Abstract:
A reliable estimation of Above Ground Biomass (AGB) in Tropical Mountain Forest (TMF)
is still complicated, due to fast-changing climate and topographic conditions, which modifies the
forest structure within fine scales. The variations in vertical and horizontal forest structure are hardly
detectable by small field plots, especially in natural TMF due to the high tree diversity and the
inaccessibility of remote areas. Therefore, the present approach used remotely sensed data from a
Light Detection and Ranging (LiDAR) sensor in combination with field measurements to estimate
AGB accurately for a catchment in the Andes of south-eastern Ecuador. From the LiDAR data,
information about horizontal and vertical structure of the TMF could be derived and the vegetation at
tree level classified, differentiated between the prevailing forest types (ravine forest, ridge forest and
Elfin Forest). Furthermore, topographical variables (Topographic Position Index, TPI; Morphometric
Protection Index, MPI) were calculated by means of the high-resolution LiDAR data to analyse the
AGB distribution within the catchment. The field measurements included different tree parameters
of the species present in the plots, which were used to determine the local mean Wood Density
(WD) as well as the specific height-diameter relationship to calculate AGB, applying regional scale
modelling at tree level. The results confirmed that field plot measurements alone cannot capture
completely the forest structure in TMF but in combination with high resolution LiDAR data, applying
a classification at tree level, the AGB amount (Mg ha??1) and its distribution in the entire catchment
could be estimated adequately (model accuracy at tree level: R2 > 0.91). It was found that the AGB
distribution is strongly related to ridges and depressions (TPI) and to the protection of the site (MPI),
because high AGB was also detected at higher elevations (up to 196.6 Mg ha??1, above 2700 m), if the
site is situated in depressions (ravine forest) and protected by the surrounding terrain. In general,
highest AGB is stored in the protected ravine TMF parts, also at higher elevations, which could only
be detected by means of the remote sensed data in high resolution, because most of these areas are
inaccessible. Other vegetation units, present in the study catchment (pasture and subpáramo) do not
contain large AGB stocks, which underlines the importance of intact natural forest stands.
