Background/Question/Methods
Light availability is known to affect recruitment, growth, and survival in tropical forest tree species. However, the effects of light limitation on tree species distributions in tropical forests are not well understood. The challenge has been that spatial patterns of irradiance beneath forest canopies, which determine where seedlings of different species can establish and grow quickly enough to survive, are difficult to characterize. Evidence exists that understorey light varies spatially beneath closed canopy, but the patterns of variation (and their mechanisms) remain poorly understood. Spatial variation in understory light could, however, simply follow spatial patterns of irradiance at the top of the canopy. Light does pass through forest canopies, so understoreys should be brightest where canopy irradiance is highest – canopies and understoreys should both be brighter on equatorial slopes, for example, than on polar slopes. We investigate this hypothesis at the Mo Singto forest dynamics plot in Thailand, by assessing correspondences between the spatial pattern of canopy irradiance and the distributions of understorey trees (assumed, conservatively, to be those < 10 cm DBH) among 105 species with at least 50 understorey individuals. We estimate the spatial pattern of canopy irradiance (cumulative annual watt-hours/m²) at Mo Singto using algorithms that incorporate the solar path, topographic position, and shading by surrounding relief.
Results/Conclusions
Randomization tests based on Monte Carlo simulations reveal that, in 63 of the 105 tree species studied, understorey individuals have strongly non-random distributions (at p < 0.001) with respect to the spatial pattern of canopy irradiance. Nineteen species are associated with low canopy level light – average canopy irradiance above understory trees in each of these species is > 3 standard errors lower than expected by chance. Forty four species are associated with high canopy level light – in these species, average canopy irradiance above understory trees is > 3 standard errors higher than expected by chance. We might reasonably conclude, given the correspondence between spatial variation in canopy irradiance and understorey tree distribution in 60% of species, that at Mo Singto spatial patterns of irradiance in the understorey follow those at the top of the canopy. Mo Singto is a relatively open forest, however, so during the next year we will replicate our analyses for other tropical forest dynamics plots at sites with varying canopy architecture.