Project work plan

From CryoClim

Image provided by the SeaWiFS Project, NASA/Goddard Space Flight Center, and ORBIMAGE http://visibleearth.nasa.gov

The project work is split up into four phases and six work packages (WPs).


Phase 1 was carried out in 2008 (the pilot project phase). The three remaining phases are each expected to last about one year. The current phase, Phase 2, which started April 1st 2009, is described in most detail, while the following phases are just outlined recognising that every project gain important experience under the course of the work that should be applied in the detailed work planning. Each phase will be described in full detail before it starts.


WP 1 : System development and verification

Phase 2

Image courtesy NASA JPL http://visibleearth.nasa.gov

All functional and technical requirements will be identified, documented and related to specific user communities within or external to the system. The initial requirements were retrieved in the pilot project and will be built upon. All functional requirements will be described by use cases and related to external requirements as formulated through Infrastructure for Spatial Information in the European Community (INSPIRE), WMO Information System (WIS), and Global Earth Observation System of Systems (GEOSS). The CryoClim system will integrate seamless with INSPIRE WIS, and GEOSS. The overall system design will be updated progressively under the course of the project.


The functional and technical requirements will be broken down into a detailed specification which is suitable for implementation. The detailed specification has to be unambiguous when used for implementation and must comply with external requirements as formulated through INSPIRE, WIS and GEOSS. The detailed specification must specify which INSPIRE, WIS and GEOSS standards (e.g., ISO 19115, ISO 23950, CSW 2.0.2) that will be supported and how the distributed framework will be set up, including onthologies or controlled vocabularies to use.


The technology and standards identified have to be implemented and used to integrate existing infrastructure. This is a stepwise task and requires strong coordination within the project to achieve the functional requirements identified. A test and verification plan has to be followed to make sure the implementation comply with the requirements. The test plan should relate directly to functional and technological requirements identified. The main items to be developed and updated are:

  1. Test and verification plan.
  2. Metadatabases describing the data.
  3. Open-source Project for a Network Data Access Protocol (OpenNDAP) server.
  4. Web Map Service (WMS), first version.
  5. Web Coverage Service (WCS), first version.
  6. Web Portal, first version.

Phase 3

Updated system requirements and design documents after evaluation by key users. Further implementation and testing, improvement of WMSs and WCSs, addition of Catalogue Services linked to the meta-databases and Web Feature Service (WFS).

Phase 4

Updated system requirements and design documents after evaluation by key users. Further implementation and testing, improvement of WMSs and WCSs, Catalogue Services for data and WFS. Addition of catalogue services for services.

WP 2 : Sea ice sub-service

Most of the sea ice sub-service, building on the European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT) Ocean and Sea Ice Satellite Application Facility (OSI SAF), was developed in Phase 1 of the project. It will be fully completed in Phase 2.


Phase 2: Completing the sea ice service

The sea ice products are only processed for areas covered by the satellite data over the ocean where pixels have not been contaminated by radiance from the land. This leaves out sea ice areas close to the coast and at the North Pole. A procedure will be implemented and used to complete the analysis for these areas. In Phase 1 of the project only the northern hemisphere was processed. The southern hemisphere will be processed with the same processing chain. The product portfolio comprises Sea Ice Concentration(SIC), Sea Ice Edge (SIE), and a selection of climate-change indicator products. The aggregated products shall be compared with similar products from NSIDC. The CryoClim sea ice service shall be maintained and run semi-operationally throughout the project period and until the final operational service takes over for this project.


Future operational sensors will have a potential for making more accurate products with higher spatial and temporal resolution. Therefore, when selecting algorithms and implementing the processing chains, it is important to have a clear understanding of future development in order to make the processing chain flexible and adaptable to new data sources. For the sea ice products the NPOESS satellite is particularly interesting.

WP 3: Snow sub-service

Fra-svalbard.jpg

The ultimate goal of the snow service development in the CryoClim project is to develop a novel operational service for multi-sensor Snow Cover Extent (SCE) based on Passive Microwave Radiometer (PMR) and optical data. If this development is successful, a daily high-resolution (1 km) product on SCE will for the first time be available. The intention is to draw on the best features of PMR and optical data for snow cover retrieval and combine these features in an optimal way in a multi-sensor algorithm. A multi-sensor algorithm for optical and SAR data is already developed by the Norwegian Computing Center (NR), so the goal of doing this for optical and PMR is not far-fetched.


