Problems of landslide chronology in the Mätra mountains in Hungary

a) marginal zones o f vo lcan ic mountains, b ) s t eep inner s lopes o f the former calderas o f eruption, c ) va l ley s lopes cut into s t ra tovolcanic Structures. In their characteristic reg ions o f occu r r ence landsl ide forms are important, in s o m e p laces even determinant c o m p o ­ nents o f morphology. T h e overwhelming majority o f the landforms are no more act ive and can b e cons ide red fossi­ lised. Consequent ly , landslide p rocesses used to play a sig­ nificant role in forming s lope morpho logy in previous stag­ es o f geomorph ic evolut ion. Research per formed heretofo­ re has provided only rather ambiguous a n s w e r to the que­ st ion concern ing the t ime o f e m e r g e n c e o f landslides on a theore t ica l basis . D u e to the a b s e n c e o f y o u n g sediments traditional stratig raph ic methods cou ld not yield satisfactory results. T h e authors therefore a t tempted to determine the t ime o f the upfilling o f lakes or b o g s loca ted in undrained depress ions c lammed by slides and the minimum age o f the landslides through the analyses o f b o r e h o l e samples. In the Mätra se­ d iments upfilling o f 7 such depress ions we re carr ied out. In severa l cases pollen diagrams were ob ta ined and in two b o r e h o l e s radiocarbon dating o f samples w e r e performed. T h e pape r summarises the results o f the b o r e h o l e analyses and evaluates pollen records supported by C'< dating in detai l . Analyses performed using various methods indicate a b e ­ ginning o f the bog formation in the Early H o l o c e n e , conse ­ quen t ly landslides occur red presumably in the w a k e o f the Preborea l or prior to that in a warm spell fol lowing the Late Glacial . Al though these results c anno t b e ext rapola ted to all the landslides identified in the t w o mountains ( their n u m b e r is ca . 2 0 ) without more a d o never theless they are orientating. [ R u t s c h u n g s c h r o n o l o g i s c h e F r a g e n i m M a t r a G e b i r g e in U n g a r n ] K u r z f a s s u n g : Im letzten Jahrzehn t wurden b e i der Unter­ suchung der Rutschungsprozesse und -formen in Ungarn in d e n Mittelgebirgen, die zumeist in der Mitte und am En­ de d e s Miozäns ents tanden sind, viele Rutschungsformen *) Adresses o f the authors: Dr. J . SZABO, DR. E. FELEGYHAZI, Depar tment o f Physical Geography , Lajos Kossu th Univer­ sity, D e b r e c e n , Hungary a u f g e n o m m e n . Als Ergebnis der Ge l ändeun te r suchungcn kann festgestellt werden , d a ß diese Formen vor a l l em auf fo lgenden Geb ie t en v o r k o m m e n : a ) in der Randzone der vulkanischen Gebi rge , b ) au f der steilen inneren Sei te der Calderen der ehema l i ­ gen Ausbruchszent ren , c ) a u f d e n Hängen der die schichtvulkanischen Strukturen tief e ino d e r du rchschne idenden Tälern. In den e rwähn ten charakter is t ischen V o r k o m m e n s b e r e i ­ c h e n s te l len die Rutschungsformen einen wich t igen , an m a n c h e n Stellen e inen en tsche idenden F a k t o r des morpho log i s chen Bi ldes dar. Die übe rwiegende Mehrhei t der F o r m e n ist heute nicht mehr aktiv, und die m e i s t e n von ihnen kann als fossil betrachtet werden. Desha lb ist e s of­ fensichtl ich, d a ß die Rutschungsprozesse in der f rüheren Per iode ( o d e r in den früheren Per ioden) e i n e wich t ige Rolle in der Gestaltung de r Form der Hänge spiel ten. Die b i sher igen Forschungen gaben nur e ine z ieml ich un­ s ichere , b l o ß prinzipielle Antwort auf die Frage n a c h der Ents tehungszei t der Formen . Da die jungen Sed imente fehlen, konnte durch d ie ge­ w ö h n l i c h e n strat igraphischen Methoden kein gu tes E rgeb ­ nis erzielt werden , desha lb versuchten die Verfasser durch B o h r u n g e n in den gesch los senen , abflußlosen mit e i n e m S e e o d e r M o o r ausgefüllten Eintiefungen der g r ö ß