Using a Multistructural Object-Based LiDAR Approach to Estimate Vascular Plant Richness in Mediterranean Forests With Complex Structure

Abstract

Artículo de publicación ISIA multistructural object-based LiDAR approach to predict plant richness in complex structure forests is presented. A normalized LiDAR point cloud was split into four height ranges: 1) high canopies (points above 16 m); 2) middle-high canopies (8-16 m); 3) middle-low canopies (2-8 m); and 4) low canopies (0-2 m). A digital canopy model (DCM) was obtained from the full normalized LiDAR point cloud, and four pseudo-DCMs (pDCMs) were obtained from the split point clouds. We applied a multiresolution segmentation algorithm to the DCM and the four pDCMs to obtain crown objects. A partial least squares path model (PLS-PM) algorithm was applied to predict total vascular plant richness using object-based image analysis (OBIA) variables, derived from the delineated crown objects, and topographic variables, derived from a digital terrain model. Results showed that the object-based model was able to predict the total richness with an r(2) of 0.64 and a root-mean-square error of four species. Topographic variables showed to be more important than the OBIA variables to predict richness. Furthermore, high-medium canopies (8-16 m) showed the biggest correlation with the total plant richness within the structural segments of the forest.CONICYT through Integration of Advanced Human Capital into the Academy Project 79110001

The spectral variability hypothesis does not hold across landscapes

“…While these links sometimes allow good local estimates, the approach is hardly transferable in space and time and depends on calibration on the spot. The correlative approach was found to work in all major domains of remote sensing like multispectral remote sensing (e.g., Foody & Cutler 2003, 2006, Feilhauer & Schmidtlein 2009, Rocchini 2007, hyperspectral remote sensing (e.g., Laurin et al 2014) as well as active remote sensing or combinations (e.g., Camathias et al 2013, Higgins et al 2014, Lopatin et al 2015, 2016, Simonson et al 2012, Hernández-Stefanoni et al 2014. Reviews can be found for example in Wang et al (2010) and Rocchini et al (2010).…”

Above ground biomass and tree species richness estimation with airborne lidar in tropical Ghana forests

“…All the previous species richness estimations are based on lidar metrics related to vegetation height variations, habitat heterogeneity or topographic information, independently from the forest type under examination. Besides the indices useful in the studies already mentioned above (Ceballos et al, 2015;Hernandez-Stefanoni et al, 2014;Simonson et al, 2012), indices related to vegetation height and structural complexity were used by Lucas et al (2010), Lopatin et al (2015), and Wolf et al (2012); while altitude above sea level, standard deviation of slope and mean canopy height were the most important predictors in the Lopatin et al (2016) research.…”

“…However, the use of lidar for tree species richness estimation has been tested in an exiguous number of studies. Successful results were obtained in marsh, meadow and woodland habitats in Mississippi (Lucas et al, 2010), in Mediterranean forests (Lopatin et al, 2015;Lopatin et al, 2016;Simonson et al, 2012), where lidar also outperformed hyperspectral data for species richness estimation (Ceballos et al, 2015); and in two tropical forest cases (Hernandez-Stefanoni et al, 2014;Wolf et al, 2012). Lopatin et al (2016) argued that lidar can be used to derive three types of information that interacts with plant species richness: micro-topographical, macro-topographical and canopy structural information.…”

A hybrid partial least squares and random forest approach to modelling forest structural attributes using multispectral remote sensing data

“…The PLSR linear multivariate model is useful for analysing datasets with many high dimensional and collinear predictors (Wold et al, 2001). The PLSR model creates orthogonal (uncorrelated) weight vectors by maximising the covariance between the explanatory and response variables while reducing the dimensionality of these x variables by sifting out the factors that explain the most information between all the x and y variables (Lopatin et al, 2015). …”