The Sustainable Development Goals (SDGs) implemented by the United Nations epitomise the need for interventions to support life on land and reduce poverty and are of particular relevance to rural communities living in the Indo-Gangetic Plain (IGP). Agriculture constitutes a significant part of India’s economy and across Northern India, high agricultural yields and strong productivity are underpinned by fertile soil and the extraction of groundwater. Unfortunately, a strong dependence on irrigation and intensifying agricultural practises are resulting in significant and continued groundwater depletion, and further environmental pressures come from the widespread burning of the leftover rice and wheat stubble, which degrades air quality locally and regionally during the post-harvest seasons (October and November).

Agricultural policy has developed to try to minimise groundwater depletion by shifting rice paddy growing to coincide with the Indian monsoon, and satellite data products from MODIS, AATSR and the Copernicus Land Services shows this shift in north Indian agriculture wheat and rice cropping cycles. However, the same satellite data indicate that an unintended consequence of this change is that the post-fire crop residue burning now occurs later in the year and is compressed into a shorter period, when smoke emissions are therefore more intense. We show that these changes have amplified winter ‘smog’ episodes in Delhi, resulting in widely reported issues of poor air quality that are increasing not only because of an overall growth in total smoke emissions related to increasing yields and intensifying crop production, but also to the shift in timing. These significant (and unforeseen) couplings between the drivers of agricultural productivity and air quality are linked to changes in state government legislation, and demonstrate how use of satellite-derived datasets in the future may better inform evidence based agricultural policy development, both in northern India and possibly in other areas of the world where similar environmental pressures are resulting from intensified agriculture.

Area A - Board 101

So far, Radar data have been mostly used as an all-weather surrogate of optical imagery. However, it is well known that microwave illumination reveals complementary facets of vegetation. A correct interpretation of the backscatter fluctuations would be key towards the synergic in-season mapping of the land cover type and of its phenological and water conditions. Leveraging on this rationale, our research focused on assimilating the C-Band Sentinel-1 data to the optical time-series through a simple but powerful Bayesian network, based on Hidden Markov Models, in order to track the crop status and the land cover changes in near real time.
In this work, our mapping efforts on the rainfed Brazilian agriculture are debated. A fertile scientific partnership has been prominently established with the São Paulo state, thanks its cooperative network of experts and its sensitivity to sustainable growth problems. Several project experiences during the last years led us to fine-tune the methodology to different large-scale and small-scale land use scenarios in this region. A dynamic mapping of the whole Sao Paulo state with 16 days revisit and 30 m resolution has been carried out using Landsat imagery from 2003 (see graphics for 2015 map) and integrating Sentinel-1 IW imagery from 2016. In a first moment, we focused on the characterization of the sugarcane, annual crops and pasture dynamics (caused by the aggressive bio-ethanol production policies) that have been heavily influencing the socio-economic and environmental debate during the last 2 decades. The study was able to show how the sugarcane area underwent a 150% expansion (from 4 Mha to more than 6 Mha), mainly replacing pasture fields, and how the automated mapping solution can achieve comparable results to more expensive human-assisted procedures (e.g. the Canasat products from INPE). It was also highlighted how the assimilation of Sentinel-1 data introduces a 5-10% accuracy increment with respect to the already satisfying Landsat-only solution.
The potential of radar and optical sensors in fine-mapping integrated crop-livestock systems (ICLS) is then further discussed as investigations are currently carried out in the south-west municipalities of the São Paulo state, where our test sites are located. The work, jointly funded by NWO and FapEsp (the organizations for scientific research in the Netherlands and in the São Paulo state respectively), is aimed at assessing the benefits of ICLS (an important asset of the Low-Carbon Agriculture Plan in Brazil) with respect to continuous pasture or crop systems. The remote sensing activities have been aimed in particular at adding relevant features to the Bayesian mapping methodology in terms of land degradation and water use efficiency.
In the presentation we will hence explain the challenges, pitfalls and achievements from our research during the last couple of years.

Area A - Board 102

The present pilot study has applied an innovate Earth Observation (EO) methodology in supporting an IFAD project impact evaluation in Georgia, with a focus on assessing the impact of irrigation scheme rehabilitation on agricultural production of the rural poor. The methodology is derived from the before/after control/impact ‘BACI’ contrast (Meroni et al. 2017). The work has been developed in the framework of the IFAD-WFP climate analysis partnership.

The objective was to have two different and independent sources for evaluating the project impact: the EO based methodology run in parallel with traditional household surveys. Convergence of results from the two approaches can strengthen the quality of the impact evaluation.

The methodology consists of a comparative method that analyses the temporal variations (before and after the intervention) of the Normalised Difference in Vegetation Index (NDVI) of the intervention areas with respect to multiple control sites that are automatically and randomly selected from a set of candidates that are similar to the intervention area.

The rationale is that the intervention will cause a different pattern of change from before to after the intervention compared with similar (control) areas not affected by project interventions. This concept forms the basis of the before/after control/impact (BACI) sampling design applied in this pilot study. The output is an estimate of the magnitude and significance of the difference in greenness (vegetation) change between the intervention area and control areas. The BACI analysis provides two important statistics: the significance level (P-value) of the BACI effect test and the BACI contrast.

Results of the geo-spatial analysis show positive impact in 7 out of 14 area samples with 4 having a significant 0.05 P-value. In three of the five rehabilitated schemes, small plots in project intervention areas performed better than similar plots in the control sites. Results of the ground (household) surveys also showed a similar impact of the intervention on irrigated areas thereby confirming the validity of the geo-spatial analysis, and indicating the exciting potential of this methodology.

The methodology has been applied in a complex environment (anthropized irrigated area) but results are promising as it could allow to better understand and describe the impact of project intervention on agricultural productivity. On the other side a well-designed control group and field visits, generally improved information from the field as well as a better integration of EO analysis in the I&E procedures could enhance the quality of the assessment.

Area A - Board 103

GEO-Essential is the strand of H2020 ERA-PLANET project focusing particularly on the European perspective of global sustainability to ensure the security of supply and the management of resources. In particular, the development of Essential Variables (EV) estimation and Sustainable Development Goals (SDG) monitoring workflows is one of the main targets of this project.
Land productivity is one of the key indicators that allow to estimate state of whole ecosystem as well as general interconnection between “Food-Water-Energy” within NEXUS concept [1, 2].
At the moment there are global land productivity products at 300 m spatial resolution developed by JRC [3] but this product based on total above-ground green biomass and could not be used in direct way for estimation food production potential estimation over cropland areas.
To address analysis of land use changes as well as cropland productivity assessment land cover maps [4-6] are essential as a proxy for understanding the processes in ecosystems and solving many applied problems of satellite monitoring, in particular sustainable agriculture.
Cropland-based land productivity is essential for sustainable and productive agriculture and corresponding workflow is developed within GEO-Essential project for indicator 2.4.1: Proportion of agricultural area under productive and sustainable agriculture.
Two methods for indicator 2.4.1 calculation for the territory of Ukraine were developed.
The first method is close to the classical approach to land productivity estimation based on the NDVI index trend change as a proxy of land productivity. Using the classification map, we estimate the total area of agricultural land (namely cropland mask) and the trend of the NDVI index change over cropland. We define productive and sustainable land as the cropland that has non-negative trend of NDVI index change. The final output of this method is a proportion of the cropland area with non-negative NDVI trend to all area of agricultural land in percentage [0..100].
The second method is based on the EVs derived from land cover and land use classification maps (LULC) as well as outputs from biophysical model WOFOST (the part of gridded CGMS model). For each point over the regular grid, that covers the territory of interest, WOFOST model is used for crop state modeling for 5 main crops. The outputs from the model are driven by meteorological conditions (namely LAI, total biomass and integral characteristics based on these parameters) and could be considered as potential crop production characteristics per specific soil type (soil profile). It should be noted that WOFOST model could not take into account soil degradation – actually soil profile is considered as static.
Sentinel-2 data are used for LAI and total biomass estimation and could be considered as baseline for comparison with potential one’s from WOFOST model. It’s possible to identify areas of soil degradation via comparison of satellite-based and modelled crop characteristics – in both cases the meteorological parameters are more or less the same while soil state could differ significantly.
With this approach land degradation could be calculated without determining of land productivity – it could be calculated as the scaled difference between the potential characteristics of the crop obtained from biophysical model and equivalents derived from satellite data. With comparison of this parameters for cropland areas over the territory of interest it’s possible to estimate land productivity trends.
Thus, productive and sustainable agriculture areas could be considered as areas with positive productivity trend. The output of workflow is a proportion of agricultural area under productive and sustainable agriculture in percentage.
With developed approach it’s possible to assess the areas in Ukraine capable to ensure productive and sustainable food production.

References
1. https://flores.unu.edu/en/research/nexus
2. Kussul, N., Lavreniuk, M., Sumilo, L., Kolotii, A., Rakoid, O., Yailymov, B., ... & Vasiliev, V. (2018, July). Assessment of Sustainable Development Goals Achieving with Use of NEXUS Approach in the Framework of GEOEssential ERA-PLANET Project. In XVIII International Conference on Data Science and Intelligent Analysis of Information (pp. 146-155). Springer, Cham.
3. JRC technical report on LPD dynamics - http://publications.jrc.ec.europa.eu/repository/bitstream/JRC80541/lb-na-26052-en-n%20.pdf
4. Kussul, N., Lavreniuk, M., Skakun, S., & Shelestov, A. (2017, July). Cropland productivity assessment for Ukraine based on time series of optical satellite images. In Geoscience and Remote Sensing Symposium (IGARSS), 2017 IEEE International (pp. 5007-5010). IEEE.
5. Skakun, S., Kussul, N., Shelestov, A. Y., Lavreniuk, M., & Kussul, O. (2016). Efficiency assessment of multitemporal C-band Radarsat-2 intensity and Landsat-8 surface reflectance satellite imagery for crop classification in Ukraine. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 9(8), 3712-3719.
6. Shelestov, A., Lavreniuk, M., Kussul, N., Novikov, A., & Skakun, S. (2017). Exploring Google earth engine platform for Big Data Processing: Classification of multi-temporal satellite imagery for crop mapping. Frontiers in Earth Science, 5, 17.

