1088 
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B4. Beijing 2008 
each other, while dealing with large amount of geospatial data 
and functions to provide spatial decision-making support. For 
example, weather department is to provide meteorological data 
around the chemical facility, such as wind speed, wind direction 
and temperature and so on. Environment department takes on 
simulation of toxic gas dispersion with meteorological 
parameters and other factors. With the support of dispersion 
map of toxic gas, more rational emergency plan would be made 
to guide evacuation and relief of the disaster. 
4.2 GIService Collaboration and Integration 
Based on the presented solution, a formal emergency 
evacuation plan about the motivating example is developed as 
service chains. Some web services, including geo-information 
services and IT web services, can be chained together with the 
help of explicit semantic description and finally translated into 
executable web services flow by BPEL4WS [Curbera, Goland, 
Klein, et al., 2002]. In our prototype system, geospatial data 
services are constructed in GeoServer, the popular open source 
GIS Server. Other IT web services and modules are mainly 
developed in lightweight J2EE Environment, Eclipse and 
Tomcat are chosen. And geographic ontologies are developed 
with Protégé 2000. 
ClScrvkc« {¡cooperations 
Inputs Ì Outputs 
<W$!H.) 
f 
GasDispersion 
Map WS 
PrePlanWS 
Get Facility 
Locaetion 
IdentifyThreal 
Area 
I ! 
j-t-0 GetFacLocRcqVlsg 
~AÊ ( jetF ac LocRes Msg 
L- -(U Get Threat AreaReq M sg 
GetThreatAreaResMsg 
locate Weather Ft® Loc WeatherS T Req Vt sg 
Station 
I oc WeatherS I ResMsg 
GctWcathcr rHüGetWeatherSuinRcqMsg 
Summ.ii. I ^ GetWcatherSutnResMsg 
Plume 
CalculateGas ftf* CalcuGasPlumeR«qMsg 
CalcuGasPlumcRevMsg 
PortrayGas Port ray Ga sD i spers ionReq M sg 
I PortrayGasDispersionResMsg 
Figure 4. Chaining GIServices based on ontological semantics 
Figure 4 describes the collaborative emergency services of 
application scenario, four application services are listed in this 
figure, including evacuation preplan service (PreplanWS), 
weather information service (WeateherlnfoWS), gas plume 
simulation service (GasPlumeWS), and gas dispersion portrayal 
service (GasDispersionMapWS). The above distributed services 
are hosted by different people or agencies within the emergency 
virtual organization. Every service is described by a WSDL 
document, within which interface is presented for remote call 
regardless of platform implemented. The input and output 
parameters of sendees are referenced to particular concepts of 
geo-data ontologies, while the operations of services referenced 
to particular concepts of geo-task ontologies. With the support 
of ontological semantics, distributed services can be chained 
according to data flow. The principle and basic process is 
briefly described as follow. . 
When gas leakage detected and reported, the chemical facility 
is located and symbolized on map rapidly. Service Preplan WS 
contains many geoOperations. IdentifyThreatArea operation 
will be called to identify the directly threatened area of the 
disaster. The weather stations around the facility can be 
searched and returned by Locate Weatherstation operation. 
Given meteorological and terrain parameters around the facility, 
Gas Plume Simulation Service is executed to calculate the toxic 
gas plume. The Gas Dispersion Portrayal Service is to produce 
visualization of toxic gas dispersion while gas plume is 
calculated. Finally, supported by gas dispersion map. 
Evacuation Plan Service can be invoked again to determine the 
secondly threatened area, and more rational evacuation plan 
would be made. The case study shows that semantic-embedded 
GIServices are more flexible to collaborate, and the proposed 
solution can integrate distributed geospatial resources and 
services in loosely coupled way and bring them into business 
process seamlessly. 
5. CONCLUSIONS AND FUTURE WORK 
Urban emergency management is a typical distributed 
collaborative GIS application, characterized by data- 
intensiveness, distribution of resources, and cooperative work. 
To facilitate geospatial information sharing and collaboration in 
emergency management activities, this paper explored 
combining Ontology and Web Services technologies to support 
geo-collaboration. Two elementary geographical concepts, 
geoFeature and geoOperation, were advanced to describe basic 
data semantics and behaviour semantics separately in GIS 
application domain, and an integrated geographical semantic 
representation framework was proposed and designed based on 
service-oriented model. Based on the proposed solution, we 
considered a simplified evacuation scenario of toxic gas 
dispersion, and the orchestration of semantic-embedded 
GIServices was conducted and illuminated. The result shows 
that the distributed geospatial data and services could be 
integrated and collaborated in loosely coupled way. The 
proposed solution is feasible and effective to enable Geo- 
Collaborative emergency services. 
The future work includes the full implementations of the 
proposed framework. Meanwhile, we are currently exploring 
incorporating Grid computing technologies to provide a more 
robust service infrastructure for collaborative GIS applications. 
More efforts will be put on the collaborative application mode 
of Grid-enabled GIServices in support of semantics. 
REFERENCES 
Akkiraju, R., Farrell, J., Miller, J., et al., 2005. A joint UGA- 
IBM Technical Note, version 1.0 "Web Service Semantics - 
WSDL-S," http://lsdis.cs.uga.edu/projects/METEOR-S/WSDL- 
S (18 April, 2005)