Monitoring”
O’ Hagan, R. G., B. Robinson, G. Swan, and D. Kinny 2008 “Web-based Visualization of Water
Information,” Commonwealth of Australia Water for a Healthy Country Flagship Report
C. W. Pathe, W., D. Sabel, M. Doubkova, and J. Basara, 2009 "Using ENVISAT ASAR Global
Mode Data for Surface Soil Moisture Retrieval Over Oklahoma, USA," IEEE Transactions on
Geosciences and Remote Sensing, 2009
Page 67
Version:
2.0
Date: 11/Feb/2011
Peters, E., & H. A. J. van Lanen 2005 Separation of base flow from streamflow using
groundwater levels- illustrated for the Pang catchment (UK). Hydrological Processes, 19(4), 921
– 936
Rizolli, Li, Athanasiadis, and Marechal 2007 “Semantic-rich interfaces for Simulation
Interoperability”
Rodell, M. “Remote Sensing of Terrestrial Water Storage and Application to Drought
monitoring,”
Sabel, D 2006 “Scaling information from ENVISAT ASAR for downscaling of scatterometer
derived soil moisture estimates down to 1 km,” (thesis), Technical University of Wien
Shafer, B.A., & L. E. Dezman 1982 Development of a Surface Water Supply Index (SWSI) to
assess the severity of drought conditions in snowpack runoff areas (Colorado). Pages 164 – 175
of: Product of 50th Western Snow Conference
Sheffield, J., G. Goteti, F. Wen, and E.F. Wood, 2004 A simulated soil moisture based drought
analysis for the USA. J. Geophysical. Res., 109, D24108, doi: 10.1029/2004JD005182.
Sheffield, J., G. Goteti, and E.F. Wood 2006 Development of a 50-yr, high resolution global
dataset of meteorological forcing for land surface modeling, J. Climate, 13, 3088-3111.
Sheffield, J., D.B.Lobell, and E. F. Wood 2010. Global Drought Monitoring and Forecasting,
based on Satellite Data and Land Surface Modeling, American Geophysical Union, Fall Meeting,
abstract #H23B-1189
Sheffield, J., K. M. Andreadis, E. F. Wood, & D. P. Lettenmaier, D.P. 2009. Global and
continental drought in the second half of the twentieth century: Severity-Area-Duration analysis
and temporal variability of large-scale events. Journal of Climate, 22, 1962 – 1981
Sheffield, J., & E. F. Wood 2007. Characteristics of global and regional drought, 1950 - 2000:
Analysis of soil moisture data from off-line simulation of the terrestrial hydrologic cycle.
Journal of Geophysical Research, 112, D17115
Stahl, K., & H. Hisdal 2004. “Hydroclimatology,” in L, M. Tallaksen & H. A. J. Lanen, (eds.),
Hydrological Drought: Processes and estimation methods for streamflow and groundwater.
Development in water science, no. 48. Elsevier
Svoboda, M 2010 “Drought: A global perspective: Efforts towards a Global Drought Early
Warning System,” presentation at the INSPIRE conference, 23-25 June 2010, Krakow, Poland
Tallaksen, L.M, & H. A. J. van Lanen 2004. Hydrological Drought: Processes and estimation
methods for streamflow and groundwater. Development in water science, no. 48. Elsevier
Page 68
Version:
2.0
Date: 11/Feb/2011
Van Lanen, H. A. J., Z. W. Kundzeciwz, L. M. Talleksen, H. Hisdal, M. Fendekova, and C.
Prudhomme 2008. Indices for Different Types of Droughts and Floods at Different Scales. EU
Water and Climate Change (WATCH) Technical report number 11
Vargas, E 2008a “Drought Management in Spain,” July 8, Zaragoza
Vargas, E 2008b “New Water Indicator System in Spain,” Thematic EIONET Workshop “Water
Quantity and Use,” Copenhagen, June 10-11
Vischel, T., G. G. S. Pegram, S. Sinclair, W.Wagner, and A. Bartsch, "Comparison of soil
moisture fields estimated by catchment modeling and remote sensing: A case study in South
Africa," Hydrology and Earth System Sciences, vol. 12, pp. 751-767, 2008
Wagner, W., C. Pathe, D. Sabel, A. Bartsch, C. Kuener, and K. Scipal “Experimental 1 kilometer
soil moisture products from ENVISAT ASAR for Southern Africa,”
W. Wagner, G. Lemoine, and H. Rott, "A Method for Estimating Soil Moisture from ERS
Scatterometer and Soil Data," Remote Sensing of Environment, vol. 70, pp. 191-207, 1999
Wanders, N., H. A. J. van Lannen, A. F. van Loon 2010 Indicators for Drought Characterization
on the Global Scale, European Union Water and Climate Change (EU-WATCH) Technical
Report 24
