In April 2021, NEO started cooperation with the Ukrainian Soil Partnership (UaSP), an alliance of several Ukrainian research institutes in the field of soil science and agronomy. The aim of the cooperation is threefold.
The first objective is to identify which agricultural crops are grown on the plots and the area of these crops within the country. The innovative aspect is to look not only at crops grown on a large scale, such as cereals, maize and sunflowers, but also at crops grown mainly locally by small family farms, known as smallholder farms. Examples include buckwheat, peas, watermelon, fruit and vineyards.
We are also building a model that forecasts the expected yield at the end of the growing season for the four main crops during the growing season. This is an agricultural planning development.
Finally, NEO and UaSP have taken up the challenge of mapping sensitive natural areas. The focus here is on wet natural areas, so-called 'wetlands'. If the method works for wetlands, it can also be extended to other natural areas, such as forests and coastal areas. In essence, it is about faster and easier identification of potential sensitive areas, which can help enforce laws to protect nature.
For these developments, NEO uses publicly available satellite images with the main data source being the Sentinel-2 mission of the European Space Agency's Copernicus programme. The sister satellites Sentinel-2A and Sentinel-2B together take an image of the entire Earth every 5 days with a pixel resolution of 10 metres in the visible light and near-infrared, and 20-60 metres for a number of other spectral bands. We will analyse this data using artificial intelligence and the results will be published on Google Earth Engine.
The two organisations NEO and UaSP complement each other well. Whereas NEO has extensive experience in operational monitoring of agricultural areas, UaSP has a wealth of knowledge about Ukrainian habitat, agriculture and the environment. Currently, teams from UaSP are collecting data in the field on crop types and wetland areas. In doing so, they take into account the differences that occur under the influence of Ukraine's main agro-climatic zones, which run from the northern forests, through the temperate central zone, to the southern steppe. The data collected by UaSP is used by NEO to train the machine learning models and validate the model results.
During July and August, the necessary data will still be collected and models developed. First results are expected in September after which NEO and UaSP will collaborate on articles describing the project's findings. This work is being carried out under the EU-funded and World Bank-implemented project 'Supporting Transparent Land Governance in Ukraine'.