Area A - Board 104

Providing access to adequate and nutritious food is one of the greatest challenges faced worldwide, especially in low-income and food-deficit countries. From the 86 countries falling in this category, 43 are located in the African continent. EU supports activities relevant to the improvement of rural livelihood in Africa through the funding of the Horizon 2020 project AfriCultuReS “Enhancing Food Security in African Agricultural Systems with the Support of Remote Sensing” (2017-2021). The project aims at using satellite images and other ancillary data to implement and demonstrate an integrated agricultural monitoring and early warning system for Africa, which will support decision making in the field of food security.

Earth Observation (EO) data are routinely being used to monitor vegetation and detect anomalies in plant growth due to water stress, drought, or pests. Moreover, using EO data, it is possible to monitor surface water availability, weather conditions, disaster risks, land use and land cover changes, and to evaluate soil degradation. Satellite data are provided on a regular basis by worldwide organizations (ESA, NASA, EUMETSAT, etc), covering a wide variety of spatial, temporal and spectral data characteristics. The aim of this work is to present the design of EO products to support food security in Africa through the AfriCultuReS project.

First, food security related services were designed in consultation with users to provide the requirements and background information based on identified real needs. The relevant services to be developed fall into the broader categories of climate, crop, drought, land, livestock, water and weather-related services. For each service, a consensus was found for the most relevant scale (local, national, regional and/or pan-African scale). For the national scale, eight African pilot countries were selected, namely Tunisia, South Africa, Ethiopia, Ghana, Kenya, Mozambique, Niger, and Rwanda, thus covering a wide variability of vegetation, climate and socio-economic characteristics.

Next, the available EO data were identified and assessed for their thematic content, accuracy and operationality, according to the needed characteristics identified during the conceptual service development. The proposed EO data were divided into several categories, according to their thematic content or the source for their acquisition. Priority was given on Copernicus’ data and ESA’s products, with various Copernicus operational products selected for use and Sentinel data proposed for integration into the AfriCultuReS EO production chain to estimate essential biophysical indices. For instance, the product “Crop Water Requirements Assessment”, part of the “Water Services” category, will estimate the amount of water needed, through evapotranspiration (ETc), by a disease-free crop (growing under non-restricting soil conditions) in order to estimate and achieve full agricultural production potential.

In general, the nature of the EO products required for the services development are of two types:
1. Ready products made available from other initiatives (Copernicus, etc.) which need to be downloaded, clipped, re-projected, and ingested/delivered from the AfriCultuReS platform. This mostly happens for low spatial resolution products.
2. Products which need the whole process from the downloading of the image to the generation of the final information. In this case, the processing chain includes downloading of the image, pre-processing, clipping, reprojection, processing for information extraction. An example is the computation of LAI at high spatial resolution. Priority was given to algorithms that can be automated to none or minimal human intervention.
The accuracy of all the EO products was reported when available. Nevertheless, the EO data need testing at different levels according to the availability of reference data, as well as the data level of processing. The reference datasets, which can be used for training the algorithms, evaluating their accuracy and fine-tuning, include:
- in-situ measurements collected through campaigns at test sites;
- historical in-situ measurements;
- EO products of high spatial resolution (e.g. Sentinel-1 and Sentinel-2 imagery);
- data coming from other reliable sources (e.g. local governmental institutes and organizations).
The criteria for choosing the data acquisition and pre-processing methodologies were formulated in order to minimize possible errors or problems that could occur during the data implementation process. Therefore, well-known methods commonly used in relevant research were proposed and the data providers are leading organizations in the field of EO data provision.

The designed EO products cover the users' requirements. After their integration into AfriCultuReS services, they are expected to provide a continuous monitoring framework for early and accurate assessment of factors that could affect food security in Africa. Overall, this work contributes towards the second UN Strategic Development Goal 2 "Zero hunger".

Area A - Board 105

The conflicting use of water is becoming more and more evident, also in regions that are traditionally rich in water. With the world’s population projected to increase to 8.5 billion by 2030, the simultaneous growth in income will imply a substantial increase in demand for both water and food (expected to increase by 70% by 2050). Climate change impacts will further stress the water availability enhancing also its conflictual use. The agricultural sector, the biggest and least efficient water user, accounts for around 24% of total water use in Europe, peaking at 80% in the southern regions, is likely to face important challenges in order to sustain food production and parsimonious use of water.
The presentation shows the development of a system for operative irrigation water management able to monitor and forecast the crop water needs reducing the irrigation losses and increasing the water use efficiency. The system couples together multiple satellite LST and vegetation data (from SENTINEL 2,SENTINEL 3, LANDSAT 7 and 8, MODIS) and ground data, with pixel wise hydrological soil water balance model using recent scientifically outcomes on soil moisture retrieval from satellite data and hydrological modelling. Discussion on the methodological approach based on the satellite land surface temperature LST, ground evapotranspiration measures, and pixel wise hydrological modelling is provided proving the reliability of the forecasting system and its benefits.
The activity is part of the European Chinese collaborative project (SIM, Smart Irrigation Modelling, www.sim.polimi.it) which has as main objective the parsimonious use of agricultural water through an operational web tool to reduce the use of water, fertilizer and energy keeping a constant crop yield. The system provides in real-time the present and forecasted irrigation water requirements at high spatial and temporal resolutions with forecast horizons from few up to thirty days, according to different agronomic practices supporting different level of water users from irrigation consortia to single farmers.
The system is applied in different experimental sites which are located in Italy, the Netherlands, China and Spain, which are characterized by different climatic conditions, water availability, crop types and irrigation schemes. This also thanks to the collaboration of several stakeholders as the Italian ANBI, Capitanata and Chiese irrigation consortia and Dutch Aa and Maas water authority
The results are shown for two case studies in Italy and in China The Italian ones is the Sud Fortore District of the Capitanata Irrigation consortium which covers an area of about 50’000 hectares with flat topography, hot summer and warm winter, mainly irrigated with pressurized aqueduct. The district is an intensive cultivation area, mainly devoted to wheat, tomatoes and fresh vegetables cultivation The Chinese one is in the Hehie Daman district covering an area of 20000 ha with fixed time flooding irrigation.

Area A - Board 106

The Earth Observation for Sustainable Development (EO4SD) ESA initiative aims at increasing the uptake of satellite-based environmental information by policy makers. Land cover and land use services are long-term, strategic geospatial information key sources for developing countries and development organizations. In the framework of this project a set of tailored services to foster local projects in Bolivia and Myanmar are being developed in Starlab Barcelona SL. The services are delta mangroves classification, urban classification and land cover for irrigation. Traditional image analysis for land cover and land use classification requires an unsustainable amount of manual labour. Even semi-automatic solutions for land cover classification of satellite images entail the incorporation of expert knowledge. In order to increase the scalability of the built solutions, methods that automate image processing and analysis pipelines are required. There are many classical classifiers with good results for some classes, but nowadays Deep Learning models solve challenging problems with great success and it is expected that these approaches will soon outperform shallow networks and other classification algorithms. In this work a Land Cover and Land Use classification service with Sentinel-2 satellite images based on a deep convolutional neural network (DCNN) is introduced, supported by conventional index-based threshold classifiers, to yield the following classes: Urban, Forest, Cropland, Grassland, Soil and Water. The Deep Learning Algorithm is adapted from the AlexNet model, using a TensorFlow framework and trained with the EuroSat database. The model has been modified to process down to 4x4-sized pixel patches of RGB and NIR bands. The threshold algorithm is based on the well-known NDWI and NDVI indices, where thresholds to separate classes are set following an empirical approach. The two algorithms are applied to the single date image of each region of interest with several tuned configurations. The results from each single classifier computation are integrated with a tree band logic scheme that integrates all the information contained in the individual classifier result. The DCNN provides the full set of class labels while the threshold classifiers support mainly the water and vegetation labels. The whole data acquisition, pre-processing and classification procedure is integrated by a custom processing chain developed in Python taking Sentinel-2 imagery as sole input. The computation is fully automated and works on a configurable tile geographical basis, improving the probability of cloud free scenes. The validation is performed with a statistical sample selected and labelled with the support of very high-resolution imagery. Besides, classification accuracy metrics (overall accuracy and Kappa coefficient) are derived.

Area A - Board 107

The estimation of artificial areas over Europe is necessary to characterise urban sprawl and contribute to indicators related to sustainable development goals (SDGs). Several land products provide information on the location, extent, structure, nature and changes of artificial land in Europe.
a) CORINE Land Cover coordinated by EEA
b) The Copernicus Imperviousness layer for which the production is managed by EEA.
c) The European Settlements Map (ESM) produced by the JRC.
d) The Global Urban Footprint+ (GUF+) produced by the DLR
e) The Copernicus Urban Atlas for which the production is managed by EEA.
f) The LUCAS point survey conducted by Eurostat.
g) Validation samples for some the above mapping products.
The four first products are maps with different scales and concepts of artificial land. The Urban Atlas maps artificial land in over 800 Functional Urban Areas (FUAs). LUCAS is a general purpose survey. It presents limitations due to a complex organization, but has a good potential to be combined with maps. The validation samples selected for the validation of specific maps can be also used as independent data sets.
The direct area estimation by pixel counting in a classified image, or equivalent approaches, is known to have a bias because commission and omission errors are not balanced. The risk of bias in pixel counting is often stronger for classes occupy a small proportion of the total area.
The bias can be estimated and corrected with the help of a sample of reference data. This study compares the results obtained from the datasets listed above to best characterize (i) the extent of artificial areas and (ii) their evolution by combining them with a suitable sampled reference dataset.
Differences on area estimation are often due to the heterogeneous concepts of “artificial”. Comparing changes is more delicate but more informative, often raising anomalies in the datasets.
The reference samples must have associated probabilities to be used for the correct extrapolation of confusion matrices. There are several methods to correct the bias of pixel counting, such as calibration and regression estimators. The reasonable choice depends on the type of data.
Initial results based on the assessment of HRL imperviousness 2012 show that the bias of the naïf (direct) estimation of impervious area from classified satellite images is above 20%, even if the overall accuracy of the classification is above 98%. This observation has an implication on the use of remote sensing for area estimation that is not new. New results will be presented based on the assessment of HRL Imperviousness 2015 and the recently re-analysis of the HRL Imperviousness time series for 2006-2009-2012 as well as ESM, GUF+, CLC , UA and LUCAS datasets.

