(1999) 11, 267–284 drainage on evolving fold-thrust belts...

Basin Research (1999) 11, 267–284

Drainage on evolving fold-thrust belts: a study oftransverse canyons in the ApenninesW. AlvarezDepartment of Geology and Geophysics, University ofCalifornia, Berkeley, USA, and Osservatorio Geologico diColdigioco, 62020 Frontale di Apiro (MC), Italy

ABSTRACT

Anticlinal ridges of the actively deforming Umbria–Marche Apennines fold-thrust belt aretransected by deep gorges, accommodating a drainage pattern which almost completely ignoresthe presence of pronounced anticlinal mountains. Because the region was below sea level untilthe folds began to form, simple antecedence cannot explain these transverse canyons. Inaddition, the fold belt is too young for there to have been a flat-lying cover from which therivers could have been superposed.

In 1978, Mazzanti & Trevisan proposed an explanation for these gorges which deserveswider recognition. They suggested that the Apennine fold ridges emerged from the sea insequence, with the erosional debris from each ridge piling up against the next incipient ridge toemerge, gradually extending the coastal plain seaward. The new coastal plain adjacent to eachincipient anticline provided a flat surface on which a newly elongated river could cross thefold, positioning it to cut a gorge as the fold grew. Their mechanism is thus a combination ofantecedence and superposition in which folds, overlying sedimentary cover and downstreamelongations of the rivers all form at the same time.

A study of Apennine drainage, using the sequence of older-to-younger transected Apenninefolds as a proxy for the historical evolution of drainage cutting through a single fold, showsthat transverse drainage forms when sedimentation dominates at the advancing coastline.Longitudinal drainage forms when uplift dominates, the folds first emerge as offshore islandsand the Mazzanti–Trevisan mechanism is suppressed.

Complicating factors include several departures from steady-state growth of the fold-thrustbelt, a possible case of precursory submarine drainage, early emergence of anticlinalculminations and the location of several transverse canyons at the structurally highest pointalong anticlinal axes.

the advancing front of deformation. The fold-thrust beltINTRODUCTIONof the Umbria–Marche Apennines is still forming, and

This paper deals with the classic geological and geo- comparison of younger to older anticlines provides amorphological problem of the origin of river gorges that proxy for the evolution of the drainage cutting a singlecut through topographic barriers. It focuses on the anticline and shows in detail how the mechanism ofUmbria–Marche Apennines of the Italian Peninsula, Mazzanti and Trevisan has operated.where transverse gorges are being cut in a very young Nineteenth-century attempts to explain deep riverfold-thrust belt. Neither antecedence nor superposition gorges transecting the broad structural uplifts of thealone can explain the cross-cutting Apennine rivers. The Colorado Plateau led to formulation of the well-knownlittle-known work of Mazzanti & Trevisan (1978) offers concepts of (1) antecedence, in which the rivers pre-datean attractive solution to the origin of the Umbria–Marche the structural deformation and simply cut down throughgorges; it invokes a combination of antecedence and the uplifts as they grow, and (2) superposition, in whichsuperposition taking place at the advancing coastline near old uplifts are buried by younger, flat-lying sediments,

and with resumed erosion, rivers wandering across theflat-sediment surface are let down onto the old, stableCorrespondence: Walter Alvarez, Department of Geology anduplifts beneath.Geophysics, University of California, Berkeley, CA

94720–4767, USA. E-mail: [email protected] The Appalachian Valley and Ridge Province is the

© 1999 Blackwell Science Ltd 267

W. Alvarez

classic area for the problem of rivers cutting through the sea during the late Cenozoic. Therefore, the rivers areno older than this, which eliminates subaerial antecedencenarrow ridges of fold-thrust belts. Mechanisms proposed

to account for transverse drainage in fold belts have been as a mechanism. There is little time available and noevidence to be found for the deposition of an overlyingclassified in various ways (Strahler, 1945; Oberlander,

1965, 1985; Thornbury, 1965), but the principal sug- flat sedimentary surface from which the streams couldhave been superposed. The peculiar stratigraphicgested mechanisms are (1) antecedence, (2) superposition,

(3) headward erosion into a ridge, following weak zones sequence of the Zagros, with an easily erodable intervalsandwiched between resistant units, which made Oberlan-along transverse faults, and (4) modification of original

consequent drainage by captures. der’s (1965) mechanism plausible, does not occur in theApennines. However, a different mechanism has beenIn a study of spectacular river gorges in the Zagros

Mountains fold-thrust belt, Oberlander (1965, 1985) proposed by Mazzanti & Trevisan (1978). It is evaluatedbelow and appears to offer a satisfactory explanation foroffered a new mechanism for generating transverse drain-

age. He noted (Oberlander, 1965, figs 59, 60, 64) that in the Apennine gorges.an area of gorges not explainable by standard mechanisms,two resistant limestone formations are separated by aunit of easily eroded flysch which is thicker than the Apennine geology and landscapeamplitude of the folding. He proposed that a relativelyflat erosion surface developed in the folded flysch interval, The Apennine Mountains, running the length of the

peninsula, are composed largely of Mesozoic and Ceno-and from this surface the streams were let down acrossthe folded limestone unit below. Oberlander’s mechanism zoic sedimentary rocks deformed by north-eastward

transport which began in the Oligocene and continues tois thus a variant on superposition.Burbank et al. (1996) have studied the response of the present. Trains of folds are rare in the Central and

Southern Apennines, which were formed by the thrustingrivers and their deposits to growing folds emerging fromaggrading alluvial plains. Such a stream may maintain its of massive platform carbonate sequences. In the northern

part of the range, and particularly in the Umbria–Marchecourse across the fold, or may be defeated and divertedaround the fold by avulsion or capture, depending on Apennines (Fig. 2), well-bedded Mesozoic and lower

Cenozoic pelagic carbonates and upper Cenozoic turbiditethe ratio of aggradation behind the fold to uplift, andthe ratio of stream power to erosional resistance. In the units have been deformed into a fold-thrust belt which

is continuing to propagate north-eastward, toward thecases studied by Burbank et al. (1996), there is noquestion that these transverse streams are antecedent to foreland of the Adriatic Sea.

