Megatrends with a global impact, such as increasing urbanization and the associated socio-economic and climate change, are leading to an increased disaster risk worldwide. The recent heavy rainfall events with severe flooding in the Mediterranean area in October 2018 showed this in a dramatic form. Deaths and damage amounting to millions are the consequences. According to the latest figures published by the African-Arab Platform for DRR, annual losses in Tunisia alone are estimated at up to USD 140-200 million.
The main challenge of sustainable risk management lies in the availability of complete up-to-date datasets of constructions, infrastructure, terrain and natural land cover. High-resolution satellite data delivers this base information and extensive archives of satellite data are an essential basis for carrying out change analyses.
The pilot project in Monastir is an example, how local data can be combined, supplemented and updated by satellite data in order to improve preparedness and planning of prevention measures
The project "Urban Disaster Resilience through Risk Assessment and Sustainable Planning (UD-RASP)" carried out in Tunisia aimed to improve urban resilience to multiple disasters as part of prevention.
The project, co-financed by the European Commission / DG Echo, was carried out by the project partners IABG (, United Nations University - Institute EHS and the city of Monastir (Tunisia) as pilot city in the period January 2017 - October 2018.
The project developed a methodology to integrate the results of the risk assessment into sustainable urban planning and thus reduce the urban disaster risk (DRR) and support the implementation of the UN Sendai Framework for Disaster Risk Reduction (2015-2030) as well as the UN Sustainable Development Goals 9 and 11. A comprehensive digital urban model for flood simulation, monitoring of coastal erosion and simulation of future scenarios leads to a multihazard risk assessment, accompanied by extensive capacity building measures.
Due to missing, outdated or inaccessible data, it was necessary to use World View 3 stereo data to derive a high-resolution terrain model to provide a reliable basis for exposure analysis and flood modelling. The land cover maps were updated by using the high-resolution imagery in the dense urban area as well as and Sentinel 2 in the surrounding area. A comprehensive Landsat archive delivers the information for an urban sprawl model. This retrospective analysis forms the basis for various future scenarios that can be taken into account in urban planning aspects. The land cover plays also an important role especially in the calculation of the runoff velocity during flash floods.
A standardized geo-database, which integrates the land use plan (Plan d' Aménagement) that is obligatory for North Africa, allows further evaluations and analyses to identify and localize hazard zones (focus on urban flash floods and coastal erosion), analysis of exposure (population, infrastructure) and vulnerability in the affected urban areas. In order to support sustainability in urban planning, a scenario analysis was used to identify and visualize potential risk areas in view of urban flash floods up to the year 2030. This task required a dedicated retrospective analysis based on a large archive of satellite imagery to simulate future urban development (urban growth modelling). In addition to the simulation of different precipitation events (e.g. 100-year flood event HQ100) for the analysis of exposed areas, socio-economic data that could significantly influence future urban development (e.g. economic data, legal situation) were considered in the vulnerability modelling and included in the risk assessment.
This project can be seen as a blueprint scalable for North Africa considers several state of the art simulation processes, the use of different satellites and in situ data and focuses on integration into existing political and operational processes of the Municipalities.
Extensive workshops and trainings for the urban planners and IT experts in the Municipality will ensure the sustainable use of the project results. The extensive training enables the Municipality staff to use the project results and to continue the process of integrating Earth observation based information into urban planning.

On a global basis, disasters strike regularly, in both developed and developing countries. Occasionally however, disasters take on catastrophic proportions, either because of particularly vulnerable populations, a dramatic natural event, or exceptionally unfortunate circumstances. Hurricane Matthew struck southwest Haiti as a Category 4 storm on 2016 October 4th, the first Category 4 hurricane to strike Haiti since Hurricane Cleo in 1964. With upwards of 1,300 lives lost across the Caribbean, and more than a 1,000 lives lost in Haiti, the storm is the deadliest hurricane to strike in the Caribbean since Jeanne in 2004.

In December 2016, CEOS triggered the first ‘Recovery Observatory’ pilot further to Matthew, to answer the needs from Haitian authorities in charge of reconstruction. Haiti’s CNIGS (Centre national d’information geospatiale) quickly took on a leadership role. Together with the CIAT (Comité interministériel d’aménagement du territoire) and ONEV (Observatoire national de l ’environment et de la vulnérabilité), the three Haitian champions defined a vision of how EO could be brought to bear to support a range of applications tied to Recovery, including agriculture, forest and environmental management, coastal zone monitoring, critical infrastructure monitoring (ports and airports) and internal population migrations.

