The Irish Land Mapping Observatory (ILMO)- Mapping and Monitoring Land Cover, Use and Change
This page displays all of the metadata information which describes this resource. This metadata information provides details of: the owners and creators of this resource; download links to any files which are available for downloading; geographical and temporal information about the datasets or project in general; other information such as a description of the project, experimental techniques used, data download restrictions, etc.
View other resources on SAFER owned/managed by the owner of this resource.
Files attached to this resource have been downloaded 24 times.
View Download Timeline Graphic
Resource or Project Abstract
Although EU-wide mapping projects such as CORINE have been undertaken since the 1990s, these have proven inadequate for many national and local scale applications given their spatial resolution, restricted class categories and infrequent updating. The Irish Land Mapping Observatory (ILMO) project is part of Ireland?s response to creating a national land cover map that can be generated on an annual basis, in an objective, efficient, scalable and repeatable manner, drawing primarily on satellite imagery supplemented by ancillary field data, to populate the Ordnance Survey Ireland PRIME2 vector database. Satellite imagery have long been used to discriminate ground features on the basis of their spectral reflectance, but to take into account the dynamics of the Irish vegetation several images per year are required. To achieve this in a frequently cloud covered country, 16-day composite MODIS vegetation index images were used to separate elements on the basis of their phenological cycle. While classification using machine learning methods proved to be very successful for homogenous regions, the Irish landscape is highly fragmented and in both Longford and Sligo, problems were encountered with multiple land covers within the 250x250m pixels. Although methods for sub-pixel content retrieval were explored, illumination and geo-location instabilities contributed a significant error component to the time series and detailed land cover mapping in Ireland using currently available medium (250x250m or coarser) resolution data is not possible. To overcome the limitations on data acquisition imposed on optical data by cloud cover, synthetic aperture radar (SAR) imagery for land cover classification were also investigated. The temporal frequency of acquisitions is lower than for MODIS and only one wavelength is available but the 20m spatial resolution of ASAR and PALSAR data is far superior and imagery is reliably available as SAR wavelengths penetrate through cloud. Machine learning algorithms were used to successfully classify the SAR imagery (this time introducing an extra level of classification for grassland; wet and dry) and these maps, with their higher resolution, were used to populate the Ordnance Survey Ireland PRIME2 database. Maps for 2008 and 1992 were produced for Sligo, with an overall accuracy at 20m resolution of 91%, and of Longford with an accuracy of 96%. These land cover maps were used to estimate greenhouse gas emission/reduction profiles for crop and grasslands, with methods developed to estimate carbon stock changes for the land management and land cover and use transitions, based on existing methodologies and newly derived biomass and soil organic carbon (SOC) activity data. Transitions between improved grassland, semi-improved grassland and scrub sub-categories within the grassland remaining grassland category resulted in a large sink (sequestration) of carbon dioxide, equivalent to 0.3 to 1 t CO2 per ha per year. Based on the outputs of this project, it is recommended that high spatial resolution (better than 30m) satellite data are used for land cover mapping of Ireland, but lower spatial resolution data acquired on a shorter update cycle can be used to inform classification of more infrequently acquired imagery. Using the PRIME2 polygons as well-defined coherent objects enables a seamless land cover map to be generated, however this cannot take sub-object variation, such as strip grazing within a field, into account. Furthermore, the polygon boundaries are not updated annually so discrepancies may arise between the apparent edge of a feature and its appearance on a satellite image. Nevertheless, the combination of raster and object datasets makes it possible to perform a more complete carbon balance estimate for land use and cover transitions between two years in Ireland than is currently available.
Go back to top of page
Top 
Contact Information for This Resource
Data, Files, Information Objects Related To This Project Resource
Please Read Carefully Before Downloading
It is important for the EPA and the Environmental Research Community to know that the data and information products we distribute are useful to the user community we support. One way of doing this is by tracking the use of data and information in publications. Parts of or all of the datasets, data, or digital information products available here can be freely downloaded for further use in scientific applications under the condition that the source will be properly quoted in published papers, journals, websites, presentations, books, etc. Whenever possible the EPA strongly urge you to cite the use of these data, information, and web resources in the reference section of your publications. For convenience an automatically generated citation is available below which can be used immediately for citation purposes.
