The fourth pilot of the agricultural showcase is focusing on the agro-industry. While many services for the farmers and for the industry already exist, the uptake of EO data remains low for a number of reasons. In this pilot, rather than a technology push, we will work closely together with a number of users, to better understand their EO needs, and how this service can best be provided. 

The starting point for Pilot 4 is the existing WatchItGrow (WIG) platform, which provides EO-based information to the potato growers and -industry. The expected outcome is to have a better understanding of how EO-based services are best tailored towards user needs, which bottlenecks still exist that hamper the uptake, as well as gain insights in the underlying processes that determine user uptake. For the latter, there will be a close collaboration with WP2.

The WIG platform is used for the daily monitoring of the fields, but the available services are extended to meet the user's needs. While timely observations are already very useful, having also predictions on crop development included can be very useful, especially when ripeness of the crops play an important role in for example determining the optimal time for harvest. This new service will be targeted in the first year.   

A beta-version of the new service was expected in June 2020, but could not be delivered due to a too limited amount of available training data, and issues with field data collection due to the Covid-19 situation. Developments are now ongoing again, including the data from the 2020 season. 

A Beta-version is now expected end of 2020, whose performance will be evaluated in the season of 2021. This beta version will only be available internally, given the low TRL of the service.

An improved version is expected in the fall of 2021. Depending on the results and the obtained confidence levels, these predictions on crop development (e.g. a 10-day forecast of fAPAR ) can be made available on the WIG platform for a number of selected users   

The outcome will be predictions of fAPAR time series, probably not longer than a 10-day forecast. These predictions will also include confidence intervals.

In an initial step, this service will only be available through the WIG platform, given the low TRL levels. However, once the confidence level of the product is high enough, it can be included in other services and platforms as well, if of interest. 

Web based application, API, Web-GIS

Seed multipliers are faced with the problem that their fields have to be located far away from each other to avoid cross-pollination, which has repercussions on their daily activities. For example, due to these large distances, it is difficult for them to closely monitor all their fields. However, as the seeds are the part of the crop of interest, the optimal harvest time is strongly linked to the ripeness of the seed pods, with the harvest window typically being very small. These two factors (low revisit time of field inspections, and narrow window for harvest), result in often sub-optimal conditions of the harvest. An EO-based monitoring service has a potential large impact on their daily operations, as the maturity of the fields can be monitored on a near-daily basis. A significant improvement for them would even be if the ripeness could be predicted shortly in advance, to enable a better organisation of the field visits and of the harvest.  The impact of this service for the seed multipliers could thus be very substantial

The impact is two-fold. The first one is on the outcomes of the process itself that is followed, and which is strongly linked to the c-design approach. Rather than the usual technology-push, where standard EO-products are offered to a range of users, this pilot will also provide information on the co-design process itself, where dedicated services can be created based on EO-data, to respond to very concrete needs from end-users. This is often encountered as a bottleneck when trying to increase the uptake of EO data by the (potential) end users.

Secondly, the service which is currently under construction, also provides many advantages outside of this specific pilot. Having information on how the crop will develop, together with for example meteorological predictions, can provide valuable information on a wide range of users, even outside of the agricultural application domain. This, however, will depend on the confidence levels with which these predictions can be delivered.     

In an initial stage, the fAPAR predictions will only be available for the users in the pilot, while in a later stage, this will be extended to other users in the WIG platform. Once the confidence of the predictions is high enough, these fAPAR predictions could be offered as a service by itself for interested users. The modalities hereof, however, are not known yet. 

As the outcome of the pilot will be an on-demand service, a direct registering in the GEOSS catalogue is not foreseen. However, we do plan to register our pilot application in service catalogues, such as the NextGEOSS service catalogue, once it's fully developed.