Area A - Board 108

In 2013 the Chinese government proposed the initiative of an integrated Silk Road Economic Belt (SREB) and the 21st Century Maritime Silk Road (CMSR), known as China’s Belt and Road (BAR) initiative to foster regional and global cooperation, and transform the economic environment in which economies in the region operate. The vision of the Belt and Road Initiative is to realize the “Five Links”, which refer to linkages in policies, infrastructure, trade, finance and people.
Cities and human settlements are key to these five links, being centres of socio-economic activities that enable access to agricultural supplies, markets and trade, and so it is very important to understand regional distribution of urban areas and urbanization patterns, to encourage sustainable development that balances the economic, social and ecological dimensions.
Hence this paper aims to map urban growth over the geography of the SREB, to identify distinctive regional development patterns. To this end, we use low-light imaging data acquired from 1993 to 2017 by the DMSP Operational Linescan System (DMSP/OLS) (1993-2013) and the Visible Infrared Imaging Radiometer Suite (VIIRS) Day/Night Band (DNB) of NASA and NOAA (2013-2017).
DMSP/OLS NTL data were corrected to improve temporal coherence using 13 empirical regression models derived from a stable light detection algorithm with adaptive texture filter; integration of the corrected NTL and DNB data, enabled identifying spatial patterns, characteristics and trends of urbanization of the main cities along the SREB.
The results show that in general the NTL intensities in the SREB geography present conspicuous characteristics; a gradually decreasing intensity trend from west to east, and each of the six economic corridors of the belt increasing its intensity pattern at its own rate, from 1993 through 2017; evidencing the great development of urbanization during this period over the geographies of the Silk Road Economic Belt. As for the types of urban development processes, the analysis of 273 cities, showed a pattern of small cities decreasing, and large and medium-sized cities increasing.
Our study shows that cartography of urban development using multi-temporal remote sensing of night-time lights products can:
a) assist in understanding regional drivers and patterns of urbanisation;
b) monitor the impact of urban planning policies over the SREB through mapping of urbanized features,
c) provide science-based evidence in support of regional socio-economic sustainable development and planning to realise the five links of the BAR initiative and of the SDG 11.3 aimed at enhancing inclusive and sustainable urbanization and capacity for integrated and sustainable human settlement planning and management in all countries.

Area A - Board 109

Managing urban development and growth patterns is a challenge for local administrations in charge of taking into account community’s needs, environmental protection, economic development and social well-being. Increased urbanisation emphasizes challenges for planners and decision makers because urban growth more often causes reduction of natural and agricultural areas, soil sealing, traffic congestions and mobility issues, increasing carbon emissions, worse quality of life. Actually, in 2015 cities and human settlements have been included in the UN Sustainable Development Goals (SDGs). Among the SDGs, SDG 11 (Make cities and human settlements inclusive, safe, resilient and sustainable) stands out as a goal that has placed explicit focus on the measurement of indicators requiring geospatial data for monitoring. Urban planning is a long journey from identification of strategic guidelines and requirements up to completion of plans, and it requires continual monitoring to check actual results and their impacts with respect to the original objectives. Thus, Earth Observation by Remote Sensing has a role to play.
Over the past few years, the ESA’s Copernicus Program and commercial data providers gave birth to a new era of Big Data in the field of Earth Observation (EO), easing the detection of changes in urban areas. Planetek Italia has great experience in urban monitoring and succeeded in creating Rheticus® Urban Dynamics, a vertical service devoted to provide continual monitoring of urban areas by integrating contents generated by its Rheticus® platform and based on VHR data. Especially, Rheticus® Urban Dynamics covers several continuous urban monitoring fields ranging from spatial planning to land use and land cover change monitoring, from soil consumption and imperviousness assessment, to urban heat island detection, urban sprawl, or infrastructural instabilities. Rheticus® Urban Dynamics simplifies the stream of data thanks to a cloud architecture and automatic processing chains, and it overcomes the difficulties and costs of field surveying. It provides a synoptic point of view over wide areas, through remotely measurements of indicators useful to assess different aspects of urban dynamics and to plan and achieve sustainable developments of urban areas. The service could also be available on EO Data and Information Access Services (DIAS) platforms, which are expected to be operational by the end of June 2018.
Among the innovations introduced by Rheticus® Urban Dynamics, shifting from a model of monitoring services on request to long-time information service subscriptions is the real disruptive innovation in the field of modelling of urban development and growth patterns. It helps national-to-local authorities, policy and decision makers to implement urban sustainable development and good practices in accordance with the UN SDGs, to support the quality of planning processes and their proper execution, and to anticipate environmental impacts of urban sprawl.

Area A - Board 110

Climate change-induced phenomena are threatening the rich marine life of the Mediterranean Sea. Climate change and human activities including overfishing, coastal pollution, climate change, eutrophication, and invasion by alien species are leading to habitat and their related ecosystem service loss. The land encompassing the Mediterranean has been inhabited by humans for thousands of years. More than 200 million tourists flock to its sunny coastline each year, increasing the population density in the coastal areas. At this zone, seagrass meadows made by the marine plant Posidonia oceanica are extending down to 40 meters of depth, providing several ecosystem services and functions such as carbon sequestration, prevention of coastal erosion, sediment burial, fishing ground, fisheries nursery, nutrient retention and recycling, and water quality control. Although many regulations include this seagrass as protected habitat today, in the year 2019, not much is known about its spatial distribution in the Mediterranean. Starting from Greece and by using open access satellite remote sensing data from USGS/NASA (Landsat OLI) and EU-Copernicus (Sentinel-2), we have mapped the spatial extent of P. oceanica seagrass meadows along the Greek coastline.
On the methodological aspects, we have used two widely used classification approaches: a) an object-based (OBIA) approach utilizing single images on local computer, and b) a cloud-based pixel approach employing a chain of pre-processing steps and machine learning algorithms over a time-series of data. The pixel approach use 1045 Sentinel-2 tiles over one month period between September 2017 and October 2017, while the OBIA employ 50 carefully selected images over the period between June 2013 and July 2015. Both approaches provide similar spatial distribution patterns; the one provided by Sentinel-2 is 2,510 km2 the Landsat OLI-based is 2,619 km2 while both are achieving accuracies over 75%. Here, we examine the differences in the results of the two methodologies and their causes. Sources of difference are the pixel size (10m vs. 30m) and related minimum mapping unit (100m2 vs 900m2), the various type of in situ data for calibration/validation, and the different methods for the pre-processing and analysis.
The used methods can be adopted for nation-wide baseline mapping of coastal marine habitats at basin level, providing updated information on the spatial extent and distribution at variable temporal scales. As such, Earth Observation can become an important ally in the monitoring of the status of coastal habitats, with special focus at the seagrass meadows, allowing effective climate change mitigation strategies while supporting SDG goals like the SDG 14.

Area A - Board 111

The issues of environmental monitoring and of the technological advances in Remote Sensing are closely related and interdependent fields. Great technological developments in Remote Sensing has led to the observation of phenomena difficult to monitor by conventional methods. Thus, the significant improvement on the spatial and spectral analysis of the available satellite imagery, has provided robust tools for environmental monitoring.
Within this framework, the United Nations proposed a mechanism for monitoring the recorded progress regarding the implementation of the Sustainable Development Goals (SDG's) as adopted in the 2030 Agenda for Sustainable Development. Specifically, the progressing towards the SDG's, is verified by the reporting of indicators, related to various environmental, social and economic phenomena.
In this context, the research's aim is to develop a system for managing voluminous satellite imagery data of various time periods, statistical data and other available spatial data to compute variables and SDG's indicators that are essential to environmental monitoring. The results of the analysis are examined with regards to their compliance with the environmental law and are visualized using thematic maps, as well as dashboards. Thus, the deliverables can be used as evidence for the enforcement of the law (Figure 1).
The prime objective of the proposed methodology is to develop a monitoring tool by using free satellite imagery. In this context, the high resolution Satellite imagery of the Sentinel Missions is the most suitable data to utilize. Moreover, to complete the time-series of satellite imagery prior to the Sentinel Missions (i.e. prior to 2014), Landsat imagery is also used. In addition, to overcome barriers to collecting, integrating and analyzing massive amount of satellite data, the proposed system aims to connect to already operational or under construction platforms that has been or will be created for accessing and delivering satellite imagery (i.e. Copernicus DIAS).
Since the system will be developed for monitoring and reporting SDG's indicators at National level (i.e. monitoring specific areas across Greece), the platform will also rely on various national or local data repositories to acquire additional spatial or thematic data. The aforementioned data are essential for the computation of the SDG's indicators (i.e. population data, social and economic data, spatial data about the extend of protected ecosystems etc.).
The final selection of the satellite imagery that will be used to define the essential variables (EVs) of each indicator, will depend mainly on the definition of the study area from the end-user (i.e. initial geocoding), the time period that the end-user needs to monitor a specific indicator, and finally, the total extend of the selected study area (i.e. specific town, ecosystem, areas that are subject to specific environmental risks etc.).The proposed workflow suggests also the use of various remote sensing techniques to pre-process the selected satellite imagery (i.e. image enhancement and sharpening, georeferencing, etc.).
The second parameter of the proposed system is the codification of the existing environmental legislation. Specifically, all existing national environmental laws, guidelines, policies, presidential Decrees etc. is organized in individual thematic areas according to the issue they address (i.e. industry, illegal waste, air pollution, water quality, legislation on specific protected areas, etc.). Main objective of the proposed methodology is to identify the requirements of each legislation and the resulting corresponding alert values.
Finally, the third parameter of the proposed platform is the SDG's Indicators and the corresponding variables that are essential for the monitoring of each indicator (EVs). The official indicators list includes 232 SDG's indicators. However, only few of them have clear methodology for monitoring and reporting, while at the same time, only 9 indicators are defined using exclusively earth observation data. The proposed methodology, suggests organizing all indicators in a common database and associate them with their corresponding essential variables. Main incentive of the study, is to investigate whether satellite imagery and earth observation techniques can be used to define new essential variables for computing the SDG's indicators.
Finally, the proposed platform will be tested and applied in several pilot areas across the country, especially, on areas that are subject to potential environmental risks, or urban, high-populated areas.
In that spectrum, the wider area of the Lefkas’ Lagoon, has been selected as a case study that represents a typical area where the enforcement of the environmental legislation and the monitoring of SDG’s indicators can be assessed using the above-mentioned automated system.
Sprawling over a 2.12 acres area, Lefkas' Lagoon, is one of the biggest protected ecosystem in Greece (Natura Ecosystem) and an area that is subject to potential risks due to the existence of the nearby town of Lekfas. Specifically, Lefkas town is one of the most popular tourist destinations of the Ionian Islands, which has led to the development of an ever-growing tourist industry that has altered the natural landscape.
The interaction of the above-mentioned natural ecosystem with the booming urban tissue of Lefkas, poses important environmental considerations for both sites. For instance, Lefkas' Lagoon has been recognized as a potential site of the Mediterranean sea, that can be affected from tidal waves in case of earthquakes and other natural disasters (e.g. floods) that will also have hazardous impact on the adjacent Lefkas Town. Recently, the earthquake that struck the neighboring Zakynthos Island (6,8 on Richter scale) created a huge tidal wave of 20 centimeters above the sea level, affecting Lefkas Lagoon in less than half an hour, while an hour later, the tidal waves had already reached the Italian coasts (Figure 2).
Finally, in the wider area of the protected Lagoon, serious violations of the environmental law are recorded systematically, due to the overuse of natural resources (e.g. overfishing and illegal aquaculture, illegal dumping, degradation of the subsoil and the natural habitat of the Lagoon due to dehydration etc.), (Figure 3)
Concluding, the study’s aim is to investigate and record the aforementioned burden of the area’s natural environment as well as the level of the environmental law enforcement, by using satellite imagery of different time periods. The processing of satellite data is carried out by modern remote sensing techniques in order to define and calculate the essential variables that best describe the studied environmental phenomena.
For instance, in the pilot case of Lefkas' Lagoon, the degradation of Lagoon's protected water-ecosystem, can be monitored via the indicator 6.6.1 Change in the extend of water-related ecosystems over time that is subject to SDG 6. "Ensure availability and sustainable management of water and sanitation for all", as well subject to SDG 15. "Protect, restore and promote sustainable use of terrestrial ecosystems, sustainably manage forests, combat desertification, and halt and reverse land degradation and halt biodiversity loss".
The information on the total extend of the water-ecosystem will be computed via the combination of satellite imagery and bythometric data of the wider area. In addition, the above-mentioned indicator, is also a robust tool for monitoring the compliance to environmental legislation related to the requirements of the protected ecosystem of Lefkas' Lagoon (i.e.Natura 2000 site, Natura Code: GR2240001).
Finally, the deliverables of the study are visualized in thematic maps and interactive dashboards, a cutting-edge data visualization tool for specialized and non-specialized users.