EPA SAFER Data Download Disclaimer: Before downloading any of the data or information objects below you are kindly asked to read the data download disclaimer Read Disclaimer (Opens in New Window). Downloading data or information objects from SAFER-Data indicates an acceptance of the data download disclaimer and the conditions outlined within it.
| Attachment Name and Download Link |
|---|
| Att 1 longford_modis.mpk (1.05 Mb) |
| Att 2 sligo_modis.mpk (1.64 Mb) |
| Att 3 SAR.mpk (8.91 Mb) |
Suggested Citation Information
The information supplied in the table below should be provided with all citations to this electronic resource. You are requested to format each of the fields below as required by the specific bibliographical style you are using.
| Author(s) | Cawkwell, F. |
| Title Of Website | Secure Archive For Environmental Research Data |
| Publication Information | The Irish Land Mapping Observatory (ILMO)- Mapping and Monitoring Land Cover, Use and Change |
| Name of Organisation | Environmental Protection Agency Ireland |
| Electronic Address or URL | https://eparesearch.epa.ie/safer/resource?id=269e8137-2e2c-11e6-ab63-005056ae0019 |
| Unique Identifier | 269e8137-2e2c-11e6-ab63-005056ae0019 |
| Date of Access | Last Updated on SAFER: 2024-10-11 |
An example of this citation in proper usage:
Cawkwell, F. "The Irish Land Mapping Observatory (ILMO)- Mapping and Monitoring Land Cover, Use and Change". Associated datasets and digitial information objects connected to this resource are available at: Secure Archive For Environmental Research Data (SAFER) managed by Environmental Protection Agency Ireland https://eparesearch.epa.ie/safer/resource?id=269e8137-2e2c-11e6-ab63-005056ae0019 (Last Accessed: 2024-10-11)
Go back to top of page Top
Access Information For This Resource
| SAFER-Data Display URL | https://eparesearch.epa.ie/safer/iso19115/display?isoID=3127 |
| Resource Keywords | Satellite remote sensing, optical, radar, land cover, grassland, mapping, carbon budget |
| EPA/ERTDI/STRIVE Project Code | 2011-CCRP-MS-1.4 |
| EPA/ERTDI/STRIVE Project Theme | Land-use, Soils, and Transport |
| Resource Availability: |
 Public-Open |
| Limitations on the use of this Resource | Any attached datasets, data files, or information objects can be downloaded for further use in scientific applications under the condition that the source is properly quoted and cited in published papers, journals, websites, presentations, books, etc. Before downloading, users must agree to the "Conditions of Download and Access" from SAFER-Data. These appear before download. Users of the data should also communicate with the original authors/owners of this resource if they are uncertain about any aspect of the data or information provided before further usage. |
| Number of Attached Files (Publicly and Openly Available for Download): | 3 |
| Project Start Date | Sunday 1st April 2012 (01-04-2012) |
| Earliest Recorded Date within any attached datasets or digital objects | Sunday 1st April 2012 (01-04-2012) |
| Most Recent Recorded Date within any attached datasets or digital objects | Saturday 31st May 2014 (31-05-2014) |
| Published on SAFER | Thursday 9th June 2016 (09-06-2016) |
| Date of Last Edit | Wednesday 15th June 2016 at 10:24:38 (15-06-2016) |
| Datasets or Files Updated On | Wednesday 15th June 2016 at 10:22:33 (15-06-2016) |
Go back to top of page
Top
Geographical and Spatial Information Related To This Resource
Description of Geographical Characteristics of This Project or Dataset
Tthe datasets are landcover maps of counties Sligo and Longford, the two test counties for this project, produced from different satellite imagery (MODIS-optical, and ENVISAT ASAR-radar) over different time periods
Go back to top of page
Top
Supplementary Information About This Resource
In this section some supplementary information about this resource is outlined. Lineage information helps us to understand why this project was carried out, what policy or research requirements did it fulfil, etc. Lineage is important in understanding the rationale behind the carrying out of a project or the collection of a specific dataset etc. Links to web sites, applications, papers, etc are outlined to provide you with additional information or supplementary reading about the project or dataset
