2022

2023

The objective of this project is to develop a prototype Digital Twin of a crop farm and use it to model and monitor Luke Jokioinen Smart Farming field infrastructure in real-time. It will be based on widely used simulation tool, soil scans, real-time data measurement and remote sensing data. Longer term aim is to develop software which can be used simulate all Luke research farms and extended to commercial farms with different types of production including livestock farming systems. Our Digital Twin will be based on APSIM (Agricultural Production Systems sIMulator) simulation software, standard Farm Management Information System (FMIS) farming plans, sensor data obtained through APIs and publicly available remote sensing data. APSIM is a state-of-the-art open source software for modelling agricultural systems allowing us to benefit decades of development in the simulation models which we will extend to fit our needs. It can be used to model whole farms and farming systems including all major components: soil, crops, weather, nutrients, soil organic matter, management actions, livestock etc. APSIM has been cited over 6500 times, is under active development and widely used in crop and farming system studies e.g. to simulate the effect of management decisions and climate change. In Finland it has been applied for cereal growth modeling (e.g. Palosuo et al. 2021). However, to our knowledge, no work focusing on combining its simulation engine with real-time measurements data from multiple sources exists. A digital twin is a collection or generation of data representing a physical object or process. When representing a dynamic object or dynamic process, digital twins consist of models that emulate the physical world. Digital twins have become more and more important in industry in various applications. Digital twins are used to observe processes that are difficult to measure directly or in planning and optimization of products or processes. In agricultural context, the concept of digital twin is recognized and used in various specific applications. For example, Daisy model has been used to model plant growth and nutrient state of soils. Precision farming technology enables site specific management (e.g. fertilization, working depth) based on e.g. yield maps, soil scans, soil sensors and remote sensing. However, many decisions are currently made individually by each farmer and the effect of different decisions are difficult to assess as each location in a field has specific characteristics (e.g. soil type) and each year is different. Use of digital twins in agricultural research opens new possibilities to observe farming systems and to develop and apply modeling for projecting the responses of the systems on the altered management or changing environment. The software is developed openly and documented on project website: https://twinyields.github.io