The Food Security Thematic Exploitation Platform is the youngest out of seven TEPs and is developed in an agile mode in close coordination with its users. It provides a platform for the extraction of information from EO data for services in the food security sector mainly in Europe & Africa, allowing both access to EO data and processing of these data sets. Thereby it targets to foster smart, data-intensive agricultural and aquacultural applications in the scientific, private and public domain. The platform has been open to the public since March 2018 and is currently in its second phase of development.
The FS-TEP builds on a large and heterogeneous user community, spanning from agriculture to aquaculture, from small-scale farmers to agricultural industry, from public science to app developers to the finance and insurance sectors, from local and national administration to international agencies. Service pilots demonstrate the platform’s ability to support agriculture and aquaculture with tailored EO based information services. The main point of access to the Food Security TEP is the Open Expert Interface providing the main functionalities of the platform and access to a variety of EO data sets as well as supplemental data. Furthermore, FS-TEP allows the mobile visualization of crop parameters and the provision of customized products and services to selected users.
The presentation will focus on the second out of three service pilots that brings together a local African company that provides micro credits to farmers with the Global Smart Farm concept using the FS-TEP. FarmDrive is a local company in Kenya that has been founded with the explicit goal to leverage technology to enable smallholder farmers in Africa to achieve self-sustenance. Since data on small holder farmers is sparse and difficult to collect, Earth Observation helps to reduce the cost of collecting data on wide geographic areas. In this pilot, we use EO data to monitor the risk of the loan portfolio at regional scale in North-Western Kenya.
Farm Drive’s mobile services expand access to credit to underserved smallholder farmers by providing modern tech-driven credit assessment tools. They build innovative comprehensive credit profiles used for real-time credit assessment to then serve creditworthy smallholder farmers through digital financial services. FarmDrive uses statistical models for credit assessment and behavioural scorecards for limit enhancement and reduction. Presently, their mobile loan service does not make use of EO data.
However, in order to access the risk and performance of agricultural loans, crop and soil monitoring throughout the crop growth cycle could be crucial. Hence, it is the aim of the ongoing pilot to show that the continuous monitoring of agricultural crops and soils at the regional scale can be used for an objective, near-real time assessment of crop development for risk assessment (e.g. are crops developing well, how does the actual growth period compare to past periods, what is the risk of crop failure?). The results of this pilot application shall demonstrate that time series of Sentinel data products, weather data and predictive models can assist in managing FarmDrive‘s portfolio risk.
The project team developing the Food Security TEP and implementing pilot services during a 30 months period (started in April 2017) is led by Vista GmbH, Germany, supported by CGI Italy, VITO, Belgium, and Hatfield Consultants, Canada. It is funded by ESA under contract number 4000120074/17/I-EF.

A cropland monitoring system that delivers timely and accurate information is critical for several national and international programmes. In Europe, the Common Agricultural Policy requests such kind of information. At global-scale, the Sustainable Development Goals (e.g., #2: Zero Hunger) require also reliable information on crops. Users of such information can be found in public institutions, but also in farmer cooperatives, and other economic fields such as insurance and trading.
The advent of the Copernicus Sentinel satellites provides an unprecedented amount of Earth Observation (EO) data. Concomitantly, the Open Data Policy (ODP) adopted in several European countries enables access to accurate information on recent crop cycles for several countries. This synergy between dense time series of EO data and accurate national crop information opens avenues to test novel methods to progress towards more accurate and timely crop information. CS SI proposes the EarthSignature AI project to test Deep Learning methods that can improve the identification of crops within farmlands, in a more accurate and timely fashion. We will present the first results from different Deep Learning techniques that will take into account the spectral, spatial and temporal dimensions of dense time-series of Sentinel-2 datasets located over three sites in Europe: southern France, Belgium, and The Netherlands. These results will be compared to those obtained with a baseline method such as Random Forests.
A goal of this project is to be able to process efficiently large volume of data, to provide added-value information in near real-time, when it is the most valuable. Deep Learning processing is well suited for parallelization, and typically takes advantage of the computing capabilities of GPUs. Rapid processing of large volume of data will benefit from Cloud computing with GPU capabilities, and CS Systèmes d’Information has developed a platform that can run such processing on any of the Copernicus DIAS, transparently, and is already used operationally to support the Copernicus Research and User Support (RUS) service platform.
This project is supported by ESA (ITT reference: ESA/AO/1-9101/17/I-NB), as part of the Open Call framework.

The volume and variety of Earth Observation (EO) data available for the polar regions is growing rapidly, providing opportunities for more complex applications by the polar research and operations communities. This trend is forecast to continue given plans for new polar focused instruments, including a potential Sentinel expansion mission.

