Godiva2 Overview

40 %
60 %
Information about Godiva2 Overview
Technology

Published on October 20, 2008

Author: jonblower

Source: slideshare.net

Description

Overview of the Godiva2 environmental data online visualization system.

Interactive visualization of four-dimensional environmental data using an enhanced Web Map Service Jon Blower, Reading e-Science Centre, Environmental Systems Science Centre, University of Reading [email_address] http://www. reading.ac.uk/godiva2

Issues we will cover Why does environmental science need geospatial web services? Online visualization of scientific data using a Web Map Service What is meant by an “enhanced” WMS? Difficulties applying geospatial web services to scientific data Ideas for future research

Why does environmental science need geospatial web services?

Online visualization of scientific data using a Web Map Service

What is meant by an “enhanced” WMS?

Difficulties applying geospatial web services to scientific data

Ideas for future research

We need to see into the future All require interdisciplinary science! Flood prediction Search and rescue Climate prediction

How can we make useful predictions? Need computer models that: encapsulate our scientific knowledge are validated by observations Output from these models needs to be disseminated: within the scientific community to government policy makers to emergency response situations ... often in (near) real time Need to compare models with other geospatial data sources e.g. land use maps, locations of assets Requires interoperability

Need computer models that:

encapsulate our scientific knowledge

are validated by observations

Output from these models needs to be disseminated:

within the scientific community

to government policy makers

to emergency response situations

... often in (near) real time

Need to compare models with other geospatial data sources

e.g. land use maps, locations of assets

Requires interoperability

The importance of visualization Detecting features in models (e.g. storms) Diagnosing problems in models Preview data before downloading Make sense of large datasets Puts data into wider context Communicate complex concepts

Detecting features in models (e.g. storms)

Diagnosing problems in models

Preview data before downloading

Make sense of large datasets

Puts data into wider context

Communicate complex concepts

Existing scientific visualization software Problem-solving environments Matlab, IDL 3-D desktop visualization MayaVi 3-D remote visualization Silicon Graphics Web-based Live Access Server Geographic Information Systems (GIS) All require expert knowledge Limited interoperability between systems

Problem-solving environments

Matlab, IDL

3-D desktop visualization

MayaVi

3-D remote visualization

Silicon Graphics

Web-based

Live Access Server

Geographic Information Systems (GIS)

All require expert knowledge

Limited interoperability between systems

Barriers to effective visualization Computer model outputs are large ... Four-dimensional rasters (x,y,z,t) High-resolution Many variables Ensembles Tera/petabyte scale ... complex ... Many file formats and conventions Many numerical grids (right) ... and distributed Too much data to hold in one place

Computer model outputs are large ...

Four-dimensional rasters (x,y,z,t)

High-resolution

Many variables

Ensembles

Tera/petabyte scale

... complex ...

Many file formats and conventions

Many numerical grids (right)

... and distributed

Too much data to hold in one place

Summary so far Predictions need models that are validated by data Numerical model data pose several practical challenges Visualization of env. sci. data is v. important but complex Need to visualize data from lots of different sources

Predictions need models that are validated by data

Numerical model data pose several practical challenges

Visualization of env. sci. data is v. important but complex

Need to visualize data from lots of different sources

Open Geospatial Web Services (plus many more!) Web Coverage Service Gridded data (rasters) Web Map Service Map imagery (PNG, JPG, GIF) Web Feature Service Simple features Complex features

Web Map Service: a closer look GetCapabilities -> metadata GetMap -> map image, based on client-selected parameters, inc: Image width/height Image coordinate reference system Style GetFeatureInfo -> information about a particular map pixel Format not standardized Supports multidimensional data Mandated by EC INSPIRE directive as means for visualizing geospatial data

GetCapabilities -> metadata

GetMap -> map image, based on client-selected parameters, inc:

Image width/height

Image coordinate reference system

Style

GetFeatureInfo -> information about a particular map pixel

Format not standardized

Supports multidimensional data

Mandated by EC INSPIRE directive as means for visualizing geospatial data

Lightweight visualization methods Most heavyweight logic is transferred to server Pros and cons! Imagery transmitted over Web in standard formats WMS interfaces (often) Simple data formats, Javascript APIs Simpler approach, easier to use But functionality often limited NASA World Wind OpenLayers Google Earth Microsoft Virtual Earth

Most heavyweight logic is transferred to server

Pros and cons!