Area A - Board 112

The International Financing Institutions (IFIs) provide technical and financial support to developing countries to reduce poverty and stimulate economic growth. This type of support is included in the classification of Official Development Assistance (ODA), which is managed and regulated by the Development Assistance Committee (DAC) of the Organisation for Economic Cooperation and Development (OECD).
However, many of the countries and regions receiving this type of development support are in areas where there are ever-increasing changes taking place in the environment, as a result of both natural processes and also human activities.
Achieving this goal of sustainable development involves dealing with many complex challenges, examples of which include: the effects of climate change, the increasing frequency and severity of extreme weather events, rapid urbanization, threats to food security, natural resource depletion, and the risk of natural disasters.
Addressing these challenges requires a detailed understanding of the state of the earth’s environment; what is happening now, where has it come from, and what is going to be the likely evolution in the future. It is in this context, that information about the environment from Earth Observation (EO) satellites can play a key role.
It will be presented the second group of activities launched by ESA in its EO4SD initiative "Larger-Scale EO Exploitation Activities in Support of Sustainable Development Initiatives". The overall objective is to integrate EO-based products & services as ‘best-practice’ environmental information in the planning and implementation of the development projects, programs and activities of the IFIs, together with their respective Client States. Main IFIs addressed by EO4SD are the WB and ADB.
It comprises four specific thematic domains: Climate Resilience , Disaster Risk Reduction, Marine & Coastal Resources Management and States Affected by Fragility, Conflict & Resilience.
The main objective of all EO4SD thematic cluster is to carry out convincing demonstrations of the benefit and utility of EO-based information of the environment in support of international development projects in different priority regions, mainly Africa, South Asia and East Asia & Pacific and Latin America.
To achieve the objectives EO4SD is based on basic premises: a) to set up and build stakeholder engagement. in order to gather their specific requirements for tailored EO services to their needs. b) Emphasis on Capacity Building activities and events (regional workshops) to build awareness/acceptance of EO derived information among all stakeholders c) delivering sustainable EO services and products that fulfill their needs and foster the EO adoption in their operational processes.
As a case study of governance, the presentation will include the results of EO-derived information with Sentinel data supplied during 2018 by the EO4SD DRR cluster to the ADB in support to the Disaster risk & Recovery Master Plan (2018-2020) after the devastating effects produced by an earthquake followed by a tsunami in Sulawesi (Indonesia).

Area A - Board 113

Sentinel-1 is the operational C-band synthetic aperture radar (SAR) satellite mission of the European Copernicus Program providing weather and sunlight independent radar imagery of the whole Earth on a 6 to 12 day cycle. However, there is still a strong reluctance by a large community to use SAR data because of its complexity in regard to sensor-specific noise, topographic effects, (pre-)processing requirements, and its unfamiliar appearance for interpretation for untrained eyes. These technical and human capacity requirements as well as the vast amount of data makes it out of reach for most non-technical user groups that could benefit from its monitoring potential in general and especially in persistently cloud-covered areas as the tropics. Under the ESA activity “Earth Observation for Sustainable Development”, this project aims to overcome these technical challenges and subjective reluctance by providing attractive, easy-to-use “Advanced Sentinel-1 Analysis Ready Data” (ASARD) imagery with a focus on African users. Such imagery needs to be timely accessible, visually attractive and easily interpretable without losing essential information for the users monitoring needs and its potential for further numerical analysis. Such information could for example be conserved by providing statistical parameters of dense time series and providing noise-reduced averaged gamma-naught backscatter mosaics over specific time periods (monthly, seasonal or yearly) instead of single Sentinel-1 scenes. By a demand-driven approach involving five African countries (D.R. Congo, Ethiopia, Ghana, Malawi and South Africa), we aim to standardize such ASARD imagery to boost the use of Sentinel-1 data for a wider non-technical user community, including policy makers, in order to support the United Nations Sustainable Development Goals. We will also investigate cloud processing possibilities and cooperate with the African Regional Data Cube (ARDC) to ease the operational provision and timely access of such data in the future.
In this presentation, we will present the first approach of ASARD products over the five study sites (Figure 1), each covering an area of [2° longitude x 2° latitude]. The products consist in monthly and yearly averaged S1 mosaics with 3-band RGB as
RGB = [VV;VH; NDI= (VV-VH)/(VV+VH)] and two 6-band images representing the first order yearly statistic of the co- and cross-polarization gamma-naught backscatter; average, variance, minimum, maximum, the number of acquisitions and an accumulated mask for SAR shadow and layover. Examples of the applicability and added-value of these products for environmental monitoring of water, forest, land cover and/or change detection will be given with a first user assessment to implement the data in their operations.

Area A - Board 114

In 2009, the Geostationary Environmental Monitoring Spectrometer (GEMS) mission was launched to monitor various environmental issues periodically over a wide area around the Korean peninsula and East Asia using remote sensing technology. GEMS will be onboard the GEOstationary Korean Multi-Purpose SATellite 2B (GEO-KOMPSAT2B), whose launch is scheduled for 2019 on an Ariane-5 ECA vehicle from French Giana-Kourou and will be the world’s first geostationary sensor for primarily monitoring the atmospheric environment. An Environmental Satellite Center (ESC) and antenna system has been established and the GEMS ground station system was designed and developed last year. The system consists of a receiving subsystem, a data pre-processing subsystem (PPS), a data management subsystem (DMS), a scientific data processing subsystem (SDPS), and an environment-portal (E-Portal) subsystem. The receiving subsystem is to receive GEMS raw data. The PPS contains the data processing system for producing L1A and an image navigation system for producing Level 1B (L1B). The DMS is to store and manage all produced data in the GEMS ground station. The SDPS is to produce Level 2 (L2) information from pre-processed L1B images using an air pollutant generation algorithm. The E-Portal is to collect data within the ground station and to provide a data distribution service. In addition, there are data analysis subsystems for displaying Level 3 (L3) and Level 4 (L4) data from L2 and metadata received, a data verification subsystem for analyzing the quality of GEMS outputs and deciding whether to reprocess, and a control and monitoring subsystem for monitoring and controlling operations, communication, security, and faults in each subsystem. Currently, all the subsystems have been developed and their integration testing is in progress at the center. GEMS operation has the following six main concepts: a non-stop operating ground station for 24 hours per day and 365 days per year, data acquisition in real-time and distribution in near-real-time, operating for 10 years according to the designed duration of GK2B, the archiving of all data that is received and produced in main storage, the construction of a back-up system for data reliability, and achieving 99% or more operational availability for high-speed processing and customized services with a Hot backup system. Through the successful operation of the ground system, we expect GEMS to improve monitoring and climate change and air quality forecasts over the Korean peninsula and East Asia.

Area A - Board 115

In 2015, member countries of the United Nations adopted the 17 Sustainable Development Goals at the Sustainable Development Summit in New York. These global goals have 169 targets and 232 indicators based on the three pillars of sustainable development: economic, social, and environmental. However, substantial challenges remain in obtaining data of the required quality. One promising and innovative way of addressing this issue is to use Earth observation (EO). The research reported here presents an update on ongoing work to develop a Maturity Matrix Framework (MMF) for assessing the suitability of EO derived data for populating the SDG indicators, with an especial focus on those indicators covering the more social and economic dimensions of sustainable development as these are relatively unexplored in an EO context. An early version of the MMF was published in 2018, and since then the key assumptions have been tested and evolved with inputs from a wide community of EO and indicator experts and conducted through semi-structured interviews with 38 respondents worldwide. Results from the survey confirm that EO can make an important contribution towards populating a wide diversity of the SDG indicators, and respondents have made suggestions for modifying the MMF. Therefore, the paper presents the new version of the MMF and how the structures relate to the interview results. A case study is present for one of the SDGs showing how the MMF can be applied to indicators to asses their suitability for assessment via EO.

