background image


Paleoecological development of the Neogene Sofia Basin

(southwestern Bulgaria) based on sedimentological and

paleontological evidence








Institute of Geology, Bulgarian Academy of Sciences, Acad. G. Bonchev str. 24, 1113 Sofia, Bulgaria;


Institute of Zoology, Bulgarian Academy of Sciences, Tsar Osvoboditel Blvd. 1, 1000 Sofia, Bulgaria

(Manuscript received June 16, 2006; accepted in revised form June 13, 2007)

Abstract: Comprehensive lithostratigraphic and biostratigraphic investigation of diatoms, chrysophycean stomatocysts,
and mollusc fauna was performed on a borehole C-14, near the village of Katina, from the peripheral zone of the Sofia
Neogene Basin. The diatom flora consisted of 302 species, varieties and forms. Principal Component Analysis (PCA)
was used to classify the diatom thanatocoenoses. The results of these different analyses presented here indicated three
successive stages in the paleoenvironmental development of the Sofia Basin, at the end of the Miocene and the
beginning of the Pliocene.

Key words: Late Miocene—Early Pliocene, Sofia Basin, paleoecology, sedimentology, siliceous microfossils, PCA.


The use of paleolimnological approaches to reconstruct
past environmental changes has increased rapidly in the
past 10 years. Analysis of biotic (e.g. diatoms, chryso-
phycean stomatocysts) and abiotic proxies (geochemistry,
sedimentology) recorded in lacustrine sediments allow the
reconstruction of environmental conditions and the im-
pact of these on aquatic systems in the past (e.g. Charles &
Smol 1994; Bradbury 1999; Smol & Cuming 2000). Re-

paleoecological and environmental history of the Sofia
Neogene Basin.

Geological setting

The Sofia graben, located in the southwestern part of

Bulgaria (Fig. 1), is 80 km long and more than 20 km wide.
The basin is filled with Neogene sediments that are cov-
ered by Quaternary deposits. A great diversity of magmat-

cently paleolimnology has rapidly developed as
a quantitative science, and new multivariate
techniques have been used largely to reconstruct
lake evolution.

In a series of studies we set out to demonstrate

the applicability of different methods used for re-
construction of paleoecological conditions in
the Sofia Basin during the period of the Late Mi-
ocene—Early Pliocene. The main steps carried
out were:

• Grouping the samples with similar diatom

assemblages (thanatocoenoses) to assess the pa-
leolimnological impact, using the minimum-
variance clustering (Ward’s method) based on
squared Euclidean distance (Ognjanova-Rume-
nova & Popova 1996);

• Reconstruction of the paleoecological

changes, inferred from siliceous microfossils (di-
atoms, chrysophycean stomatocysts) and mol-
luscs, preserved in the sediments studied (Ogn-
janova-Rumenova 1997).

The goals of the present study were to sum up

the results gained from sedimentological and
paleontological analyses using Principal Com-
ponent Analysis (PCA) and to reconstruct the

Fig. 1. Location and lithostratigraphic col-
umn of core-drill C-14, village of Katina, Neogene
Sofia Basin. Time scale after Harzhauser & Piller (2007).

background image



ic, volcanic, sedimentary and metamorphic rocks aged
from Precambrian up to Cretaceous occur in the basement
of the graben. The sedimentary rocks that fill in the graben
are of Late Miocene to Pliocene age (Yaneva 2001; Nakov
et al. 2001). The lithostratigraphic units were introduced
by Kamenov & Kojumdgieva (1983) and Angelova &
Yaneva (1998), and include:

• Variegated terrigenous formation – composed of an

irregular alternation of multicoloured compact claystones
and sandy claystones with mica-rich siltstones, sand-
stones, and sands; 350—400 m in thickness; probably of
Late Miocene (Maeotian-Pontian) age;

• Gniljane Formation – composed of green or brown

gravel conglomerates, sandstones, and sandy claystones
with lignite-bearing beds (the Balsha Member); 100—150 m
in thickness; of Late Miocene (Pontian) age;

• Novi Iskar Formation – composed of grey to grey-

bluish clays, diatomaceous clays, silty clays and siltshales
as well as varved siltshales, composed of alternating car-
bonaceous and clayey siltshales; 100—400 m in thickness;
of Late Miocene—Early Pliocene (late Pontian—early Da-
cian) age;

• Lozenec Formation – composed of an irregular alter-

nation of sandstones, sandy claystones, and claystones;
with coal beds in the lower part (Novihan Member) and
silty and sandy-micrite limestones (Bogjovci Member) in
the upper part of the formation; exceeding 400 m in thick-
ness; Pliocene (Dacian—Romanian) in age.

The last three Formations form the Sofia Group.