| Lineage information about this project or dataset |
| A number of projects had been undertaken by the UCC group on land cover mapping from satellite imagery, giving them a wealth of experience in both optical and radar land cover classification. The EPA have long recognised the need for reliable land cover maps of the country for a variety of policy and administrative purposes, not least carbon accounting, and therefore with the production of the Ordnance Survey PRIME2 vector database, the necessary components were available for the research project. |
| Supplementary Information |
| Other collaborators on this project: Dr Brian Barrett Department of Geography, University College Cork, Cork, Ireland E-mail: bbarrett@ucc.ie Ingmar Nitze Department of Geography, University College Cork, Cork, Ireland E-mail: initze@ucc.ie Stuart Green, Dept. of Agrifood Business and Spatial Analysis Teagasc Food Research Centre, Ashtown, Dublin 15 Email: stuart.green@teagasc.ie Dr Kevin Black FERS Ltd, 117 East Courtyard, Tullyvale, Cabinteely, Dublin 18 E-mail: kevin.black@ucd.ie Peter Hallahan Ordnance Survey Ireland, Phoenix Park, Dublin 8 E-mail: peter.hallahan@osi.ie The key outputs were the land cover maps for Counties Sligo and Longford, at both the 250m resolution using MODIS imagery and 20m resolution using SAR imagery, as well as a national scale map using the MODIS images. The methods developed proved the benefits offered by the PRIME2 vector database for detailed and consistent land cover mapping, which could be adopted at a national scale. Carbon stock changes were calculated for two counties, demonstrating the benefits offered by high resolution imagery. More than 10 oral presentations were given to national and international audiences, with two peer reviewed publications in international journals. the landcover maps were generated by a classification process known as extremely randomised trees, which was found during the project to be the most robust for creating the classifications, the 250m resolution MODIS data are at classified to a general land cover equivalent to Fossitt level 1 with a distinction between improved and semi-improved grassland, the 20m SAR data, which just cover 2008, are classified to greater detail equivalent to Fossitt level 3 as a result of the higher spatial resolution the landcover maps were generated by a classification process known as extremely randomised trees, which was found during the project to be the most robust for creating the classifications, the 250m resolution MODIS data are at classified to a general land cover equivalent to Fossitt level 1 with a distinction between improved and semi-improved grassland, the 20m SAR data, which just cover 2008, are classified to greater detail equivalent to Fossitt level 3 as a result of the higher spatial resolution |
| Links To Other Related Resources |
| https://landmapping.wordpress.com/ilmo/ (Opens in a new window) |
Any links supplied by the resource owner are stored such that they will open in a new window. Following such a link may lead to a 3rd party website over which EPA has no control in regards to contents or suitability.
Go back to top of page
Top
The following is a list of similiar projects and resources on SAFER. Usually these resources share a similiar thematic area to the resource The Irish Land Mapping Observatory (ILMO)- Mapping and Monitoring Land Cover, Use and Change you are currently viewing. You can view the full description for these projects and resources by using the links supplied.
- Irish SIS Final Technical Report 12: Multiple Classifier System View Full Metadata Description
- Irish SIS Final Technical Report 1: Interim classification, Harmonisation and Generalisation of County Soil Maps of Ireland View Full Metadata Description
- The effects of chromium VI on behaviour and on gene expression on the nematode Steinernema feltiae. View Full Metadata Description
- Quantification of the components of the carbon budget at farm scale View Full Metadata Description
- Presentation of the Irish Soil Information System Project View Full Metadata Description
Go back to top of page
Top