Keywords: Earth Observation, Sustainable Development Goals, indicators, Maturity Matrix Framework

Area A - Board 116

The Amazon is home to many indigenous communities. In Ecuador about 80% of these groups are characterized by high levels of poverty and further lack access to basic services including energy supply. Located in remote areas that are difficult to access strengthen the problem of energy access facilitation in these regions. The project Sustainable Energy for Amazonian (SE4Amazonian) aims at bringing clean energy to six indigenous communities in the Ecuadorian Amazon.
Earth Observation supports the monitoring of environmental change in and around these communities. Satellite imagery depending on availability and quality of the images is collected for all study regions. Processing includes calculation of multi-temporal metrics such as tasseled cap components and vegetation indices with optical data such as Sentinel 2 and Landsat. Further water indices, backscatter coefficients and coherency is based on Sentinel 1. Due to high cloud coverage the study relies on SAR data derived from Sentinel 1 allowing analysis between 2015 until today. Long term analysis of the region is conducted via Landsat imagery observing changes between 1990-2018. Especially changes between 1990, 2001 and 2010 are of high relevance as these represent the collection periods for the Census Data from the Instituto nacional estadística y censos (INEC) which provides socio-economic information available from the Province to the Parrish level. Linking environmental and socio-economic change helps to understand the coupled social environmental dynamics in the selected regions. Further, data from the Global Urban Footprint (GUF) and the NASA VIIRS Night-time lights dataset support identification and analysis of areas with access to electricity.
Resulted mapping is used as input for the participatory mapping conducted during community workshops to identify needs and wishes for sustainable energy supply among the communities. Data collection during the community workshops is conducted in close collaboration with the indigenous population. No sensitive data is used without their clear agreement to fully secure their rights and to not invade their privacy now and in the future.
This study gives insights in the use of earth observation to support the Global Agenda. In combination with a participatory mapping approach it reaches out to those who are in need of clean energy. The extension of basic services should be facilitated to reduce poverty levels and improve livelihoods of indigenous communities in the Ecuadorian Amazon.

Area A - Board 117

Sustainable planning of affordable distributed energy installations considers many factors and requires draft but up-to-date analysis of the local communities potentially affected. It should provide a picture of the actual situation, its potential effect on local development, and with this can support a well-balanced improvement in line with the United Nations Millennium Development Goal agenda (i.e. SDG 7 - Affordable and Clean Energy).
The presented project “Integrated applications for micro grids in developing countries” (ESA 4000124252/18/UK/AD) reviews state of the art processes and addresses the potential of EO services towards supporting the site identification process within the Indian and African (selected countries) context. It will show, on how Earth Observation (EO) Services can support site identification and assessment by providing up-to-date information layer, such as urban extent, infrastructural and agricultural elements as well as information related to relief and potential hazards, among others. Overall aim is to provide ready-to-run cost-efficient, standardized solutions suitable for a scalable site identification process following different aspects, such as remoteness of a region, economic viability, vulnerability regarding natural hazards or other.
India and Sub-Saharan Africa
Electrification as well as availability of essential public services, such as health and education (SDG 3 and 4), can act as first steps towards realization of better life and sustainable development. However, fast population growth results in rapidly growing cities as well as widespread distribution of small settlements/ hamlets not connected to any main grid. The latter show disadvantage regarding their development potential towards higher sophisticated production methods.
Whereas India managed to connect 87.4% of its population to national grid power (2016), the situation especially within the Sub-Saharan Africa is far less convenient. Here, grid-connectivity was still well below 50% in 2016 [1]. Within the addressed project, both regions are taken into consideration, concentrating on identifying parallels and differences.
Micro grid developers and EO services
Essential element of the presented project is a comprehensive questioning of numerous micro/ mini grid developer regarding their needs and strategies. Within this process, EOservices are introduced and presented, based on Sentinel 1 and 2 data. The discussion leads to essential and case specific/ more complex EO information services, tailored by the feedback given by the users.
The site identification process underlies typical decision criteria’s which are often dependent on natural circumstances, and consider the developers intent. This will include parameter such as type of micro/mini grid (solar, wind, water, biomass or other), number of potential users, productivity and type of income, exposure to hazards (long and short term), and others. The better these elements can be described, the more precise the site identification process can be supported. EO data series of the Sentinel missions provide a suitable and very valuable base for highly automated medium scale applications. In combination with historic Landsat data, we can look back at the development of artificial structures and the natural environment at reasonable costs for the users in need.The site identification process gets applied following a multi-scale strategy. Draft site characterization gets done on national level, using open source information layer. Whereas this is suitable on national level, the fast development due to rapid population growth results in significant blurredness within regional scale, for data actuality becomes highly important. Here, Sentinel 1 and 2 information suits as highly suitable multi-temporal data source for more detailed analysis. On local level, just considering the area of a low number of villages/ limited extent, detailed information is also of interest to the users. However, here it became obvious that data costs (VHR satellite data) and the related services do not correspond to the requirements the developer has regarding cost and complexity of the results. Most of the missing information, identified within on-site investigations will not be solved by EO only and are therefore not of interest.
Dense interaction with micro-grid practitioner, providing valuable input regarding decision strategies and criteria, and is an essential element of identifying and forming the right services. Overall aim lays in providing EO information layer at a reasonable cost-benefit structure, appropriate for the users addressed (thematic and geometric accuracy, extent, spontaneous availability …).
The services proposed go beyond pure data analysis, but deliver advanced geo-spatial analysis as well as are implementation ready for the environment set-up by the developer itself. This includes technical support, maintenance of the service platform as well as efforts to expand the service to new geographies.
Impact assessment
At later stage, it is envisaged to implement EO data analysis to monitor newly electrified villages over several years in order to detect changes which can be attributed to the electrification effort, e.g. increased agricultural and economic activity, growth of settlements or the development of new public infrastructure could be performed using satellite imagery. This can establish a valuable feedback loop for development agencies and can thus lead to more impactful investments.
Present findings and actions towards financing
Global investments, such as of World Bank (WB) or the African Development Bank (AFDB) see micro grids playing bigger role in Africa’s electrification as solar costs fall. This means that this will “incorporate interconnected or stand-alone ‘mini grids’ and ‘microgrids’ serving small concentrations of electricity users, and off-grid home-scale systems,” WB states in its latest Africa’s Pulse publication, regarding the sub-Saharan Africa (SSA).
Micro grids represent an interesting middle ground between home-scale systems and the extension of national grids in accelerating electrification in many SSA countries. In several SSA countries, administration is already revising their power-sector regulations to reduce barriers to micro grid development, mainly in solar systems.
So far, the need for supporting the increase the number of well-designed and soundly operated private mini grid investments in sub-Saharan Africa is identified by World Bank and other development partners as very valuable in shedding light on the benefits they can provide and the costs incurred. Within the The Africa’s Pulse [2] report points out, that “over time, as incomes rise and populations agglomerate in higher-productivity locations, the national grid can spread out. In the meantime, household and community quality of life in off-grid areas can be substantially improved at relatively low cost through the provision of basic electricity services using solar home systems.”
Summary
Within the presentation, the project and some of its major outcomes will be addressed. EO service set-up is done in close reconciliation with the affected micro/ mini grid developers. Spatial modelling results, based on the multi-thematic and temporal EO information layer will be presented and its qualitative impact on the site identification process worked out.

Area A - Board 118

Nowadays South America is facing a gold rush due to rising gold prices in recent decades. Illegal extraction of gold has grown to be a great problem in many countries, causing degradation of the environment and general deterioration of living conditions. One of the biggest problems related to this crisis is a lack of information about the scale of damage and pollution caused by illegal and informal gold mining. This makes official processes of remediation, safety management and formalization of mine sites much more difficult. As the problem is growing and mostly occurs in remote areas, the usage of remote sensing for detection of illegal mine sites is attracting more attention.
Currently Peru produces more gold—both legal and illegal—than any other country in South America. In 2013, illegal gold mining was reported in 21 of the 26 provinces in the country. Moreover, in Madre de Dios Province, 90% of gold (up to 18 tons of gold per year) is mined illegally. In this region, miners have also entered the Tambopata National Reserve, an area containing globally unique diversity and many endemic species that are listed by the International Union for Conservation of Nature and Natural Resources.
Deforestation is the issue threatening the Amazon rainforest, with an estimated area loss of 1,680 km2 between 2001 and 2013. In September 2012 in Madre de Dios Region, deforestation reached the level of 50 thousand hectares. Huge damage is also done to soils and general morphology. Satellite images are probably the easiest way to obtain data relating to vegetation loss in remote areas such as the Amazon rainforest. Loss of tree cover can be seen on the raw unprocessed images, but to obtain accurate estimations the NDVI (Normalized Difference Vegetation Index) can be used. Research has shown disturbing changes in vegetation and damage to the Amazon forest.
Illegal and informal mining in this region involves the transport of large volumes of material (mostly gravel); this damages the natural vegetation and soils, thereby causing erosion and the reduced possibility of re-establishing natural plant coverage. Moreover, drainage of acids related to weathering of sulphide minerals like pyrite can lead to a rise in the acidity of water in surrounding areas. Analysing geology indexes can provide a variety of information that helps on a basic level to both understand the current situation at mine sites and predict the risk of pollution propagation.
The main purpose of this research is to investigate surface pollution and to detect its type and propagation in the environment. Development of a specific mining site has been observed and analysed with images obtained by LANDSAT and Sentinel missions. The results of the study showed clear changes in the state of the environment, revealing threats that could lead to serious contamination of water with heavy metals, thus endangering the health of local communities.
The results of research’s calculations can reveal the basic environmental conditions in any chosen part of the Amazon forest. The NDVI ratio can be helpful in many tasks; for example, when determining the boundaries of an illegal mine site and observing changes in its shape. Massive tree loss is a real threat to the Amazon forest and needs to be monitored and limited. The validity of remote sensing methods to observe the development of individual mining sites and their character was confirmed. All this information is enough to relatively quickly and with no additional cost determine which areas require immediate protection and which are at risk of pollution. Moreover, the collected information could help in the planning of more detailed field research.

