Designing profitable and climate-smart farms using virtual reality

with half of the total farm area being on slopes that are 27 degrees or more, while less than 10% of the farm is located on slopes between 0 and 10 degrees. The pasture-productivity map (Fig 4B) derived from NDVI data locates the highest produc- tive pasture on the flat areas along the hill tops and the flat areas in between the steep gullies that extend along a North to South plane. Other areas of the farm have low and very low pas- ture productivity, most likely due to combined factors such as the steep slope, shallow soils and low soil fertility which is typical of landscapes in this region. As expected from the maps above, there is a negative trend between the NDVI values and slope i.e., pasture productivity tends to be greatest in flat areas (Fig 5). Considering the carbon and profit values used in this study (which reflect the national average) our model highlighted that 8% of the farm would need to be converted to forestry if the user only considers planting pine trees on very low productivity classes. This number increased to 22% of the farm if the user only consider planting native trees (Fig 6A). Future iterations of this tool could reproduce the approach of previous studies that have used remote sensing data to generate robust estimates of pasture productivity over large areas [30–32]. This more in-depth analysis might also support a refinement of the stocking rates [33], farm financial returns and GHG emissions that would increase the level of detail consid- ered into farm planning strategies. This approach was not used here, as the focus was on devel-

oping a prototype. Fig 5. NDVI values for sentinel-2 pixels (20m) according to the slope in degree. NDVI values have been used as an estimate of relative pasture productivity (most producing pixels versus least producing pixels). The horizontal red dashed lines represent the four productivity quantiles used to classify the pasture productivity from very high productivity to very low productivity. As slope increases, pixel productivity tends to decrease. The black dashed line represents the linear equation describing this negative correlation between NDVI and slope. The grey shade colours (from light grey to black) used to represent the individual pixels within the scatter plot represent the pixels that would need to be converted from pasture to forest for the farm to be carbon neutral when converting the least productive pixels first. Light grey pixels (n = 798) up to NDVI values equal to 0.69 are the pixels that would need to be converted if only planting pine forest. Light grey pixels plus grey pixels (n = 2257) up to NDVI values equal 0.73 are the pixels that would need to be converted if only planting native forest. https://doi.org/10.1371/journal.pone.0286723.g005 PLOS ONE Designing profitable and climate-smart farms using virtual reality PLOS ONE | https://doi.org/10.1371/journal.pone.0286723 June 2, 2023 8 / 16