Italian geologists have long known that deformation inthe folds.the fold belt of this part of the Apennines migratedprogressively north-eastward (Migliorini, 1948; Merla,TRANSVERSE DRAINAGE IN THE1951). This is clear from the age of the synorogenicUMBRIA–MARCHE APENNINESturbidites deposited just in front, and quickly deformedby, the advancing thrust front. These turbidites dateThe present paper calls attention to another way in which

transverse drainage may form. In the Umbria–Marche from the late Oligocene in Tuscany (Macigno Formation),the early Miocene around Perugia (Cervarola Formation),Apennines of northern peninsular Italy, anticlinal ridges

of a young fold-thrust belt are transected by gorges the middle Miocene around Gubbio (Marnoso-arenaceaFormation) and the late Miocene east of the main ridgesseveral hundred metres deep, allowing the rivers to flow

straight to the sea, across the structural and topographic of the fold belt (Piceno Formation, south of the studyarea). North-eastward migration was well understood bygrain of the mountains (Fig. 1).

These canyons have been important as routes of the date of the first major synthesis in English (Bortolottiet al., 1970, fig. 39). Modern ideas on the tectonics ofcommunication since antiquity. One such canyon, the

Gola del Furlo, forms a bottleneck on the Via Flaminia, fold-thrust belts were just starting to emerge at that time(Bally et al., 1966; Dahlstrom, 1969) and had not yetthe road that connected the imperial capitals of Rome

and Ravenna in late Roman times, and was the site of a reached Italy, where, in the early 1970s, there was stilldebate over autochthony vs. allochthony in the Apenninesbattle during the Gothic War of the 6th century AD. In

that battle, the Byzantines defeated the Goths and (Abbate et al., 1970; Alvarez, 1972).In a remarkable tectonic pattern, a north-eastward-reopened communications between Rome and Ravenna

by means of a geological tactic – they pushed boulders propagating extensional front is following about 100 kmbehind the compressional front (Elter et al., 1975), soover the cliffs and down onto the fortifications below,

until the Goths surrendered in terror (Procopius, #560, that normal faults tear apart the fold-thrust edifice shortlyafter it is generated. At present the extensional frontBook VI, Ch. xi).

These gorges are difficult to explain by standard has reached Gubbio (Fig. 2). The propagating com-pressional–extensional couple inferred from structuralmechanisms. The presence of Miocene turbidites across

the region and Pliocene marine sediments in the north- observations is confirmed by focal mechanisms of earth-quakes (Lavecchia et al., 1994) and by detailed studieseast demonstrates that the fold belt emerged from the

© 1999 Blackwell Science Ltd, Basin Research, 11, 267–284268

Transverse canyons in the Apennines

Fig. 1. View east into the Frasassi Gorge. This valley is representative of the transverse canyons of the Umbria–MarcheApennines, although steeper sided than some others, because it cuts through the massive Jurassic platform limestone of theMassiccio Formation.

Fig. 2. Map of northern peninsular Italy, showing the asymmetric drainage pattern and the present positions of the north-eastward-migrating compressional and extensional fronts which apparently control the asymmetric drainage. The Umbria–Marche Apennines trend NW–SE between Perugia and the Adriatic Sea. CROP-03 and Bally 2 (Barchi et al., 1998, plate 4;Bally et al., 1986, section 2) are the lines of the contrasting interpretive structural cross-sections in Fig. 7.

of the CROP-03 seismic reflection profile (Pialli et al., these stresses would not change from extensional toextensional in only 100 km. The pattern suggests that1998).

The reason for the propagating tectonic couple is not the block between the extensional and compressionalfronts is moving north-eastward, generating compressioncertain. It clearly cannot be due to applied stresses

transmitted horizontally through the lithosphere, because in front and extension behind. Yet the block is too narrow

© 1999 Blackwell Science Ltd, Basin Research, 11, 267–284 269

W. Alvarez

to be driven by viscous coupling to flowing mantle at primary questions have been addressed in this literature.From the broadest to the most detailed, the questionsits base.

Therefore, various authors have suggested that some are: (1) Why is the drainage of this part of the Peninsulaasymmetric? (2) Why is the main drainage divide locatedportion of the crust and/or upper mantle of the Apen-

nines is delaminating, rolling back toward the north-east, in relatively low country, offset to the south-west of thehighest ridge? (3) Why are many of the anticlinal ridgesand sinking into the mantle. This rollback would require

the overlying block to move north-eastward, as observed. cut by transverse gorges?Italian geomorphologists working in this area haveIt is not clear which entity is delaminating. Reutter et al.

(1980) identified the sinking entity as the subcontinental been most interested in the first two questions (Bonarelli,1891; Marinelli, 1926; Castiglioni, 1934; Merla, 1938;lithospheric mantle; Castellarin et al. (1982) considered

it to be the lower continental crust; Malinverno & Ryan Giannini & Pedreschi, 1949; Sestini, 1950; Selli, 1952;Ghelardoni, 1958, 1962; Gonsalvi & Papani, 1969; Cat-(1986) showed most or all of the continental crust

detaching. Despite these differences, all these authors tuto, 1976; Mazzanti & Trevisan, 1978; Dramis & Bisci,1986; Cencetti, 1988; Cattuto et al., 1989). The origin ofagreed on the basic geometry. Pialli & Alvarez (1997)

argued that the delaminating entity is the lower continen- transverse gorges has received less attention. Neverthe-less, one paper (Mazzanti & Trevisan, 1978) proposed atal crust and lithospheric upper mantle, and proposed

that the sinking and rollback is driven by the density novel and important solution to this problem, which isin accord with recent concepts of the evolution of fold-increase that accompanies the inversion from basaltic

composition to eclogite when the lower continental crust thrust belts and with new observations on the Apenninespresented here.reaches about 50 km depth. Careful study of the results

of the CROP-03 profile (Pialli et al., 1998) should allowfurther progress in understanding this mechanism for

THE MECHANISM OF MAZZANTI ANDgeneration of the Apennines.TREVISANThe rivers of the Umbria–Marche Apennines form

the eastern half of an asymmetrical drainage pattern in In the course of a paper which deals with all three classicproblems of landscape in the Umbria–Marche Apenninesthis part of the Italian Peninsula (Fig. 2). Drainage to

the Tyrrhenian Sea, through the region behind the – drainage asymmetry across the Peninsula, displacementof the drainage divide and transverse canyons – Mazzantiextensional front, is collected into a few long trunk

streams – the Tiber, the Ombrone and the Arno – in a & Trevisan (1978, fig. 5) proposed an ingenious expla-nation for the transverse gorges (Fig. 6). Although theytrellis pattern, with long segments following grabens

parallel to the peninsula, interrupted by steps to the did not develop this idea at length, it appears to agreewith what is seen on maps and in the field, and withsouth-westward. In contrast, the many short rivers that

drain north-eastward across the compressed but not-yet subsequent theoretical understanding of fold-thrust belts.The stratigraphic sequence of this part of Italy isextended area (Fig. 3) are closely spaced, parallel and

follow relatively straight paths to the Adriatic Sea (Maz- exclusively marine. Limestone and marl dominate fromthe Jurassic to the Lower Tertiary and are replacedzanti & Trevisan, 1978; Hovius, 1996).

