FFL-JA-Forskningsmidlene for jordbruk og matindustri

Grassland-based forage production plays a critical role in Norwegian agriculture for producing milk and meat and is crucial for the farm economy. With increasing temperatures due to climate change, perennial ryegrass, can be grown further north and in more continental regions. This requires high rates of nitrogen fertilisation, leads to increased GHG emissions, and nitrogen leaching. To address this challenge, it is crucial to breed future perennial ryegrass cultivars with a higher nitrogen use efficiency (NUE). The DLT-Farming project will develop knowledge for breeding perennial ryegrass cultivars with improved forage quality and lower environmental footprints.Precicely, the project objectives are 1) to conduct field trials across two locations in Norway, with different amounts of fertilizer content. 2) Designing a sustainable IoT (Internet of Things) sensor network that combines low-power communication and data storage with AI/ML techniques to assess soil quality and nitrogen use efficiency in perennial ryegrass. 3) Enhancing the accuracy and robustness of phenomics protocols for collecting dry matter yield and nutritive traits using advanced robotics and sensing technology coupled with AI/ML models. 4) Identifying and characterizing the genes related to nitrogen use efficiency through genome wide association studies. 5) Developing an AI/ML data analytics platform for autonomous processing of big data collected from field sensors, creating a real-time reporting application for N uptake, crop yield and quality data of perennial ryegrass. The integration of high throughput phenomics and genomics in precision agriculture offers a sustainable approach for producing high-quality and high-yielding forage grass with reduced fertilizers. Thus, this project supports several of the United Nations sustainable development goals (e.g., particularly Goal 2, Goal 9, Goal 13). Field Trials at NMBU and Graminor: Extensive planning for field trials and data collection using sensors on drones and robots was initiated during our kick-off meeting and further developed through literature review. We selected 40 perennial ryegrass cultivars, including both diploids and tetraploids, adapted to Norwegian and other European climates, for the trials. These cultivars, selected from our previous PPP perennial ryegrass project, were planted in field trials with three fertilizer treatments (low, normal, high) and three replicates, totaling 360 populations. Trials were established at NMBU (Vollebekk, Ås) in May and at Graminor (Ridabu, Hamar) in July. Soil Sensors Integration and Preliminary Testing: We integrated MODBUS-RTU RS485 soil quality measurement sensors to rapidly assess soil nitrogen, phosphorus, and potassium levels across different fertilizer treatments. Additionally, sensors for soil moisture, pH, temperature, and electrical conductivity were used to monitor salinity and nutrient availability. These sensors were integrated with Arduino Uno microcontrollers and ROS (Robot Operating System) in a lab environment. The data collected from these sensors will be analyzed in preparation for field deployment next year. Image-Based Field Data Collection using UAVs: We utilized UAVs equipped with RGB and multispectral cameras, integrated with DJI software, to capture weekly image data of the ryegrass fields at varying flight heights, both pre- and post-harvest. This data will be used to generate NDVI-based datasets for monitoring crop growth and health, vegetation greenness, height, density, and changes in plant health and nutrient content. The data collected so far will be analyzed shortly. Field Data Collection with Ground-Based Mobile Robots: The Thorvald mobile platform, equipped with a 3D LiDAR and RGB-D camera, was used to collect range and image data of the ryegrass fields at NMBU (Vollebekk, Ås) and Graminor (Hamar). The goal is to create an image-based dataset to support AI models that estimate nitrogen uptake, nitrogen use efficiency, and plant biomass. The data gathered this year will be analyzed to inform model development, which we aim to deploy before next year’s field trials.

Grassland-based forage production plays a critical role in Norwegian agriculture for producing milk and meat and is crucial for the farm economy. With increasing temperatures due to climate change, perennial ryegrass, which is a high yielding and nutritious grass species, can be grown further north and in more continental regions. Expansion of the cultivation area of perennial ryegrass, which is mostly grown in monoculture for leys and requires high rates of nitrogen fertilization, leads to increased GHG emissions, and nitrogen leaching. To address this challenge, it is crucial to breed future perennial ryegrass cultivars with a higher nitrogen use efficiency (NUE). To speed up the development of improved cultivars, Norwegian perennial ryegrass breeding must be more efficient by incorporating advanced breeding methods and techniques. This project will utilize ecotypes and modern cultivars characterized in a previous Nordic/Baltic “Public-Private-Partnership (PPP) project for pre-breeding in perennial ryegrass”. NUE of these populations will be evaluated at two locations in Norway by integrating phenomics and genomics, to select the best NUE populations to incorporate in the breeding program. Phenomics creates huge amounts of data from various field sensors by drones and robots. Currently, no platform can process such big data in real-time. This project aims to establish an AI/ML-based data analytics platform for autonomous processing of big data, creating a real-time reporting application for dry matter yield and forage quality in grasses assisting breeders with rapid selection of superior genotypes and farmers with optimal harvest time decisions. The project will develop knowledge for breeding perennial ryegrass cultivars with improved forage quality and lower environmental footprints. Precision farming aided by utilization of the AI-based big-data platform will increase the economic output for farmers.