Material and methods

This study was carried out on sediments from the bore-

hole C-14 near the village of Katina, 15 km northern of

Sofia, 42

°50’01” N and 23°14’13” E. It was located in

the northern part of the Sofia Basin, where the
sedimentation process was the longest.

The depth of the borehole was 271 m; 75 samples were

collected, 44 of which contained diatom flora.

Samples were examined by the following methods to

obtain information about their structural, mineralogical
and geochemical composition – grain-size analysis; mi-
croscopic study of thin-sections; X-ray diffraction analysis
(XRD) on Siemens D-500 under the following conditions:
40 kV, 30 mA, with Cu tube, scans were 2 = 3—80

° for dis-

oriented samples, and 2 = 2—20

° for sedimented samples

heated up to 600 

°C and glycerin saturated; Differential-

thermal analysis (DTA) on Derivatograph MOM system
Paulik-Paulik-Erdey with DTA and DTG 1/10, TG-100 mg,
heat rate 10 

°C/min; and X-ray fluorescence analysis

(XRF) on multichannel X-ray fluorescence Spectrometer
Type CPM-25—16-C. The classification of shales, pro-
posed by Potter et al. (1980) has been used.

The samples for diatom analysis were cleaned according

to the method described by Ognjanova-Rumenova (1991).
The relative abundance of diatoms was estimated by
Schrader’s scale (Schrader 1973). Preparation for scanning
electron microscopy followed Hasle & Fryxell (1970) and

samples were examined with a Jeol Superprobe 733 and
Jeol JSM T300. Ecological spectra were composed on the
basis of studies of recent diatoms by the method of Abbott
& Van Landingham (1972). The basic environmental fac-
tors interpreted in the paleoecological analyses were the ac-
tive water reaction (pH), temperature, trophic state, type of
habitat and salinity. Index B was used to assess trends in
acidification and to derive equations for inferring lake-wa-
ter pH (Renberg & Hellberg 1982). The extent of diatom
preservation provided additional information about paleo-
environmental change (Flower 1993). The percentage ratio
of diatom frustules to chrysophycean stomatocysts was ap-
plied as an index of trophic status (Smol 1985).

Principal Component Analysis (PCA) was used to sum-

marize the major patterns of variation of the data. In the
biplot, variables with high positive correlation generally
had small angles between their biplot arrows. Variables
with long arrows had high variance and generally the
more important within the data (ter Braak 1995; ter Braak
& Prentice 1988). The ordination was implemented by the
computer program CANOCO 4.0 (ter Braak & Šmilauer

Results and discussion

Sedimentological characteristic

The borehole C-14 (sampled interval between 6—199 m)

cuts across the sediments assigned to all the formal lithos-
tratigraphic units in the Sofia Basin except the variegated
terrigenous formation. Deposits between 6 and 20 m were
assigned to the Lozenec Formation, and those between 20
and 199 m – to the Novi Iskar Formation. The section
comprises clayshales (64 %), siltshales (17 %), coal
(14 %), sand (4 %) and chalk (1 %) (Figs. 1, 2). An alterna-
tion of clayey siltshales and silty clayshales, green-grey-
ish in colour, with random coarse sandy grains of quartz,
feldspars and micas and plant detritus represented the up-
permost interval between 6 and 16 m. In that sediment py-
rite and organic matter occurred. A coal layer was found
between 16 and 17 m. The next 3 m were composed of
thin layered sandy and silty clayshales.

Light grey-greenish calcareous clayey siltshales alter-

nating with dark grey-brownish silty and sandy clayshales
occurred within the depth interval of 20—57 m.

The siltshales included shell detritus, pyrite and fishbone

fragments. Shales in the interval 46—57 m were more sandy.

The next interval encompassed the depth range of 57 to

115 m and was composed of light grey laminated shales.
Lamination resulted from the alternation of clay and silt
layers with thicknesses of about 0.8—1.2 mm. Silt layers
were often enriched in micrite.

Dark grey-greenish calcareous silty clayshales constitut-

ed the next 10 meters of the section. These sediments were
followed downwards by 20 m grey-greenish clayshales
with abundant organic matter, pyrite and nestlike heaps of
sand-sized quartz, feldspars and muscovite.

background image



Light-grey diatomaceous clayshales with pollen and py-

rite microcrystals composed the interval of 145—199 m.
Sand-sized grains of quartz, feldspars and micas appeared
in the lowermost part of that interval. The lowermost part
of the section comprised an alternation of coal clays and
lignite coals and a 2 m thick bed of lacustrine chalk en-
riched in organic matter and plant detritus on the bottom.
These sediments belonged to Balsha Member of Gniljane
Formation. The terrigenous fraction was composed of
quartz, plagioclase, K-feldspars, muscovite, biotite, and
traces of chlorite. Smectite was the dominant clay mineral
(12—35 %), while illite (1—18 %) and kaolinite (1—6 %) oc-
curred in subordinate quantities. The smectite content de-
creased downwards whereas illite and kaolinite slightly

The co-occurrence of smectite, kaolinite and illite in

shales indicates a clay minerals association typical for
open freshwater lakes and supposedly a detritic origin of
these minerals. The prevalence of smectite and illite over
kaolinite in the upper part of the sequence may be inter-
preted as the climate cooling during the late Pontian and

Dacian. As a result, the bedrock in the source area was sub-
ject to partial weathering.