The fold structure in the Umbria–Marche Apennines upward by terrigenous turbidites in the Late Tertiary(Bortolotti et al., 1970; Cresta et al., 1989). Marineis expressed in elongated anticlinal ridges (Fig. 4). The

ridges culminate in peaks with elevations from 1500 m deposition continued until approximately the time ofemergence of each anticline, with the emergence in anyin the north to 2400 m south of the study area. Because

the rivers of the Umbria–Marche Apennines pass through given place dated to approximately the nearest chrono-stratigraphic stage.these anticlines in deep gorges, the network of major

rivers seems not to reflect the geological structure, even Mazzanti & Trevisan could not yet know or portraythe deep structure of the anticlinal ridges, but theythough on a smaller scale differential erosion has sculpted

prominent strike valleys in weaker units like the middle correctly showed the ridges forming progressively towardthe north-east (Mazzanti & Trevisan, 1978, fig. 5), andCretaceous Fucoid Marls. The lack of structural control

on major streams can be seen by comparing a geological they realized that as each anticline emerged from the sea,its erosional debris would be shed into the synclines onmap (Fig. 5) to the drainage map (Fig. 3) of the same

area. From the drainage map alone, it would be difficult either side. On the seaward side, some of the sedimentwould be trapped behind the next incipient anticline,to infer the presence of a belt of large parallel folds.thus extending the coastal plain seaward. As a result, theAdriatic rivers were progressively elongated in the down-stream direction. In addition, the newly formed coastalGeomorphological questionsplain provided a surface from which each new segmentof each river could progressively cut down through theWhile there is a rich literature on the evolution of the

landscape of this part of the Apennines, it has not underlying anticline as it grew.The Mazzanti–Trevisan mechanism is thus a combi-received the international recognition it deserves because

it has been published almost entirely in Italian. Three nation of superposition and antecedence, but quite



Fig. 3. Drainage pattern of the study area, showing prominent geographical features and the locations of the profiles shown inFigs 7 and 8. Note that the anticlinal ridges trending NW–SE (Figs 4 and 5) are almost unrecognizable in the drainage pattern.

different from the Colorado Plateau ‘anteposition’ of RELEVANT ASPECTS OF THEHunt (1956, p. 65). In the Apennine case, the river is TECTONICS OF FOLD-THRUST BELTSsuperposed on the anticline from a surface that forms atessentially the same time as the initiation of folding, and Great strides have been made in the study of fold-thrust

belts since the 1960s, and some aspects of this newthus the river is antecedent to all but the very first stageof deformation. It avoids the insuperable problems of (1) understanding are relevant to evaluating the Mazzanti–

Trevisan hypothesis. The major breakthrough came withsimple superposition, where there is no evidence of andno real possibility for unconformable sedimentary cover the recognition that reverse faults cutting the fronts of

anticlines do not steepen with depth, but flatten into soleor higher thrust sheets from which streams could be letdown across the Apennine folds, and (2) simple anteced- thrusts. Seismic data from the Canadian Rockies showed

the entire fold-thrust belt to have been been pushedence, where rivers could not pre-date folds in sedimentswhich were below sea level even after the folding began. forward over its basement, up a gently dipping basal

detachment (Bally et al., 1966). More detailed studies ofThe heart of the mechanism is that the fold, the coastal-plain surface and the downstream growth increment of this deeply exposed and intensely drilled mountain range

made it possible to recognize that flat thrusts in weakthe rivers were all forming at essentially the same time.As discussed in the next sections, the mechanism of stratigraphic units ramp up through stronger units to

flatten again in weak horizons above, with anticlinesMazzanti & Trevisan agrees well with the understandingof fold-thrust belt evolution that has emerged since their forming above the ramps, a pattern now recognized as

characteristic of fold-thrust belts (Boyer & Elliott, 1982).paper was written, and with detailed geological obser-vations in this part of the Apennines. Before turning to Thrust arrays propagate toward the foreland, with new

active thrusts taking over from older, inactive ones, andan evaluation of the Mazzanti–Trevisan mechanism, itwill be useful to describe the structures across this part thus the belt of resultant anticlines also propagates toward

the foreland. This progressive appearance of new folds,of the Peninsula in the light of current understanding ofthe tectonics of fold-thrust belts. characteristic of fold-thrust belts, is what Apennine


W. Alvarez

Fig. 4. Topography of the study area, showing the geological/topographic belts discussed in the text. Squares, triangles, two-letter abbreviations and thin lines show towns, peaks, gorges and profile lines identified in Fig. 3.

geologists recognized on the basis of biostratigraphy long driven forward. In the continuum-scale view of fold-thrust belts, the detailed erosional features of the surfacebefore its explanation was understood, and it was essential

to the Mazzanti–Trevisan hypothesis. slope are below the limit of resolution. However, thesedetails – the transverse gorges of the Apennines, forIt is now clear, however, that orderly, in-sequence

formation of new thrusts during overall forward propa- example – may provide a record of historical conditionssuch as the uplift–sedimentation balance at the coastline,gation may be modified in some cases by out-of-sequence

thrusts (Morley, 1988). There is presently an active debate or of structural conditions at depth, including the pres-ence of out-of-sequence thrusts.over the extent of out-of-sequence thrusting in the

Umbria–Marche Apennines (Ghisetti et al., 1993). Becausethe Mazzanti–Trevisan mechanism operates right at the TRANSECT ACROSS THE DEFORMEDcoast, it would not be invalidated by out-of-sequence BELTthrusting. However, the orderly progression of higheranticlines with deeper gorges shown in their diagram The present study area includes 15100 000 map sheets