Area A - Board 119

"The best development aid aims to eliminate the very need for aid" (Margaret Chan, WHO Director-General). It is necessary to strengthen the capacities of the countries of the South so that they can move towards autonomy. South-South cooperation aims to ensure that developing countries are the actors in the actions carried out and provide each other with mutual assistance through their knowledge, technical assistance and/or investment. This process is a departure from the traditional cooperation framework in which aid is provided in one direction only by donors to recipients. In this model, the emphasis is on capacity building, self-reliance and sustainable development in the South (WHO, 2018). We present here the establishment of a cooperation and development network with three partner countries (Burkina Faso, Ivory Coast and Rwanda). In order to meet the objectives of capacity building, autonomy and sustainable development of these countries, collaboration and training are built in the form of applied research prototypes in remote sensing and Earth observation, mainly based on Copernicus data and services. These prototypes are specifically dedicated to local needs, capitalizing on training and the results of previous scientific research (Nicolas et al., 2009) with the aim to end poverty, promote prosperity and people’s well-being while protecting the environment. This approach limits the need for field measurements and surveys, which are generally costly in terms of equipment and time. Finally, they offer a global vision where the loss of spatial resolution is often compensated by the higher temporal resolution of the data (Drusch et al., 2012).
The three prototypes were defined with the EO scientific community in order to ensure operational applicability:
1. Mapping and description of the location of the measurement sites of the future air quality monitoring network in Abidjan according to land use, relief and local climatic conditions;
2. A (semi-)automated methodology for characterizing water retention by remote sensing (determination of envelope curves) and mapping soil erosion risk at the level of the watershed (RUSLE method) (Hallot et al., 2017);
3. The development of mapping tools based on Sentinel-1 and Sentinel-2 data to support environmental management in the context of climate change in the high altitude equatorial zone (Kayiranga et al., 2016).
However, in addition to basic training in the acquisition, pre-processing and initial applications of these data, it is mainly at the level of data access and processing capacities that efforts must be undertaken. In most projects or prototypes developed by research centres or companies in European countries, data access via the Internet and bandwidth problems are often negligible. In addition, most students or course participants have personal computers, or the host structure is able to provide them with all the necessary equipment. Unfortunately, in most African countries, access to a sufficiently powerful and stable network to display, download or process online data is very limited or expensive. Moreover, the computer equipment available to students or in training structures is often not very powerful or defective. The objective has been also to create a “Copernicus-based training package for African countries” to build and train these users to specific Copernicus-based services, taking a particular attention to the technical challenges they are facing. The training are essentially based on open source tools developed by ESA (SNAP Software) or derived from OpenGIS and Open Street Map. The use of remote platforms such as the Collaborative Ground Segments or DIAS are also preferred depending on costs and the possibility of access with technical means and local Internet connections. The “training package” includes pocket PCs with preinstalled and preconfigured software’s and preloaded training data in order to reduce the implementation time and increase the possibilities of training locations in less well-equipped regions.
The further objectives are to reach other African countries using the training package created in the following years.

• Drusch, M., Del Bello, U., Carlier, S., Colin, O., Fernandez, V., Gascon, F. , Hoersch, B., Isola, C., Laberinti, P., Martimort, P., Meygret, A. , Spoto, F. , Sy, O., Marchese, F. & Bargellini, P., 2012. Sentinel-2: ESA's Optical High-Resolution Mission for GMES Operational Services, Remote Sensing of Environment, 120, 25-36. https://doi.org/10.1016/j.rse.2011.11.026.
• Hallot, E., Guyon, F., De Thysebaert, D., Diarra Galley, B. (2017). Gestion intégrée des petites retenues hydro-agricoles au Burkina Faso. Journée du développement. Arlon, Belgique.
• Kayiranga, A., Kur - ban, A., Ndayisaba, F., Nahayo, L., Kar a- mage, F., Ablekim, A., Li, H.W. and Ilniyaz, O., 2016. Monitoring Forest Cover Change and Fragmentation Using Remote Sensing and Landscape Metrics in Nyungwe - Kibira Park. Journal of Geoscience and Enviro n- ment Protection, 4, 13 -33. http://dx.doi.org/10.423 6/gep.2016.411003
• Nicolas, A., Radja, K. & Schembri, P., 2009. Quelles formations pour un développement soutenable dans les pays en développement ? Une approche par les compétences. Mondes en développement, 2009/3, 147, 29-44. 10.3917/med.147.0029.

Area A - Board 120

Avoiding, reducing and reversing land degradation and restoring degraded land is an urgent priority to protect the biodiversity and ecosystem services that are vital to life on Earth. The latest Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services (IPBES) Landmark Assessment Report highlighted that land degradation through human activities is undermining the well-being of at least 3.2 billion people, and is pushing the planet towards a sixth mass species extinction. Currently, more than 75 % of Earth’s land areas are substantially degraded. If this trend continues, more than 90 % of the Earth’s land areas could become degraded by 2050, potentially exacerbating climate change, loss of biodiversity and leading to mass migration, conflict and major food security concerns.
In order to halt and reverse the current trends in land degradation, there is an immediate need to enhance national capacities to undertake quantitative assessments and corresponding mapping of their degraded lands, as required by the Sustainable Development Goals (SDGs), in particular the SDG indicator 15.3.1 (“proportion of land that is degraded over total land area”) as well as by the adoption of Land Degradation Neutrality (LDN) targets under the auspices of the United Nations Convention to Combat Desertification (UNCCD). Indicator SDG 15.3.1 has been classified as a Tier 2 indicator requiring regular data production by countries to inform this indicator. Earth Observations (EO) can play an important role both for generating the indicator as well complementing or enhancing national official data sources.
We will present a workflow using EO data and various components of the Global Earth Observation System of Systems (GEOSS) platform to leverage EO resources for informing SDG 15.3.1. This workflow is developed in the frame of the GEOEssential project (http://www.geoessential.eu), an EU-funded H2020 project, that is aiming to address the need for trusted sources of data and information to monitor the progresses made on environmental conditions towards policy targets. Using various Essential Variables, cross-thematic workflows are created to evaluate, predict and monitor natural resources to inform via Earth Observations the Sustainable Development Goals.
The proposed workflow follows the Data-Information-Knowledge pattern using the Trends.Earth model (http://trends.earth) and various data sources to generate the indicator. It also implements additional components for model execution and orchestration (GEOEssential Virtual Laboratory), Knowledge management (GEOEssential Knowledge Base), and visualization (GEOEssential Dashboard). It is able to generate the SDG 15.3.1 indicator at different spatial and temporal scales using the GEOSS platform and the Google Earth Engine for data provision and analysis at global and regional scales; and the Swiss Data Cube (http://www.swissdatacube.ch) to generate medium to high spatio-temporal data necessary to generate the indicator at a country scale.
This workflow advances the vision of (1) the establishment of federated collaborative platforms with high computing capacities and big data analytics tools (e.g., EO data cubes) that would allow countries to easily select, access, process, analyze, interpret and quality control large datasets associated with EO and geospatial information; and (2) building capacity at the country level and ensuring national ownership of EO data and information on the use of practical tools to support countries in accessing, interpreting and validating EO data.

Area A - Board 121

Straddling the countries of Perú and Bolivia, the Andean Altiplano is a high elevation plateau, enclosed by two branches of the Andean Mountain range. Two large lakes are located in the Altiplano: Lake Titicaca, in requirethe Northwest, and Lake Poopó, in the Southeast.

Lake Titicaca overflows into the Desaguadero River, which ends at Lake Poopó. The latter is an endorehic shallow lake, receiving approximately 65% of its inputs from the Desaguadero River (Pillco and Bengtsson, 2006). For decades, the extent of Lake Poopó closely mimicked the catchment precipitation patterns (Satgé et al., 2017). However, in the last few years, the lake has experienced an unprecedented rapid shrinkage, defying the previously observed correlation with precipitation. This shrinkage had devastating effects for the autochthonous fauna and for the local communities, which lived on fisheries and livestock for millennia.

Several factors have been postulated as possible causes for the lake reduction: increased evaporation due to climate warming, expansion of crops in the catchment, droughts induced by El Niño event, aquifer depletion, etc. The understanding of the driving forces behind Lake Poopó’s drying is compelling to address its restauration and sustainability.

This study investigated changes in the catchment of the Lake Poopó in the period 2000-2017 that could be linked to its shrinkage, within the frame of the project Integrated Water Resources Management in the Titicaca-Desaguadero-Poopo-Salar de Coipasa (TDPS) System, funded by the United Nations Development Program.

The study area includes the catchment draining into the Desaguadero River, from its upstream end on the southeast bank of Lake Titicaca, to the lowest terrain elevation area in the Salar Coipasa. The basin extends over 103.878 km2 and is known as the Desaguadero-Poopó-Salares (DPS) system. NASA´s Moderate Resolution Imaging Spectroradiometer (MODIS) Normalised Difference Vegetation Index (NDVI) and evapotranspiration (ET) products were used to reconstruct vegetation and evapotranspiration dynamics in the catchment for the period 2000-2017. Rainfall data were acquired from local meteorological stations and from the Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks (PERSIANN) data set. The abovementioned products were accessed and processed through the Google Earth Engine platform. Additionally, the data collection included temporal series of: flow discharge of the main water courses, surface elevation and extent, and water volume in Lakes Titicaca and Poopó, water surface temperature in Lake Poopó, gravity maps derived from the Gravity Recovery and Climate Experiment’s (GRACE) measurements, and soil moisture maps obtained from the ESA Climate Change Initiative portal.