The chemical composition of the samples studied was

characterized by rocks with average to low alkalinity, with
higher than average iron-hydroxide content and with rela-
tively high degrees of chemical alteration according to
CIA of Ibbeken & Sleyer (1992) (Table 1).

Diatom flora

The taxonomic diversity of the identified diatom flora

was very high, in total 302 diatom taxa have been identi-
fied. Twenty taxa could be determined to the generic level
only. The diatom flora consisted of 86 % recent and 14 %
extinct species. A detailed taxonomic list of the diatom
flora from the Sofia Basin has been published elsewhere
(Ognjanova-Rumenova 1994). Numerous new and inter-
esting species have been described (Temniskova-Topalo-
va & Ognjanova-Rumenova 1990, 1992; Ognjanova-Ru-
menova 1996; Lange-Bertalot & Metzeltin 2006;
Ognjanova-Rumenova & Metzeltin 2006). The class Cen-
trophyceae accounted for 20.3 % of the diatom flora. The
profile studied was generally dominated by the genus Aula-
coseira (17 species, varieties and forms), Cyclotella (12),
Stephanodiscus (11) and Actinocyclus (11). Some of the
established centric genera had limited stratigraphic range
and could be used as index genera. These included Alveo-
lophora,  Concentrodiscus,  Mesodictyon and Ectodictyon.
Pennate diatoms were generally the most abundant
(79.7 %).  Navicula sensu lato was the most species-rich
genus (44 species, varieties and forms), followed by Fragi-
laria (35), Cymbella sensu lato (17), Achnanthes (15),
Gomphonema (12), Surirella (12), Pinnularia (11) and
Diploneis (10). The remaining genera were represented by
less than 10 species each.

Diatom biostratigraphy

The results of diatom analysis have been synthesized in

the form of a diagram of the species attaining relative
abundance 4—5 according to the Schrader’s scale in at
least one sample (Fig. 3). In the profile studied four diatom
thanatocoenoses were distinguished (Ognjanova-Rume-

Fig. 2. Ternary diagram of grain-size composition of samples
from core-drill C-14, village of Katina, Neogene Sofia Basin.

Table 1: Oxide composition and geochemical modules of samples from C-14, village of Katina (alkaline module ALK = Na





bulk normalized alkalinity BNA= (Na


O+ K


O)/ Al




; chemical index of alteration CIA= Al




/ (Al




+ CaO + Na


O + K



background image



Fig. 3.

 Diatom diagram from core-drill C-14, village of Katina, showing the succession of the most common taxa.

background image



nova & Popova 1992). The results of the diatom
biostratigraphy were calibrated against the
stratigraphic data provided by vertebrate faunas,
molluscs and carpoflora (Kamenov & Kojum-
dgieva 1983; Kojumdgieva et al. 1984; Palama-
rev 1991).

Ecological analysis of diatom flora

Ecological spectra were established for 225

taxa (74.5 % of the determined diatom taxa). The
results of ecological analysis were presented in
Fig. 4.

Periphytic forms predominated, with epiphytic

ones ranging between 17.6 and 76.3 %, and
benthic deep water forms ranging between 10
and 52.9 %. Planktonic species ranged between
3.3 and 66.7 %, in the uppermost part of the pro-
file their abundance increased – more of the
Aulacoseira species showed a mass development
(Fig. 5).

With respect to the salinity, the indifferent oli-

gohalobous taxa predominated, ranging from
81.8 to 100 %, with halophobous taxa account-
ing for 1.8 to 16.7 %, and halophilous taxa (1.5
to 9.8 %) as less abundant. A small percentage of
mesohalobous forms (1.7—5.1 %) occurred. A rel-
atively weak salinity increase was observed in
the depth intervals between: 155.00—190.00 m;
90.00—125.00 m; 47.00 m; 41.00—43.00 m and
34.00—35.00 m (Fig. 6).

In relation to pH of the water, the content of al-

kaliphilic species ranged between 47.1 and 85 %,
and of indifferent ones between 10 and 44.4 %.
These two groups strongly dominated in the dia-
tom flora. The content of acidobiontic forms
ranged from 0—11.1 %, and of alkalibiontic forms
from 0—24.1 % (Fig. 7).

