Future-fit orchards

Profiles of pome-fruit areas

The first project, which has already been completed, profiled the climate of the EGVV, Koue Bokkeveld, and Langkloof, specifically in the context of pome-fruit production. It drew on weather data for eight seasons, from 2016/7 to 2023/4.

For the pome-fruit climate profiles, TerraClim generated terrain maps at 40-metre and 2–5-metre resolutions. Terrain includes characteristics such as aspect, elevation and slope.

Long-term climate layers at 90-metre and 1-km resolutions were added to the terrain maps, allowing for detailed analyses of daily and hourly temperatures for each growing season. Based on this, Southey could report on the spatial and annual variability in bioclimatic indices such as chill and heat units.

Southey illustrated the results with maps of growing degree days (see images on facing page) and hours for the three areas. She highlighted the differences between maps of the mean values and maps of the standard deviation. Means show differences between seasons, and standard deviations indicate consistency within seasons.

“It was interesting that areas with less heat – fewer growing degree days – tended to have more variation from season to season,” she noted. “The Langkloof had the lowest number of heat units but the highest variation between seasons.”

Cooler winters, warmer summers

The climate of the three regions responded differently in each season. “When looking at growing degree hours, 2023 was the hottest year in the EGVV,” said Southey. “Looking at the Koue Bokkeveld, 2021 was the hottest, and in the Langkloof, 2022 and 2023 were the hottest.”

A comparison of the three regions shows that the EGVV has the most growing degree days and the Koue Bokkeveld the fewest. However, whereas the Langkloof used to be intermediate, the number of growing degree days appears to be creeping up, overtaking the EGVV in 2023.

“The Langkloof has a lot more hours above 35 °C than the other regions, but the figure jumps from more to less across seasons,” said Southey. Hours below 18 °C, which reflect unfavourable conditions during bloom, also vary significantly between seasons. But there is less regional variation than for hours above 35 °C.

“The trend we’re seeing is cooler winter conditions but slightly warmer summers over the last few seasons,” commented Southey.

She pointed the audience to the Climate Information page on the Hortgro website, where they can view or download monthly climate reports prepared by Hortec, iLeaf, and TerraClim.

The mean number of growing degree days over eight seasons from 2016– 2017, for different regions:

Right plum, right place

The second project that Southey discussed is an ongoing study of the suitability of different areas for Japanese plums.

“Our main aim was to integrate FruData and TerraClim data,” said Southey. She singled out the contribution of Pietman Wessels, senior economist at FruData, an agricultural economics company, and acknowledged the inputs of Provar and ExperiCo.

“The project is about using the know-how of what’s working to identify what will work,” explained Southey. “We wanted to identify the plum cultivars that yield the most export fruit and then find new regions to grow them.”

A database of existing orchards, including their yields and quality, was compiled with the assistance of FruData and a panel of experts. Seven plum cultivars were selected for inclusion in the project: African Delight, Angeleno, Fortune, Laetitia, Ruby Star, Ruby Sun, and Songold.

Five sites that consistently perform well in terms of pack-outs, and five that consistently underperform, were identified for each cultivar based on five years of data.

The orchard yield and pack-out data were aggregated with terrain and growing-season climate data, and machine learning was applied to model the features associated with tree performance. The model ranks the climate factors according to importance and assigns a weight to each.

“For example, humidity between 40 and 60% was the top variable driving performance for Angeleno and Songold, but it had a much lower ranking for African Delight,” said Southey. “Temperatures between 15 and 20 °C were highly important for African Delight, less so for Angeleno, and even less for Songold.”

Using the model’s output of important features, TerraClim created maps of the similarity between different areas and those areas where a cultivar currently performs well or poorly. When comparing the maps for African Delight, Angeleno, and Songold, it’s evident that the most suitable areas for these cultivars differ.

“These are preliminary results but provide a proof of concept that combining data sets can provide valuable insights for making decisions,” concluded Southey. “And it shows the value of integrating multiple data sets.”