Imagery transmitted over Web in standard formats

WMS interfaces (often)

Simple data formats, Javascript APIs

Simpler approach, easier to use

But functionality often limited

Limitations of WMS for science Map-oriented Scientists want to slice data in lots of ways Need extra metadata for scientific data Format not standardized Clients and servers often don’t support z and t (but it’s in the specification) Server implementations often slow for high-res raster data Can’t use interactively

Map-oriented

Scientists want to slice data in lots of ways

Need extra metadata for scientific data

Format not standardized

Clients and servers often don’t support z and t

(but it’s in the specification)

Server implementations often slow for high-res raster data

Can’t use interactively

A new system: Godiva2 Interactively explore 4D geospatial raster datasets on the web ~40 datasets Research data, operational forecasts, satellite products Images generated dynamically for maximum flexibility OGC Website of the Month, January 2008 http://www.reading.ac.uk/godiva2

Interactively explore 4D geospatial raster datasets on the web

~40 datasets

Research data, operational forecasts, satellite products

Images generated dynamically for maximum flexibility

OGC Website of the Month, January 2008

Selection of depth Select from all the depth levels of the model

Selection of time (range) Select from all the timesteps in the model Selection of a time range leads to an animation

Finding the data value at a point Click on the data layer, data value and precise position is shown Lon: -64.08 Lat: 36.21 Value: 19.27

Timeseries plots If a time range is selected, can create a timeseries plot at a point

Vector plots

Selection of colour palette

Contrast-stretching Manual or automatic

Manual or automatic

Polar projections

Choice of background images

Export to Google Earth Allows visualization of multiple data sources Hurricane Katrina, August 2005 Storm track positions (analysed from ECMWF vorticity data) by Lizzie Froude, ESSC Sea surface temperature data from UK Met Office FOAM model Combination shows cooling of surface waters on right-hand side of cyclonic storm track High winds cause upwelling of cool, deep water

Allows visualization of multiple data sources

Hurricane Katrina, August 2005

Storm track positions (analysed from ECMWF vorticity data) by Lizzie Froude, ESSC

Sea surface temperature data from UK Met Office FOAM model

Combination shows cooling of surface waters on right-hand side of cyclonic storm track

High winds cause upwelling of cool, deep water

Architecture of Godiva2 system Java Web Application (Spring, JSP) Data abstraction layer NetCDF files Other files GetCapabilities GetMap GetFeatureInfo Custom metadata Godiva2 website JSON PNG, GIF Generic WMS client XML Remote data OPeNDAP PNG, GIF

Visualizing distributed data: the MERSEA project OPeNDAP DATA North Atlantic data centre OPeNDAP DATA OPeNDAP DATA WMS @ Reading Dynamic Quick View website (= rebranded Godiva2) Uses existing OPeNDAP-based architecture Single point of failure http://www.resc.rdg.ac.uk/mersea Baltic data centre Arctic data centre Background imagery (from NASA etc)

Removing the bottleneck: Federated visualization OPeNDAP DATA WMS OPeNDAP DATA WMS OPeNDAP DATA WMS Each data centre must install the WMS Less network traffic More robust Third-party WMS Background imagery (from NASA etc)

What is the best use for this? Have an idea Discuss/explore Do the work Formally publish Godiva2 Matlab, IDL etc Disseminate Godiva2

Who’s using Godiva2? 100,000 GetMap requests served in 3 months From 5 continents Customized versions of Godiva2 site set up for MERSEA and ECOOP projects Major EU framework projects – INSPIRE compliance important! Will be used in MyOcean UK National Centre for Ocean Forecasting Server software installed by: Plymouth Marine Labs AIMS, Australia NOAA, US Code contributions from: MeteoGalicia, Spain TPAC, Tasmania AIMS, Australia

100,000 GetMap requests served in 3 months

From 5 continents

Customized versions of Godiva2 site set up for MERSEA and ECOOP projects

Major EU framework projects – INSPIRE compliance important!

