Dr. Marcela DOUBKOVÁ (Pessl Instruments GmbH)

Management Zone maps are the first prerequisite for practicing precision farming and are defined as regions within the farmer’s field with similar yield potential. Treating regions with different yield potential differently during operations such as seeding, application of nutrition and protection or during irrigation was demonstrated to save resources, decrease nutrient run-off and increase profit. While it’s still the farmers’ knowledge that counts the most, geospatial technologies proved to be an important asset helping farmers to assure or even improve their knowledge of management zones.

At EUXDAT, we developed a state-of-the-art big data exploitation platform that enables users to work in one space with data of different formats and sizes originating from a variety of sources and provides them with user-friendly tools for their analyses. To demonstrate the benefit of the EUXDAT platform in the agriculture sector we are building a prototype of a management zones delineation tool. In particular, we provide to the farmer a) a map of the management zones based on the Sentinel-2 data using Jenks natural breaks algorithm as well as b) a collection of layers that are known to generate these zones. As such, the tool should support not only the process of the management zone delineation but also the process of understanding the origin of these zones.

While the management zones come as a result of a variety of geophysical parameters (i.e. soil type, water holding capacity, soil depth, nutrients and organic matter, weather, elevation), we rely on our work on an assumption that significantly simplifies their delineation. In particular, we claim that satellite-derived biomass measured during intervention-appropriate periods can function as a surrogate of all the above-mentioned input variables and thus can directly reflect the management zones. This assumption is based on similar research activities [1, 2].

In the first phase, the user sends API requests providing field borders and seeding dates. Upon this request, he/she is provided management zones computed from the Sentinel-2 data from the i.e. first 3 weeks after seeding (this period was demonstrated as the most relevant for N fertilizer application and can be adjusted) along with a DEM layer (Figure 1).

Figure 1. Orthophoto of the field of interest (upper left) and 2 layers available upon EUXDAT API request: (upper middle) Management zone computed from Sentinel-2 images available 3 weeks after seeding (red: biomass is below field average, to dark green: biomass is above field average) and (upper right) DEM SRTM model. The lower layer displays the cloud-free images used to compute the management zone map above.

In the next phase, multiple other datasets will be displayed to the user in one platform upon the API call:

  • Weather data forecasted as well as historical available for a given location from Meteoblue and Pessl Instruments companies
  • DEM-related parameters such as topographic wetness index (TWI). TWI quantifies the distribution of on a slope.
  • Chemical and topsoil physical properties for Europe based on LUCAS topsoil data

Pessl Instruments, as one of the EUXDAT consortium partners, leads the development of the management zone tool. Pessl Instruments has been providing software and hardware weather-related decision-making tools for the last 35 years. ‘While the primary use of the management zone maps is to generate a variable rate application maps, we at Pessl Instruments consider also many other ways of usage for these maps, i.e. as a soil and tissue sampling map for the iMETOS MobiLab (https://metos.at/imetos-mobilab/) or as a distribution map for crop coefficients to provide daily water balance, or disease models estimates for both low- and high-yielding regions.’ says the remote sensing expert at Pessl Instruments, Dr. Marcela Doubkova.

Literature:

[1] F. M. Breunig et al., “Delineation of management zones in agricultural fields using cover–crop biomass estimates from PlanetScope data,” Int. J. Appl. Earth Obs. Geoinf., vol. 85, p. 102004, 2020.

[2] M. Vizzari, F. Santaga, and P. Benincasa, “Sentinel 2-based nitrogen VRT fertilization in wheat: Comparison between traditional and simple precision practices,” Agronomy, 2019.