EO has a unique and important role in the polar regions as the only reliable and consistent source of regular, calibrated regional scale data. Collecting year round data in the remote, vast and harsh conditions of the Arctic and Antarctic would be logistically and financially impossible using surface, airborne and ship instruments.

Scientific studies are increasingly using user larger volumes of multi-sensor data, and safety critical operations require speedy delivery of up to date information requiring efficient processing of data.

This situation is not limited to the polar regions and other domains face the same challenges posed by increasing data volume and how to fully process, analyse, disseminate and exploit them. As a result, numerous cloud-platforms have emerged to provide online work environments collocating data, processing capabilities, development tools and ICT resources. Among these are the Exploitation Platforms established by the European Space Agency, including the Polar Thematic Exploitation Platform (PTEP).

This presentation will describe the current status of the Polar TEP. It will cover the rich set of polar EO datasets, toolboxes and processing capabilities, plus functionality to allow deployment of user defined workflows and processing environments. We will also outline new PTEP features which allow easier processing of large data volumes and automated scheduled data processing, both using the European DIAS data platforms. We will also describe current scientific and operational projects supported by PTEP. This includes the original pilot project which demonstrates the use of Polar TEP to investigate current and future iceberg risk in Baffin Bay using a diverse set of data, processors and models to investigate linkages between iceberg trajectories, changes in ice sheet velocity and glacier calving rates.

In parallel to PTEP, a growing number of similar generic or domain focused platforms have been sponsored by numerous entities, creating a crowded and potentially confusing situation for the user community. The technical features and capabilities of most platforms are broadly similar. This raises a significant challenge of how a domain specific platform distinguishes itself and sustains operations when there is an apparent abundance of competition.

We will present options for how to distinguish the Polar TEP and make the available data and functionality targeted to the polar community. We will also propose ideas about how common open source development of exploitation platforms may lead to easier maintenance and sustainability of similar niche exploitation platforms.

The PAVICS (Power Analytics for Visualization of Climate Science) platform is a collection of climate analysis services served through Open Geospatial Consortium (OGC) protocols. These services include data access, processing and visualization. Both data and algorithms can be accessed either programmatically, through OGC-compliant clients such as QGIS or ArcGIS, or a custom web interface. The platform is entirely modular and consists of a number of servers that can be launched independently or together. External servers offering data or additional services can easily be connected to the platform, enabling the creation of loosely coupled federations of data and service providers. What makes these generic interconnections possible is the compliance with OGC standards, a common file format among climate scientists (NetCDF) and strict conventions regarding metadata (CF-Convention).

The current version of PAVICS offers the following features to support the work of climate researchers: computation of close to 50 different climate indices, a process for dataset regridding, a process for bias-correction, spatial and temporal subsetting processes (where region contours can be provided by shapefiles), calculation and visualization of spatial analogue regions from climate datasets and publication of analysis results. There are also processes to query and analyze remote sensing imagery, work with biodiversity datasets and another in development to analyze large-scale weather patterns.

An ongoing project is currently aiming to integrate hydrological modelling and prediction tools in PAVICS to better understand the impacts of climate change on water resources availability in Canada and in the United-States, and eventually on a global scale. Research groups that require such information but might not have the required expertise to perform hydroclimatic analyses, such as in ecology, biology and forestry can leverage the tools undergoing development.

Currently, performing even simple hydroclimatic simulations is a daunting task and requires the following steps to be completed:

- collecting observed climate and streamflow data;
- calibrating a hydrological model on these data;
- gathering climate model outputs and post-processing them;
- using the hydrological model to simulate future climate-driven streamflow.

Each of these steps is labor-intensive, some act as a barrier to hydrologists wanting to use climate data to run their models, while others as a barrier to climatologist wanting to better understand the impacts of climate processes on hydrology. The PAVICS framework simplifies the process of accessing gridded observations and climate model simulation data, but there would be benefits for hydrological researchers to perform hydrological model calibration and simulations within PAVICS.

The proposed improvements to PAVICS would be very valuable to scientists, engineers and researchers in a myriad of disciplines. An example of a related field of research that would be interested in the PAVICS - hydroclimatology platform is related to an area of current work, which is the modeling of river temperature, because the projected impacts of climate change would be to increase river temperature. This has obvious impacts on not only for fish habitat but also water quality in general. For the energy sector, water from rivers is also used to cool equipment, and there are regulatory restrictions on the temperature of water returned to the environment. Prolonged heatwaves can thus raise river temperature enough to force power station curtailments. The current project will set the groundwork to embed water temperature models that can be driven by future climate scenarios to facilitate the evaluation of those types of impacts.

