Table 3. Summary statistics of lineament analysis on spring mound distributions with different directional 
filters. 
Filter direction 
No. of 
lineaments 
identified 
Lineament distance 
(m) 
min. max. mean 
Dominant 
directions 
Minor 
directions 
None 
42 
0.90 
8.00 
2.89 
NE; E-SE 
N-NNE 
N 
48 
0.49 
7.99 
2.52 
E-SE 
ENE 
NW 
26 
0.89 
5.83 
2.86 
ESE; SE-SSE 
- 
W 
36 
0.56 
7.68 
2.31 
E-ESE; SE 
N-NNE 
SW 
32 
0.30 
5.80 
1.91 
SE-SSE 
NNE; ENE; ESE 
None* 
12 
0.72 
7.19 
3.07 
ESE 
* Only known rock outcrops analysed. 
directions parallel to the long axis of the spring 
mound area than across it. This will become apparent 
in the results to a certain extent. After all 
lineaments have been identified full statistics 
(starting and finishing pixel coordinates, distance 
in Km and directions) are tabulated; in addition 
summary statistics are displayed as a semi-circular 
rose diagram. 
The statistics summarising the results of the 
lineament analyses with the different directional 
filters are shown above (Table 3). The number of 
lineaments detected with different treatments ranged 
from 26 to 48. The length statistics of the 
lineaments was quite similar under all filters except 
the SW filter. The smallest lineament was always 
less than 1 Km; ranging from 300m in the SW filter 
image to 900m in the unfiltered image. The longest 
lineament detected under different filters fell into 
two internally consistent groups. The NW and SW 
filtered images had longest lineaments of 5.83 and 
8.8 Km respectively. The other group consisted of 
the N and W filtered and unfiltered images; here the 
longest lineament ranged from 7.68 Km (W filtered 
image) to 8.00 Km (unfiltered image). The directional 
data on the lineaments was more useful than the number 
of lineaments and their lengths in assessing the 
effects of the different filters and for geological 
interpretation. The directional data are tabulated 
(Table 3) and summarised in semi-circular rose 
diagrams, (Fig. 5) 
The effect of directional filtering on the Band 3 
imagery seems to have a marked effect on the linea 
ments in the N to NNE sectors, when compared to the 
unfiltered image. In the latter image NE was a 
dominant lineament direction, but in the filtered 
images all of the lineaments with directions 
between N and ENE were minor when compared to those 
with directions between E and SSE. Generally 
however it can be seen from Table 3 and Fig. 5 
that the dominant lineament directions fell between 
E-W and SE-NW and that a secondary direction - NNE- 
SSW - perpendicular to the dominant direction can 
be identified. The other directions, N-S; ENE-WSW 
and SSE-NNW, are unimportant. This information is 
summarised in Figure 6. 
5. DISCUSSION 
It was suggested earlier that the distribution of 
spring mounds and aioun in the Chotts el Djerid and 
el Fedjadj might be related to Alpine folding and 
faulting of sedimentary strata on the Saharan 
Platform. This can be examined by comparing the 
directions of the lineaments to the structural trends 
in the region. 
5.1. Structural controls on spring mound distribution 
If underlying geological structure controls the 
distribution on spring mounds on the Chott el 
Fedjadj the distribution of lineaments should reflect 
the stress patterns associated with the folding of 
the Chott el Fedjadj anticline. Reconstructed 
folding of this anticline (Fig. 7a) suggests that 
two parallel E-W fold axes were present during 
folding. Stresses in the brittle strata of the 
region associated with the folding would probably 
have created a series of faults parallel to the 
fold axes, i.e. approximately E-W. The present-day 
geological structure of the Chott el Fedjadj is 
similar to the dome and basin structural association 
suggested by Hobbs et al., (1976) which would fit 
the folding and faulting patterns. This can partly 
be seen in Fig. 7b. This shows that the Chott el 
Fedjadj anticline has been breached as a result of 
two major faults parallel to the fold axis which 
have led to the downfaulting of the sediments in 
the crest of the southernmost fold. Undoubtedly 
other minor parallel faults exist within the 
structural association, particularly tensional 
faulting in the upper strata of the fold where the 
hinge angles were far less acute than in the deeper 
strata. Faulting parallel to the fold axes probably 
therefore partially accounts for the dominant 
direction of alignments of spring mounds in the area. 
However in the core of the anticline, under 
Quaternary playa sediments, sedimentary strata, 
undercrop with a strike parallel to the fold axes 
as well. Some of these horizons particularly the 
sandstones, are known aquifers. It is likely 
therefore that some of the linear alignments relate 
to spring lines along junctions between aquifers 
and aquLcludes under the playa sediments. The strata 
subcrop is also probably responsible for the extent 
of the spring nound field in the south-west Chott 
el Fedjadj. Particularly the fact that it is found 
only in the southern part of the chott and that there 
is no equivalent to the north. 
The secondary lineament direction is orthogonal to 
the main direction. It is a well known physical 
phenomena that jointing and faulting patterns related 
to stress release occur in brittle materials 
perpendicular to the main stress directions (Park, 
1983; Whalley, 1976). If this is the case in this 
area on a large scale then the secondary lineament 
direction can also be explained in terms of the 
folding of the Chott el Fedjadj anticline. There is 
evidence to support this hypothesis along the 
southern limb of the anticline. Here the resistant 
dolomitic limestones which form a number of cuestas, 
the highest of which is the Djebel Tebaga, display 
a regular series of wind gaps orientated 
perpendicular to the fold axes and parallel to the 
secondary lineament direction (Fig.3). These can be 
seen as continuations of the lineaments on the 
imagery and therefore are related to the hypothesised 
faulting and jointing beneath the playa sediments. 
610