Wind kills far more trees in Europe than pests, diseases or drought [26]. In recognition of this problem, the forestry industry has worked with researchers to model wind-damage risk and integrate these models into decision-making processes. By contrast, very little is known about the impacts of wind on natural tropical forests, except for at a few sites in the hurricane belt [27]. Wind speeds are generally lower in equatorial regions than at higher latitudes, but tropical trees tend to be taller and wind damage is still a significant cause of tree mortality [28]. Given the significance of tropical rain forests to global biogeochemical cycles and likely changes in wind flow patterns under climate change, this knowledge gap needs to be addressed. This research will deliver a step change in our understanding of the drivers of tropical rainforest height by bringing together an interdisciplinary team to calculate windthrow risk maps over contrasting tropical landscapes.
Our working hypothesis is that differences in forest height between tropical regions is primarily driven by wind regimes and by fundamental mechanical properties of species, while topography has important local effects. Local variation in rainforest height and dynamics is primarily related to topography, but we postulate that wind has important but hitherto unrecognised secondary effects.
Topography is important because mineral nutrients are transported downhill over millennial timescales and accumulate in low-lying areas. The nutrient rich soils in bottomland create opportunities for fast growth and can support tall forests. We contend that wind is also a major contributor to forest height patterns, via its spatially variable impact on mortality (See Table 1). We predict that trees in bottomlands are more susceptible to windthrow, because their trunks grow tall and narrow in response to intense competition for light in these nutrient-rich sites, making them mechanically weak when storms occur.
Table 1: Hypothesised influences of topographic position and wind on growth and mortality of trees, demographic processes that ultimately determine forest height.
                       Nutrient-depleted hill               Nutrient-enriched bottomlands
Exposed     Slow growth + high mortality →       Fast growth + high mortality →
                    Short forest                                    Intermediate height
Sheltered    Slow growth + low mortality →        Fast growth + low mortality →
                     Intermediate height                        Tall forest

Our project will use repeat-survey ALS in Sabah (Malaysia) and French Guiana lowland rainforests to measure the height, growth and mortality of at least 200,000 tall trees (>30 m), providing us with an unprecedent dataset with which to evaluate these hypotheses. We will coordinate activities with Prof Uriarte, who ALS-surveyed Puerto Rican forests immediately before and after hurricane Maria. Forest height in these three regions is known to vary strongly with topography. As shown in Fig. 3, bottomlands in our Sabah site support the tallest of all tropical forests (>60 m) growing within a kilometre of stunted heath forests on hilltops (~20 m). Similar correlations between topography and height are apparent in French Guiana and Puerto Rico, but there the tallest forests reach only 30-40 m in height. These contrasts make these landscapes ideal for evaluating how topography and wind affect forest height and dynamics in tropical regions.
Evaluating the interplay between topography and wind on forest height requires an interdisciplinary approach, calling on expertise in wind modelling, tree biomechanics and ALS to produce wind susceptibility (critical wind speed) maps. We expect tree height, growth and mortality to be correlated with critical wind speeds – and with topography - in these rainforests, which we can address with regression analyses. Finally, storm simulations will be used to assess whether differences in wind regime drive the marked differences in forest height and dynamics observed among the regions.