Other examples of uses for the PAVICS platform with hydrology tools include:
- Water resources management: Irrigation for agriculture, hydropower generation, flood control, drinking water supply.
- Ecology: Determining the risks posed to fish and aquatic species based on water shortage risk and ultimately on water temperatures.
- Environmental impacts from mining operations under modified hydroclimate,
- Biodiversity: Migration patterns, ecosystem functioning and transformation.
- Dendrochronology
- Northern studies (Arctic research)
The PAVICS platform is intended to be a scalable resource making use of Compute Canada’s computational network to parallelize and distribute workload when the need arises, and for this transition to be completely transparent to the user. Workflow management using JupyterLab instances will also make it easier for researchers to save, share and reuse previous workflows to support their research. PAVICS has proven to be a powerful and useful tool for climate data management, and will continue to be scaled to include other sources of data, namely from remote sensing and satellite imagery, in the near future.

With Canada’s hydrographic conditions being highly variable and increasingly affected by impacts of climate change, populations residing in flood-prone areas continue to experience more risk of being affected by severe floods. A rich time series of satellite imagery exists in the archives from a variety of sensors acquired across major river basins where larger numbers of inhabitants are exposed to potential risks of impacts from severe flooding.

A methodology is being developed that can process large volumes of multi-sensor imagery to generate Analysis Ready Data (ARD) from these rich image archives (both satellite optical and SAR imagery) available across Canada. Deriving risk assessments in near real time by leveraging historical imagery will assist emergency management officials in making best use of limited resources where they are most needed. It is within this context that the Canada Centre for Mapping and Earth Observation (CCMEO) is currently working with PCI Geomatics to develop a Big Geospatial Data Analysis framework (i.e. Data Cube) for the thematic classification of a variety of earth-observation imagery (Landsat, Sentinel 1 and 2, RADARSAT-2). The system is implemented using the CEOS Open Data Cube open source code repository and leverages existing algorithms for mapping water extents (Water Observation from Space -WOfS) in addition to custom algorithms developed at CCMEO that leverage machine learning algorithms such as See5 and Support Vector Machine (SVM).

This prototype system will enable a time-series analysis of earth-observation data over the Saint John River Basin located in eastern Canada to provide an assessment of historical water observation locations and frequency which can be used to generate near real time surface water extents to inform operational decision-making. This presentation will discuss the solutions under development with a focus on the creation of at scale multi-sensor Analysis Ready Data and time series analyses to detect changes in both historical and near real time surface water extents.

In the framework of the French research infrastructure “Pôle Système Terre” under construction, the Solid Earth data and services centre named ForM@Ter develops an on-demand service for Sentinel-1 InSAR processing for researchers.
ForM@Ter was launched in 2012 by the French national space agency (CNES) and the National Centre for Scientific Research (CNRS). Its perimeter, designed with the scientific community, focus on the determination of the shape and movements of the Earth surface. It aims to facilitate data access, provide processing tools and value-added products with support for non-expert users.
The first target focuses on surface deformation from SAR data. The associated services are implemented considering the needs expressed by the French scientific community to support the use of the massive amount of data provides by radar satellites of the European Sentinel-1 mission. In fact, with a worldwide systematic background acquisition strategy over tectonically active regions and a 6-days revisit time, the Sentinel-1 mission offers new horizons for a wider use of SAR interferometry to measures ground deformations. However, the new default mode of acquisition (TOPSAR) and the amount of data available in near real time, require new processing schemes, and significant computing and storage facilities not available to every researcher.
With the service proposed by ForM@Ter, that will be presented here, users can launch a computation of single interferogram (both wrapped and unwrapped version are available geocoded and in radar geometry) or a network of coregistred interferograms. In a second phase, it aims to propose computing its associated time-series of unwrapped phase.
The user interacts with a web application that queries the web service, which in turn submits jobs to a dedicated HPC cluster. Input data are Sentinel-1 images downloaded from PEPS which is the French Collaborative ground segment for Copernicus Sentinel program operated by the CNES (French Spatial agency), that provides a mirror site distributing all the Sentinel data in near real time. InSAR processing is done through the NSBAS chain (Doin et al. 2011). Each web service is implemented according to the OGC and INSPIRE recommendations in order to offer a normalized access to the service.
This project is done in coordination with EPOS (European Plate Observatory System) project, as part of the GDM service (Ground Deformation Monitoring) proposed in the work package 12 (satellite data). All the GDM products will be available from the ESA platform GEP (Geohazards Exploitation Platform).
A demonstration of the service will be given as well as a presentation of its architecture and operation.