In biogeographical distribution, the abundance

of cosmopolitan forms predominated (44.4 to
87.5 %), followed by boreal (0 to 33.9 %) and
north-alpine (0 to 33.3 %). Tropical species repre-
sented by e.g. Melosira undulata (Ehr.) Kutz.*,
Cyclotella iris var. charetonii and Aulacoseira
agassizii var. malayensis constituted only 1.6 %
of the diatom assemblage and they occurred in
the lower part of the investigated sequence
(Fig. 8).

pH reconstruction by index B

The inferred lake water pH values ranged be-

tween 7.02 and 8.24 (Fig. 9). The pH values
were subject to relatively low variability, but it
is worth noting that the highest values were
inferred at the depth of: 180.00 m, 145.00—
155.00 m, 45.00—46.00 m and 40.00 m. The
lowest value was inferred at 18.00 m, below the
coal layer.

Fig. 4.  Histograms-Percentage ratio of the ecological groups diatoms. Halo-
bion spectrum: 1 – halophobous, 2 – halophilous, 3 – indifferent, 4 –
mesohalobous;  “Habitat” spectrum: 1 – planktonic, 2 – periphytic (epi-
phytic), 3 – periphytic (deep water form); pH spectrum: 1 – alkaliphilic,
2 – acidophilic, 3 – alkalibiontic, 4 – indifferent; Geographical
distribution: 1 – cosmopolitans, 2 – north-alpine forms, 3 – boreal
forms, 4 – tropical forms.

Fig. 5. Distribution diagram of the habitat diatom groups from core-drill C-14,
village of Katina.

background image



Differential diatom dissolution

Additional information about paleoenvironmental

changes was provided by the assessment of the extent
of diatom preservation (Flower 1993). Diatom dissolu-
tion was determined by water chemistry and time relat-
ed factors, such as the sediment accumulation rate.

As long as water remains undersaturated with regard

to dissolved silica, diatoms will tend to dissolve and
the dissolution rate depends on pH and temperature.
Our findings were similar to the results of Flower
(1993) based on the rich data set in a variety of lakes.
According to Flower (1993) diatom preservation de-
clines with increasing pH, temperature, coarseness of
sediment, grazing and bioturbation, water depth and
exposure. The rate of preservation generally increased
with diatom robustness, diatom and sediment accumu-
lation rate and diatom concentration.

Measurement of the extent of dissolution was diffi-

cult and required visual recognition and quantification
of different preservation stages for a particular diatom
taxon. The most useful frustules for determination of
the diatom preservation stages included representa-
tives of genera Cyclotella and Campylodiscus. Our
analysis was based only on SEM examinations of the
frustules. Dissolution occurred progressively and cen-
tripetally, the final stage was – the corroded silica ma-
trix of the central area (Fig. 10).

Multivariate analysis

The PCA analysis included all diatoms with relative

abundance exeeding “4”—“5” according to Schrader’s
scale in at least one sample (a total of 60 species, vari-
eties and forms, see Table 2). The result was presented
as a PCA correlation biplot (Fig. 11). The first two
principal components (


= 0.220, 


= 0.113) account-

ed for 33.3 % of the total variance and effectively cap-
tured the main patterns of variation in the species data
thus separating them in three groups. The first axis was
related to sample group corresponding to the interval
of 150.00—198.00 m – including sediments deposited
above the coal layers of the Balsha Member, Gniljane
Formation and were plotted on the right of the dia-
gram. The dominant species was Concentrodiscus sp.,
as well as species belonging to genera Cyclotella  (C.
iris,  C. iris var. charetonii,  C. iris var. combiriensis,  C.
ovata,  Cyclotella sp.) and Actinocyclus  (A. makarovae,
A. aff. lobatus,  Actinocyclus sp. 4). Subdominants of
the centric diatoms were Melosira undulata var. nor-
mannii and Ellerbeckia kochii. The genus Aulacoseira

Fig. 6.  Distribution diagram of the halobion diatom groups
from core-drill C-14, village of Katina.

Fig. 7.  Distribution diagram of the pH diatom groups from
core-drill C-14, village of Katina.

background image



Fig. 8. Distribution diagram of geographical distribution type of diatoms
from core-drill C-14, village of Katina.

was represented by a low number of species i.e. A.
ambigua and A. agassizii var. malayensis  occurring
only in the lowermost samples (195.00—198.00 m).
There were common benthic deep water taxa of the
class Pennatophyceae – Diploneis elliptica and D.
mauleri,  Navicula hasta and N. placentula,  Gy-
rosigma acuminatum.

