OR/15/061 EO service sustainability
| Grebby, S, Jordan, C J, and Dijkstra, T. 2015. Risk information services for Disaster Risk Management (DRM) in the Caribbean - Service utility document. British Geological Survey Technical Note, OR/15/061. |
|
|
EO service recommendations
Following the user assessment, it is recommended that future implementation of Service 1 should involve some refinement of the land use/land cover classes, with input from the local users. The required classes would undoubtedly vary between applications, but it should be possible to define a common set of classes to meet the overall needs of the range of users. Additionally, it might also have been useful to demonstrate to the users the capacity of Service 1 to generate land use/land cover maps that can be used for monitoring purposes.
One of the main recommendations for Service 2 is the extension of the time period covered by the landslide mapping inventories. Only one landslide was identified for Grenada during the specified period in the SOW, which made it difficult to accurately represent previous landslide activity in the hazard modelling. A further recommendation is to always seek to utilise the available satellite imagery with the highest spatial resolution. Although the RapidEye imagery has a relatively high spatial resolution (5 m), the pan-sharpened Pleiades imagery (0.5 m) provides an order of magnitude increase in the amount of detail that can be resolved. Consequently, the Pleiades satellite imagery provided a step-change in the capability to map landslides in this complex terrain, by enabling smaller occurrences to be detected and the style of landslide activity (e.g., rotational, translational, complex) to be determined. Ultimately, having access to very-high resolution satellite imagery acquired prior to 2010 would have aided the generation of a more comprehensive inventory, and subsequently helped to gain a better understanding of the controls on landslide activity.
Services 2 and 3 could be improved by providing national-scale DEMs that represent a significant increase in spatial resolution over the existing data. The specification of 30 m national-scale DEM would probably need increasing to at least 10 m to enable more accurate landslide and flood risk modelling undertaken by the CHARIM project. This project aimed to address this for St. Lucia and Grenada by delivering 1 m DEMs derived from Pleiades stereo satellite imagery. However, this was not possible within the timeframe of the project because of difficulty in acquiring usable cloud-free imagery. For Service 3, one user recommended that the precise 1 m DEM would be extremely useful for the whole of Belize. Moreover, based on user feedback, Service 3 could also be updated to assess the ability extract shallow bathymetric data from very high resolution optical satellite imagery.
Access to relatively cloud-free imagery has been one of the primary constraints on this project, hampering both the land use/land cover mapping and DEM generation aspects of the Services. One general improvement for the future would be to adjust the timing of the project to ensure that the tasking of new satellite imagery does not coincide with the hurricane season. This would help to significantly improve the chances of acquiring imagery with acceptable levels of cloud cover.
The three Services have been designed primarily to meet the needs and requirements of the CHARIM project. For this reason, contact with the local users has been restricted. Accordingly, it has been somewhat challenging to gain a better understand of needs and expectations of local users in order to help ensure that the delivered EO products are also fit for their purposes. More dedicated interaction with the entire range of users would help to address this, while also allowing a more rigorous assessment of the Services to be undertaken.
Relevance to wider bank operations
This project demonstrates the ability to utilise EO data to provide risk information services for disaster risk management. Specifically, the Services have provided fundamental geospatial information directly for input to the landslide and flood risk mapping that has been undertaken for the Caribbean region by the WB CHARIM project. Nevertheless, the type of EO-derived land cover/land use maps, landslide inventories and DEMs delivered by Services 1, 2 and 3 are also relevant for other WB disaster risk-related projects, such the sub-Saharan African landslide risk project that has recently commenced, and the WB Regional Disaster Vulnerability Reduction project. In fact, information on land use/land cover and DEMs should be considered a mandatory requirement of any hazard mapping project, whether that be related to the disaster risk or the risk posed by climate change. The generation of landslide inventory maps using EO data, as demonstrated in Service 2, is especially useful in rugged or mountainous terrain as this approach helps to overcome accessibility issues that limit traditional field surveys. Recently, BGS has employed this approach for mapping co-seismic and monsoon-induced landslides in Nepal following the 25 April 2015 earthquake (e.g. http://www.bgs.ac.uk/research/earthHazards/epom/Nepalearthquakeresponse.html).
Service 1 has the capability to generate temporal land use/land cover information which would be of considerable benefit to other WB Urban Development, Forestry, and Coastal/Ocean monitoring projects being undertaken as part of the eoworld2 initiative. Specifically, the approach used in Service 1 could be used to support ongoing WB forestry, biodiversity and conservation projects in South America, Africa and Indonesia. Moreover, the combination of land use/land cover, river/stream information and DEMs would be applicable to watershed management projects in areas such as Nigeria and India, and sustainable land and water management projects in Ghana and Mauritania. The EO data and techniques utilised in Service 1 can also be modified to contribute information in support of projects concerning food security. For example, such approaches can be used to provide information on agricultural productivity in areas with harsher climatic conditions, in particular Africa and the Middle East.
Service 3 demonstrates the ability to generate accurate and high-resolution DEMs from very high resolution optical satellite imagery. Accordingly, such an approach provides a valuable cost-effective alternative for acquiring elevation data in parts of the world where airborne LiDAR surveys are not feasible due to financial or political constraints. However, if terrain elevation information is required, then airborne LiDAR is perhaps more suitable in areas with dense forest, such as the Caribbean, South America, Indonesia and forested regions of Africa.
Conclusions
This project successfully demonstrates how information derived from EO data can contribute directly to flood and landslide disaster risk management in the Caribbean by providing fundamental geospatial information on land use/land cover, landslides and DEMs. The users appreciated the ability to produce this information from EO data with a higher degree of detail and accuracy than already available, in a more time- and cost- effective manner. Another major strength of the EO approach is the ability to produce consistent information, both historical and in the future (as soon as imagery is acquired). This provides users with the means to undertake various monitoring activities that are not possible or viable using current practices. The main limitations of the EO approaches in the Services were perceived to be the cloud cover and the validation of the EO products. Cloud cover and associated shadows in the imagery made it challenging to achieve complete coverage of the areas of interest during the hurricane season. Although unavoidable here, this issue can be mitigated in the future by either extending the tasking window or planning the acquisition of new imagery around the hurricane season. All the EO products delivered by the Services were validation using conventional practices where possible. However, some users felt that it was necessary to further validate the products through their own additional ground-truthing. This may be partly due to the relative unfamiliarity of these users to EO based products and conventional validation procedures. Accordingly, in future additional interaction with the local users may be required to increase their level of confidence in the reliability of the EO products. Nonetheless, in summary, all users expressed acceptance of the Services, with the delivered EO products appearing to either meet or exceed their expectations.
|
|