109 ‘Pesaro’, 110 ‘Senigallia’, 116 ‘Gubbio’, 117 ‘Iesi,’(Fig. 6) could be interrupted by out-of-sequence thrusting.Further understanding of fold-thrust belts came by and 118 ‘Ancona’, covering the entire Adriatic drainage

of this part of the Italian Peninsula.backing away from the detailed geometry of the structureswithin the deformed belt and considering the entire belt The subsurface structure of the Umbria–Marche

Apennines fold-thrust belt is still controversial. Figure 7as a wedge-shaped continuum bounded by a rearward-dipping basal detachment, a forward-sloping upper sur- shows two contrasting interpretive sections across the

Apennines in the study area. The lower section (redrawnface and an advancing driver at the rear (Davis et al.,1983; Dahlen et al., 1984). In this view the fold-thrust from Bally et al., 1986; plate 8, fig. 55) assumes thin-

skinned deformation, a gently dipping basal detachment,belt is seen as a steady-state system in which basal slidingcauses internal deformation which is accompanied by and an orderly progression of thrust and anticline forma-

tion from south-west to north-east, based on the Canadianshallow-level adjustments in order to maintain a criticaltaper between basal detachment and surface slope, so Rockies model for evolution of fold-thrust belts. Seismic

reflection data at that time were not good enough tothat all stresses stay in balance as the fold-thrust belt is



Fig. 5. Generalized geological map of the study area, showing the lack of correspondence between geological structure anddrainage lines. Dashed line shows location of Fig. 9. Black squares and straight lines are towns and profiles as in Fig. 3.

image the structure or the basement in the complicated drainage pattern of the Tyrrhenian side of the Peninsula.To the north-east is a hilly belt underlain by the turbiditeswestern half of the section. The upper section (redrawn

from Barchi et al., 1998; Plate 4) is based on the recent of the middle Miocene Marnoso-arenacea Formation,which were deposited in a synorogenic foredeep, in orCROP-03 seismic reflection profile, which provided evi-

dence for involvement of basement in the thrusting and around the thrust front, and quickly deformed as thethrust front advanced (Damiani et al., 1983). Most ofa greatly reduced shortening across the fold-thrust belt.

The differences between these sections, although criti- this hill country is drained by short streams flowingsouth-west to the Tiber.cal for understanding the origin of the Apennines, do

not strongly affect the interpretation of landscape evol-ution in the present paper, for in both cases the thrusts 2. Gubbio valley and faulted half-anticlineand folds propagate north-eastward across the belt offolds from the Monte Nerone – Monte Catria anticline Gubbio is built on a fault escarpment that cuts a small

anticline, dropping the south-west half down to make ato the present coastline. The two viewpoints are admir-ably contrasted in a ‘civilized debate among friends who depression occupied in Pleistocene time by a lake. The

Gubbio half-anticline exposes Cretaceous and Lowerdisagree’ (Ghisetti et al., 1993).Starting in the south-west part of the map area and Tertiary pelagic limestones. The fault that cuts the

anticline is the north-easternmost major extensional faultproceding to the north-east, the following major geo-logical and drainage belts appear (Figs 3–5). and marks the present position of the advancing exten-

sional front. The Gubbio Valley represents the newestpart of the Tyrrhenian trellis drainage. The half-anticline1. Belt of Miocene turbidites between theis presently being dissected by short steep valleys (Con-Tiber Valley and Gubbiotessa Valley and Bottaccione Gorge) which are apparentlyvery young, dating from the initiation of the exten-The Tiber Valley, about 10 km south-west of the map

area, is the easternmost well-developed part of the trellis sional faulting.


W. Alvarez

this belt, lying at an elevation of about 850 m, which islower than the anticlinal crests which flank it on bothsides. This rather surprising situation has been discussedin the Italian literature, as noted above.

4. The Umbria–Marche ridge

The most prominent topographic and geological featureon this transect through the Apennines is a ridge formedby closely spaced, parallel anticlinal mountains and knownas the ‘Umbria–Marche Ridge’, because it follows theboundary between these two regions. In this anti-clinorium, two major anticlines (Monte Nerone, MonteMontiego) or three of them (Monte Cucco, Monte Catria,Monte della Strega) may occur side by side, separatedonly by narrow, faulted synclines. Exposures are notdeep enough to reveal what role, if any, was played bythrusts and thrust ramps, and the deep Burano well, onthe crest of the main anticline, did not intersect anymajor thrust fault. The anticlines typically end at steeplyplunging terminations, where they may be continued enechelon by the next anticline. Local axial culminations(structural highs along the trend of the fold axis) mayreflect the difference between horst and basin facies inthe Jurassic limestones (Alvarez, 1989a,b), or may resultFig. 6. The concept for the origin of the Apennine transversefrom the pattern of Miocene deformation.drainage evaluated in this paper (Mazzanti & Trevisan, 1978,

Major streams flowing to the Adriatic transect thisfig. 5). The original caption reads: ‘Model of mixed process ofanticlinal complex via a number of deep gorges (Biscubio,superposition followed by antecedence. In A, a fold which is

beginning to form creates a gentle rise on the ocean floor; the Gorgo a Cerbara, Bosso, Burano and Sentino). The Bossodepression between the rise and the coast fills rapidly with and Burano gorges pass through a broad saddle betweensediment, largely of molassic type. In B, the original depression, the peaks of Monte Nerone and Monte Acuto, but Gorgowith gradual uplift, emerges as a flat surface inclined toward the a Cerbara cuts through the culmination of the MontiegoAdriatic, while new folds continue to ripple the sea floor. In C Anticline, in a remarkable pattern also seen elsewhere inand D, the process continues to propagate, extending the

the study area. Subsequent tributaries and longitudinalemergent land, while the accentuation of the preceding foldsportions of the trunk streams flow parallel to the axialand a general uplift lead to the deepening of the gorges whichtrends for short distances, following synclinal axes or thecut the anticlines’ (transl).strike valleys of weaker units (particularly the middleCretaceous Fucoid Marls, whose prominent valley is aclear guide to the structure throughout the anticlinal belt).

