Outline for a review paper to the Ramsar STRP on the application of remote sensing data and low-cost GIS for wetland inventory
Comments from UNEP-WCMC GIS and remote sensing staff, with only a few comments from STRP focal point CZ
GIS and Remote sensing for Wetland Inventory and Assessment work.
"to review further the application of remote sensing data, low-cost GIS and classification systems in wetland inventory"
First comments relate to 'low cost', Is there is such thing as a low-cost especially when it comes to remote sensing? We need to make it very clear that remote sensing is a long process - time is an issue. We have a high standard of expertise in GIS in the centre but less so in remote sensing. We could help with the remote sensing work but are at present still limited in our knowledge. We have participated in a project relating to the remote sensing of coral reefs but even in this we were only involved in a small stage of the processing (in identifying reef/non reef areas), and this is for just one marine ecosystem. We would not have the skills on our own to use remote sensing for wetland identification and analysis but it would be good if we could work with experts to help undertake some of the processing in order to build up our skills base here at UNEP-WCMC. We do have staff with experience using both the ENVI remote sensing software and ERDAS image software. Which software would be used? (Currently we only hold 1 licence for ERDAS IMAGINE and would need to purchase ENVI (costs can be provided). Remote sensing is not low cost, initially there is the purchase of images (which may not be an issue if the bulk of the images are available to this project at no-charge?) and then there is processing time. For gathering of non ESA images, to reduce costs it will be important to look for people/organisations who have already purchased images and perhaps set agreements for image exchange. For example, The University of South Florida's Institute for Marine Remote Sensing (IMaRS) has already purchased a large proportion of the coastal images for the tropical world in landsat 7 /landsat 5 for its Millennium Coral Reef mapping project. If agreements / collaborations with such groups that would benefit both parties could be set up, this would minimise the need to purchase images directly from the suppliers thus reducing image purchase costs.
"One priority recommendation (2002: CoP8 Doc 16, E52.vi) was the compilation and dissemination of models for effective wetlands inventory, assessment and monitoring, using appropriate remote sensing and ground techniques"
What would be the methodology for processing the images? There are numerous ways to do this. We need to make sure there is a consistent methodology in order to produce a consistent and comparable dataset between sites, countries, regions etc. Ground techniques can be very specific and we would need to collaborate. Ed Green Head of Marine Programme) may have a wider knowledge of appropriate techniques and may comment further.
"address the role of remote sensing and GIS for baseline inventory, assessment and monitoring at both the site, national and international scales, and for varying knowledge needs. Data characteristics, suitability, availability, accessibility and cost at multiscales will be described. A range of currently available GIS software applications will be reviewed, using case study examples where appropriate. The review will reduce gaps in existing guidance by clarifying the role of remote sensing and GIS for wetland inventory, assessment and monitoring."
We could use some of our IMS applications as examples. Especially those complex ones where we have integrated image information, spatial databases and non-spatial database. Also with the storage of our data in SDE that allows people access to our GIS datasets online. These could be good examples and could be an area which we could be involved in.
"For site assessments in particular, the review will describe the use of remotely sensed data for implementing a cost-effective approach to reduce the intensity of fieldwork required to identify physical, physico-chemical and biological features of wetland ecosystems."
Using fieldwork to aid a supervised classification to help interpret the images. For extrapolation we need to look at how effective this has been - some habitats are easier than others.
Wetland Assessment "An outline of remotely sensed data, including data requirements, advantages and cost of such data implementation for the identification of status and threats to wetland environments"
For status and threats would we need coverage over time? Need to obtain some older images too (preferably from same sensors for comparable data).
"An outline of the advantages of coupling remote sensing and GIS data for wetland assessment, particularly an environment with which to construct models to test hypotheses or to predict, undertake complex spatial analyses, and ecological risk assessments"
Perhaps we could get involved in some of the complex analysis. Have much experience and expertise in GIS analysis. Lera brings along expertise in models/hypothesis testing/risk assessments. Also WRI's Reefs at Risk - what methodologies are used there? Could these apply to wetlands assessment?
GIS platform selection "A review of the capabilities and limitations of a range of GIS software applications (including ability to integrate and display different remotely sensed data sources; and to edit and create additional new datasets)."
This is exactly what UNEP-WCMC does as a global data centre. Harmonisation of datasets and integration of data from different formats.
"A comparison / evaluation of different GIS applications for undertaking wetland inventory, assessment and monitoring tasks described in the preceding sections (1-3)"
We need more detail to comment on this. Has this already been done? By whom?
Identification of hardware and operating system requirements for specific applications
Remote sensing needs significant amounts of storage space for images and PC with enough processing power, disk space and memory to process the images. For example, we had to have new PC's better than any in the GIS room already in order to process images (we can provide details if required).
"Considerations for managing and maintaining spatial datasets obtained for wetland inventory including: The storage requirements of different data types; Data maintenance issues (to ensure standardisation, and compatibility and ability of exchange of datasets); The requirement for data management through the creation, and maintenance of metadata"
Hold in an SDE system like UNEP-WCMC and output via an IMS. Easy access for all to view the data and everything held in a central location. Alternatively, An IMS accessing data held in multiple locations but in a specific way so that a central IMS can access it.
reviewed, using case study
Apart from our wetlands dataset, which is not really very good, what other wetlands data exist in GIS format. It would be good if there was an inventory of what GIS or Remotely Sensed data everybody involved in wetlands research holds so that it could all be brought into a central location. Some of it indeed may help with processing of images etc. There must be something better out there than our global wetlands dataset!
Finally before making too many suggestions about us undertaking lots of remote sensing we need to liase with UNEP's remote sensing group in Sioux Falls. Perhaps we could work with them.
Also I am a little concerned with the IMS idea that Ramsar is planning to use the people who created the ReefBase IMS. The trouble with this is that it is not realtime so the dataset gets out of date and is reliant on regular transfers of data from us. I don't know whether this could be changed so that it could directly access our data. Also if an IMS was done through us we hold many other datasets that may be useful. It would be good if we could host the Ramsar IMS. [CZ: I see this as part of our contribution to the Ramsar database with the GIS component].