Axis 2 reflected two gradients. The first one run

to the top and separated sample group of 45.00—
150.00 m. For these samples the dominant taxa
were  Actinocyclus variabilis,  Cyclotella iris var. in-
tegra and C. radiosa var. pliocaenica,  Campylodis-
cus sp. Species of Stephanodiscus showed mass de-
velopment, and these were: S. carconensis var.
pusillus,  S. hantzschii f. tenuis,  S. minutula. The
dominants and subdominants of genus Aulacoseira
were:  A. granulata,  A. islandica,  A.  lirata and Aula-
coseira sp. The second gradient ran to the bottom
left and separated the third group of samples
(16.00—40.00 m). The planktonic species of Aula-
coseira including A. muzzanensis,  A. distans,  A.
valida and A. italica, dominated in these samples.

Relative proportions of diatom frustules and chryso-
phycean stomatocysts

Tracing the ratio of diatom frustules/chryso-

phycean stomatocysts several peaks in the develop-
ment of the latter were determined (Fig. 9). Chryso-
phycean stomatocysts were most abundant at
140.00 m (77 %), followed by 183.00 m (40 %),
125.00 m (41 %), 56.00 m (41 %), and 46.00 m
(55 %). On the basis of these percentage ratios three
trophic stages in the lake history were distinguished
(Ognjanova-Rumenova 1997).

Fig. 9. Percentage ratio of the diatom frustules/chrysophycean stomatocysts with inferred lake-water pH curve.

background image



Fig. 10. Stages of sequential dissolution of Cyclotella sp. (C-14, v. Katina, depth 195.00 m) and Campylodiscus sp. (C-14, v. Katina,
depth 120.00 m). 1 – Cyclotella sp., Stage 1 – the morphology not significantly altered by dissolution; SEM 7500.  2 – Cyclotella
sp., Stage 2 – corrosion of radiate striae clearly visible; SEM 7500.  3 – Cyclotella  sp., Stage 3 – the mantle has become detached
from the valve face and corrosion of the striated zone is progressing; SEM 7500. 4 – Cyclotella sp., Stage 4 – inter striae costae be-
come shorter and striae lost; SEM 20,000. 5 – Cyclotella sp., Stage 5 – inter striae costae almost lost and central area processes have
fallen away leaving holes; SEM 5000. 6 – Campylodiscus sp., Stage 1 – diatom preservation is generally excellent; SEM 860. 7  –
Campylodiscus sp., Stage 2 – apparently dissolved valve showing loss of inter costae; SEM 1200.

background image



Paleoenvironmental interpretations

Neogene sedimentation in the region of Katina started

during the early Pontian with deposition of alluvial sedi-
ments (Gniljane Formation) – coarse sand with gravel
and pebble lens-like intercalations, typical for the facies
of braided river systems. Later, the river system changed to
a meandering type and the partial or complete separation
of some meanders changed the environment to marsh and
lacustrine chalk was deposited followed by a coal forma-
tion. The deposition of chalk requires alkaline conditions
and water enriched in dissolved CaCO


. Under such con-

ditions the Balsha Member was formed.

The continuing subsidence during the middle Pontian

led to the appearance of a lacustrine environment. As a re-
sult the thick shale beds of the Novi Iskar Formation were
deposited. The paleolimnological interpretation faces dif-
ficulties when based on various analyses. The different
aquatic organisms’ had high mobility and migration rates,
which made the correlation problematic. Comparison of
the changes in biological remains (algae and molluscs), as
well as in the chemistry, coincides remarkably with three
distinct stages in the evolution of the Novi Iskar Formation.

In the beginning of the sediment formation (middle

Pontian age) the tropical diatoms Aulacoseira agassizii

Fig. 11. Biplot based on Principal Component Analysis (PCA) of diatoms assem-
blages with relative abundance greater than “4”—“5” according to Schrader’s scale
in at least one sample. Polygons wrap the position for each sample group consid-
ered. The species arrows are not shown and species points coincide with the head of
each species arrow. For diatom species codes see Table 2.

var.  malayensis,  Melosira undulata, and
Cyclotella iris var. charetonii occurred
abundantly. The development of these
tropical elements indicated relatively
higher temperature, which coincided with
the climate reconstruction based on macro-
flora remains (Palamarev 1967). Prevalence
of freshwater, planktonic, oligohalobous
species of Cyclotella  was usually character-
istic of the early stages of oligotrophic
lakes, as reported by Bradbury (1999). The
salt content ranged between 0.2—0.3 ‰, but
the presence of mesohalobous diatoms
proved that it reached up to 0.5 ‰. Most
characteristic for this stage was the appear-
ance of Paradacna aff. okrugici Brus. and
Dreissena aff. superfoetata Brus., molluscs
– typical for the Pannonian Basin (Ka-
menov & Kojumdgieva 1983). Up to now
there has not been any hypothesis for their
penetration and distribution in the Sofia
Basin. The first change in the stomatocysts
proportion is observed at 183.00 m (40 %),
and coincided with the increase of diatom
diversity. Dominant were again species, be-
longing to genera of Cyclotella and Acti-
nocyclus, but there were subdominants of
pennate deep water species of the  genera
Navicula,  Diploneis,  Gyrosigma,  Surirella
and  Campylodiscus. The diatom spectrum
showed great diversity of abundant fresh-
water, planktonic forms, which were very
common in clean, alkaline waters of oligo-

