Within a forest canopy, plant productivity depends on two suites of factors: 1) the functional diversity of plants at the level of functional type, species, individual, or leaf, and 2) the physical structure or canopy architecture, and therefore the within-canopy light environment, of the forest. Predictive models of plant productivity, from ‘green slime’ to ‘big leaf’ to multi-layer as well as size and age-class structured demographic models all assume that photosynthetic rates can be lumped into generalized classes (‘plant functional types’) and that explicit handling of the three-dimensional structure of the forest is not essential to estimating its productivity. Yet we know these assumptions are inadequate – important plant functional traits like foliar nitrogen concentrations ([N]L) and leaf mass per area (LMA) can vary significantly within a single species and through a canopy, where the radiation regime and the amount of light a leaf receives is not due to its general canopy position but is due to the locations of the leaves and branches that surround it.
Here we use hyperspectral imagery (HSI) and LiDAR point returns from the National Ecological Observatory Network’s (NEON’s) Airborne Observation Platform and NASA’s G-LiHT system to model the three dimensional distribution of [N]L and LMA through the forest canopy.
Results/Conclusions
This pilot study uses data from two NEON sites – the Smithsonian Environmental Research Center (SERC, Maryland, USA) and Harvard Forest (HARV, Massachusetts, USA). We show that LAD above a given sample location can be used to predict LMA, though relationships are somewhat site and species specific. Multiple regression models using HSI-derived top of canopy [N]L improve the relationship. Analyses of landscape patterns of total leaf mass and total canopy nitrogen follow expectations from other studies. We estimate total foliar carbon and total foliar nitrogen for these two sites and compare them to estimates from the Community Land Model. Overall these preliminary results suggest that HSI and LiDAR data can be combined to improve estimates of leaf traits within and among forest canopies.