3. Miocene turbidite belt and the Apenninedrainage divide

5. The Internal Marche FoldsNorth-east of the Gubbio half-anticline, the Mioceneturbidites continue in a generally synclinal belt. This North-east of the major anticlinorium is a 10-km-wide

belt of lower, shorter anticlines exposing Cretaceousarea has a complicated internal structure involving thrustsand folds whose geometry and kinematics are still being limestones, separated by synclines occupied by Miocene

terrigenous units. Some of the synclines were the depos-debated, particularly as to whether they formed in anorderly sequence progressing to the north-east (Bally itional sites of synorogenic turbidites that flowed north-

eastward, across and around the growing anticlines (Cen-et al., 1986) or in a complicated, irregular sequenceinvolving backward, out-of-the-synclines motions (de tamore et al., 1976).

Continuing the pattern seen in the Umbria–MarcheFeyter & Menichetti, 1986; de Feyter et al., 1990). Atthe north-east edge of this belt, the youngest turbidites Ridge, the rivers of this belt cut through anticlines in

several places, avoiding the easier routes around thesehave been folded into the prominent Monte Vicinosyncline which closely reflects the geometry of the last small folds. Particularly striking are the Acqualagna

Anticline, where two streams enter the fold, join at thedeep synclinal basin in this part of the advancing fold-thrust belt (Centamore et al., 1977). axial culmination, and flow out through the other side of

the fold, and the narrow gorge of Frasassi (Fig. 1).The drainage divide of the Peninsula passes through



Fig. 7. Two contrasting interpretive cross-sections across the Umbria–Marche Apennines fold-thrust belt in the study area,showing progressively younger thrusting toward the north-east, and the advancing extensional front, after Bally et al. (1986) andBarchi et al. (1998). The sections are drawn along lines roughly 20 km apart (Figs 3–5), and the fold-thrust structure plungesnorth-westward from the Bally section to that of Barchi. A topographic profile has been added to each section, showing thehighest ridge line within 4 km of the line of section. Note that the grey unit represents a thicker stratigraphic interval on theBally section. The differences between the two sections are not relevant to the landscape evolution discussed in this paper, sincethe shallow structure of anticlines A through F is not greatly different.

synorogenic Miocene and Pliocene terrigenous clastics.6. Paired anticlines of Furlo-Cesana, andAlthough there is not much topographic grain in thisthe External Marche Ridgearea (Fig. 4), the geological map (Fig. 5) shows a patternof gentle anticlines and synclines, whose intensity dimin-The large anticline of Furlo, cored by Jurassic limestones,

is flanked by the lower Cesana Anticline, which exposes ishes north-eastward. This belt is crossed by manystreams flowing north-eastward to the Adriatic (NanniCretaceous limestones. The Furlo Anticline has a double

plunge away from a sharp axial culmination, and the et al., 1986). There is an asymmetry to the pattern oftributaries, with the larger ones commonly entering theFurlo Gorge cuts exactly through the culmination. The

Cesana Anticline is transected at Fossombrone, near its trunk stream from the north side. The major streamscrossing this belt occupy floodplains 1–2 km wide.south-east plunge termination, by the river emerging

from Furlo, which is joined by another major streamthat follows the syncline between the two anticlines. 8. Incipient folds of the Adriatic coastal

South-east of a broad axial depression at Pergola, the plain and offshoreFurlo structure is continued by the External Marche

Along much of its length, the Adriatic coast in this areaRidge, a long anticlinorial ridge that rises southward todoes not show an obvious structural control. Exceptionsthe Sibillini Mountains. This ridge is also cut throughoccur between Cattolica and Pesaro, where the coastby a number of deep gorges, including the Gola dellafollows the outer edge of an anticline reaching an elevationRossa. In front of it at one place is the Cingoli Anticline,of about 200 m, and south of Ancona, where the anticlinein a position analogous to that of the Cesana Anticline,of Monte Conero, exposing most of the Cretaceous, isand cut by two transverse gorges.unusually high (572 m) to be this far in front of the mainmountain front.7. Foothill belt of minor folds

The Adriatic Sea in this area is shallow and is occupiedby as much as 5 km of Pliocene–Quaternary terrigenousBetween the Furlo-Cesana Anticlines and the Adriatic

Sea is a belt of lower hills exposing easily eroded clastics, deposited in a foredeep basin in front of the


W. Alvarez

advancing Umbria–Marche fold belt (Ori & Friend, shortening, rock uplift, erosion and deposition. In thebelt studied here there are at least five departures from1984). Seismic data and exploration drill holes show that

the foredeep sediments are deformed into a set of gentle steady-state evolution.1 The older anticlinorium of the Umbria–Marche Ridgeanticlines, some of which produce oil or gas, without

giving rise to significant relief on the shallow sea floor appears to be more complex, with multiple adjacent folds,than the younger pair of Furlo-Cesana, and the still(Fig. 7). The present deformation front is roughly in the

middle of the Adriatic Sea. younger single anticlines near the coast. This may indicatethat the style of folding changed through time, or thatout-of-sequence thrusting continues to modify earlierEVALUATION OF THEanticlines, or it may simply be an artefact of depth ofMAZZANTI–TREVISAN EXPLANATIONexposure.OF CROSS-CUTTING DRAINAGE2 There are not any higher, more eroded folds west ofthe Umbria–Marche Ridge, so that structure marks theThe purpose of the present study is to examine evidence

relevant to the Mazzanti–Trevisan hypothesis, in order west boundary of the transect that can be used as a proxyfor time sequence.to evaluate whether it is a satisfactory explanation of the

cross-cutting drainage in the Umbria–Marche Apennines. 3 Through at least the Tortonian, the Adriatic forelandbasin was a deep-sea environment in which turbiditesMazzanti & Trevisan offered their hypothesis in the form

of a single drawing and a brief description (Mazzanti & were deposited on and around the advancing thrust front.This basin was filled, probably diachronously, during theTrevisan, 1978; Fig. 6), and did not report detailed

observations from the Apennines. However, a careful Messinian and Pliocene. By Quaternary time the Adriatichad its present configuration with a shallow, sediment-examination of relevant evidence shows their hypothesis

is in detailed accord with what is seen on geological maps filled sea or coastal plain in front of and above theyoungest folds.and in the field.4 The Messinian salinity crisis, in which the isolatedMediterranean water evaporated, involved a sea-levelDrainage evolution reconstructed usinglowering of more than 1 km (Cita & Ryan, 1978).spatial sequence as a proxy for timeIncreased erosion and transport of sediments furthersequencefrom their site of erosion would have interrupted thesteady state in the fold-thrust belt, and some isostaticBecause of the propagation of the deformational front

