Predicting growth and future yield in Eucalyptus grandis x...

The aims of this research were to (1) partly parameterise and initialise and (2) test the CABALA process-based model for hybrid Eucalyptus grandis x Eucalyptus urophylla clones in the Zululand region of South Africa. To achieve these aims, detailed data was obtained from a set of 18 sites at which permanent sample plots had been monitored for seven or more years. The model was run using two sources of weather data and the output from CABALA predictions were compared to observed stand volume, mean diameter at breast height (DBH), and mean height for the 18 study sites. Overall, when the long-term mean monthly weather data was used, CABALA overestimated stand mean DBH, mean height, and stand volume. When the daily weather data was used, the model gave better estimations of mean DBH but tended to underestimate mean height and stand volume, especially on higher productivity sites that generally received high mean annual precipitation. The leaf area index (LAI) estimated by CABALA was compared to estimates from the MODIS LAI product and while CABALA overestimated LAI in many instances, it predicted the LAI drop trend observed during the drought period 2014-2016 as soon as the daily weather conditions were introduced. Additionally, results showed that the model, with its modified parameter set, underestimated mortality in all sites irrespective of weather conditions.