Data and technology

Rapid development of digital technology is having an impact on the long-term goals of the foreign trade and development cooperation (BHOS) policy of The Netherlands. The BHOS digital agenda [1] aims to achieve BHOS policy goals faster and more effectively by responding to the opportunities and threats presented by digitalisation. The G4AW programme fits within this digital agenda. A wide variety of interesting showcases that highlight the role of digitalisation for development has resulted from Dutch development cooperation [2].

Information services

Earth observation data, plays a central role in all created information services. The role that satellite data plays in the products has been one of the selection criteria in the programme. Most of the services also make use of Global Navigation Satellite System(GNSS) [3]. A wide variety of Agservices and Fintech services have been created in the different G4AW projects. Creating new data and technologies has not been a priority; rather the focus has been on creating new use-cases with existing data and technologies. Important aspects in the domain of satellite data and related data flows include: data selection and correction; data processing (value-adding), converting data to services and establishing data flows.

Satellites

The most used satellite data in the G4AW program has been data from weather satellites, followed by Sentinel and Landsat. Satellite data from other sources has also been used in nearly half the projects. For most of the uses in the G4AW program, open source data from US and European satellites such as MODIS, LANDSAT and Sentinel has been satisfactory. Commercial satellites are used in a few cases where they provide clear added value: 1) by making available more sensors to create products with higher temporal resolution [4]; and 2) by improving the usability of data by using very high-resolution datasets [5]. Each project has provided the sensors they expected to use in the proposal, and have provided updates on the satellite data that is used in the operational products. The biggest difference when it comes to expected vs actual use can be seen in the use of Sentinel-1 SAR data. Few projects expected to use Sentinel-1, while it became the single most used instrument in G4AW Facility.

Service types

The most important Agtech and Fintech service categories that have been addressed in the program (outside of crop monitoring) have been weather-related (67%) and finance-related (33%). Each service type requires a specific use of satellite data and technologies. Within the crop monitoring categories, the emphasis has been on crop management (crop cycles), pest and disease warning, soil-related advice and fertilizer application. Crop monitoring is still frequently done by creating vegetation indices such as NDVI; improvements in spatial and temporal resolution continue to make this highly relevant [6].

Weather-related service require bundling of a large number of satellite missions and complex models (numerical weather prediction). Ground-stations and crowd-sourcing also play an important role in weather now- and forecasting. Datasets on precipitation and evapotranspiration are also used for drought index insurances. For services such as crop-based index insurance, the most important selection criteria of the data is the availability of long (>15 year) time-series combined with operational reliable satellite data. A dataset that has been created in many projects is soil moisture. Soil moisture data is used in irrigation advice [8] and more general crop management, but also in index based insurance [7].

The high temporal resolution of datasets due to a) increasingly frequent satellite overpasses; and b) the possibility to combine different sensors (sensor fusion), enables the use of accurate change-detection algorithms [9] to check for relevant changes in the field (deforestation, irrigation, harvesting, etc.).

AI and traceability

The use of artificial intelligence (AI) in G4AW the programme has been relatively new and has mainly been used for land-use classification (machine learning). Much of this technology is still being created and validated in developed countries where training data is readily available. Lack of in-situ (survey, soil sampling) data to train models is a problem in many G4AW partner countries. Several sources of training data have been made available in recent years. These include BigEarthNet [10], and LandCoverNet. Such datasets are available on platforms such as Radiant Earths’ MLHub [11]. Technologies that enable traceability have been applied in one project (SpiceUp). Blockchain and space-based technologies can strengthen each other and contribute to tracking food supplies [12]. Technologies such as blockchain have not yet been implemented in the G4AW projects, but could be a useful addition in service bundling.

Data management

Data-management and cloud-computing platforms have played an important role in making processed satellite data embedded in services available to farmers in a structured way [13]. The role of data protection (see CSR) is becoming increasingly recognized in different countries, creating new challenges to the consortia. National data laws might restrict use of international cloud-based raster processing and storage infrastructure [14], resulting in new opportunities for local service (cloud) providers.

Data-integrators and data-platforms play an important role in data-management. These parties need to ensure all aspects of the data flows fit within the existing legal framework. At the same time, these platforms should provide enough structure, speed and support to provide good services. Services are not only valued by the provided advice, but also by the uptime, response time, and ability to reach support. The distribution between Dutch-based and local organizations that have played this role as data-integrator in different G4AW projects has been 50-50. Benefits of local solutions is the local embedding, the ability to provide support (same time-zone) and generally the lower costs. Benefits of Dutch solutions have been a stronger established technical base, resulting in quicker deployment.

Open data

Interesting (open) developments related to data & technology for digital advisory services include the expanding datasets (incl. WaPOR) in Google Earth engine [15], FAO’s Hand-in-Hand initiative [16] and SEPAL platform [17], the CGIAR data-harvesting tool GARDIAN [18] and different EU projects focused on using open-access tools to process remote sensing data: OpenEO [19] and Big EO Analytics [20].

Read more:

1. Digital Agenda for Foreign Trade and Development Cooperation (BHOS).
2. 33 Showcases - Digitalisation and Development - Inspiration from Dutch development cooperation.
3. Satellite Navigation for Digital Earth..
4. The Effect of Three Different Data Fusion Approaches on the Quality of Soil Moisture Retrievals from Multiple Passive Microwave Sensors.
5. Scouting Watering Holes from Space.
6. A comparison of global agricultural monitoring systems and current gaps .
7. Remote sensing for index insurance Findings and lessons learned for smallholder agriculture.
8. The value of satellite remote sensing soil moisture data and the DISPATCH algorithm in irrigation fields..
9. An Evaluation and Comparison of Four Dense Time Series Change Detection Methods Using Simulated Data.
10. http://bigearth.net/
11. https://www.mlhub.earth/
12. http://interactive.satellitetoday.com/blockchain-the-next-big-disruptor-in-space/
13. An Overview of Platforms for Big Earth Observation Data Management and Analysis.
14. 2020 is a crucial year to fight for data protection in Africa.
15. A planetary-scale platform for Earth science data & analysis.
16. FAO Hand-in-Hand Initiative.
17. Sepal. System for earth observations, data access, processing & analysis for land monitoring.
18. GARDIAN. The Global Agricultural Research Data Innovation & Acceleration Network.
19. openEO.
20. Big EO Analytics. Increasing capacity to combine optical and RADAR Sentinel satellite data for rapid change detection.
21. World Bank - Data-driven Digital Agriculture: Knowledge and Learning Platform