to mesotrophic lakes. Some diatoms were dissolved during
diagenesis of sediments containing a large amount of

The second stage of Novi Iskar Formation development

was the beginning of natural eutrophication in the paleo-
lake during the late Pontian age. At 140.00 m depth of the
investigated borehole there was a sharp increase in the
quantity of the chrysophycean stomatocysts (77 %). This
peak coincided with the mass extinction of Dreissena aff.
superfoetata Brus. (Kamenov & Kojumdgieva 1983). The
layer with this fossil mollusc was used for designation of
the boundary between the middle and late Pontian sub-
stages. The eutrophication in the paleolake was confirmed
with the observed increase of planktonic diatoms repre-
sented by Stephanodiscus carconensis var. pusillus,  S.
hantzschii f. tenuis,  S. minutula and Aulacoseira granula-
ta,  A. islandica,  A. lirata and Aulacoseira sp. A change in
the taxonomic composition of the genus Cyclotella is also
observed here. Cyclotella radiosa var. pliocaenica and
Cyclotella sp. are becoming the most abundant taxa. The
salt content ranged between 0.2—0.3 ‰, hence the active
water reaction of the lake water was neutral to slightly al-

A change of the sedimentary environment happened at

the boundary between Late Miocene and Early Pliocene.
This was the third stage in the development of the Novi

background image



Table 2: List of the diatoms with relative abundance greater than “4” and
“5” according to Schrader’s scale in at least one sample, included in PCA

Iskar Formation. The quantity of chrysophycean
stomatocysts increased in several layers in the
uppermost part of the investigated borehole. It was
preceded by abundant development of numerous
Aulacoseira taxa, including A. distans,  A. lirata,  A.
islandica,  A. muzzanensis, and diverse Fragilaria
flora. Changes in species composition of planktonic
diatoms might have resulted from periodical water
level oscillations, evident in the lithology – alter-
nation of clayey carbonaceous siltshales with silty
clayshales. The ecological investigations of the
boreal lakes showed that genus Fragilaria
dominated in sand and silty habitats in open lake.
On this substratum Fragilaria pinnata,  F.  parasiti-
ca, and numerous Navicula  species were abundant
(Kingston 1984). The high abundance of north-
alpine species – Aulacoseira distans,  A. islandica,
A. lirata,  A. valida,  Navicula scutelloides, indicated
a drop in the temperature. The salt content ranged
between 0.2—0.3 ‰. An increase in the acidity was
proved by the prevalence of acidophilous diatoms
–  Aulacoseira distans,  A. lirata and typical for
marsh environments – Aulacoseira islandica mor-
photype  curvata. The natural eutrophication in the
lake was in progress, indicated by the expansion of
Aulacoseira  species. During this period several
layers were deposited with Viviparus bulgaricus
Brus., a mollusc typical of marsh sediments. These
levels mark the boundary between the Pontian and
Dacian stages (Kamenov & Kojumdgieva 1983). At
the same period the lake started to shallow and
gradually lacustrine conditions changed to lacus-
trine-marsh and alluvial with deposition of the coal-
bearing sediments of the Novihan Member of the
Lozenets Formation.


Diatom analyses, complemented by sedimento-

logical investigation could demonstrate the impact
of different paleoenvironmental proxies on the de-
velopment of the diatom flora. The primary goal of
this study was to determine which of the paleoenvi-
ronmental variables explained best the variation in
the fossil diatom assemblages of the sediments stud-
ied. We found that the process of eutrophication,
the salt content, temperature and the changes in the
depth of the lake correlated significantly with pat-
terns in the diatom assemblages.