The project’s approach of reaching the ultimate goal is to develop progressively more advanced SCE algorithms in a stepwise way and demonstrate the corresponding sub-services under the course of the project. Phase 2 is to develop a sub-service for global SCE products based on PMR, while Phase 3 will do the same for optical data. Both phases will deliver current products as a Near Real-Time (NRT) service, and a time series of products going back in time as far as practically possible. Phase 4 will build on the two previous phases and develop the multi-sensor sub-service. A comparative analysis of the products from the three sub-services will then be carried out, complemented with retrieval and analysis of user experiences of the services. A final decision will then be taken which sub-service that will be continued operationally, which most likely is the multi-sensor service.


Phase 2: Global SCE based on PMR data

A new processing chain for daily global snow cover retrieval will be developed and implemented. This will include:

  • Definition of area: Ocean areas and areas were snow never occurs will be masked out.
  • A first, basic version of a SCE algorithm will be implemented for initial production-chain verification.
  • Preparation of level 1b swath projection of the PMR Special Sensor Microwave/Imager (SSM/I) data.


For this development SSM/I data operationally received at the Norwegian Meteorological Institute (METNO) will be used. Atmospheric data from the European Centre for MEdium-Range Weather Forecasts (ECMWF) will be interpolated to the SSM/I swath data files.


Apply the selected baseline SCE algorithm on a level 1b data set for a given demonstration period (TBD), containing SSM/I brightness temperatures in swath projection and relevant atmospheric data. The result will be used for algorithm evaluation and selection.


The V0 SCE product set will be evaluated against other available product sets. Based on this evaluation, an updated or improved algorithm will be selected for implementation. The previous task will be redone using the new algorithm. This will produce a version 1 product set.


Data will be back-processed as far as PMR data are available to the project (optionally from 1987 for SSM/I and from 1978 for Scanning Multichannel Mircrowave Radiometer (SMMR)).


The production chain will be updated for aggregated SCE products (monthly, seasonal and annual). The production chain will also be extended with necessary sensor cross-calibration and quality control. The products will be stored in a database hosted by METNO and integrated into the CryoClim service.


The CryoClim sub-service for PMR SCE products shall be maintained and run semi-operationally throughout the rest of the project period until the final operational service takes over for this project.


Future operational sensors will have a potential for making more accurate products with higher spatial and temporal resolution. Therefore, when selecting algorithms and implementing the processing chains, it is important to have a clear understanding of future development in order to make the processing chain flexible and adaptable to new data sources. For the snow products the Sentinel-3 and NPOESS satellites are particularly interesting.


Phase 3: High-resolution SCE sub-service based on optical data

A new widely applicable algorithm for fractional snow cover (FSC) area retrieval from optical data was developed by NR in the EnviSnow, SnowMan and EuroClim projects. The algorithm is currently running in a laboratory environment. In the project it will be evaluated for regional and global operational use. During the evaluation it will be compared with other existing snow cover algorithms, including those with METNO (EUMETSAT NWCSAF PPS and OSI SAF AVHRR_ICE). The Bayesian approach applied in the OSI SAF AVHRR_ICE algorithm will also be adapted and tested for use over land. Based on the evaluation, the selected algorithm will be tailored for regional or global operational use of AVHRR data. It will be implemented in the production chain and validated. Products of 1 km spatial resolution will be produced based on the AVHRR time series (FCDR) available at this time (intentionally from 1981), and the optical SCE sub-service will be run semi-operationally throughout the project period.


Phase 4: High-resolution SCE based on multi-sensor data

A new approach based on modelling and assimilation to combine SAR and optical data for snow cover area mapping has been developed by NR. A somewhat similar approach is assumed to work for fusion of optical and passive microwave data and is therefore suggested as the approach combining the best features of the PMR SCE and optical SCE sub-services developed in this project. Current results indicate that it is possible to obtain consistent results of high accuracy from a combination of the two sensors. Based on initial algorithm tailoring and evaluation, it will be implemented in the production chain and extensively validated. Products of 1 km spatial resolution will be produced based on the PMR and optical data time series available, and the multi-sensor SCE sub-service will be run semi-operationally throughout the project period.
Progressively more advanced SCE sub-services have been developed under the course of the project – each sub-service building upon the previous concerning experience, Information and Communication Technology (ICT) and algorithms. If all are successful, the ultimate sub-service will be that based on the multi-sensor algorithm, which represents a novel contribution to snow monitoring in general. Therefore, close to the end of the project, a comparative analysis of the products from the respective sub-services will be carried out. This will be complemented with retrieval and analysis of user experiences of the three services. A decision will then be taken which sub-service that will be continued operationally. The comparative analysis will be made available in the final, operational sub-service(s).

