(5) in the counties not completely occupied 
by the LUMZ, the part that is closest to 
Lisboa is connected by the LUMZ, 
the remaining part of the 
while 
county is 
characterised by a large number of small 
and scattered UMZs. 
For a detailed characterisation of the UMZs 
of AGL in 1984 we present Table 2. The most 
important characteristics of the 1984 UMZs 
are the following: 
in 1984 there are 182 UMZs; 
the area of two whole 
Lisboa and Oeiras, 
UMZ, the LUMZ; 
89% of the Amadora county belongs to 
UMZs, and this county has a low number 
of UMZs; 
a: little bit more -than half of- the 
Cascais county belongs to UMZs, but the 
number of UMZs is relatively large (11 
UMZs) ; 
in three counties, Loures, Sintra and 
V.F.: de Xira, less than 40 $ of. the 
county belongs to UMZs. These counties 
are also characterised by a large number 
of UMZs. 
counties, i.e. 
belong to a single 
Table 2. Characteristics of the 1984 UMZs 
  
  
  
  
  
  
  
  
of AGL. 
County Percentage of the | Number of 
county area UMZs 
occupied by UMZs 
V.F. de 1352 35 
Xira 
Loures 39.9 52 
Sintra 2755 79 
Amadora 89.2 3 
Lisboa 100 1 
Cascais 62.2 11 
Oeiras 100 1 
  
  
  
  
  
For a detailed characterisation of the UMZs 
of AGL in 1995 we present Table 3. The most 
important characteristics of the 1995 UMZs 
are the following: 
696 
in 1995 there are 274 UMZs; 
the area of three whole counties, i.e. 
Lisboa, Oeiras and Amadora, belong to a 
single UMZ, the LUMZ; 
most of the Cascais county (71%) belong 
to. UMZs, but. the. number. of  .UMZ is 
relatively large (19 UMZs); 
in three counties, Loures, Sintra and 
V.F. de. Xira, less than 50 % of the 
county belongs to UMZs. These counties 
are also characterised by a large number 
of UMZs. 
Table 3. Characteristics of the 1995 UMZs 
  
  
  
  
  
  
  
  
of AGL. 
County Percentage of Number of 
the county area UMZs 
occupied by UMZs 
V.F. de 15 62 
Xira 
Loures 47 70 
Sintra 34 120 
Amadora 100 1 
Lisboa 100 1 
Cascais 71 19 
Oeiras 100 1 
  
  
  
  
  
A comparison of the UMZs for 1984 and for 
1995 reveals a difference on the area of 
the AGL occupied by UMZs and on the number 
of UMZs, namely: 
In 
to 
to 
the area of the UMZs increased, because 
many rural areas in 1984 become built-up 
areas in 1995; 
the number of UMZs increased from 1984 
to 1995, revealing new built-up areas 
that are not connected to the ones that 
already existed in 1984. 
should say that this increase may be 
overestimated because of the methodology 
used for 
However, one 
generating the binary mask 
urban/rural. In both years, we used as 
reference the  urban/rural mask from 
1997, which was generated by an 
automatic procedure. The 1991 
rural/urban mask was then updated, by 
to generate the 1984 and 
1995 masks by overlaying the 1991 mask 
over the 1984 and 1995 satellite images. 
In the generation of the 1984 mask, many 
small urban areas were removed from the 
mask because 
visual means, 
could not be 
most of the small 
urban areas in 1991 were maintained to 
1995. For this reason, the number of 
UMZs may be overestimated in 1995; 
many areas that became built-up in 1995 
were located between different UMZs in 
1984. As a consequence, distinct UMZs in 
1984 were aggregated and became a single 
UMZ in 1995. 
they 
identified. However, 
relation to the UMZs expansion from 1984 
1995, at a county level, it is important 
point out the following: 
the LUMZ expanded greatly from 1984 to 
1995; 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, Budapest, 1998 
CH à Sir 5 
ci 
a0 un » Q0 
In 
for 
EURC 
metr 
proc 
were 
the 
algc 
prir 
app] 
the 
reve 
buil 
This 
mair 
rat! 
Thi: 
EUR( 
Sta: 
Bari 
~ © In