The Neogene sediments from Sofia Basin were de-

posited during the Late Miocene and Pliocene. The
sedimentation started in the early Maeotian with the
subsidence of the central and central-northern parts
of the Sofia graben where the alluvial sediments of
the variegated terrigenous formation were deposit-
ed. In the early Pontian a new faulting and subsid-
ence in the central and central-eastern parts oc-
curred. The alluvial sediments from the lower parts

background image



of the Gniljane Formation were deposited. The lignite in-
terbeds from Balsha Member are a result of paludification
caused by continuing subsidence during the early Pon-
tian. The further fast deepening changed the environment
from paludal to lacustrine. The newly formed lake expand-
ed to the South and to the East. During that time the sedi-
ments of Novi Iskar Formation in the central parts of the
basin were deposited. Three stages in its development
could be recognized:

• The first stage was oligotrophic, characterized by a

relatively higher temperature;

• The beginning of natural eutrophication in the pale-

olake occurred during the second stage (late Pontian sub-

• The third stage in the development of the Novi Iskar

Formation is marked by a change of the sedimentary envi-
ronment at the boundary between Late Miocene and Early
Pliocene. The natural eutrophication in the lake was in
progress, indicated by the expansion of Aulacoseira  spe-
cies. The high abundance of north-alpine species indicat-
ed a drop in the temperature. At the same period the lake
started to shallow and gradually lacustrine conditions
changed to lacustrine-marsh and alluvial with deposition
of the coal-bearing sediments of the Novihan Member of
the Lozenets Formation.

The inference model, we developed may also be useful

in describing the colonization of diatoms in the continen-
tal basins from graben structures in the region of the Bal-
kan Peninsula during the Late Neogene. Reconstruction of
paleoenvironmental variables from fossil diatom assem-
blages is now possible, and might reveal the paleolimno-
logical history of this system.

Acknowledgments: We would like to thank Dr Martin
Kernan, Environmental Change Research Center, Univer-
sity College London for his help with the multivariate


Abbott W. & Van Landingham S. 1972: Micropalaeontology of

Miocene nonmarine diatoms from the Harper District, Malheur
County, Oregon. Nova Hedwigia, Beih. 23, 4, 847—907.

Angelova D. & Yaneva M. 1998: New lithostratigraphical data

about Neogene of Sofia Basin. Rev. Bulgarian Geol. Soc. 59,
2, 37—40 (in Bulgarian).

Bradbury J.P. 1999: Continental diatoms as indicators of long-term

environmental change. In: Stoermer E.F. & Smol J.P. (Eds.):
The Diatoms: Applications for the environmental and Earth
Sciences. Cambridge University Press, Cambride, 169—182.

Charles D.F. & Smol J.P. 1994: Long-term chemical changes in

lakes: quantitative influences using biotic remains in the sedi-
ment record. In: Baker L. (Ed.): Environmental chemistry of
lakes and reservoirs. Advances in Chemistry, Series 237. Amer.
Chem. Soc., Washington DC, 3—31.

Flower R. 1993: Diatom preservation: experiments and observa-

tions on dissolution and breakage in modern and fossil materi-
al.  Hydrobiologia 269, 270, 473—484.

Harzhauser M. & Piller W. 2007: Benchmark data of a changing

sea. Palaeogeography, palaeobiogeography and events in the

Central Paratethys during the Miocene. Palaeogeogr. Palaeo-
climatol. Palaeoecol. doi:10.1016/j.palaeo 2007.03.031.

Hasle G. & Fryxell G. 1970: Diatoms: Cleaning and mounting for

light and electron microscopy. Transact. Amer. Microsc. Soc.
89, 4, 469—474.

Ibbeken H. & Sleyer R. 1992: Source and sediment. Springer-Ver-

lag, 1—286.

Kamenov B. & Kojumdgieva E. 1983: Stratigraphy of the Neogene

in Sofia Basin. Palaeontology, stratigraphy and lithology. Bul-
garian Acad. Sci. 18, 69—85 (in Bulgarian).

Kingston J. 1984: Palaeolimnology of a lake and adjacent fen in

southwestern Labrador: Evidence from diatom assemblages.
In: Mann D. (Ed.): Proceedings of the 7


International Diatom

Symposium, Philadelphia, O. Koeltz, Koenigstein, 443—453.

Kojumdgieva E., Nikolov I. & Mein P. 1984: Les associations de

grands mamif

ères du Miocene superieur en Bulgarie et leur

correlation avec l’echelle regionale de la Paratethys. Comptes
rendus de l’Academie bulgare des Sciences, 37, 3, 341—343.

Lange-Bertalot H. & Metzeltin D. 2006: Comparison of seven

species of Navicula sensu stricto (Six species described as
new to science from Miocene lacustrine deposits in Bulgaria
and Romania) In: Ognjanova N. & Manoylov K. (Eds.):
Advances in phycological studies, prof. D. Temniskova-
Topalova, Festschrift, Pensoft, Sofia, 55—73.

Nakov R., Birchfiel B.C., Tzankov T. & Royden L.H. 2001: Late

Miocene to recent sedimentary basins of Bulgaria with ex-
planatory notes. Geol. Soc. Amer.,  Map and chart series
MCH088, 28.