WP 4 : Glaciers mainland Norway sub-service

Image courtesy Image courtesy NASA/GSFC/MITI/ERSDAC/JAROS,and U.S./Japan ASTER Science Team http://visibleearth.nasa.gov

The glacier products for mainland Norway that will be delivered as a service are:

  • Glacier Area Outline (GAO)
  • Glacier-dammed Lake Outline (GLO)
  • Glacier Periodic Photo series (GPP)


The products will be based on Landsat, ASTER, air-borne sensors and topographic maps, as well as terrestrial photography. For each Landsat scene (or image from other optical sensor) the glacier products GAO and GLO will be derived. Topographic maps will be used to extend the GAO and GLO time series prior to 1984. For the products retrieved from satellite sensors the algorithms will be implemented and validated against in situ observations and orthophotos when available, which also will be used to tune the algorithms. The spatial resolution will be ~30 m. The products will be conformant with the recommendations for climate monitoring developed by the Global Climate Observing System (GCOS).


The processing chain will be prepared for use of future operational satellite/sensor systems, like the Sentinel series of satellites.


Phase 2: Algorithms for satellite sensors

The main focus of this phase is to implement the algorithms in the production chain and testing and validating them on a few years of data.


For GAO we will implement the method and derive products for all of mainland Norway. We will also test and adjust the semi-automatic mapping algorithm in areas with special challenges as steep topography, debris cover and lakes of varying turbidity adjacent to glaciers. To increase the temporal resolution (frequency) of the products, images acquired with less favourable conditions (snow, partly cloud cover, low sun angle) could also be used. GLO will be retrieved by using GAO combined with Normalized Difference Water Index (NDWI). The algorithms will be validated by aerial photographs where available.


The CryoClim mainland Norway glacier service shall be maintained and run semi-operationally throughout the project period and until the final operational service takes over for this project.


Future operational sensors will have a potential for making more accurate products with higher spatial and temporal resolution. Therefore, when selecting algorithms and implementing the processing chains, it is important to have a clear understanding of future development in order to make the processing chain flexible and adaptable to new data sources. For the glacier products the Sentinel-2 satellite is particularly interesting.


Phase 3: Creating time series of satellite products

This phase focuses on extending the short time series of products from the previous phase back to 1984 (the Landsat TM era). Furthermore, climate-change indicator products (scalar variables) will be derived from the map products (spatial products).


For suitable Landsat scenes (or images from other optical sensor) the glacier products GAO and GLO will be derived. Orthorectification and thorough quality check of the georeferencing will be crucial for creating a time series that can be used for glacier change assessments.


Three glacier climate-change indicator products are currently foreseen: glacier length, area and mass balance. However, these might be changed or the list extended according to user needs. The indicator product glacier area will be derived directly from the GAO product. The product will be derived for single glaciers as well as on a regional scale. Glacier flow lines need to be created to derive glacier lengths. The length product will be compared with in situ measurements. Glacier mass balance will be taken from in situ measurements on a selection of glaciers.


Phase 4: Extending time series to pre-satellite era

The final phase is to extend the time series of products based on satellite data further back using other information sources. This phase will also include periodic photos into the glacier product portfolio.


Map information will be used to extend the GAO and GLO time series prior to 1984. The sources are (1) digital glacier outlines from the national N50 topographic map series, and (2) digitised glacier outlines in old maps and aerial photographs. For (1) routines for checking digital glacier maps (assign mapping year and quality) are needed. For (2) work will focus on processing chains for geo-referencing, quality checking, digitizing glacier products from analogue maps and aerial photographs.


Glaciers suitable for deriving GPP series will be selected and photos scanned from a selection of glaciers where data are available. It will also be possible to include drawings and paintings (depending on copyright). Guidelines for composing the products also have to be established.