Werner, M.G.F., H.C. Winsemius, Y.A.Iglesias, Morales, S. Maskey, and D. Love2010
“DEWFORA: Drought Forecasting and Early Warning for Africa,” 11th
WaterNet/WARFSA/GWP-SA symposium, Victoria Falls, Zimbabwe.
Yirdaw, S.Z, K. R. Snelgrove, & C. O. Agboma 2008 GRACE satellite observations of terrestrial
moisture changes for drought characterization in the Canadian Prairie. Journal of Hydrology,
356, 84 –92.
Yu, L. (2007) Semantic Web and Semantic Web Services. Chapman and Hall/CRC
Page 69
Version:
2.0
Date: 11/Feb/2011
10. EuroGEOSS Drought Vocabulary Keywords
desertification
discharge
DMCSEE - Drought Management Centre
for Southeast Europe
Drought
drought control
drought duration
drought early warning
drought end
EDO - European Drought Observatory
European drought product
drought forecast
GPCC data
drought frequency
Hydrolog
y
Drought
hazard
drought impact
Meteorology
NDWI - Normalized Difference Water
Index
drought index
drought indicator
National/multinational drought product
Natural hazard
drought intensity
drought management
PDSI
Regional/local drought product
drought map
drought mitigation
Remote sensing
SPI
drought monitoring
drought monitoring system
Soil
drought onset
Statistics
drought overview
alert
drought plan
drought product
anomaly
arid climate, desert climate, dry climate
arid zone, dryland, dry zone
drought region
drought resilience
climate
climate change
drought risk
drought severity
climate variability
drought spatial extent
composite drought indicator
drought status
cumulative departure from normal or
climatologically expected precipitation
cumulative precipitation deficit
drought stress
drought threshold
Page 70
Version:
2.0
Date: 11/Feb/2011
Dry
season
snow
snow pack
emergency
soil moisture
evaporation
evapotranspiration
soil moisture deficit
fAPAR - Fraction of Absorbed
Photosynthetically Active Radiation
susceptibility to drought
spatial assessment of drought
Time series
trend
groundwater
heat stress
hydrological drought
type of drought
vegetation productivity
hydrological drought index
hydrological status
vegetation state index
vulnerability to drought
low flow
water deficit
meteorological drought
water
runoff
meteorological drought index
meteorological state
water scarcity
normality
piezometric level
water stored in reservoir
water stress
weather extremes
potential evapotranspiration
pre-alert
precipitation
precipitation anomaly
precipitation deficiency (amount, intensity,
timing)
precipitation deficit
precipitation percentile
rainfall
rainfall anomaly
remote sensing product
reservoir
reservoir volume
semiarid climate
semiarid zone
Page 71
Version:
2.0
Date: 11/Feb/2011
11. EuroGEOSS Water Societal Benefit Area Keywords
Biogeochemistry
Climate prediction
Drought prediction
Ecosystem
Fisheries and habitat
Flood prediction
Human Health
Impacts of Humans
Land use planning
Management
Production of Food
Resource management
Telecommunications-navigation
Water Cycle
Weather prediction
12. Acknowledgments
The role of the USA National Integrated Drought Information System and the USA
National Oceanic and Atmospheric Administration (NOAA) is greatly appreciated in
extending manpower and data and services hub capacity towards hosting the Global
Drought Monitor Portal. The manpower of the European Drought Observatory staff in
setting up OGC Web Mapping Service-enabled Map Servers on the Princeton server and
establishment of interoperability with the NIDIS server is also gratefully acknowledged.
The role of the Japanese Aerospace Exploration Agency (JAXA) in helping
support the GEO Water Community of Practice is acknowledged and appreciated. Such
support was crucial in establishing the initial impetuous for setting up the global drought
monitor through GEO.
Page 72

final document

Transcription

Similar documents

Scientific Name: Cercidium floridum Common Name: Blue Palo

Nate Engle

Beating the Drought with Jamur Zoysia

Challenges Abound as Dry Conditions, Low Reservoirs Take Hold

Isaac Drought - Scholastic New Zealand

slides - DRY Project

Will Pozzi

PowerPoint - Range Beef Cow Symposium

El Nino Induced Drought 2015

Inaugural Address: P. Goswami, Director, CSIR-NISTADS