Improved precision in forage crop management

The overall aim of the project IMPRESS was to contribute to increased returns as well as environmentally sound production strategies in Norwegian forage farming. The basic idea was that principles and technologies from precision agriculture in cereal cropping can be adapted and developed further for perennial grasses. Development of technology, tools and knowledge that may constitute the basis for precision fertilization with nitrogen has been the core activity in the project. A prototype of a tractor mounted and low-cost sensor package for dry matter estimation has been constructed. A spin-off project financed by NFR in 2022 will investigate the interest among developers and manufacturers for further commercialization. Numerous drone missions with hyperspectral and multispectral cameras above heterogenous fields with near simultaneous yield and quality recordings on the ground, have also allowed extended modelling with spectral data as inputs. The conclusion from this work was that it is possible to predict forage yield and quality with high accuracy from air with sensor and drone solutions affordable for farmers. Another important delivery from the project is a process-based prediction model for yield development and soil-plant nitrogen dynamics in grass-clover crops. The model, called NORNE, can be updated with remotely sensed data and run for georeferenced subareas of farmers’ fields. This model has the potential to be the basis of a decision support system integrating site-specific soil and weather data, soil and crop information, nitrogen fertilization rates and farmers’ aims for the production. Several field experiments have been conducted to investigate dose-relationships for nitrogen fertilization rates and dry matter yield, as well as how protein quality may be affected by split application of nitrogen within growth periods. There appeared to be no reason to revise present norms for nitrogen demand in grass dominated crops. Within recommended rates of application, there seemed to be low risk of high nitrate levels and low protein quality at harvest under split-fertilization regimes, even with late top dressing. Lastly, the project has delivered a web-based decision support system to identify causes for lower yield in subareas of fields. This is a kind of questionnaire that farmers can tick off from their mobile phones while inspecting their fields with a digital yield map in their hands. IMPRESS has been financed by the Research Funding for Agriculture and the Food Industry (FFL/JA) and the industry partners Yara Norge AS, Yara Research Centre Hanninghof Germany, Felleskjøpet Agri SA and Strand Unikorn AS. Lantmännens forskningsstiftelse has also contributed financial support. Research partners have been the Norwegian Institute of Bioeconomy Research and the Swedish University of Agricultural Sciences/Hushållningssällskapet. The Norwegian Agricultural Extension Service has also been an active project partner.

Modeller og sensorløsninger fra IMPRESS er utvikla for å innen kort tid kunne brukes i praktisk jordbruk. De vil dermed være nyttige for utviklere som ønsker å lage kommersielle produkter og for bønder som ønsker å bruke innsatsfaktoren nitrogengjødsel med større presisjon og økt avkastning. En har videre sett og formidla hvilke muligheter og begrensninger som ligger i andre typer sensorer og sensorplattformer, slik at utviklere og brukere har et godt grunnlag for å prøve andre alternativer i årene som kommer. På lenger sikt vil resultatene og modellene kunne bidra til en mer bærekraftig grovfôr- og husdyrproduksjon ved at klimagassutslipp og andre miljøproblemer knyttet til dårlig nitrogenutnytting reduseres. De vil også legge til rette for at grovfôr kan utgjøre en stor andel av fôrrasjonen til norske drøvtyggere og at det fremdeles vil være den viktigste proteinkilden til disse.

The project aims to contribute to increased returns as well as environmentally sound production strategies in Norwegian forage farming. The basic idea is that principles and technologies from precision agriculture in cereal cropping can be adapted and developed further for the more challenging and heterogeneous perennial grasses. Improved precision in management will be developed along two axes. We intend to optimize nitrogen (N) fertilization practices in space and time, and we aim for enhanced yields in within-field problem areas by identifying and removing growth constraints. Our approach to reach these goals is to develop a platform that integrates new knowledge of plant responses to N, novel algorithms for sensor based yield estimation, simple models for crop growth and soil N mineralization, and real-time and predicted weather data. The main deliverables will be i) hardware and software for prototypes of a tractor mounted and an airborne sensor system for grass yield estimation, ii) a model providing recommendations for N fertilization with high temporal and spatial resolution, and iii) user-friendly and practical guidelines for identification of low-yielding areas. Several field experiments will be conducted in swards managed as to yield highly digestible forage. Here soil heterogeneity and treatments with differing rates and timing of N fertilization will cause yield differences needed for sensor evaluations and algorithm development. The yield data will also be basis for refinement of yield response curves to be integrated parts of the model providing site-specific fertilizer recommendation, also with regard to the feed quality of the crude protein fraction at grass harvest. There are several industry partners with financial contribution to the project that are interested in developing the suggested solutions to decision support services. All prototypes, algorithms, models and protocols will, however be published and freely available for everyone.