Will be used in MyOcean

UK National Centre for Ocean Forecasting

Server software installed by:

Plymouth Marine Labs

AIMS, Australia

NOAA, US

Code contributions from:

MeteoGalicia, Spain

TPAC, Tasmania

AIMS, Australia

Enhancements to WMS Piecemeal metadata-serving avoids large Capabilities document Extra metadata for science data e.g. units of measurement New parameters in GetMap for styling: Choose colour palette Set contrast range Linear or logarithmic scaling Far simpler than Styled Layer Descriptor Generation of timeseries plots via GetFeatureInfo … but fully backward-compatible with WMS1.1.1 and 1.3.0

Piecemeal metadata-serving

avoids large Capabilities document

Extra metadata for science data

e.g. units of measurement

New parameters in GetMap for styling:

Choose colour palette

Set contrast range

Linear or logarithmic scaling

Far simpler than Styled Layer Descriptor

Generation of timeseries plots via GetFeatureInfo

… but fully backward-compatible with WMS1.1.1 and 1.3.0

Interoperability 3rd-party clients can’t use the custom WMS extensions NASA World Wind Cadcorp SIS Google Earth

Godiva2 summary Godiva2 site is useful for exploring and previewing data Users need to download data for more sophisticated analysis Available as open-source software (http://ncwms.sf.net) Have focussed on marine data but applicability is much wider Use of WMS standard enables wide adoption and helps to build a community Successful example of delivering an application via the web

Godiva2 site is useful for exploring and previewing data

Users need to download data for more sophisticated analysis

Available as open-source software (http://ncwms.sf.net)

Have focussed on marine data but applicability is much wider

Use of WMS standard enables wide adoption and helps to build a community

Successful example of delivering an application via the web

Elephants I have ignored: The three “S”s Security Semantics Scalability

Security

Semantics

Scalability

Conclusions…

Why is it hard to reconcile scientific data and open GIS standards? Data volumes often too large for XML XML is primary exchange mechanism Weird and wonderful coordinate systems (spatial and temporal) Well-known, stable coordinate systems Fully four-dimensional data Map-oriented, i.e. 2.5D ( although things are changing, slowly ) Geographic location is an attribute of data Everything is an attribute of a geographic location Science GIS

Future work Support non-raster data E.g. In-situ observations Support non-map slices x-t (Hovmuller) x-z, y-z (depth sections) Visualize multiple datasets at once Add capability for simple data processing Integrate with existing community software THREDDS, GeoServer, ERDDAP

Support non-raster data

E.g. In-situ observations

Support non-map slices

x-t (Hovmuller)

x-z, y-z (depth sections)

Visualize multiple datasets at once

Add capability for simple data processing

Integrate with existing community software

THREDDS, GeoServer, ERDDAP

More research needed… Scalability of servers Key disadvantage of service-oriented software! “ Science profile” for Web Map Service? Earth Observation profile already exists How best to link with Processing Services? E.g. for data intercomparisons Service chaining Appropriate security methods? Redesign of OGC services? Reveal information (esp. metadata) piecemeal Implementation of standards!

Scalability of servers

Key disadvantage of service-oriented software!

“ Science profile” for Web Map Service?

Earth Observation profile already exists

How best to link with Processing Services?

E.g. for data intercomparisons

Service chaining

Appropriate security methods?

Redesign of OGC services?

Reveal information (esp. metadata) piecemeal

Implementation of standards!

Some final thoughts Geospatial Web Services are all about interoperability Interoperability is almost always lossy Law of diminishing returns applies In science we usually can’t lose any information Hence what is the practical limit for application of GWS technology in science?

Geospatial Web Services are all about interoperability

Interoperability is almost always lossy

Law of diminishing returns applies

In science we usually can’t lose any information

Hence what is the practical limit for application of GWS technology in science?

Add a comment

Related presentations