toward the foreland in a fold-thrust belt, a transect across uplift may have occurred, depending on water depthprior to the desiccation.a set of anticlines can be taken as a rough proxy for the

time sequence of deformation and erosion in a single 5 During Quaternary time, there have been repeatedsea-level oscillations of about 100 m. On the steep Tyrrh-anticline. The Umbria–Marche fold-thrust belt provides

an excellent opportunity for this approach, displaying a enian continental margin, this produced repeated chan-nelling and filling of river valleys near the coast (Alvarezsequence of progressively younger and less eroded anti-

clines toward the north-east, in the direction of fold- et al., 1996; Karner & Marra, 1998). On the flatterAdriatic continental margin the channelling effects werethrust propagation. A landscape history reconstructed on

this basis is in agreement with the Mazzanti–Trevisan probably less, but the migration of the coastline wouldhave been greater, and the broad floodplains of themodel, but some caution needs to be kept in mind.

A first caution is that the approach used here is valid Adriatic rivers may reflect incision during Quaternarysea-level lows. In addition, because the maximum rate ofto the extent that the fold belt and its internal structures

propagate forward in an orderly sequence. In the basic sea-level change during the Quaternary glacial oscillationsprobably greatly exceeded the rate of rock uplift, sea-tectonic model, with thrusts propagating in sequence

(Bally et al., 1986), a single anticline would be folded level history may well have been the controlling factorin the development of transverse or longitudinal drainagerapidly as the front of thrusting passed, and subsequently

would retain its folded shape while being slowly uplifted in the youngest anticlines (Robert S. Anderson, writtencommunication, 1998).as it was passively pushed over the upward-sloping basal

detachment. In a more complicated model, with out-of-sequence and rearward-directed thrusts (Morley, 1988),anticlines would continue to change their profile, as well Drainage crossing the sequence of anticlinesas being uplifted, after the initial wave of deformationhad passed. Which of these two models is applicable to Transverse drainage is impressive and puzzling where it

chops through tall anticlines in deep, narrow gorgesthe Apennines is still under debate (Ghisetti et al., 1993).A second caution is that spatial pattern as a proxy for (Bosso, Burano, Gorgo a Cerbara and Furlo). It is less

noticeable where it passes through the gentler anticlinestime sequence will be valid to the extent that fold-thrustbelt evolution has been a steady-state process, with a between Fossombrone and the coast, where the folds are

barely reflected in the landscape of gentle hills. It isroughly constant structural and topographic profile shift-ing toward the foreland through a balance between instructive to examine drainage–structure relations from



north-east to south-west, as a proxy for the evolution of also apparent that they are being let down across gentleanticlines affecting the Pliocene sediments, as shown bya single anticline through time (Fig. 8).the situation at Cattolica (Fig. 5). Here the Conca Riverflows straight into the Adriatic across the north-westAdriatic coast. Trunk streams flow almost straight to the

sea across the belt of coastal hills, on floodplains up to plunge termination of a low-relief anticline (Casabiancaet al., 1995; De Donatis et al., 1998) that exposes Miocene4 km wide, cut slightly below the level of the adjacent

hills. Within a few kilometres of the coast, the hills are sediments and controls the position of the coastline. Afew kilometres north-west, at Riccione, on the other sidemainly composed of Pliocene foraminiferal marine clays,

with some sandstones and rare conglomerates (Castellarin of the Conca, the gentler continuation of the anticlineslightly warps Pliocene sediments (15100 000 map sheet& Stewart, 1989). Quaternary sediments are restricted to

the incised floodplains (Nanni & Vivalda, 1986) and to 109, section I). This anticline will probably continue torise as the fold-thrust belt is driven forward and, if so,discontinuous coastal terraces and narrow beaches (Dal

Cin & Simeoni, 1987). The coastal Pliocene is deformed the Conca will cut down through the rising fold.This situation conforms very well to the model Maz-into several anticlines which expose lower Pliocene, upper

Miocene and rarely middle Miocene. zanti & Trevisan (1978, fig. 5) proposed as the initialstage in the evolution of the deep gorges. The olderThe throughgoing rivers cut across these anticlines in

valleys with floors 100–200 m below the flanking hills. anticlines emerged first at the edge of a deep Adriaticwhereas the new anticlines face a shallow Adriatic, butBecause they cross Pliocene sediments of marine origin, it

is clear that these rivers have lengthened in the downstream this does not change or invalidate the mechanism. Atpresent no anticlinal islands exist offshore, a situationdirection as the marine Pliocene sediments emerged. It is

Fig. 8. Topographic profiles along the axes of six anticlines cut by the Metauro River and its tributaries (for locations, seeFig. 3). These are not sections taken along a single straight line; they are projections of the ridge line onto a N50W–S50Evertical plane and thus approximate the topography as seen from a distance. From near the coast (profile F), south-westward tothe highest anticline (profile A), there is a progressive increase in topographic elevation and a progressive deepening of thetransverse river valleys, except for the abnormally low Acqualagna anticline. The white band (grey where eroded) is the profile ofthe Middle Cretaceous Fucoid Marls along the anticlinal axes; it shows the plunge pattern of the folds and the progressivelyhigher structural uplift from the coast inland, again with the exception of the Acqualagna anticline.


