OR/14/064 EO information services production, quality-checking and delivery
| Kessler, H, and Dearden, R. 2014. Scoping study for a Pan-European geological data infrastructure: D3.4: technical requirements for serving 3D geological models. British Geological Survey, OR/14/072. |
 |
EO data procurement plan
All EO data required to generate the products for the three services is detailed in Table 7.
| Platform | Product | Date | Scenes | Primary service | TPM? |
| Pleiades | Bundle (0.5 m Pan + 2 m MS) | 25/02/2014 | 2 | 1 - LCM (AOI-A) | Yes |
| Pleiades | Bundle (0.5 m Pan + 2 m MS) | 06/08/2013 | 2 | 1 - LCM (AOI-B) | Yes |
| Pleiades | Bundle (0.5 m Pan + 2 m MS) | 11/04/2014 | 1 | 1 - LCM (AOI-B) | Yes |
| Pleiades | Bundle (0.5 m Pan + 2 m MS) | 18/04/2014 | 1 | 1 - LCM (AOI-B) | Yes |
| Pleiades | Bundle (0.5 m Pan + 2 m MS) | 15/08/2013 | 1 | 1 - LCM (AOI-C) | Yes |
| Pleiades | Bundle (0.5 m Pan + 2 m MS) | 28/01/2014 | 2 | 1 - LCM (AOI-C) | Yes |
| Pleiades | Bundle (0.5 m Pan + 2 m MS) | 23/02/2014 | 1 | 1 - LCM (AOI-C) | Yes |
| RapidEye | MS 5 m L3A | 12/02/2014 | 4 | 1 - LCM (AOI-C) | Yes |
| RapidEye | MS 5 m L3A | 27/12/2014 | 1 | 1 - LCM (AOI-C) | Yes |
| RapidEye | MS 5 m L3A | 27/02/2011 | 1 | 1 - LCM (AOI-C) | Yes |
| RapidEye | MS 5 m L3A | 27/09/2012 | 1 | 1 - LCM (AOI-C) | Yes |
| RapidEye | MS 5 m L3A | 03/01/2014 | 1 | 1 - LCM (AOI-C) | Yes |
| Landsat-8 | OLI/TIRS | 01/02/2014 | 1 | 1 - LCM (AOI-A) | Yes |
| Landsat-8 | OLI/TIRS | 16/01/2014 | 1 | 1 - LCM (AOI-B) | Yes |
| Landsat-8 | OLI/TIRS | 05/03/2014 | 1 | 1 - LCM (AOI-C) | Yes |
| SRTM | SRTM DEM | 2000 | - | 1 - LCM (AOI-ABC) | No |
| ASTER | ASTER GDEM | - | - | 1 - LCM (AOI-ABC) | No |
| RapidEye | MS 5 m L3A | 18/08/2010 | 4 | 2 - LSM (AOI-A) | Yes |
| RapidEye | MS 5 m L3A | 03/01/2011 | 4 | 2 - LSM (AOI-A) | Yes |
| RapidEye | MS 5 m L3A | 29/09/2012 | 4 | 2 - LSM (AOI-A) | Yes |
| RapidEye | MS 5 m L3A | 15/01/2013 | 2 | 2 - LSM (AOI-A) | Yes |
| RapidEye | MS 5 m L3A | 14/02/2013 | 2 | 2 - LSM (AOI-A) | Yes |
| RapidEye | MS 5 m L3A | 03/01/2014 | 2 | 2 - LSM (AOI-A) | Yes |
| RapidEye | MS 5 m L3A | 19/02/2014 | 2 | 2 - LSM (AOI-A) | Yes |
| RapidEye | MS 5 m L3A | 10/03/2011 | 6 | 2 - LSM (AOI-B) | Yes |
| RapidEye | MS 5 m L3A | 12/02/2012 | 6 | 2 - LSM (AOI-B) | Yes |
| RapidEye | MS 5 m L3A | 03/01/2014 | 1 | 2 - LSM (AOI-B) | Yes |
| Pleiades | Fresh stereo | 2014 | 2 | 2 - DEM (AOI-A) | Yes |
| ASTER | VNIR stereo | 2009/11 | 2 | 2 - DEM (AOI-A) | No |
| Pleiades | Fresh stereo | 2014 | 2 | 2 - DEM (AOI-B) | Yes |
| ASTER | VNIR stereo | 2010 | 2 | 2 - DEM (AOI-B) | No |
| SPOT | Elevation30 DEM | - | - | 3 - DEM (AOI-D) | No |
| ASTER | VNIR stereo | 2009-13 | 19 | 3 - DEM (AOI-D) | No |
| Pleiades | Fresh tri-stereo | 2014 | 3 | 3 - DEM (AOI-D) | Yes |
To date, all RapidEye, Landsat-8, ASTER, SRTM and ASTER GDEM data have been procured. The Spot Elevation30 DEM covering ~40% of the total landmass of AOI-D (i.e. Belize) has been ordered and will be delivered shortly after all end-users have been approved by Airbus.
Acquiring Pleiades imagery has been problematic because the ESA TPM contract with Airbus is still officially under negotiation. Nevertheless, the requested archived Pleiades data for use in land use/land cover and landslide inventory mapping has now been obtained ahead of the ESA TPM contract being finalised. Additionally, our request for the acquisition of fresh stereo and tri-stereo Pleiades imagery was approved by Airbus on 6 August 2014, and will be acquired if/when a suitable window of opportunity arises. Stereo and tri-stereo Pleiades imagery is essential for producing the DEMs in service 2 and precise DEM in service 3. Failure to procure the data within the planned schedule due to factors such as cloud coverage will affect the delivery of the associated products.
