1. 1 Final Report on WTF-CEOP JAXA Prototype system Satoko Horiyama MIURA JAXA/Earth Observation Research Center

2. 2 Contents System Overview Lessons & Learned

3. 3 System Ovserview Purpose : to provide user-friendly access to satellite, in-situ and global gridded model output data. Operational since June 2005, upgraded in March, 2008. Two entry points on the JAXA WTF-CEOP web site (http://jaxa.ceos.org/wtf_ceop/),one for the “Flood monitoring service for Hue, Vietnam” and the other for the “Distributed Data Integration Prototype System”.http://jaxa.ceos.org/wtf_ceop/

4. 4 Flood monitoring service for Hue, Vietnam Focusing on a serious flood that occurred in 2004. Users can acquire various information such as the area covered by water, precipitation amount, stream flow information and population density during the flood period in order to understand the damage and also estimate the risk of future floods.

5. 5 System Configuration Web Browser Client Software WMS WCS TRMM 3B42 data (precipitation) WCS Gateway (creating a time series data) Jblob Model Output Data (precipitation) Base map GIS data etc

6. 6 In-situ Data Station (Kim Long) Example

7. 7 Distributed Data Integration Prototype System Handle satellite data, in-situ data, model output data and MOLTS data. Basic functions Access data from distributed archive sites Select data type, location (Reference Site, Station name), variable name, by height or depth, by time range (start time / stop time) “See the data” (plot, animation and view values on the screen) Download data (NetCDF, csv format) Compare in-situ data and MOLTS data Registration (to satisfy CEOP data policy) Tutorial videos and user’s manuals are available on the web site.

8. 8 (i) Satellite Data 250 km x 250 km subset scenes, regridded to Lat/Lon grid, centered over the 35 in-situ data sites Band selection, period selection and download multiple scenes available (ii) In situ Data Four types of CEOP in-situ data from 35 Reference Sites Surface (19 variables), Tower (9 variables), Soil (temperature, moisture), Flux (Sensible Heat, Latent Heat, CO2, Soil Heat Flux) Download of QC flags available (iii) Model Output Data MOLTS – Model Output Location Time Series (time series at model grid point nearest center of in-situ data sites) Gridded Model Output – 2D/3D time series. Model Output, convert from GRIB to NetCDF Available Data

9. 9 Final System Configuration (1/2) Perl Module Ferret NCAR Archive [In-situ] (U.S.) MPI Archive [MOLTS, Model Output] (Germany) CSDIC Archive [Satellite] (Japan) CEOP Data Archives OPeNDAP Interface Custom Interface Web Browser JAXA Prototype: “Distributed Data Integration Prototype System for CEOP ”

10. 10 Final System Configuration (2/2) JavaScript (google maps API).gif Perl Module Backend Application (Ferret) Catalog Meta data (XML) DODS format data Browser DODS serverUser Interface Software Commnand Product http XML request (http) Port “80” Post method

11. 11 Demonstration

12. 12 Lessons & Learned (1/8)

13. 13 Lessons & Learned (2/8) 1.How to collect requests from scientists [Proposal] Focused questions are more effectivewe. For example; Ask scientists what systems they are using for their research Study those systems and made a list of the characteristics, such as the types of data analysis they provided Send the list to them and ask to select or prioritize what they want

14. 14 Lessons & Learned (3/8) 2.Selecting the interface [Proposal] Selecting the “standard” interface is very useful.

15. 15 Lessons & Learned (4/8) 3.Cooperation with each data center [Proposal] Frequent discussions with data centers, especially about interface, formats and protocols are very important and useful to system development.

16. 16 Lessons & Learned (5/8) 4.Open Source software [Proposal] A kind of “limit test” is recommended before selection of open source software.

17. 17 Lessons & Learned (6/8) 5.Transition to operation How to transition from a prototype system to an operational system is a big task. This next step is very important for us and is still unresolved.

18. 18 Lessons & Learned (7/8) 6.Schedule [Proposal] To emphasize gathering all required information as early as possible in the project, in order to minimize the impact and avoid the worst case that the last information causes serious impact on everything.

19. 19 Lessons & Learned (8/8) Summary 1.Focused questions are more effective to collect scientists’ requests 2.Selecting the standard interface (i.e. OpenDAP) is very useful 3.Discussions with data centers are very important. 4.On using Open Source software, a kind of limit test is important. 5.Transition from a prototype system to an operational system is a big task. 6.Collect required information as soon as possible