Site visits of the study area were conducted, and high spatial resolution, multi-temporal images from Google Earth were visually examined to relate the land cover type and land use to the observed trends. A preliminary analysis examined the rainfall and NDVI annual patterns over the DPS system. No significant change in the spatio-temporal distribution of the rainfall was detected. The greenness of autochthonous flora showed a tight positive correlation with precipitation accumulated for the two months preceding the NDVI measurement. This result points to soil moisture as the main limiting factor for the photosynthetic activity, as opposed to temperature, light or other environmental factors. The relationship between NDVI and precipitation was significantly weaker over wetland and agriculture areas. Most wetlands in the study area are groundwater or river flow fed, which partially decouples their response from the local precipitation. In the case of the crops, anthropogenic actions, such as irrigation, and fertilization would presumably explain the decoupling from rainfall.

A linear trend was interpolated for the NDVI and ET MODIS products on pixel basis, for the entire study period and for specific seasons within this period. The gradient was mapped, revealing the spatial distribution of different temporal trends (Fig. 1). A clear, statistically significant NDVI and ET increasing trend was observed in the central and eastern part of the DPS system for the whole study period. When analyzed per season, virtually no temporal trends surfaced during the dry season, from June to August, while all the increments occurred in the wet period, from December to March.

The analyses per land cover type revealed that in the wet season, the ET temporal gradient was slightly negative or null over the western mountain ridge and slightly positive for the autochthonous vegetation on the plateau. The highest ET gradient was observed on the central and eastern agriculture areas of the DPS system. A statistical analysis corroborated the significance of this trend. In the dry season, the ET temporal gradient was approximately flat over most of the study area, and only exhibited slightly negative values at the northwest end of the system. Regarding the NDVI, significant increasing trends were only observed over agricultural areas in the wet season, as for the ET losses.

This study evidenced a significant increase in vegetation and associated ET losses in the extensive croplands of the central DPS system during the wet seasons between 2000 and 2017. These results strongly suggest that crop practices may have exacerbated the water losses in the system, in detriment of the final recipient, which is the Lake Poopó. However, the observed increase in ET over autochthonous vegetation in the plateau, where NDVI and precipitation values did not augment, suggests that environmental factors, such as temperature, may have also played a role. These factors and their links to Lake Poopó’s water storage are currently under investigation. Data such as gravity measurements and water storage in Lake Titicaca and Poopó, are being used to approximate the DPS water balance throughout the study period.


References:
Zola, R. P., & Bengtsson, L. (2006). Long-term and extreme water level variations of the shallow Lake Poopó, Bolivia. Hydrological Sciences Journal, 51(1), 98-114.
Satgé, F., Espinoza, R., Zolá, R. P., Roig, H., Timouk, F., Molina, J., & Bonnet, M. P. (2017). Role of climate variability and human activity on Poopo Lake droughts between 1990 and 2015 assessed using remote sensing data. Remote Sensing, 9(3), 218.

Area A - Board 122

Land degradation is a complex issue which needs an interdisciplinary approach to determine its spatial extent considering the effects of anthropogenic activities and weather conditions from the last decade. The impact of the different degradation processes vary across Europe. Southern European countries are affected by the most severe water erosion linked to the extreme and intermittent rainfall whiles in the Northern Europe losses are moderate. The lowlands from the Western Europe wherein the agricultural production is most intensive have been stressed by the diffuse contamination. On the other side, Eastern Europe is characterized by a rapid intensification of agriculture with over exploitation of the most fertile soils and abandonment of less productive lands.
Depending on the degradation degree, European countries were encouraged to adopt measures tailored to the local needs through the elaboration of programs able to reduce the risks, inventorying contaminated sites and establishing national strategies for its remediation. In this perspective, Romania has aligned with the United Nation Convention to Combat Desertification (UNCCD) and has adopted a long term action program (2001-2020) [1]. The measures foreseen in this program contribute to reduction of the land degradation processes associated with severe erosion problems (water erosion, wind erosion, crusting and salinization) that affect most of the Romanian arable areas [2], [3]. Another soil problem has arisen from slope of the arable areas which favor the erosion processes by creating optimum conditions for the concentration of surface runoff and increasesof the kinetic energy in soil [4], [5]. Moreover, land use changes and uses of heavy machinery in the intensive agricultural areas, poorly correlated with the need and capabilities of the environment, contribute to the soil degradation [6].
The study focuses on assessing spatial extent of the land degradation from Banat region, Romania. The test area is located in the South-West of Romania bordered by the Mures River in the north, the confluence between Danube and Cerna rivers in the south, Romanian – Serbian border in the west and the Jiu River in the east. From the morphologically point of view, this plain is a flat relief unit, uniform appearance but complex in lithology and soil, flat surfaces are frequently separated by abandoned meanders. The climate is moderately continental temperate with Mediterranean influence that is manifested by weak mild winters and hot summers, with an average annual temperature by 10.8 °C and a mean yearly rainfall of 550 mm.
Based on multi-temporal Sentinel 2 data, the support vector machine (SVM) classification algorithm was applied to retrieved land use/ land cover classes. Soil adjust vegetation index contributes in delimitation of the soil / vegetation while Sentinel 1 SLC polarized data offers information about soil degradation in terms of erosion or crusting. In this sense, differential interferometric technique was applied on Sentinel 1 SLC, VV polarized data. Sentinel 1 and Sentinel 2 dataset covers September 2016 – September 2018 period, highlighting the land degradation during two years of consecutive monitoring.

References:
[1] Ministry of Water, Forest and Environmental Protection, "National Report on the Implementation of the UNCCD in Romania," Bucharest, (2000).
[2] Motoc, M., Ionita, I., Nistor, D. and Vatau, A., "Soil erosion control in Romania. State of the art," Soil Erosion Prevention and Remediation Workshop, US Control and Eastern European Agro-Environmental Program, Budapest, 111-133 (1992).
[3] Nistor, D. and Ionita, I., "Development of soil erosion control in Romania," Man and Soil at the Third Milenium, vol. 1, Rubio Jl, Morgan RPC, Asins S, Andreu V (ed), GeofarmaEdiciones, Logrono, 299-309 (2002)
[4] Bucur, D., Jitareanu, G and Ailincai, C., "Soil Erosion Control on Arable Lands from North East Romania," Soil Erosion Issues in Agriculture, Dr. Danilo Godone (Ed.), InTech, 295-314 (2011).
[5] Nistor, D. and Nistor, D., "Land Degradation by Erosion and Its Control in Romania," Proc. 12th ISCO Conference, 621-626, Beijing (2002).
[6] European Commission, "Study on Soil and Water in a changing environment," Final Report, 271 p (2014). [9] Copacean, L. and Oncia, S., "Estimate soil loss through surface erosion using Geographical Information Systems and satellite images," Research Journal of Agricultural Science, 47 (4), 47-54 (2015).

Area A - Board 123

The Paris agreement to avert catastrophic climate change and limit global warming to below 1.5 °C compared to pre-industrial levels poses a big challenge to governments and societies around the world. At the same time, growing global demand for agricultural products results in agricultural expansion at the expense of environmental frontiers, and such demand is likely to increase in the future. Agricultural land abandonment is widespread globally and has various implications for biodiversity, global biogeochemical cycles and societal well-being. Over the time though abandoned lands can be encroached by young forests, thus providing valuable areas for restoration and defragmentation of existing forests, harboring umbrella species but also alternative land use (e.g., agroforestry). European part of former Soviet Union (EP of fSU) is a prime example of widespread agricultural land abandonment after the collapse of the Soviet Union. Statistics and exemplary studies suggest 20 to 40 Mha of abandoned croplands in EP of fSU alone. However, it is unclear the actual patterns of abandonment and how much abandoned lands were afforested after 1990.
Here using different public available datasets such as forest and cropland masks and wall-to-wall Landsat-based three-year time-series of Normalized Difference Water Index (NDWI) for European Russia using an expert defined thresholds for NDWI we reconstruct ongoing afforestation across EP of fSU at 30-m resolution by 2017. Dataset includes three classes: active arable (croplands), tree-less arable abandoned lands (3 years without croplands), fallow lands (forest or wooded land on former arable or tree-less lands). Area of active arable is about 29Mha and area of fallow lands is about 17Mha or 40% of all abandoned lands (44Mha).
We implemented a partial accuracy assessment of our map based on random sampling and field visits (area 1.8 Mha) in three European Russia regions. Overall accuracy for validated part was 90 ± 1.5%. Area of abandoned lands is about 387 thousand ha (20%), include 227.3± 24 thousand ha (12%) arable abandoned and 159.7± 21 thousand ha (8%) fallow (forested) lands. Area of active croplands is 91± 14 thousand ha or 5.4 % of all area. Our results shows great potentials for afforestation, especially in European part of Russia, but now forests growing on abandoned lands don’t have any legal status in Russia. These lands typically, are not used for agriculture due to lack of resources. More promising is to use them for sustainable forestry that could help to solve problem in covering demands for timber and decrease, minimize or even eliminate logging pressure on natural forests, including intact forest landscapes.
Data is available online https://peatfires.nextgis.com/resource/175/display?panel=info.

