Leaf area index inferred from solar beam transmission in mixed conifer forests on complex terrain

Forest process models are used to predict forest growth in a broad range of conditions. Because more than a quarter of the world's forests are in mountainous regions, leaf area index (LAI) must frequently be estimated for forests in complex terrain. LAI is often inferred from canopy light transmission using various canopy models. Choice of a canopy model becomes especially problematic in complex terrain because the complex topography changes the path length of the solar beam through the canopy. For example, if solar elevations are expressed relative to the inclined surface, low elevations can occur on steep slopes at any time of the day, especially on surfaces facing north (in the northern hemisphere). We inferred LAI with various models at 36 plots in a mixed conifer forest in northern Idaho, USA, over a wide range of altitude and solar insolation. We compared these inferences to allometric estimates of LAI. We also tested a theoretical solution for path length on complex terrain. We conclude that the use of this path length correction does not improve agreement between ceptometer LAI and allometric LAI. However, residuals of the fit between transmission and allometric LAI were weakly related to path length (P=0.052) and squared path length (P=0.049) in a quadratic multiple linear regression. The best fits between ceptometer LAI and allometric LAI were both based on extinction coefficients corrected only for solar zenith angle; we recommend the Beer-Lambert model for its simplicity. We analyzed residuals of ceptometer estimates of LAI and determined that they were correlated with species composition. These results suggest that species composition might be used to predict extinction coefficients, though the current data set will not support such prediction. Thus, terrain complexity had only a minor influence on model predictions of leaf area index in these stands, simplifying the prediction of LAI in mountain forests.