HIGH RESOLUTION SATELLITE IMAGERY FOR FORESTRY STUDIES: 
THE BEECHWOOD OF THE PORDENONE MOUNTAINS (ITALY) 
G. Mauro 
University of Trieste and CETA (Centre for Theoretical and Applied Ecology), Gorizia , Italy 
mauro@pug-univ.trieste 
KEYWORDS: Forestry, Estimation, Sampling, Quickbird, Experimental, Advancement 
ABSTRACT: 
In this paper we used high resolution satellite imagery in order to study the Pordenone Mountain forests (Friuli Venezia Giulia - 
North-East of Italy). Our aim is to develop a method able to estimate wood biomass using a vegetation index (NDVI - Normalized 
Difference Vegetation Index). In order to do this we analysed a Quickbird satellite image, dated 24 June 2003, and we compared our 
results with biomass data in weight, cut from two small forest areas. 
The adopted methodology is subdivided in two principal stages: 
I. Checking of the precision of the traditional forestry methods in order to estimate the weight of the wood biomass. These 
methods use dendrometric tablets to estimate wood biomass from measured tree diameters and heights. This stage is 
developed in the following steps: 
a. Identification of two small testing areas (around 300m?) in the beechwood. This forestall typology is the most 
common in this territory. 
b. Collection of some forestry information (height and diameter of plants). 
c. Cutting of all the timbers in the two training areas. The cutting operation has been realized soon after the acquisition 
of the satellite image. 
. Weighing of the cut logs and of the leaves. 
e. Estimation of the wood biomass using some different dendrometric tablets. 
f. Comparison of the biomass weighted data with the tablet-estimated values, in order to verify the precision of the 
standard forestry methodology. 
2. Analysis of the correlation between NDVI and weighted wood biomass. This stage is developed in the following steps: 
a. Rectification of the Quickbird scene. 
b. Analysis of the correlation between NDVI values and weighted biomass data (cut logs and leaves). 
Further developments of this research will employ dendrometric tablet estimated biomass values and NDVI values to improve their 
relationship. When the correlation will be enough satisfying, only the NDVI values will be used to estimate beechwood biomass. 
1. INTRODUCTION To do this we need test correlations among these variables. We 
choose NDVI (Normalized Difference Vegetation Index) as 
1.1. Remote Sensing and Forestall Biomass vegetation index, because it is generally used in this kind of 
research. 
In order to use the forestall biomass as renewable resource of 
energy, we need to know its amount. At the present time we can 2. STUDY AREA 
only do this by measuring with external survey in a systematic 
sampling way. In this case, the instrument used (dendrometric Two small areas test, located in the Pordenone Mountains forest 
tablet) allows predicting the amount of a forestall biomass (Friuli-Venezia Giulia — North East of Italy) (fig.1), have been 
sampled measuring height and diameter of the trees. However, studied in this paper. 
to get global information, this way is too much expensive and it 
takes a lot of time. That's why we don't know the total amount S. 
of our forestall resources (Jodice and Nassimbeni, 1999). E A 
A new way to have an updated database about this renewable es „Klagenfurt 7 
resource could be represented using satellite imagery. In this RS f 
work we use the high-resolution image to try to define the C Jo LTT n 
relationship between wood biomass and a vegetation index. Fri 
enezia m Ljubljana 
Giulia e SLO . 
1.2. Estimate Biomass by Remote Sensing e 
NA 
Many Authors (e.g.: Tucker et al., 1985; Cook et al., 1989; 
Benedetti et al., 1991; Rondeaux et al, 1995; Clevers and 
Leeuwen, 1996; Borfecchia et al, 2001) point out the 
relationship between indexes vegetation (from satelite data) and 
biophysical variables of vegetation, as productivity or healthy. 
We would test the following flux process: 
   
  
Satellite image > Vegetation index > Leaf crown > Forestall ; u.s ; "e ta ant 
; N e E s Figure 1. Pordenone District location (Friuli-Venezia Giulia 
Biomass : 
Region - Italy). 
1074 
In 
  
To 
(NI 
dat: 
(arc 
fore 
con 
ima 
bee 
Frit 
Fur 
con 
WOC 
of t; 
The