Sen4CAP’s Main Focus in 2020 was on Transfer Activities to the Paying Agencies

The main objective of the 1-year project extension was to facilitate the transfer of the Sen4CAP system operation to the Paying Agencies. Some of the Sen4CAP pilot Paying Agencies, i.e. Castilla y Leon, Italy and the Netherlands, decided to operate the system by themselves during 2020, either on the cloud or their local infrastructure. They have been fully supported by the Sen4CAP consortium and all of them succeeded in getting the expected results.

In parallel, we continued to run the Sen4CAP system for the other pilot countries. We can say that this additional demonstration year allowed each country to go one step forward in its uptake of the system and has been beneficial for their future activities.

Support was also provided to the wider user community through the forum and the organization of regular webinars. The Sen4CAP system was downloaded 260 times since November 2019 and more than 20 external Paying Agencies were granted a Virtual Machine deployed on CREODIAS to discover Sen4CAP and assess its fitness-to- use. The general feedback was very positive and this experience also provided us useful inputs to decide which functionalities we should improve to make them relevant also out of our pilot countries.

Despite the COVID situation, 2020 turned out to be quite a successful year showing even more the critical role of Earth Observation data in a circumstance when the full free movement is limited. The Sen4CAP system was used to monitor the harvesting of winter cereals at a national scale in Spain continuously throughout the lock-down providing transparent information on the agriculture sector.

Sen4CAP also participated to the RACE platform which demonstrates how the use of Earth observation data can help shed new light on societal and economic changes currently taking place owing to the coronavirus pandemic.

A Vibrant Virtual Sen4CAP Community

With the first wave of COVID-19 and restrictions coming along, we realized quite soon we would have to adapt to the new reality of working from home and virtual meetings. But no matter how awkward it might had felt at the beginning, Sen4CAP was already used to organize online events and we continued on our path. The online events turned out to be a great way of sharing with you our progress and plans and even more importantly, you sharing with us your experience with the Sen4CAP system.

We organized 2 trainings, 5 webinars and 2 Q/A sessions this year. We initially planned a limited number of participants per event to allow active interactions of the interested users. Due to the high demand, we had to organize 2 sessions of the on-line training for beginners in April which made us realize we needed to remove the limit and record the sessions. The highest number of participants was over 150 which completely exceeded our expectations. We plan to continue with the online events also in the next year, starting with a webinar on 12 January. Book the date!

Sharing of the experience with the Sen4CAP system: the Danish Agricultural Agency used Sen4CAP for maize harvest detection in season 2019 and presented the results compared to the in-situ data obtained by the roadside observations at the 5th Sen4CAP webinar.

New Sen4CAP Processor – Tillage Detection

The tillage detection processor is based on the same approach as tools for agricultural practices monitoring implemented within the Sen4CAP system. Similarly to harvest monitoring, an automated analysis of EO based temporal profiles is applied to select the week in which the tillage was detected.

The aim is to provide a generic solution that is not tailored to any specific country and follows the continuous monitoring concept. It is not feasible to demonstrate all the physical and temporal variability of tilling & ploughing practices so during the development of the processor, we focused on one of the most common scenarios – tillage applied after the harvest of the main crop. Nevertheless, the existing solution can be customized in a way to address the abovementioned variability of these practices.

The processor is being developed in close cooperation with the two Sen4CAP pilot Paying agencies, the Lithuanian and of Castilla y Leon. They have provided reference data to support processor development and validation. Also, their support included the insight into the local tillage practices, e.g. the information when the tillage is usually applied in their country and how frequently, what is the parcel status before and after the tillage (bare soil vs. vegetation cover) and what are the possible activities accompanying this practice.

The temporal profiles extracted from Sentinel-2 NDVI and Sentinel-1 SAR coherence data are analysed to detect parcel conditions typical for the application of tillage. The decision is made through the subset of markers developed for harvest monitoring.

See some examples explained during the presentation of the processor at the 5th Sen4CAP webinar.

Farmers’ reports about the dates of applied practices have been provided by the Castilla y Leon Paying agency. The tillage was applied at the end of August (red line), the harvest was detected by the agricultural practices processor (blue strip). Evolution of the markers was the following: NDVI values dropped after the harvest and remained low after the tillage, coherence values were high and stable after the harvest and decreased during the tillage.

The Latest Sen4CAP System Version and Future Release Plans

The latest version 1.3 was released on 13th November. There were no new features delivered but we addressed the Sen4CAP community suggestions for corrections, adaptations and improvements of the system.

The next version 2.0 is planned for January 2021 and will include:

•      possibility to access the markers database,

•      the new processor for tillage detection,

•      different system components will be separated through docker containers,

•      support of the new version of MAJA (4.2.1),

•      support of ASTER DEM for the pre-processing, which will enable the use of the Sen4CAP system also to the European Northern countries.

Another version 3.0 is planned for April 2021. This release will focus on major upgrades to the web interface.

Version 3.0 in April 2021 will bring significant upgrades to the web interface

Sen4CAP Community after the End of the Project

The Sen4CAP project will end in March 2021 but since we received requests throughout the community for further investment in the development of the system, we feel committed to continue supporting the Sen4CAP users in the future. We will remain active on the Sen4CAP forum and we also plan to continue the bi-monthly webinars at least for one more year.

In order to enable the reliable use of Sen4CAP, its system maintenance (adaptation to Sentinels format evolution, major bug fixing and potentially integration of the new external modules) will be ensured in the long-term by the ESA Agricultural Virtual Lab. The AVL will also allow and actively support the community to further develop the algorithms in the future following an open science approach (see below article).

In terms of major evolutions, we will have a more flexible approach: one option is to partner up with other projects recognizing the Sen4CAP as a useful tool, like the H2020 project NIVA. Evolution can also be driven by the open science community on GitHub. You are kindly invited to contact us if you identify yourself as a member of a project, organization or network, which would also see a benefit in additional major upgrades of the Sen4CAP system.

We invite you to keep posting your questions on the Sen4CAP forum.

ESA Agriculture Virtual Laboratory (AVL)

In September 2020, ESA launched a new Virtual Lab focusing on Agriculture. Virtual labs are platform services especially intended for scientists to share data resources and create an enhanced research environment. AVL aims at being an online community open science tool: it will be designed to support scientist to access and share EO data, high-level products, in-situ data, and open-source code (algorithms, models, tools) to carry out scientific studies and projects, sharing results, knowledge and resources. The development will be organized in different steps, and Early Adopters are expected as soon as 2021.

A user consultation was launched in December and is open until 18th of January 2021.

This consultation tries to identify the most useful datasets to be included as a matter of priority and to define what is needed to make the AVL a successful open science tool: what is needed to foster the sharing of data and algorithms, which environment, functionalities, animation, etc. Would you be interested to participate in this user consultation, don’t hesitate to contact Sophie Bontemps.

Virtual labs are platform services especially intended for scientists to share
data resources and create an enhanced research environment.

UPCOMING EVENTS

January 12, 2021 – The 6th Sen4CAP webinar will be led by Sophie Bontemps, the Sen4CAP project coordinator.

February/March, 2021 – The closing event of the project will be an online user workshop with a review of the entire 2020 season and a discussion on the future of the Sen4CAP.

Do not hesitate to send us any feedback to info@esa-sen4cap.org. We look forward to hearing from you.

The Sen4CAP team