e r e n Rutschungen die Aufschüt tungsdauer der Eint iefungen, und dadurch das minimale Alter der Rutschungen zu b e ­ s t immen. Im Mätra-Gebirge wurde das Auffüllungsmaterial der Rutschungseint icfung untersucht. In m a n c h e n Fäl len konn te e i n e Pol lenanalyse durchgeführt werden , in zwei B o h r u n g e n k a m es auch zu e ine r »C-Al te r sbes t immung. In der vor l iegenden Studie w e r d e n die Ergebnisse d e r B o h ­ rungen zusammengefaßt , und ausführlich werden d i e auch durch d ie i '<C-Altersbestimmung unterstützten Po l l end ia ­ g r a m m e interpretiert. Die mit d e n versch iedenen Methoden durchgeführten Ana­ lysen ze igen, daß die Moorbi ldung in den un te r such ten Fällen in der ersten Per iode des Holozäns begann , u n d die Rutschungen erfolgten wahrschein l ich am Anfang d e r präborea len Zeit, ode r früher, während einer K l imave rbes se rung nach i rgendeiner Kal tper iode der Spätglazialzeit . D i e e rha l tenen Ergebnisse k ö n n e n w o h l auf die ande ren ins­ gesamt e twa 20 Rutschungsformen nicht extrapol ier t wer ­ den, a b e r s ie stellen doch e i n e Orientierungshilfe im B e z u g auf ihr Alter dar. I n t r o d u c t i o n A detailed survey and mapping of the landslide sur­ face forms in the volcanic middle mountains o f Hun­ gary was started in the past decade. Geomorpho lo Problems o f landsl ide chronology in the Mätra mountains in Hungary 121 gists studying volcanic ridges previously either did not pay particular attention to features o f slumping or considered them insignificant additional morpho­ logical elements o f the landscape. In our earlier pa­ pers (SZAB6 J . 1992, 1 9 9 3 ) short preliminary reports we re presented on the results o f field works eluci­ dating the high importance of the part played by landslide processes in the geomorphic evolution in several regions o f the Tertiary volcanic mountains (notably the Visegräd Mountains, the Mätra Moun­ tains and the basalt volcanos north o f Lake Balaton) where slumps are essential, in some cases even de­ terminant components o f the morphology. Their oc ­ currence is mainly confined to the areas as follows: a) marginal zones o f volcanic mountains, b ) s teep inner s lopes o f the former calderas of erup­ tion, c ) valley slopes cut into stratovolcanic structures. a) This is the most frequent case. W h e r e the bed o f loose material underlying the extrusive or loose vol­ canic matter is still on the surface as for exemple in the mountain margins, and is probably situated higher than ist surroundings, it has lost stability under the weight o f the volcanic layers and func­ tions as sliding plane for volcanites on a sufficiently high and steep s lope. T h e bed underlying acid and neutral volcanites formed in the Badenian stage o f the Miocene is usually the schlieren o f the Carpathi­ an stage (or occasional ly oligocene sediment) while basalts o f the late Miocene generally overlie sandyclayey Pannonian deposits. The probability o f lands­ lides was enhanced by a (sometimes asymmetric) uplift o f the volcanic margin, such increasing the ele­ vation over the surrounding area. Sliding planes could develop within the volcanic sequence . It was made possible by the stratovolcanic structure for sli­ des could occur along the surface o f loose tuff hori­ zons especially if the latter was increasingly weathe­ red. On the northern slopes of the Mätra crest the geomorphic evolution and the tectonic structure created favourable conditions for landslides so a fre­ quent series o f slumps can be observed on the slo­ pes. In spite o f the spatial discontinuity the major forms of slumping make up a distinct (maximum 1 km wide) zone o f slides, differing morphologically from the foothills o f the mountains dissected by valleys. b ) T h e previous centres o f eruption have survived in remnants only. (So . e.g. the more or less intact parts o f the double caldera in the Visegräd Mountains has s