engleski

Advanced and predictive agriculture for resilience to climate change

Climate change is one of the greatest concerns of today’s world and has significantly altered or is continuously altering Earth’s ecosystems. It can be explained as the phenomenon involving changes in environmental factors like temperature, humidity and precipitation through a long period of time. The unprecedented temperature rise has resulted in increased events of droughts, floods, heat waves and other extreme weather conditions on a global scale. Far reaching effects of climate change are now greatly visible on agriculture sector, on which relies the food production, food safety and economy of the most parts of the world. Current and future climate change is projected to have significant impact on the cultivation of agricultural crops resulting in lower yields and higher costs. Climate change also have notable impact on pests of these agricultural crops by influencing their reproduction, development, survival and spread as well as the relation between pests and their environment. As a significant driver of agricultural crops and their pest population dynamics, climate change will require adaptive management strategies. This means that we need to take proper measures to mitigate harmful effects of climate change and adapt to its consequences. In order to do so, the key is transition to sustainable and modern agriculture which must be driven by new technologies, research and innovation. In 2020, the European Regional Development Fund has supported the project "AgroSPARC - Advanced and predictive agriculture for resilience to climate change" which is carried out by the Innovation Centre Nikola Tesla, Faculty of Electrical Engineering and Computing and the Faculty of Agriculture at the University of Zagreb in Croatia. Therefore, the key priorities of this project are building resilience and adapting to present and future climate risks. The main focus of the project is to use artificial intelligence to develop mathematical models for different phenophases of wheat and use these models to forecast crop yields and forewarn of insect pests in the terms of climate change predictions. Analysis of a large data sets will be carried out in relation to various climatic conditions, which will be artificially generated and permuted in prototype climate chambers. These data sets will be correlated with established indicators of wheat growth and development at different phenophases and different intensities of the insect pest infestation. A system based on artificial neural networks will be developed. Artificial neural networks classify and select data on climatic conditions in prototype climate chamber, real-time weather forecasts and indicators of crop development and then learn and verify numerical models of different phenophases of wheat based on large experimental data sets. This activity involves determining neural network structure ; inputs, outputs, number of different layers, number of neurons per layer, etc. When determining the structure, the available domain knowledge in agronomy is expected to significantly increase the efficiency in the implementation of activities and reduce time frame for obtaining results as well as the accuracy of the final results. The mathematical models will be available publicly and interactively via an online portal to forecast different phenophases of wheat and forewarn of the most significant insect pests of wheat under real and hypothetical climatic

climate change ; mathematical models ; artificial intelligence ; prediction

nije evidentirano