Leveraging Earth Observation (EO) data to Enhance Tree Crop production and their resilience in a changing climate in Australia- FruitSense

Agriculture is a major economic activity in Australia and contributes up to 12% of the goods and services exports. Furthermore, the gross value of agricultural production has been increasing, rising from $5 billion in 2001 to $75 billion in 2020/21. One of the main drivers of this increase has been the diversification into tree crop production. For example, over the past 20 years, fruit and nut production had the fastest growth in the agricultural sector, with a 24% increase in production. Despite these successes in tree crop production, the sector still faces a number of challenges, with the main one being impacts of climate change. Australia has a largely arid climate and is highly susceptible to climate change, which impacts tree crop production. For example, increase in temperature can reduce fruit quality through changes in the sugar content and chemical composition and a reduction in precipitation coupled with shrinking ground water reserves present a challenge to availability of water to irrigate tree crops. Moreover, prolonged droughts due to climate change, will exacerbate the occurrence and frequency of bushfires, which can decimate tree crops. Climate change will also lead to changes in occurrence of tree crop pests and diseases. To address these challenges, there is a need to develop a robust tree crop management system that will ensure resilience of tree crop production under a changing climate in Australia. The recent availability of high-resolution free Earth Observation (EO) data (e.g., ESA's Sentinel mission) and cloud computing, offers an opportunity for creating a tree crop management system in Australia that would ensure their resilience in the face of climate change. Therefore, this project aims to leverage the potential of EO data to develop, demonstrator web-based tree crop management system that promote resilience of tree crops in changing climate. To achieve this aim, the project will undertake a number of activities. Firstly, we will develop a countrywide map of the major tree crops across Australia. To do this we will use the state-of-the-art classification algorithms (e.g., machine learning algorithms) and exploit multi-temporal/spatial EO data and ground reference data (e.g., from the Australian Land Use and Management (ALUM) dataset and the Australian Tree Crop Map (ATCM) Dashboard) to generate the countrywide tree crop maps. Secondly, the project will develop an EO based indicators for tree crop health and condition. We will do this by identifying and generating key biophysical variables from EO data and use these variables to determine tree crop response to multiple stressors (e.g., drought, pests, and diseases). Thirdly, the project will assess the feasibility of using existing crop yield models to predict tree crop yield. We will do this by evaluating the CSIRO Graincast model infrastructure and identify relevant tree crop biophysical variables that can be incorporated in the model to enable it predict tree crop yields. Fourthly, the project will identify current and future climate impacts on tree crop production and health. We will do this by evaluating the response of tree crops to long term changes in temperature and rainfall. Additionally, we will evaluate the vulnerability of tree crops to climate induced changing wildfire risks by intersecting the tree crop extent with a modelled Forest Fire Danger Index. Finally, the project will integrate all these datasets and information into a web-based tree crop management system that will be used by relevant stakeholders to promote climate-smart tree crop production and ensure resilience of tree crops to a changing climate in Australia.