lumped in several places along the s teep inner side and at the base o f the caldera walls a series of huge displaced blocks can b e found.) c ) T h e slumps occurring on the s lopes o f valleys dissecting the inner portions o f the volcanic ridges are as a rule slides along the tuff layers cut during in­ cision. If the whole volcanic sequence had been cut in valley evolution the underlying b e d o f volcanites could serve as sliding plane. T h e majority o f the landslide forms are stable and are in a stage o f deg­ radation. Active parts still in motion occur sporadi­ cally and are o f insignificant size as compared to the earlier slides. Consequently, these landslide forms o f natural origin have relevance primarily not to the prensent-day surface dynamics but be ing remnants o f intensive past processes they are indispensable for the understanding of the morphology in the con­ cerned zones o f volcanic ridges. T h e majority o f the mapped landforms are much more older than to be dated directly (on the basis o f historical records, written documents or maps, perhaps by the deformed growth o f trees). At the same time they are much younger than the cover on them or in their immediate surroundings (in the studied areas no loess series that might have been relevant to dating were found) so the traditional stratigraphic methods o f age determination could not be used. Consequently, dating o f the landslide events has been carried out on a general theoretical basis (e.g. by founding periods o f climates favourable for the development o f landslides) and was highly am­ biguous. A final and general solution o f this problem necessitates further and substantial investigations o n landslides and the present paper reports on the initial efforts which might contribute to a more exact dating o f occurrence o f particular slumps. Extrapo­ lation o f these results is to be carried out by due cau­ tiousness. P r i n c i p l e s , r a n g e a n d m e t h o d s o f i n v e s t i g a t i o n s T h e determination o f the relative age o f landslide forms and an approximation o f their absolute age are made p

of investigations it can be stated that thier occurrence is mainly confined to the areas as follows: a) marginal zones of volcanic mountains, b) steep inner slopes of the former calderas of eruption, c) valley slopes cut into stratovolcanic Structures.
In their characteristic regions of occurrence landslide forms are important, in some places even determinant compo nents of morphology.The overwhelming majority of the landforms are no more active and can be considered fossi lised.Consequently, landslide processes used to play a sig nificant role in forming slope morphology in previous stag es of geomorphic evolution.Research performed heretofo re has provided only rather ambiguous answer to the que stion concerning the time of emergence of landslides on a theoretical basis.
Due to the absence of young sediments traditional stratigraphic methods could not yield satisfactory results.The authors therefore attempted to determine the time of the upfilling of lakes or bogs located in undrained depressions clammed by slides and the minimum age of the landslides through the analyses of borehole samples.In the Mätra se diments upfilling of 7 such depressions were carried out.In several cases pollen diagrams were obtained and in two boreholes radiocarbon dating of samples were performed.
The paper summarises the results of the borehole analyses and evaluates pollen records supported by C'< dating in detail.
Analyses performed using various methods indicate a be ginning of the bog formation in the Early Holocene, conse quently landslides occurred presumably in the wake of the Preboreal or prior to that in a warm spell following the Late Glacial.
Although these results cannot be extrapolated to all the landslides identified in the two mountains (their number is ca.20) without more ado nevertheless they are orientating.

