No abstract

[1] LiDAR data are increasingly available from both airborne and spaceborne missions to map elevation and vegetation structure. Additionally, global coverage may soon become available with NASA's planned DESDynI sensor. However, substantial challenges remain to using the growing body of LiDAR data. First, the large volumes of data generated by LiDAR sensors require efficient processing methods. Second, efficient sampling methods are needed to collect the field data used to relate LiDAR data with vegetation structure. In this paper, we used low-density LiDAR data, summarized within pixels of a regular grid, to estimate forest structure and biomass across a 53,600 ha study area in northeastern Wisconsin. Additionally, we compared the predictive ability of models constructed from a random sample to a sample stratified using mean and standard deviation of LiDAR heights. Our models explained between 65 to 88% of the variability in DBH, basal area, tree height, and biomass. Prediction errors from models constructed using a random sample were up to 68% larger than those from the models built with a stratified sample. The stratified sample included a greater range of variability than the random sample. Thus, applying the random sample model to the entire population violated a tenet of regression analysis; namely, that models should not be used to extrapolate beyond the range of data from which they were constructed. Our results highlight that LiDAR data integrated with field data sampling designs can provide broad-scale assessments of vegetation structure and biomass, i.e., information crucial for carbon and biodiversity science.

We simulated a shelterwood forest regeneration treatment by reducing basal area, and monitored the response of an avian community in oak-hickory forest on the southern Cumberland Plateau, northern Alabama, USA. We used five treatments: control (no removal), clear-cut (100% removal), and 25, 50, and 75% removal of basal area. Territory mapping was used to quantify bird community between mid-April and July of both 2002 and 2003. Microclimate variables were recorded at each plot. The residual basal area and canopy cover showed three distinct conditions after treatment: closed canopy, open forest, and clear-cut. The microclimate varied among treatments: air temperature was highest in clear-cut plots and lowest in control plots, whereas soil moisture had the opposite pattern. A total of 71 bird species were detected, with 36 of them defending territories. Territory density, species richness, and Shannon diversity index differed among the treatments; the relationship between these bird community indices and the level of basal area removal was quadratic, lowest in the clear-cut plots and highest in the intermediate levels. Although species richness was similar among the control, 25, 50, and 75% removal treatments, species composition varied. The richness difference among treatments became smaller in the second year post-treatment (2003) with an increase in bird density and richness occurring in the clear-cut plots.

Light detection and ranging (LiDAR) is increasingly used to map terrain and vegetation. Data collection is expensive, but costs are reduced when multiple products are derived from each mission. We examined how well low-density leaf-off LiDAR, originally flown for terrain mapping, quantified hardwood forest structure. We measured tree density, dbh, basal area, mean tree height, Lorey's mean tree height, and sawtimber and pulpwood volume at 114 field plots. Using univariate and multivariate linear regression models, we related field data to LiDAR return heights. We compared models using all LiDAR returns and only first returns. First-return univariate models explained more variability than all-return models; however, the differences were small for multivariate models. Multiple regression models had R2 values of 65% for sawtimber and pulpwood volume, 63% for Lorey's mean tree height, 55% for mean tree height, 48% for mean dbh, 46% for basal area, and 13% for tree density. However, the standard error of the mean for predictions ranged between 1 and 4%, and this level of error is well within levels needed for broad-scale forest assessments. Our results suggest that low-density LiDAR intended for terrain mapping is valuable for broad-scale hardwood forest inventories.