WP 5: Glaciers Svalbard sub-service

The glacier products available for Svalbard in the CryoClim service will be Glacier Surface Type (GST) and Glacier Balance Area (GBA). The production chain includes the following main components:

  • Selection of input product(s).
  • Ortho-rectification.
  • Estimation of relevant parameters.
  • Storage in glacier database.


To include many of the Svalbard glaciers in the project, the production chain has to be as automated as possible, with manual steps only where it is strictly necessary. For the service to become operational the different components of the production chain has to be developed, implemented and integrated in a chain.


The selection of input products is assumed to be handled manually. The configuration of the processing is then set up according to the actual input available and output products needed. The Norwegian Polar Institute (NPI) has a partly in-house-developed software for accurate ortho-rectification of SAR images. This software has to be adapted and validated to the missions selected as baseline for the project. The geo-corrected product will be used as baseline for the estimation of the defined glacier variables.


Phase 2: Algorithm development and validation

The main focus of this phase is to improve, test and validate the algorithms for a few glaciers in Svalbard. The algorithms will be implemented in the production chain making products for a few years. Also, the optical algorithms being applied in mainland Norway will be tested in Svalbard in order to determine their potential for this geographical region.


The algorithms have to be tuned for ASAR (Advanced Synthetic Aperture Radar). The algorithms will be validated against field measurements and other image products available. The GBA product will need a detailed validation to ensure that the method is valid. In the operational product production, field measurements should be used to control the quality of the product regularly.


Future operational sensors will have a potential for making more accurate products with higher spatial and temporal resolution. Therefore, when selecting algorithms and implementing the processing chains, it is important to have a clear understanding of future development in order to make the processing chain flexible and adaptable to new data sources. For the glacier products the Sentinel-1 satellite is particularly interesting.


Phase 3: Implement the production chain for GST and GBA

In phase 3 the tested algorithms will be implemented for semi-automated operation, completing a production chain that can retrieve the information with no or minimal human interaction. The database and the distribution interface will complete the system. When the production chain is operational, the initial step will be to extend the short time series of products from the previous phase back to 1992.


Phase 4: Glacier climate change indicator products at various scales

The final phase comprises the derivation of climate-change indicator products (scalar variables) from the map products (spatial products).


The previously defined products will be extended to the whole SAR satellite era (ERS and Envisat). Based on the spatial glacier products developed and in situ measurements, some climate-change indicator products are foreseen. These include products derived from the GST and GBA.


Glacier Surface Velocity (GSV) is a variable of particular interest in Svalbard as GSV changes much over time for many glaciers. GSV should therefore be considered for inclusion in the service. Estimation of GSV could be based on correlation techniques. This technique can be used both for SAR and optical images. We will have to implement the algorithm together with a validation step to remove anomalies in estimated values. If there will be an operational tandem mission in the future, a GSV product based on an InSAR technique will be a possible future upgrade.

WP 6 : Project management, coordination, promotion and outreach


International coordination

Climate monitoring is currently being developed within several projects and coordinated by initiatives and organisations like WMO, GCOS, WCRP, EEA (ETC/ACC), GMES, GEO and CEOS. In order to make the project results becoming a contribution to the system of systems for global climate monitoring being developed, it is crucial to coordinate the CryoClim project with these initiatives. Special links will therefore be established to and maintained with key users and other relevant initiatives and organisations. This task will also cover coordination and possible contributions from other projects where the project partners are not already involved.


Promotion and outreach

The CryoClim service will not be successful if it not widely accepted and recognised by key users, organisations representing various user groups and organisations representing producers and providers of other climate products. This work package is to develop promotional material (like an open web page and a brochure), expose the project in the media as well as establishing and running a dialogue with key organisations crucial for the success of the service to be developed.


Key user contact

The project established contact with a group of key users within Phase 1 of the project. It is important to keep the dialogue with the users and involve them as close as practically possible in the service development. This task will facilitate meetings with the users and retrieve their feedback on new versions of the system as it is progressively developed.


Project management

The management of the project consists in:

  • Leading the process of carrying out the work plan.
  • Monitoring project progress.
  • Reporting project progress.
  • Carrying out PM and STAG meetings.
  • Plan and prepare each new project phase.