Strong temporal variation in canopy disturbance rates in a tropical forest in relation to rainfall and windspeed: results from 5 years of monthly drone data for a 50 ha plot

Abstract:

Natural tree mortality determines changes in structure, diversity and carbon storage of tropical forests and is influenced by environmental and biotic drivers. Assessing shifts in mortality and identifying their causes is limited mostly by the size and distribution of permanent plots, as well as the long interval between re-measurements. Here, we used approximately monthly drone images and the digital aerial photogrammetry products to analyze temporal variation in canopy disturbances creation in the 50 ha forest dynamics plot on Barro Colorado Island, Panama, since October 2014. We identified new canopy disturbances by comparing successive canopy elevation model image pairs and inspecting the associated orthomosaics. We also classified disturbances as being due to either treefalls, branchfalls or decomposition of standing dead trees. We analyzed the temporal variation in canopy disturbance rates with respect to 15-minute rainfall and horizontal windspeed data from BCI. Specifically, we tested the frequency and summed intensity of events above the 90^th to 99^th percentiles (in 0.1 percentile increments) at 15-minute, 1-hour and 1-day scales. Results demonstrated strong temporal variation, with only a few months accounting for a majority of the total disturbances. There was a trend of higher canopy disturbance during wet seasons. The single highest disturbance rate was observed between June 1 and July 13, 2016, when 2.3% of the canopy was disturbed in just 42 days. Treefalls, branchfalls and decomposition of standing dead trees accounted for 71%, 22% and 7% of the total canopy disturbance area, respectively. The variation in the frequency of hourly rainfall events above the 99.4^th percentile was the best predictor, with an R² equals to 0.45, and the windspeed events had no significant relation. This means that extreme rainfall events matter to creation of canopy disturbances in BCI. These results can contribute to the long-term monitoring of canopy disturbances and the development of a predictive understanding of how tropical forests will respond to a future climate change.