Major earth observation agencies from CEOS such as ASI, DLR, CNES, and ESA have provided technical support to ensure satellite data are available to track recovery challenges and successes in South West Haiti. In addition, the EC’s Copernicus Risk and Recovery Programme was triggered, offering a range of recovery products to support user needs. The RO was set up in collaboration with WB, GFDRR, UNDP, UNEP and others, with the aim of making the Recovery Observatory pilot a demonstration of the value of using satellite Earth Observations to support recovery on a recurring basis. Working with the recovery community to define a sustainable vision, we want to establish institutional relationships between CEOS satellite data providers and international stakeholders for the benefit of people affected by major disasters worldwide.

The RO has already demonstrated success by
• Providing key information (analytical, geospatial) to support decision-making processes and progress monitoring;
• Obtaining access to regular imaging of affected area over a long period;
• Compiling key data sets in a single framework and using them seamlessly thanks to a specialized web-infrastructure;
• Identifying where EO can bring useful information in the recovery phase and defining a generic and replicable concept of Recovery Observatory.

This paper presents two years of activities of the RO, along 6 different themes : Protected natural areas, Landslides-Seismic risks, Watersheds-flooding, Land use, Infrastructure and road communication, Health. Maps and other EO derived information have been produced and validated through workshops with end users. These workshops consolidated strong interest from Haitian partners at the national and local level, and are the first step towards reinforcing capacity to ensure sustainable local solutions

Cyclone Matthew struck southwest Haiti as a Category 4 storm on October 4th, the first Category 4 hurricane to strike Haiti since Hurricane Cleo in 1964. With upwards of 1,300 lives lost across the Caribbean, and more than 1,000 lives lost in Haiti, the storm is the deadliest hurricane to strike in the Caribbean since Jeanne in 2004. The impact of Matthew will be lasting. While flooding caused significant damage and loss of life, the main impact was felt from the wind, which in some regions has destroyed more than 95% of buildings and has completely destroyed trees and agriculture. In addition, widespread environmental damage occurred.
In this context, the Recovery Observatory project, set up by the Committee on Earth Observation by Satellites (CEOS), and co-directed by the Center National d'Etudes Spatiales (CNES), is supporting the post-Matthew recovery phase since the beginning of 2017 over the South and Grande-Anse provinces. Among the objectives of the Recovery Observatory, there is the establishment of a pre-event reference mapping as a baseline for both socio-economic and environmental issues, through the exploitation of high-resolution (i.e. SPOT6/7) and very high-resolution (i.e. Pléiades) optical Earth Observation data. In addition, the setup of a regular acquisition plan enables a monitoring of the recovery process, by providing an up-to-date information for each thematic of interest.
At this time, an important analysis work based on VHR data (0.5m) has been performed, leading to a complete building damage assessment over the cities of Jérémie and Les Cayes and their surroundings just after the passage of Matthew cyclone, and to a monitoring of the reconstruction status one year after the event. Also, a detailed mapping of settlements for 2014, 2017 and 2018 over the protected area of Makaya Park has been realized.
Furthermore, a complete SPOT6/7 imagery coverage (1.5m) acquired in early 2017 over the whole Recovery Observatory study area (South and Grande-Anse provinces), allowed to precisely delineate the coastline between Jérémie and Les Cayes (about 250km). The updated position of the shoreline was compared with the archive datasets for 1970, 2002 and 2010, in order to highlight places where erosion or progradation occur.
An another concern was the lack of an overview concerning the national Makaya Park status. In this way, a vegetation classification just before the passage of Matthew cyclone has been performed. Starting from this, a damage assessment of forest stands was realized from Pléiades imagery (0.5m) acquired soon after the event. A first monitoring of the regeneration status one year after the crisis was also delivered.
The mangrove located close to Pointe-Abacou (between Port-Salut and Les Cayes) is also an area of interest in the frame of the Recovery Observatory project. Two delineations of the mangrove extent have been performed from VHR imagery (0.5m) for early 2016 and early 2018. The comparison between the two resulting layers revealed that the loss of mangrove is dominant but rather moderate and occurs mainly at the periphery.
A final thematic of interest is the impact of Matthew cyclone over the main agricultural areas located in the surroundings of Jérémie, Port-Salut and Les Cayes. A first status of agricultural activities has been derived from pre-event SPOT6/7 data. Then, after the extraction of second post-event status, the cross-analysis revealed that for each area, no changes in cropland extent and activity was detected. The main features that were impacted by the cyclone are related to the agro-forestry landscape, since most of isolated trees or woodland disappeared.