W. Alvarez

that would constrain later rivers to go around them; the clines (Acqualagna, Bellisio Solfare, Gola di Frasassi) orthe axial depressions that flank them.common occurrence of transected anticlines indicates that

the offshore was commonly free of islands. However,Umbria–Marche Ridge. The Umbria–Marche Ridgedeviation of rivers around former islands evidently did(Fig. 8A,B), with its several deep gorges, can now beoccur in the past, from time to time, as discussed below.understood as the most advanced case in a sequence ofprogressively more uplifted and eroded anticlines. TheFoothills belt. Ten to 20 km inland the anticlines (Fig. 8F)transecting drainage, which is a puzzle when viewed inare more continuous, reach higher elevations (300–500 m)isolation, seems well explained by the Mazzanti–Trevisanand are more deeply eroded (Coward et al., 1999) – inmechanism. West of Piobbico, the Biscubio River cutsone case exposing the Eocene–Oligocene Scaglia Cinereaacross the plunge-out of the Monte Nerone Anticline,(near Fontecorniale). Nevertheless, the major rivers tran-providing a probable analogue, at a more advanced stagesect these anticlines as well as the younger coastalof evolution, to the plunge-crossing streams at Cattolicaanticlines. There can be little doubt that here we areand Montelabbate. The emergence of the Umbria–simply seeing a more advanced stage in the evolution ofMarche Ridge was not complete before the end of thethe drainage. A subsequent stage in the evolution of theMiocene because, as discussed above, the MessinianCattolica situation is probably represented at Montelabb-turbidite basin of Montaiate was fed laterally across itate, where a NW-plunging anticline is less eroded on the(Centamore et al., 1976). Its emergence was probablynorth-west side of the Foglia River than on the south-east.complete by the early Pliocene, which compresses theentire erosional history of the fold belt from here to theCesana Anticline. The same conclusion can be extendedAdriatic into the last 5 Myr. The proxy historicalto the still higher, more deeply eroded anticlines furthersequence ends at the Umbria–Marche Ridge, because ofinland. The Cesana Anticline, 25 km from the coast, risesthe absence of larger, more deeply eroded folds to theto about 650 m, broadly exposing the Upper Cretaceoussouth-west.Scaglia Rossa, and is transected at Fossombrone by the

Metauro River in a gorge cut down to the top of theConclusion. These observations not only support theLower Cretaceous Majolica Formation (Fig. 8E). Whenhypothesis of Mazzanti & Trevisan, but they suggest thatone travels south-westward from the coast along the Viatransecting gorges formed at chance locations along theFlaminia, following the Metauro River, Fossombrone isanticlines, wherever a newly elongated river happened tothe first noticeable gorge, because the Cesana Anticlinecross an incipient anticline. This disagrees with the viewis the first fold to have risen high enough to expose thethat the gorges are controlled by pre-existing faultserosion-resistant limestones below the weaker marlscrossing the anticlines (Dramis & Bisci, 1986, p. 100).

deposited from the late Eocene on. Thus the abruptSuch cross faults along the lines of transverse gorges are

north-east limit of the gorges at Fossombrone is due todubious in any case, for of 11 gorges included in the

depth of erosion in a stratigraphic sequence that contains most detailed geological-map quadrangles availablean abrupt change in erosional resistance; the uplift and (1550 000 sheets 290 ‘Cagli’ and 291 ‘Pergola’), only twoerosion pattern simply continues what is seen closer to (Belisio Solfare and the southern cut through the Acqual-the coast. agna anticline) show even modest offsets of geological

contacts and minor inferred cross faults.Furlo Anticline. The same pattern continues still further Within the study area there are no obvious examplesinland. The Furlo Anticline (Fig. 8D) reaches almost of ‘wind gaps’ – palaeovalleys through which the incising1000 m elevation, and the Furlo Gorge cuts down through river no longer flows. The Apennine rivers have evidentlythe Jurassic, where the massive lower Liassic Calcare been able to cut down at a rate at least as great as theMassiccio has produced a dramatic chasm. Here for the rate of surface uplift. Considering that the Umbria–first time the exposure is deep enough to cut below the Marche Ridge, which emerged late in the Miocene,characteristic stripping surface where the soft middle around 5–7 Ma, presently reaches elevations of aroundCretaceous Fucoid Marls rest on the resistant Majolica 1500 m, with several hundred metres of original coverFormation. The Furlo Gorge has an inner canyon in the removed, the surface uplift rate is roughly 0.5 mm yr−1.Jurassic and the Majolica, a shoulder on the top of the By contrast, the Wheeler Ridge anticline in the southernMajolica, and an outer canyon walled by the Scaglia San Joaquin Valley of California, on which a wind gapRossa. The impressive chasm at Furlo is due simply to developed (Burbank et al., 1996), has been uplifted atthe depth of erosion in the stratigraphic sequence, and about 3–3.5 mm yr−1 (Medwedeff, 1992; Burbank et al.,evidently results from a continuation of the progressive 1996) – almost an order of magnitude faster than thevalley deepening whose initial stage is seen at the coast. Apennine anticlines.

Internal Marche folds. Next inland is the belt of lowLongitudinal river tractsanticlines that interrupt the gradual trend of progressively

higher anticlines toward the south-west (Fig. 8C). Here Although most of the major rivers flow nearly straightnorth-east toward the Adriatic, transverse to theagain the rivers pass indifferently across the short anti-



structure, there are a few places in the study area where longitudinal tract of the Bosso between Secchiano andCagli follows an anticlinal crest, which at first is puzzling.the trunk stream or a major tributary flows south-east

for a few kilometres, parallel to the structural trends, However, the large anticline to the south-west is veryhigh and steep-flanked, and the Bosso has apparentlybefore returning to the usual transverse direction. These

longitudinal tracts are present along the Conca, Foglia been displaced away from the tight synclinal axis towardthe much smaller anticline to the north-east. The tight-and Metauro, and especially along the lower Musone

(Fig. 3). ness of the synclines is of ambiguous significance. Itcould be due to original spacing of the anticlines, toLongitudinal drainage, parallel to the fold axes, is what

would be expected if the Mazzanti–Trevisan mechanism erosion cutting deeper into a V-shaped synclinal profile,or to late tightening of the synclines as a result of out-were not at work, so that an incipient fold emerged from

the sea as an island before it was covered by the advancing of-sequence thrusting.In this interpretation of the drainage pattern, thecoastal plain. In this situation of diminished sediment

supply or enhanced fold uplift, a longitudinal stream, distribution of longitudinal tracts and transverse gorgesin a deeply eroded fold-thrust belt reflects the balance oftrapped behind the rising coastal anticline, would flow

parallel to the coast until it reached an axial saddle uplift and sediment supply in its early history, providinginformation of structural importance that could not other-through which it could reach the sea.wise be recovered. Whichever of the mechanisms fortightening of the synclines has acted, the result is thatLongitudinal drainage near the coast. A situation that

supports this interpretation is seen at Monte Conero, by longitudinal tracts in older parts of the fold-thrust beltflow more straight to the south-east, rigidly parallel tofar the largest of the coastal anticlines (Figs 4 and 5).

