Designing profitable and climate-smart farms using virtual reality

(approximately 4.6 and 6.5 SU/ha for the low productive and high productive categories, respectively). Carbon emissions and profit have been distributed across the four pasture cate- gories based on stocking rate assumptions and the profit and carbon emission rate described previously (Fig 3). The initial greenhouse gas emissions (kgCO2eq /ha) and profitability (NZD/ha) balance were established by considering the existing land use of the farm. Regarding profitability, the initial farm profit before interest, tax and rent has been set to $124,784 per year (Fig 3), indicat- ing an average profitability of $317/ha pasture. This value is representative of NZ beef and sheep farm profitability that has ranged from $266 to $401 per hectare for the season 2019/ 2020 and 2020/2021 respectively [20]. The initial carbon balance has been set up to 823,495 kgCO2eq per year. Results and discussion Using the pilot farm as a test case, we now report the findings of our experimentation with the tool and its ability to compare different tree planting strategies based on profit, GHG and a visual appreciation/assessment of the tree (fully grown). One of the most challenging aspects of the work, but one of great importance and highlighted next, was our need to understand the spatial variability of pasture production across a property. For this project, it was important because we assumed that farmers will be looking to plant trees on areas of the farm where pas- ture productivity is low and consequently profit is also. A second challenge for us was to create

a virtual environment that accurately represented the farm. Our assumption here is that farm- ers will expect to be able to move across their whole property and that the planting strategies that they design in the tool provide a realistic impression of the changes ‘in the landscape’ alongside environmental and financial targets, giving them added confidence in their final decisions. We note here that farmers have a great deal of (tacit) knowledge about their proper- ties that is relevant for strategic planning, so the application was developed with this in mind, providing a bridge between modelling science and farmer experience. This latter point features in our discussion. Data integration The farm terrain is a mix of steep hill country and flat areas that extend across the tops of the hills and within large gullies (Fig 4A). Steep hill country comprises a large portion of the farm Fig 4. Slope and very high pasture production areas. (A) Map of the farm border (red line) with the slope. Yellow and bright colour indicates steep slopes while light green colour indicates flat areas. (B) Map of the farm border (red line), with the most productive areas highlighted in green. Relative grass productivity has been estimated based on two NDVI images from Sentinel-2 taken during the 2019 lambing season. https://doi.org/10.1371/journal.pone.0286723.g004 PLOS ONE Designing profitable and climate-smart farms using virtual reality PLOS ONE | https://doi.org/10.1371/journal.pone.0286723 June 2, 2023 7 / 16