Ognjanova-Rumenova N. 1991: Neogene diatoms from sediments

of Sofia Valley and their stratigraphic significance. Ph.D. The-
sis, Geol. Inst., Bulgarian Acad. Sci., 1—330 (in Bulgarian).

Ognjanova-Rumenova N. 1994: Taxonomical composition of fossil

diatoms from Neogene nonmarine sediments, Sofia Basin. An-
nuaire de l’Universite de Sofia “St. Kl. Ohridski”, Fac. Biolo-
gy, livre 2, Botanique, 85, 55—58.

Ognjanova-Rumenova N. 1996: Cyclotella iris Brun & Herib.

group from the Upper Miocene sediments of the Sofia Basin,
Bulgaria.  Geol. Carpathica 47, 5, 301—310.

Ognjanova-Rumenova N. 1997: Lake trophic evolution

determined by fossil diatoms, Chrysophycean stomatocysts
and mollusca. Ann. Geol. Pays Hellen., Athenes, Premiere
Ser. 37, 1996—1997, 97—117.

Ognjanova-Rumenova N. & Metzeltin D. 2006: A new Cymbella

from the Neogene in Bulgaria and its stratigraphic signifi-
cance. In: Ognjanova N. & Manoylov K. (Eds.): Advances in
phycological studies, prof. D. Temniskova-Topalova,
Festschrift, Pensoft, Sofia, 91—97.

Ognjanova-Rumenova N. & Popova E. 1992: Diatom biostratig-

raphy and comparative core correlation within Sofia basin,
Bulgaria. In: J. Eder-Kovar (Ed.): Palaeovegetational devel-
opment in Europe and regions relevant to its palaeofloristic
evolution.  Proceedings of the I Pan-European Palaeobotan-
ical Conference, Vienna, 1991, 197—203.

Ognjanova-Rumenova N. & Popova E. 1996: Palaeolimnological

reconstruction of the Sofia Neogene basin, South Bulgaria – a
case study of palaeoenvironmental diatom research. Phytolo-
gia Balcanica 2, 2, 43—51.

Palamarev E. 1967: Xerotherme Elemente in der Tertiärflora Bul-

gariens und Aspekte zum Problem der Formierung der medi-
terraner Flora auf der Balkanhalbinsel. Abh. Zntr. Geol. Inst.
10, 165—175.

Palamarev E. 1991: Composition, structure and basic stages in the

evolution of the Miocene palaeoflora. Doctor Thesis, Botani-
cal Inst., Bulgarian Acad. Sci., 1—497 (in Bulgarian).

Potter P., Maynard J. & Prior W. 1980: Sedimentology of shales.

Springer-Verlag, 1—288.

background image



Renberg I. & Hellberg T. 1982: The pH history of lakes in South-

western Sweden, as calculated from the subfossil diatom flora
of the sediments. Ambio 11, 1, 30—33.

Schrader H.-J. 1973: Proposal for a standardized method of clean-

ing diatom-bearing deep sea and land-exposed marine sedi-
ments.  Nova Hedwigia, Beih. 45, 403—409.

Smol J. 1985: The ratio of diatom frustules to chrysophycean

statospores: a useful palaeolimnological index. Hydrobiologia
123, 2, 199—204.

Smol J. & Cumming B. 2000: Tracking long-term changes in cli-

mate using algal indicators in lake sediments. J. Phycology 36,

Temniskova-Topalova D. & Ognjanova-Rumenova N. 1990: New

species of the genus Pontodiscus Temnisk. et Sheshuk. (Bacil-
lariophyta).  Botanical J., Sankt Peterburg, 75, 9, 1286—1287
(in Russian).

Temniskova-Topalova D. & Ognjanova-Rumenova N. 1992: A

new nomenclature combination in the genus Actinocyclus (Ba-
cillariophyta).  Botanical J., Sankt Peterburg 77, 11, 136 (in

ter Braak C.J.F. 1995: Ordination. In: Jongman R.H.G., ter Braak

C.J.F. & Tongeren O.F.R. (Eds.): Data analysis in community
and landscape ecology. Cambridge University Press, Cam-
bridge CB2 2RU, 91—173.

ter Braak C.J.F. & Prentice I.C. 1988: A theory of gradient analysis.

Adv. Ecol. Res. 18, 271—317.

ter Braak C.J.F. & Šmilauer P. 1998: CANOCO Reference Manual

and User’s Guide to Canoco for Windows: Software for Ca-
nonical Community Ordination (version 4). Microcomputer
Power, Ithaca, NY, USA, 1—352.

Yaneva M. 2001: Sedimentology of the Neogene deposits in Sofia

Basin. Ph.D. Thesis, Geol. Inst., Bulgarian Acad. Sci., 1—203.