Introduction
A detailed survey and mapping of the landslide sur face forms in the volcanic middle mountains of Hun gary was started in the past decade.Geomorpholo-gists studying volcanic ridges previously either did not pay particular attention to features of slumping or considered them insignificant additional morpho logical elements of the landscape.In our earlier pa pers (SZAB6 J. 1992, 1993) short preliminary reports were presented on the results of field works eluci dating the high importance of the part played by landslide processes in the geomorphic evolution in several regions of the Tertiary volcanic mountains (notably the Visegräd Mountains, the Mätra Moun tains and the basalt volcanos north of Lake Balaton) where slumps are essential, in some cases even de terminant components of the morphology.Their oc currence is mainly confined to the areas as follows: a) marginal zones of volcanic mountains, b) steep inner slopes of the former calderas of erup tion, c) valley slopes cut into stratovolcanic structures.a) This is the most frequent case.Where the bed of loose material underlying the extrusive or loose vol canic matter is still on the surface as for exemple in the mountain margins, and is probably situated higher than ist surroundings, it has lost stability under the weight of the volcanic layers and func tions as sliding plane for volcanites on a sufficiently high and steep slope.The bed underlying acid and neutral volcanites formed in the Badenian stage of the Miocene is usually the schlieren of the Carpathi an stage (or occasionally oligocene sediment) while basalts of the late Miocene generally overlie sandyclayey Pannonian deposits.The probability of lands lides was enhanced by a (sometimes asymmetric) uplift of the volcanic margin, such increasing the ele vation over the surrounding area.Sliding planes could develop within the volcanic sequence.It was made possible by the stratovolcanic structure for sli des could occur along the surface of loose tuff hori zons especially if the latter was increasingly weathe red.On the northern slopes of the Mätra crest the geomorphic evolution and the tectonic structure created favourable conditions for landslides so a fre quent series of slumps can be observed on the slo pes.In spite of the spatial discontinuity the major forms of slumping make up a distinct (maximum 1 km wide) zone of slides, differing morphologically from the foothills of the mountains dissected by valleys.
b) The previous centres of eruption have survived in remnants only.(So.e.g. the more or less intact parts of the double caldera in the Visegräd Mountains has slumped in several places along the steep inner side and at the base of the caldera walls a series of huge displaced blocks can be found.)c) The slumps occurring on the slopes of valleys dissecting the inner portions of the volcanic ridges are as a rule slides along the tuff layers cut during in cision.If the whole volcanic sequence had been cut in valley evolution the underlying bed of volcanites could serve as sliding plane.The majority of the landslide forms are stable and are in a stage of deg radation.Active parts still in motion occur sporadi cally and are of insignificant size as compared to the earlier slides.Consequently, these landslide forms of natural origin have relevance primarily not to the prensent-day surface dynamics but being remnants of intensive past processes they are indispensable for the understanding of the morphology in the con cerned zones of volcanic ridges.
The majority of the mapped landforms are much more older than to be dated directly (on the basis of historical records, written documents or maps, perhaps by the deformed growth of trees).At the same time they are much younger than the cover on them or in their immediate surroundings (in the studied areas no loess series that might have been relevant to dating were found) so the traditional stratigraphic methods of age determination could not be used.Consequently, dating of the landslide events has been carried out on a general theoretical basis (e.g. by founding periods of climates favourable for the development of landslides) and was highly am biguous.A final and general solution of this problem necessitates further and substantial investigations on landslides and the present paper reports on the initial efforts which might contribute to a more exact dating of occurrence of particular slumps.Extrapo lation of these results is to be carried out by due cau tiousness.

Principles, range and methods of investigations
The determination of the relative age of landslide forms and an approximation of their absolute age are made possible by the analysis of their degrada tion (transformation).Degradation has well recognis able distinct phases.Among them especially impor tant are the characteristic trends of development of depressions formed through damming by the slump ed masses.These depressions have basically two different ways of evolution.
1. Their closed (undrained) character might be alter ed by a complete upfilling.In this case after certain period only rounded steps remain on the place of depressions.