Area A - Board 124

The reindeer husbandry is a traditional land-using form of northern part of European Russia. Pastures of European North-East of Russia cover near 17681.0 th. ha of Nenets District and 9482.7 th. ha of Komi Republic (near 72.8% and 22.8% of total area, respectively). The reindeer population density at husbandries varies from 0.4 to 2.3 (average for all region is 1.2) animal per km2. In present time we developed and actively use the technology of reindeer pastures inventory at on-line regime by optical satellite data (Landsat and Sentinel) (patent of Russian Federation №2521755). Results were verified for individual husbandries. The obtained results of satellite inventory are represented by modern GIS-system related with attributive tables of land-using orientation. The primary results of processing are the botanical and landscape maps and quantitative characteristics of vegetation. The attribute data of vegetation cover were calculated using the principles of SMA or linear regression methods on basis of field research. The outcome results of economic estimation (reindeer capacity, number of pasture days, etc.) was calculate through the previously established and regionally adapted coefficients. Results of inventory take to account the seasonal and inter year’s variability of pastures conditions. We note that the initial stages of phytocenosis climate modifications at territory of European Russia fixed for the period 2000-2009 after growth of the surface air temperature in 2000. The materials obtained form spatially organized database includes polygonal objects (vector layers of reindeer husbandry units) and attribute data related through personal identifier. The received materials are actively required and used by reindeer herder’s leaders for calculation of industrial damages related with Timan-Pechora carbohydrate province development (organization of corridors of landcars movement and pipeline services, excess of land area).
The vegetation of Timan tundra (near 164.7 thousand hectares) was most changeable region of the European North-East of Russia The analysis of vegetation cover changes on the territory was carried out using Landsat, Sentinel and MODIS satellite images of 1985-2017 years and field observation results of 2017-2018. The Northern Timan highlands (about 164.7) are characterized by a highly intensive transformation of phytocenoses in the region. The vegetation index changes reflect growth potentials for traditional natural resource management. There was an increase in the reindeer-feeding capacity by green foodstuffs from 1985 to 2009 by 29% and by 15% for adjacent areas. After 2007 changes became stabilized. The NDVI interannual change trends were confirmed by combination of mean-resolution MODIS and high-resolution Landsat 5 and 7 images for same year intervals. The values of NDVI change calculated linear trends (β) based on survey materials of different sensors had equivalent dimensions with a significant positive correlation among them (r = 0.72, p≤0.05). The method of converting NDVI values into chlorophyll index values and aboveground green phytomass parameters by multi-seasonal images was developed. There was an increase of chlorophyll reserves in phytocenoses of the study Timan region (by 152 tons) from 2000 to 2009. Expressed as dry green phytomass of low shrubs (Betula, Arctous, and Empetrum), that increase made 1.85 centers per hectare. The maximum changes were observed in communities of low shrub lichen tundra (chlorophyll content growth by 1.26 kg/ha, dry green phytomass by 2.51 centers/ha). The main factors responsible for the changes in plant communities are surface air temperature rise since 2000 in autumn-winter period and precipitation amount in winter period. In addition, significant reduction of reindeers (two-fold decrease) could favor vegetation production in the region.
The work was financed by comprehensive program of fundamental research of Ural Department of RAS 2018-2020. Subprogram Arctic. Project № 18-9-4-5.

Figure captions:
The location of Timan tundra region and SRTM DEM (A). The field observation sites, border of reindeer husbandry and helicopter observations route are noted. (A). Trends of NDVI changes per 2000-2009 years (β) on basis MODIS interannual (MOD13Q1.005) data processing (B). The part of vegetation map (classification of Landsat 2013.07.02, 2014.07.19 and 2016.09.19) (C). Dominated classes are presented: 1.1 – low shrub mossy tundra, 1.2. - low shrub lichen tundra, 2.1 – shrub mossy, 3.1- yernick mossy undra, 3.2 – salix grass, 4 – wetland complexes, 5- rocks, 6 – water.

Area A - Board 125

Mongolia is a large country (over 6 times the size of the UK) with approximately 30% of its population dependent on livestock herding who are exposed to extreme weather events, known as dzuds, which are increasingly exacerbated by climate change and are highly damaging to Mongolia’s economy and devastating for the poorest herders. A typical dzud can impact tens of thousands of herders many of whom can lose all their livestock leaving them in extreme poverty, with associated impacts for the wider economy.
The SIBELIUs project, supported by the UKSA’s International Partnership Programme (IPP), will deliver greater dzud-resilience for herders by providing Mongolia’s National Agency for Meteorology and Environmental Monitoring (NAMEM) with improved capacity for distributing new and upgraded environmental products to key stakeholders who are supporting the herding communities. Products contain information regarding pasture, snow and drought. Improved products will provide complete-country coverage and will be provided more frequently than is currently possible, for longer periods of the year, more accurately and with improved spatial resolution.
SIBELIUs will channel its benefits through three existing networks: (1) the insurance sector, (2) government agencies coordinating livestock and land use, (3) directly to the herders through text messaging and television broadcasts.
A Mongolian Data Cube will form a key component of the SIBELIUs infrastructure which will allow time series of satellite data and derived products to be queried by staff at NAMEM. Output products from the Data Cube will be ingested by desktop applications, web apps and dashboard front ends for stakeholders to utilise further. Initially, the Mongolian Data Cube will cover six county sized regions of Mongolia, before expanding to cover the whole of Mongolia before the project’s end in 2021.
As a vital component of the project, SIBELIUs will work with herders at selected case study sites to analyse their information requirements, to better understand barriers to uptake of previous products, and to ensure their voices and priorities are heard in the development and distribution of new satellite-based environmental products.
The project’s impact is directly related to SDG1.5, aiming to build the resilience of those in vulnerable situations and reduce their exposure and vulnerability to climate-related extreme events and other economic, social and environmental shocks and disaster. The project’s impact will be measured by the economic benefits achieved and by the number of herders with increased dzud protection. It will also consider the latest research methods for addressing resilience in nomadic pastoralists, recognising that there is no single metric for, or meaning of, ‘resilience’ and it is therefore essential that metrics are developed in a participatory way with local communities and to reflect local contexts.

Area A - Board 126

D-MOSS, Dengue MOsquito Simulation from Satellites, is a dengue fever early warning system for Vietnam being developed in a project funded by the UK Space Agency’s International Partnerships Programme. The D-MOSS project is developing a suite of innovative tools that will allow public health authorities to identify areas of high risk for disease epidemics before an outbreak occurs, in order to target resources to reduce spreading of epidemics and improve disease control. Beneficiaries will be enabled to issue alerts for dengue fever and provide assessments of vector-borne disease risk under future climate and land-use change scenarios.

Before 1970 only nine countries had experienced severe dengue epidemics. Today the disease is endemic in 141 countries, affecting 390 million people and with a global annual cost estimated at almost US$9 billion, which is three times that of cholera and over four times that of gastroenteritis. Since 2000, there has been an increase of over 100% in the number of cases of dengue fever in Vietnam, with ~185,000 cases occurring in 2017 alone, and there is currently no system for forecasting future dengue outbreaks.

One of the major tasks identified in Vietnam’s five-year health sector development plan (2016-2020), issued by the Ministry of Health, is “To actively prevent epidemics; to forecast, detect early and prevent the occurrence of epidemics, especially major ones”. According to a 2016 World Health Organisation report on Dengue Prevention and Control, although significant progress has been made in strengthening regional and country capacities to detect, assess, report and respond to dengue, the Asia-Pacific Region has not seen the degree of success expected. Limited tools and resources have made it difficult to sustain efforts to control dengue and its burden continues to increase. According to the same report, early adoption of new tools, investing and undertaking the development of new methods and making better use of available interventions will contribute to mitigating the challenge of dengue. The D-MOSS project will help the Vietnamese Government to achieve the following targets of SDG3 - Good Health and Wellbeing: Ending the epidemics of AIDS, tuberculosis, malaria and neglected tropical diseases and combat hepatitis, water-borne diseases and other communicable diseases by 2030; Strengthening the capacity of all countries, in particular developing countries, for early warning, risk reduction and management of national and global health risks.

The D-MOSS early warning platform includes a water availability component. Water availability directly impacts dengue epidemics due to the provision of mosquito breeding sites. These dynamics are often non-linear; too much rainfall can fill outdoor containers, while too little can lead to people storing water in open containers within their homes. Both increase the population of Aedes aegypti mosquitoes and in turn the risk of dengue outbreaks. However, water availability or water resource management is rarely accounted for in dengue prediction models. This project generates monthly water stress assessments and uses them as inputs to a component of the dengue early warning system which also improves the skill of the system’s predictions. In addition, these forecasts of water stress will help to improve Vietnam’s water management. Vietnam’s Sustainable Development Strategy for 2011-2020 identifies one of the major challenges facing Vietnam as the issue of transboundary water management, because 63% of the surface water comes from upstream countries. The D-MOSS early warning system includes a water assessment module that will assist the Vietnamese Ministry of Natural Resources and Environment to access Earth Observation (EO) - based information on water availability and improve water management in Vietnam’s transboundary river basins where there is a paucity of hydro-meteorological information. This secondary benefit, arising from the project, will help Vietnam to meet SDG6 - Clean Water and Sanitation by helping to implement integrated water resources management at all levels, including through transboundary cooperation as appropriate by 2030.

The D-MOSS project is developing a forecasting system in which EO datasets are combined with weather forecasts and a hydrological model to predict the likelihood of future dengue epidemics up to eight months in advance. D-MOSS integrates multiple stressors such as water availability, land-cover, precipitation and temperature. The approach uses a common spatio-temporal analysis ‘grid’ with a ‘Polygon Series’ structure to integrate historical stressor datasets with each other and with historic dengue fever incidents, which are then input into a statistical model to provide forecasts based on future seasonal weather and hydrological forecasts.

UK Met Office seasonal forecasts are used to drive the VIC hydrological model, whose outputs are combined with EO datasets and translated to indicators of water availability such as commonly used indicators of drought (e.g. Standardized Precipitation Index). The water availability system includes three EO-based components (land-use, weather and water resources) and generates monthly forecasts of water stress. The system is calibrated against historical data. The water availability forecasts are subsequently fed into statistical forecasting models of disease incidence, based on a spatio-temporal Bayesian hierarchical mixed-modelling approach. This dengue early warning system model integrates the water stress forecast with a range of other covariates important for dengue transmission to forecasts of dengue incidence, up to eight months in advance.

The D-MOSS architecture relies on open and non-proprietary software, where possible, and on flexible deployment into platforms including cloud-based virtual storage and application processing.

The tools produced will also be used to increase the understanding of climate change-related health risks during a period when Vietnam is developing an updated National Adaptation Plan in line with its Paris Agreement obligations.
The D-MOSS project is within the first year of its three-year term and is currently focused on platform and model development, while gathering the key input data and engaging with the Vietnamese government to ensure that all components are fit for purpose. The portrayal system is designed to communicate the dengue and water availability forecasts to the Vietnamese Ministries of Health and Natural Resources and Environment, respectively. A user interface will also incorporate supporting information on recommended actions, provided by the decision makers and based on the forecasts and associated uncertainty.

The D-MOSS project is led by HR Wallingford, working with the London School of Hygiene and Tropical Medicine, the Met Office and Oxford Policy Management in the UK, and with the following international partners: the United Nations Development Programme, the World Health Organisation, the Vietnamese Institute of Meteorology, Hydrology and Climate Change, the Pasteur Institute Ho Chi Minh City, and the National Institute of Hygiene and Epidemiology in Vietnam.

Area A - Board 127