National Land Cover Database (NLCD) is one of most widely used GIS datasets for mapping and quantifying wildlife habitat, biological diversity, and ecosystem services in the US. However, current land cover classification scheme is often too coarse for many in-depth habitat modeling. For example, there are only three broad forest types (deciduous forest, coniferous forest, and mixed forest) delineated in the NLCD, limiting its use for many forest-dwelling species that require specific forest structural/compositional characteristics, which cannot be readily extracted from the NLCD dataset. With the recent technological developments, many states have begun to offer publicly available state-wide LiDAR point cloud data. Such LiDAR datasets may be beneficial to deriving forest canopy height and cover metrics relevant to wildlife habitat modeling. In this study, we explore the potential to integrate the LiDAR point cloud data and NLCD to map the habitat for ruffed grouses, a state species of conservation that thrive in areas of 5 to 20-year-old forest growth in eastern Kentucky. We use freely available packages in R software for downloading, manipulating and analyzing both datasets. We also apply the principles in landscape ecology to characterize the habitat in terms of spatial mosaics of different land cover and forest structure types.
Results/Conclusions
Because most state-wide LiDAR program were initiated for the purpose of terrain mapping, the point cloud datasets were often acquired during leaf-off season with relatively low point density compared to the LiDAR dataset for the forest mapping. Our analysis show that such dataset is not suitable for estimating canopy coverage and canopy tree density, but the estimation of canopy height is still very robust at a fine spatial resolution. When integrating the estimated canopy height with NLCD data, we are able to accurately identify young forest throughout the entire region. Due to lack of active forest management, ruffed grouse habitat is generally lack in eastern Kentucky, with scattered hotspots driven by past disturbances such as mining/reclamation, harvesting, and tornado. Our study highlights the use of freely available big data and geocomputational software programs in mapping specific wildlife species at regional scales.