The work presented in this abstract is framed within the activity of the CEOS-DRM project (Committee on Earth Observation Satellite – Disater Risk Management). In this joint project, the University of Genova, the Italian Space Agency (ASI) and Fondazione CIMA provide a contribution to the environmental monitoring of the recovery phase in the aftermath of Hurricane Matthew in the Haiti area, which struck the Carribean during October 2016. This task is tackled here through a change-detection perspective: given a series of remote sensing images acquired in different times it is possible to identify the occurred changes in the area affected by the disaster providing a better understanding of the observed data. The availability of very high resolution (VHR) images acquired by last-generation radar and optical sensors and characterized by a high correlation among neighboring pixels, makes this task very challenging from both an application and a processing viewpoint (Moser et al., 2013).
In particular, the proposed approach integrates multisensor, multisource and contextual information associate with COSMO-SkyMed and Pléiades images, thus formalizing the change-detection task in terms of a data fusion problem. The literature related to data fusion of multitemporal data refers to two main categories: multi-time fusion “at the feature level” and multitemporal fusion “at decision level” (Bovolo and Bruzzone, 2015). The former group is mainly associated with unsupervised decision methods able to generate only binary maps of change allowing the identification of the “change” and “unchanged” classes. However, the monitoring of the recovery phase of Haiti claims for a more detailed characterization of the different land covers. For this reason, the proposed approach belongs to the latter category and, in particular, the change detection problem is operatively addressed through a multitemporal supervised classification approach for multisensor COSMO-SkyMed/Plèiades data. Indeed, the proposed method is able to explicitly identify land cover classes and land cover transitions, characterizing the typology of the occurred change and matching the requirements of the application. In particular, the method makes use of Markov Random Field (MRF) theory to integrate the spatial context and the temporal correlation associated with the images acquired at the different dates (Kato and Zerubia, 2012). This model is fully integrated with a region-based approach able to characterize the geometrical structures in the images through the employment of multiple segmentation maps at different scales (Blaschke et al., 2008).
Given the input images acquired at different times, the output of the proposed approach is represented by two classification maps and by the related change-map. It is worth noting that these maps are jointly generated by a single application of the algorithm. Indeed, in this process the spatial and temporal correlation between the images at different dates are simultaneously considered to generate the change-detection results. On the contrary, other supervised approaches in literature, classify each image of the multitemporal data set per se and the change map is obtained through a simple comparison of the thematic map (Pacifici et al., 2007). Therefore, these methods are not able to take into account the temporal correlation between the images and the overall performances are strictly connected to the results generated on each map.
The approach has been applied to a date set composed of image pairs acquired in November 2015 (before the event), in October 2016 (few days after the event) and a year later, in October 2017.
In order to exploit the potential of the methodology and the spatial information of VHR imagery, the proposed approach is implemented in two following steps:
- a first analysis conducted on the acquired imagery subsampled at a resolution of 10 m;
- a second analysis performed on particularly interesting areas highlighted during the first step making use of the imagery at the maximum available resolution.
The proposed approach achieved very high classification accuracies on the test set (e.g over 98% in the case of the data set acquired in October 2016). A visual analysis of the classification maps confirmed a good behaviour at the spatial border between distinct classes.
The reached overall results highlight significant changes and transactions between the selected thematic classes, confirming the effectiveness of the proposed approach to obtain relevant information for the recovery process.


REFERENCES
Moser, G., Serpico, S. B., & Benediktsson, J. A. (2013). Land-cover mapping by Markov modeling of spatial–contextual information in very-high-resolution remote sensing images. Proceedings of the IEEE, 101(3), 631-651.

Bovolo, F., & Bruzzone, L. (2015). The time variable in data fusion: A change detection perspective. IEEE Geosci. Remote Sens. Mag, 3(3), 8-26.

Kato, Z., & Zerubia, J. (2012). Markov random fields in image segmentation. Foundations and Trends® in Signal Processing, 5(1–2), 1-155.

Blaschke, T., Lang, S., & Hay, G. (Eds.). (2008). Object-based image analysis: spatial concepts for knowledge-driven remote sensing applications. Springer Science & Business Media.

Pacifici, F., Del Frate, F., Solimini, C., & Emery, W. J. (2007). An innovative neural-net method to detect temporal changes in high-resolution optical satellite imagery. IEEE Transactions on Geoscience and Remote Sensing, 45(9), 2940-2952.