2. Closed depressions can be eliminated through drainage.Following landslide events water provided by springs at various levels of the failure surface fills up the depression and (depending on the degress of recharge) is overflowing its edge gradually cuts it and opens the basin (certainly water of the springs can play an important role in the upfilling of the de pression).Chances for the discharge are even better when this basin is formed in the upper section of the  To unfold details of the above mentioned processes and to determine an approximate age of slumps bore holes were drilled in closed depressions (filled with lakes, bogs or ones already upfilled and dry) found in volcanic ridges.7 holes in the Mätra were drilled down to max.6.5 m depth.
In all but 2 cases they reached the surface of a relatively hard volcanic rock (mostly tuff).Pure but not undisturbed samples were collected through 10-20 cm for the subsequent analyses.Granulometric composition, humus and calcium carbonate content and pH were determined.Palynological analysis was carried out in each samples.Absolute dating of vegetation remnants of some layers from two boresholes i. e. «C investigations were performed by Ede I Iertelendi in the Radiocarbon Laboratory of the Institute of Nuclear Physics (Debrecen).

Landslides in the Mätra Mountains
The overwhelming majority of landslides found here are resulted from the stratovolcanic structure of the ridge, the presence of primary volcanic forms (calderas, lava fields) the slope conditions brought about by subsequent tectonic movements and the specific character of the contact of the volanic sequence with the less consolidated underlying bed (SZABÖ J. 1991, 1992, 1993, DAVID L. 1992).Most of them have developed on the steep northern side of the mountain crest (Fig. 1).The western part of the crest from Agasvär to the Piszkes-tetö and it eastern portion with the Kekes group to the Tarna Valley are the margins of the remnants of a volcanic klippe with a steep northern slope.Its central part, however (between the Piszkes-tetö and the Kekes) is a fragment of the tectonically disairbed outer side of a caldera steeply sloping northward (SZEKELY A. 1983(SZEKELY A. , 1987)).During the landslides pyroxene-andesite lava, agglomerate and tuff (partly dacitic rhyolite tuff) were moving on the unconsolidateci bed underlying the volcanites, generally on Lower and Middle Miocene aleurolite (in some cases weathered tuff layers) as sliding plane.Already fossilised forms of landslides along the northern side of the Mätra crest make up a discontinuous zone with distinct morphological features.In the 7 selected depressions dammed by slumps two boreholes (M/3, M/5, Fig. 2) provided profiles perspective for dating.

Fekete Lake
The basin of the Fekete Lake located on the eastern flank of the Galya-tetö at an altitude 731 m a.s.l. was created by a single rhyme of a large-scale rockslide (Fig. 3).Based on the studies of the exposed failure front and the nearby road cut the sliding plane was heavily weathered dacitic rhyolite tuff.According to the information gained from the borehole (M/3) drilled in the depression with a diameter of 100 m the evolution of the lake can be summarised as follows (Fig. 4).
In the depression dammed by the slide a dystrophic lake (pH=5) rich in humus with low carbonate and nutrient content developed more than 9.000 years ago under cool and moist climate.At that time in the vicinity a mixed grove forest (linden and conifers with hazel in the bush level) existed.With the gradual upfilling of the lake by the end of the Boreal phase in the coastal zone bulrush (Typha) and reed (Phragmites) appeared, then with the advancement of moss and fern the lake started to turn into bog.It became a typical swelling bog in the Atlantic phase indicated by an enrichment in spores of acidophilous mosses, especially Sphagnum species, those of the sedge family (Cyperaceae) and ferns (Athyrium.Cystopteris, Dryopteris).The above age has been confirmed by a radiocarbon dating of 6,159±60 yr for the horizon between 4.2 and 4.4 m.In this phase there was a relative decrease in the share of the arboreal pollen for it was a swelling bog with no trees.Pre sently this type is very characteristic under moist oceanic climate (ombrogenic bog).In the following phase the reduced amount or the temporary disap pearance of beech (Fagus) and hornbeam (Carpinus) indicate a change in climate with drier summers and increased continental influence.According to 14 C da ting this occured 3.300 years ago i.e. in the subboreal phase.Aridisation of summers resulted in the appea rance of pine (Pinus silvestris) and in the emergence of a pine-bog.During this dry phase the pollen con taining ability of the accumulating sediments became heavily reduced and only the most resistant spores and pollen survived in low quantity not suitable for analysis.The advancement of pine is documented in other regions, too: e.g.E. KRIPFEL (1986) carrying out stratigraphic analyses of the bogs in Slovakia pointed on a similar phenomenon having taken place in the subboreal phase.Drier and moister spells were alter nating in the Subatlantic phase.The plant cover in the vicinity was represented by hornbeam-beech forests.
