Sampling effort and uncertainty in leaf litterfall mass and nutrient flux in northern hardwood forests

Designs for litterfall sampling can be improved by understanding the sources of uncertainty in litterfall mass and nutrient concentration. We compared the coefficient of variation of leaf litterfall mass and nutrient concentrations (nitrogen, phosphorus, calcium, magnesium, and potassium) at different spatial scales and across years for six northern hardwood species from 23 stands in the White Mountains of New Hampshire, USA. Stands with steeper slopes (P = 0.01), higher elevations (P = 0.05), and more westerly aspect (P = 0.002) had higher interannual variation in litter mass, probably due to a litter trap design that allowed litter to blow into traps in windy years. The spatial variation of nutrient concentrations varied more across stands than within stands for all elements (P < 0.001). Phosphorus was the most spatially variable of all nutrients across stands (P < 0.001). Litter nutrient concentrations varied less from year to year than litter mass, but the magnitude of difference depended on the element and tree species. We compared the relative importance of variation in mass vs. concentration to estimates of nutrient flux by simulating different sampling intensities of one while holding the other constant. In this dataset, interannual variability of leaf litter mass contributed more to uncertainty in litterfall flux calculations than interannual variation in nutrient concentrations. Optimal sampling schemes will depend on the elements of interest and local factors affecting spatial and temporal variability.