Using freely-available official statistics, satellite data (Landsat and Sentinel-2) and open-source software (QGIS), we analysed the correlation between urban growth and material flows trends in the Hanoi Province (Vietnam) for the period 2007-2018, when population increased from 3,2 to ~7.4 million people.
Base on this correlation, the research attempted to model the increase in demand for construction materials in Hanoi resulting from the predicted urban development and to understand the implications for future economic development planning, natural resources allocation and environmental management in the light of the rapid urban expansion planned over the next 10-20 years.
The results represent a preliminary analysis of material supply and demand in Hanoi based on availability of data from the General Statistics Office for Vietnam (GSOV) on the production, trade, consumption and demand for construction-related mineral commodities at a national, regional and city level. A semi-automatic classification scheme from satellite imagery has been used to constrain the urban development and refine national information on construction materials by mapping artificial surfaces, that have extended from ~ 700 km2 in 2013 to ~ 950 km2 in 2018. Spaceborne data has proven to be efficient and low cost tool for deriving land cover changes and particularly efficient in developing countries, where the rate of transformation of rural areas neighbouring to urban areas can be very fast.
The main findings of the research are:
• Statistics on housing and population show Hanoi is likely to continue to rapidly grow in the short to medium term.
• Sand consumption shows a more complex trend. This is most likely due to the effects of national legislation, international trade and issues around illegal mining.
• Despite a dip around 2009-2012 (due to a slowdown in the Vietnamese economy and construction market) consumption of materials shows a clear increase, often in the region of between 10 and 20% year-on-year.
• Projections for future consumption of five key construction materials predict between a twofold increase in requirements for cement and sand and gravel to fivefold increase for steel by 2030 based on 2016 consumption levels.
The work has been developed and supported within the Official Development Assistance Programme of the British Geological Survey which is engaged in providing solutions to the major challenges recognised in the UN Sustainable Development Goals and the Sendai Framework for Disaster Risk Reduction critical to underpinning economic development of a country.

Turkey, as a developing country, is designing and performing huge construction projects around Istanbul, now that it is the most populated city of the country. Beginning from the 1960s rapid urbanization has been taking place due to the industrialization, which also brings a huge increase in the population. Yet, construction projects have been accelerated especially during the last decade, and many new projects are scheduled to be finished in a couple of years. Thus, the environmental change that is caused by these ongoing engineering projects should be defined by the monitoring techniques as a function of time. Ground based observations are generally carried out in order to monitor the deformations within the construction sites, especially through geometric levelling, and GNSS techniques. These monitoring techniques provide vertical and horizontal control respectively, thus they need to be performed simultaneously, in order to fully understand the behaviour of the deforming area. However, in most cases these monitoring measurements are only scheduled within the period of the construction process, and ensuing deformations are not generally considered. Alternatively, space based interferometric technique can also be used to define vertical and horizontal surface displacements with high accuracy, through using the phase difference between SAR images. In particular, Persistent Scatter Interferometry (PSI) is one of the interferometric methods that is capable of defining the three-dimensional deformation for a desired epoch with high temporal resolution, and thus it can be used to validate ground based observations like levelling and GNSS techniques. In this context, recent time series analyses on radar line of sight velocities over Istanbul indicate that since last two decades human-derived changes have more significant effect on coastal subsidence compared to the natural factors. Therefore, the ongoing unsustainable urbanization projects that are progressing in Istanbul should be monitored continuously using an effective technique, such as PSI, that would provide high precision displacements.

The study also focuses on another key Turkish site: Rize. Indeed, Rize, a city with about a hundred thousand inhabitants on the Black Sea coast in the north-east of Turkey, has been reported on the 21st of February 2018 to be at high risk of collapse. For this reason, the local municipality and the government are planning radical solutions such as the relocation of the entire town centre in another place. Subsidence is occurring since the city centre has been built on artificial ground reclaimed to the sea in the 1960. Since then, many multi-floors buildings have been built despite the original planning that was forecasting a maximum of three floors per building. Moreover, the foundations of many buildings, below water level, are reported to be damaged.

In this study, the deforming areas related to the large scale construction projects in Istanbul and in critical infrastructures in Rize are investigated between 2015 and 2018 through the analysis of Sentinel-1 space-borne SAR data, through using PSI technique, and further the time series of deforming objects are also analysed. In Istanbul, the study is focusing on the subway construction projects which are still in progress, and further investigating the eastern coastlines of Istanbul, using velocity maps in comparison with recent studies. In Rize, the study presents many findings in different areas of the city.

Sentinel-1 has been demonstrated to be a suitable system to detect and monitor small changes or buildings and infrastructures. These changes are highly indicative of imminent damage which can lead to the loss of the structural integrity and subsequent failure of the structure in the long-term. Generally, only a few known motion-critical structures are monitored daily with in-situ technologies. SAR interferometry and early-warning systems can assist to save expensive inspection and monitoring services, especially in highly critical cases such the one evaluated in this study.


Reference: E. Erten, C. Rossi; The worsening impacts of land reclamation assessed with Sentinel-1: The Rize (Turkey) test case; International Journal of Applied Earth Observation and Geoinformation; 74, Feb. 2019, pp. 57-64.