icates no 
polygons 
nes. The 
1981 fire 
moderate 
‘compare 
4. DESERTIFICATION RISK ASSESSMENT 
The risk of desertification at a site increases the longer it 
remains without a protective vegetation cover after a fire 
and the higher its susceptibility to soil erosion. Therefore, 
desertification risk is assessed by combining the 
estimates for potential for natural regeneration (PNR) 
and risk of soil erosion (ROE). 
4.1. Potential for Natural Regeneration 
The main restrictive factor for natural regeneration after a 
fire is the availability of water for plant regrowth in the dry 
period. In addition to the amount of rainfall, this depends 
on the water storage capacity of the soil and the 
topographic aspect for a given site. Deep soils generally 
store more water and carry denser vegetation than 
shallow soils, and south-facing slopes receive higher 
amounts of solar radiation than north-facing slopes and 
are therefore drier. 
Based on decision rules provided by NARF, the following 
decision table was used to estimate the potential for 
natural regeneration. 
Table 2: Potential for Natural Regeneration Decision 
  
  
  
  
  
water storage capacity of an area. However, this 
estimate is based on the assumption that the maximum 
possible biomass for a given soil is present, which is only 
the case when a forest / wildland has not been affected 
by fires for a long period of time. For the test areas in 
Attica, the maximum cover estimate was based on the 
maximum vegetation cover derived from TM scenes from 
1984, 1987, 1990, and 1993, which were acquired in the 
dry season of the year. The algorithm was specifically 
developed to estimate the water storage capacity in 
wildlands / forests growing in arid and semiarid regions, 
and gives no useful estimate for agricultural or urban 
areas. 
The accuracy was assessed by comparing the obtained 
PNR estimates for 14 validation sites with field-based 
estimates of NARF. For 13 of the 14 validation sites, the 
estimates correspond within + 1 category, and one site 
giving a discrepancy of 2 categories (for category 
characterisation see table 2). 
4.2. Risk of Soil Erosion 
After a wildfire, the vegetation cover that protects the 
underlying soil substrate is removed, thus exposing the 
soil to the forces of wind and rain and, thereby markedly 
increasing its ability to be denuded. The following main 
factors govern the risk of soil erosion in a burnt forest / 
e Permeability 
e Water storage capacity 
e Topography 
  
  
  
Permeability is mainly determined by surface geology 
and was grouped into permeable rocks (hard limestones, 
Rules. 
wildland: 
Water Storage | Aspect PNR 
Capacity Class | Class | Characterisation 
moderate to North 1 no limitation 
high South 2 slight limitation 
low North 2 slight limitation Permeability: 
South 3 moderate limitation 
very low North 4 strong limitation 
South 5 severe limitation 
  
  
  
  
  
  
The problem of obtaining water storage capacity data: 
The water storage capacity of a soil depends on its type 
and depth. As no detailed soil maps are available for 
most wildlands / forests in the Mediterranean region, 
other information sources had to be considered. An 
approach, using ERS-1 data was not successful, due 
mainly to the complex signal response, which depends 
on soil grain size, distribution of dead branches after a 
fire, and rocks-stones-bare soil distribution (Banninger et 
al, 1994). As the water storage capacity is a main factor 
for Potential of Natural Regeneration and Risk of Soil 
Erosion assessment, a rough estimate, using TM 
Imagery acquired in the dry period, was applied. In arid 
and semiarid regions, the amount of green biomass in 
the dry period is mainly dependent on the quantity of 
water available to the plants. Since there is a high 
correlation between MSAVI and green biomass or 
vegetation cover, MSAVI-derived values from imagery 
acquired in the dry period can be used to estimate the 
243 
calcareous Tertiary deposits, siliceous Tertiary deposits, 
and colluvium), and impermeable rocks ( mica schists 
and other metamorphic rocks). 
Water Storage Capacity: 
The approach used to obtain this parameter is described 
in section 4.1 (Potential for Natural Regeneration). 
Topography: 
For the derivation of this parameter, the topographic 
factor (LS-factor), as defined by the Universal Soil Loss 
Equation (USLE) and modified by Hensel (1991), was 
used. 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996