The bog was settled by alder (Alnus) and a slightly acidic (pH=5-6) dystrophic bog developed.Climates in the Subatlantic created favourable conditions for the climax of the submontane beech forests.Fekete Lake is still situated in the submontane beech wood land zone with alder grove forests, oak and elm.Based on the above records on the bog development the landslide is assumed to have occurred in the wa ke of the Preboreal phase or even earlier, during the improvement of the climate subsequent to a cold spell of the Late Glacial.

Köris Bog
Its basin is situated at an altitude (705 m a.s.l.) similar to that of the Fekete Lake, and the fonning landslide by the prensent morphological features (Fig. 5.) oc curred in an main single phase.The deep sliding pla ne of the slide was probably a weathered pyroxeneandesite tuff but it might have been the unconsolida ted underlying bed of volcanites.The landslide of the Köris Bog forms an organic part of the landslide zone flanking the northern crest of the Mätra Mountains.Its dating was based on the analyses of the (M/5) bore hole profile (Fig. 6).
Pollen record shows a dominant coniferous forest with deciduous species during the deposition of the lowermost horizon.Pine (Pinus silvestris) and spruce (Picea) were associated with some elm (Illmus) and linden (Tilia).This vegetation was characteristic for the late Preboreal.Abundant atmospheric precipi tation created favourable conditions for the emergen ce of ombrogenic bogs.On the acidic debris a dystro phic bog developed poor in carbonates and nutrients with acidophilous bryum and fern.It developed into a swelling bog in the Boreal.In this phase the conifers were retreating from the surroundings and a deci duous forest emerged with linden, elm, oak and hazel in the shrub level.Radiocarbon analyses of these ho rizons (7.250 and 7.720 yr) seem to corroborate re sults of the palynological analysis (Fig. 6).The rate of sedimentation slowed down in the Atlantic (from an earlier 40 cm to 20 cm per 1000 yr).The reason might have been that the peat-bog had turned into a moist boggy meadow of sedge vegetation surrounded by high stalk.With the emergence of grass (Gramineae) a bog meadow started to develop.The Subboreal was dominated by beach, hornbeam, oak and linden.In creasing continentality in climate can be documented with the drop in the share of beech pollen and by the disappearance of hornbeam.In the Subatlantic the high sedge meadow survived which (after a recurring aridisation) was characterised by the enrichment of paludal elements.Also this is the present-day situa tion.Pollen of Larix already indicates coniferous plan tations.
On the basis of the bog development the occur rence of the landslide (similar to that of the Fekete La ke) can be put to the Early Holocene or to a transitio nal spell from the Late Glacial to the Preboreal.

Conclusions
First detailed studies on the time of the occurrence of landslides in the volcanic ridges of Hungary C 14 datings point out 8.000 to 10.000 yr B.P. as a minimum age of landslides.Theoretically they might even be ol der but the fact that paludal sediments of the Köris Bog immediately overlie the rock debris of the slump seems to limit the age of this phenomenon.Though these results cannot be extrapolated to all the fossil landslides identified in the mountains (their number is ca.20) without more ado nevertheless they are orien tating.