Ancillary data procurement plan
Due to the request not to contact the users directly, access to ancillary data is largely restricted to that already acquired as part of the WB/ITC project. Luckily, a wealth of ancillary data has been collated for each area of interest (AOI) and uploaded to the ITC FTP site, to which we have been granted access. The main data include limited high-resolution satellite imagery and LiDAR data, cadastral GIS layers, DEMs, geological maps and existing landslide inventories. More detailed description of the available ancillary data for each AOI is provided van Westen (2014)[1]. Additionally, existing land use/land cover maps produced for AOI-A (St. Lucia), B (Grenada) and C (St. Vincent and the Grenadines) by the US Department for Agriculture Forest Service and The Nature Conservancy have been acquired. A summary of the main procured ancillary datasets and their intended use is provided below.
| Service | AOI | Dataset | Intended use |
| 1 - LCM | A | Land use/land cover map | Production/Quality check/Validation |
| 1 - LCM | A | Cadastral GIS (e.g. roads, settlements) | Production/Quality check/Validation |
| 1 - LCM | B | Land use/land cover map | Production/Quality check/Validation |
| 1 - LCM | B | Cadastral GIS (e.g. roads, settlements) | Production/Quality check/Validation |
| 1 - LCM | C | Land use/land cover map | Production/Quality check/Validation |
| 1 - LCM | C | Cadastral GIS (e.g. roads, settlements) | Production/Quality check/Validation |
| 2 - LSM | A | Existing landslide inventories | Production/Quality check |
| 2 - LSM | A | Geological map | Production/Quality check |
| 2 - LSM | A | Landslide susceptibility map | Quality check |
| 2 - LSM | B | Geological map | Production/Quality check |
| 2 - LSM | B | Soil map | Production |
| 2 - DEM | A | 50 m DEM and contour map | Quality check/Validation |
| 2 - DEM | B | 10 m DEM and limited LiDAR data | Quality check/Validation |
| 2 - DEM | B | Cadastral GIS (e.g. roads, buildings) | Quality check |
| 3 - DEM | D | Cadastral GIS (e.g. roads, buildings) | Quality check |
Additional elevation data is required for a robust validation of the DEMs to be produced for service 2 and 3. It is intended to seek access to a set of ground-based GPS data via the WB and ITC.
Production and delivery schedule
The planned production and delivery schedule for the three services is outlined below.
| Jul | Aug | Sep | Oct | Nov | Dec | Jan | |
| Data compilation for all services | |||||||
| SERVICE 2: Initial landslide inventory mapping | |||||||
| SERVICE 2: Landslide inventory validation | |||||||
| SERVICE 2: Revise landslide inventory validation | |||||||
| SERVICE 1: Land use/land cover mapping and validation | |||||||
| SERVICE 2: DEM generation and validation — subject to successful, timely acquisition | |||||||
| SERVICE 3: DEM generation and validation — subject to successful, timely acquisition | |||||||
| SERVICE 1: Product delivery | |||||||
| SERVICE 2: Product delivery | |||||||
| SERVICE 3: Product delivery |
Quality checking
The basic quality checking will involve confirming that all products satisfy the defined criteria in terms of spatial resolution, geometric accuracy and spatial coverage. More specific quality checks will also be undertaken for each product. For instance, the initial quality of the land use/land cover maps in service 1 will be visually compared to the existing maps and also layers such as roads, buildings, water bodies contained in the Cadastral GIS. Any major inconsistencies not attributed to temporal offsets will be noted and then appropriate revisions will be made. The landslide inventory maps in service 2 will be quality checked by visually comparing them to the exiting landslide inventory maps, geological maps and landslide susceptibility maps to ensure they are coherent. The DEMs to be generated in both service 2 and 3 will be automatically quality checked as part of the elevation extraction process of the ERDAS Imagine LPS module. Additionally, the initial quality of the DEMs will be visually assessed through comparison with features contained in the cadastral GIS, and manually compared to existing DEMs and contour maps.
Initial validation
Once quality checked, the products will undergo a more rigorous validation process. For the land cover/land use maps in service 1, this will involve a conventional classification accuracy assessment through the calculation of a confusion matrix. Wherever possible, the land cover/land use at several sample locations will also be noted during field trips to St. Lucia and Grenada. These observations will help to provide an additional level of control on the performance of the land use/land cover mapping. The landslide inventory map in service 2 will also be validated during field trips to St. Lucia and Grenada. Validation will primarily consist of a verifying landslides (and their attributes, where possible) recorded in the inventory with evidence on the ground. Field-based validation will be confined to accessible areas, which are anticipated to be those largely proximal to the road network. The DEMs produced for service 2 and 3 will be validated using accurate elevation data, preferably obtained on the ground using a GPS. Attempts will be made to obtain such data from local users through the WB/ITC. As a fall-back option, the accuracy of the DEMs will be assessed using the existing LiDAR data (where available), DEMs and contour maps. The vertical accuracy will be reported using a set of summary statistics, such as the mean, standard deviation and root-mean-square error.
Necessary revision of the products will be subsequently undertaken to ensure that they all meet the desired thematic accuracy. The products will then be supplied to ITC for initial feedback on their suitability (i.e. specifications have been met, format). This feedback will then be considered and appropriate amendments will be made prior to delivery of the final products.
References
- ↑ Van Westen C J (2014). Preliminary Assessment Report: CHARIM Caribbean Handbook on Risk Information management. ITC, University of Twente.
 |