A high percentage of Mediterranean forest is covered by dense, low shrub. The difficulty involved in shrub management 
and the lack of information about shrub behavior explain the necessity of developing efficient tools to improve the analysis 
of these areas. The LiDAR system (Light Detection and Ranging) has been widely used in forest studies to estimate 
variables that characterize the forest structure. However, little research has been conducted into shrub vegetation. In these 
studies, previously, a digital elevation model (DEM) is calculated. The goals of this research were: to adapt an iterative 
algorithm to select minimum elevations from LiDAR data in a steep mountain area with shrub vegetation, and to develop 
prediction models for estimating forest variables of shrub vegetation in plots and subplots.

For computing a DEM three parameters were analyzed: window size for selecting minimum elevations, height threshold, 
and input data type. Combining these parameters, several DEM were computed. They were assessed by 1379 ground-
surveyed checkpoints measured with a RTK-GPS system. For estimating the height, the biomass and the volume of shrub 
vegetation several statistics were derived from LiDAR data and a spectral image that were used as independent variables in 
the regression models. The results showed that the lowest root mean square error (RMSE) was obtained for the DEM 
computed with an analysis window size of 10, 5, and 2.5 m, rasterized data as input data, and height thresholds equal to or 
greater than 1.5 m. These parameters produced a RMSE of 0.19 m. Relationships between RMSE and slope, point density 
and vegetation were also analyzed. For these parameters, it has been detected that when terrain slope varied from 0-10% 
to 50-60%, the RMSE increased 0.11 m, and that when point density was increased from 1- 4 to more than 8 points/m2, the 
RMSE decreased up to 0.06 m. From 8 points density upwards, RMSE remained stable. In shrub vegetation areas, RMSE 
increased 0.05 m.

The highest determination coefficients in the estimation of height, biomass, and volume of shrub vegetation were obtained 
when a plot was considered as study area, being the values of R2 of 0.73, 0.77 and 0.84, respectively. For subplots, it was 
found that the best results were obtained when height LIDAR were selected from concentric areas with different radii 
between 1.5 m and 2.25 m, around the point measured at field. These analysis reveal the importance of computing an 
accurate DTM and using a density of LiDAR data greater than 8 points/m2 to estimate with accuracy height, biomass, and 
volume of shrub vegetation in subplots. 

These results show the potential of LiDAR data to characterize shrub structure and make it possible to estimate and map 
the biomass and volume of this vegetation improving the knowledge and the management of this vegetation very often in 
Mediterranean forestry.