Geology Reference
In-Depth Information
Its greatest thickness is about 50 m and Dullo
(1983) recognized 10 microfacies types (bioclas-
tic algal debris facies, bioclastic rhodolite debris,
bioclastic algal mollusc, bioclastic, baffl estone,
foraminiferal algal mollusc, pavement, fora-
miniferal rhodolite, foraminiferal algal debris,
foraminiferal), which are broadly comparable
to modern Arabian Gulf facies (Riegl & Piller,
2000). These microfacies types were also applied
to similar rocks in the Styrian basin by Friebe
(1990, 1991), who grouped the Leitha Limestone
together with other shallow-water sediments into
the 'Weissenegg Formation'. However, the term
Leitha Limestone is more widely recognized.
A characteristic element is coralline algae that
occur as rhodolite facies or calcarenites consist-
ing mainly of branch fragments. Coral buildups
of limited size occurred only locally as patch
reefs (Piller & Kleemann, 1991) and biostromes
(Riegl & Piller, 2000). The best developed coral
buildups are present at the southern tip of the
Leitha Mountains where the limestones reach
their greatest spatial extent and thickness (about
50 m). During the Badenian, the tips of the Leitha
mountains extended above sea level as a chain of
islands (Fig. 1c). Due to the islands' position and
absence of continuous terrigenous infl ux, coral
growth was relatively prolifi c. The position in a
marginal basin at high latitude suggests a sim-
ilar temperature regime to today's Arabian Gulf
(Riegl & Piller, 2000) with repeated death of corals
due to temperature extremes.
Purkis & Pasterkamp (2004) after masking of
pixels that were outside the known depth-
resolution of the images (~8 m). Extensive ground-
truthing illustrated that the classifi ed Ikonos
map had an overall accuracy of 81%, and the
classifi ed Landsat map had an overall accuracy
of 72%. Many errors occurred in distinguishing
between algae and seagrass, which could not be
satisfactorily resolved. Some errors in classifi ca-
tion existed between coral bioherms/biostromes
and algae, since dead coral areas (many had been
killed in thermal anomalies of 1996-1998; Riegl,
1999; Purkis & Riegl, 2005) had dense algal over-
growth. Accuracy differences between the images
were largely due to different spatial resolution
(pixel size) and the size of the studied areas. Also
bathymetry was extracted from the Landsat image
using a modifi cation of the algorithm by Stumpf
et al
. (2003).
At the Fenk Quarry in Burgenland, Austria, two
complete top-to-bottom and several incomplete
sections in the Badenian Leitha Limestone were
measured at a centimetre-scale to obtain a good
understanding of present facies and then, to the
highest possible resolution, the entire outcrop was
mapped since Lehrmann & Rankey (1999) stress
the importance of incorporating two-dimensional
information into Markov chain analysis. The
measured sections were superimposed on a
photograph of the outcrop and facies units were
followed laterally for the production of polygons
showing the outlines of facies over the entire out-
crop. This outcrop was chosen because Riegl &
Piller (2000) previously demonstrated the equi-
valence of the Leitha Limestone to the modern
Arabian Gulf by qualitative facies analysis.
METHODS
Image basis for landscape analysis
Ecological background data
The basis of all optical classifi cations and maps
were spaceborne images (Ikonos and Landsat
ETM7 11-bit multispectral satellite image, 4 m
and 30 m pixel-size respectively). Details about
image processing, ground-truthing and the map-
ping approach are given in Purkis (2005) and
Purkis
et al.
(2005). Full bathymetric datasets
for depth correction of spectral refl ectance val-
ues of the Ikonos were obtained from acoustic
surveys and were tidally corrected against data
obtained from
in situ
loggers deployed during
fi eld work. The bathymetry data were then used
for calibration of the Landsat image. Image clas-
sifi cation was conducted using an unsupervised
k
-means clustering routine and the multivariate
normal probability driven classifi er described by
Data regarding the biology of the system include
series of 155 line transects of 10 m and 50 m taken
in 1995, 1999, 2000, 2002 and 2005. One hun-
dred and eighty point observations for ground-
truthing maps and satellite imagery, noting the
species composition and taphonomic status of
1
2 m sample areas, were taken in 1995,
1996, 1998, 2000, 2002, 2003, 2005. Taphonomic
status and breakdown sequence were evaluated
by sampling corals and skeletons in 1995, 1996,
1999, 2000 and 2005. Since these data are already
published elsewhere (Riegl, 1999, 2002; Purkis &
Riegl, 2005; Purkis
et al.
, 2005), they are not shown
in detail but used here as background information
for the calibration of models.
1 m or 2