With a present relief of over 500 m, it is highly probable the fold axes, whereas younger longitudinal tracts closerto the coast wander in a generally south-east but obliquethat this anticline emerged before it was covered by the

advancing coastal plain. This would explain the strongly direction across the broader synclines.deflected longitudinal drainage south-west of MonteConero. Both the Musone River and its near-coast

COMPLICATIONS IN THE DRAINAGEtributary, the Aspio, flow south-eastward, around theobstacle of the mountain, and enter the sea where the EVOLUTIONsouth-east plunge termination of Monte Conero elimin- Precursory submarine drainageates that obstacle. In contrast, the major rivers directlynorth-west and south-east of Monte Conero, the Esino One would expect that the drainage lines in any part of

the fold-thrust belt would come into being only whenand the Potenza, flow unimpeded to the sea on trans-verse paths. that place emerged from the sea through uplift and

coastal-plain advance. In one place, however, there isIf this interpretation is correct, the paths of thestreams, transverse or longitudinal, provide a sensitive evidence for submarine sediment transport that may have

influenced the location of a later cross-cutting river.record of the fluctuating balance between sedimentsupply, anticline uplift rate and sea-level change at the On both sides of the Umbria–Marche Ridge the

youngest marine turbidite units fill small basins whoseadvancing coastline. In a general way, the prevalence oftransverse drainage indicates that usually enough sedi- present outlines and palaeocurrent patterns indicate that

they were shaped by the growing folds (Fig. 9). Directlyment was supplied to the advancing coastal plain for theMazzanti–Trevisan mechanism to dominate. The isolated south-west of the Umbria–Marche Ridge is a narrow

syncline, 38 km long and 3 km wide, filled with lower tocases of longitudinal drainage record episodes in whichreduced sediment supply or enhanced anticline emerg- middle Tortonian turbidites and associated deep-marine

sediments of the Monte Vicino Sandstone. The petrogra-ence locally defeated that mechanism.This interpretation is supported by the fact that the phy of the sand grains (Centamore et al., 1977) indicates

that these sediments were not derived from the under-longitudinal drainage tracts on Map 109 ‘Pesaro’ allfollow synclinal belts (Fig. 5). The longitudinal tract lying middle Miocene turbidites of the Marnoso-arenacea

Formation, which were transported south-east from analong the Foglia River lies in a very broad, gentlesyncline, whereas the tract along the Metauro follows the Alpine source (Ricci Lucchi & Pialli, 1973; Ricci Lucchi

& Valmori, 1980); in particular, the abundant Alpine-narrower, tighter syncline between the Cesana and FurloAnticlines. sourced dolomite grains of the Marnoso-arenacea are

absent in the Monte Vicino Sandstone. Palaeocurrentstudies (Centamore et al., 1977) show that the MonteLongitudinal drainage further inland. A few longitudinal

drainage tracts can be seen further inland, on Map 116 Vicino turbidites entered the basin from the south-west,built a deep-sea fan, and flooded the elongated basin‘Gubbio’, making it possible to infer the later evolution

of these tracts as uplift continues. These longitudinal floor toward the north-west and the south-east (Fig. 9).Although the present synclinal remnant is surely narrowertracts include the Metauro near Urbania, which lies in a

broad syncline, as well as the Candigliano north-west of than the original basin, distal palaeocurrents are rigidlyparallel to the synclinal axis (Centamore et al., 1977,Piobbico which occupies a tight syncline. The short


W. Alvarez

Fig. 9. Main part of the Umbria–Marche Ridge (for location see Fig. 5), showing palaeocurrent and sandstone-petrographyevidence for north-eastward sediment transport into minor turbidite basins. These basins probably developed as the north-eastward-advancing thrust front deformed the slightly older, large-scale turbidite units with Alpine (north-west) sources –especially the Marnoso-arenacea Formation – which filled the foredeep before arrival of the thrust front. Contours on the presenttopography show that the lateral turbidite feed is likely to have passed through the area of the present saddle occupied by theBosso and Burano transverse gorges. It is thus possible that the transverse drainage developed along the line of a submarineprecursor. The original studies of these small basins were done by Centamore et al. (1976, 1977).

plate 1), which provides evidence for an originally narrow Whether the saddle had emerged or was still below sealevel at that time is not yet clear.basin moulded into a syncline between growing anticlines.

Across the Umbria–Marche Ridge, to the north-east,another small synclinal basin holds the upper Messinian Drainage response to axial culminations ofsandstones, in places conglomeratic, of the Monte Tur- the foldsrino Formation (Centamore et al., 1976). As in the caseof the slightly older Monte Vicino Sandstone, the Monte The axis of the largest anticline in the Umbria–Marche

Ridge, extending from Monte Nerone to Monte Catria,Turrino sandstone lacks dolomite and represents a tur-bidite fan derived from the south-west. The sandstone has a broad saddle in its middle portion, with bedding

appearing nearly level in a longitudinal section (Fig. 8A).petrography and palaeocurrents strongly suggest that thesediment transport that fed the Monte Turrino Sandstone The relatively thick Jurassic pelagic limestone sequence

in this saddle area was deposited in a deep-marine settingcrossed the major anticline through the present saddle.



(Colacicchi et al., 1970; Centamore et al., 1971; Alvarez, Anticlines transected at culminations1989a). This saddle is the site of the cross-cutting gorges

A remarkable feature of the transected anticlines is thatof the Bosso and Burano Rivers.in several cases the transverse gorges cut through axialNorth-west of the saddle, the anticline rises to theculminations, almost as if the stream were attracted tobroad culmination of Monte Nerone and then plungesthe most difficult obstacle to cross. This is seen in theout and disappears further north-westward. The Monteanticlines at Furlo, Gorgo a Cerbara and Acqualagna,Nerone culmination occurs in an area with a thin Jurassiceach of which plunges out in both directions, not farlimestone sequence deposited on a pelagic fault-blockfrom the anticlinal culmination.high, discussed in the papers just cited, but differences

By contrast, axial culminations crowning long anti-in Jurassic palaeobathymetry were probably flattened outclines are not transected by river canyons; this is seen atby Cretaceous deposition. The axial rise of the MiddleMonte Nerone, Monte Catria and Monte Cucco. AsCretaceous Fucoid Marls from saddle to culminationdiscussed above, the latter culminations probably formedindicates that the culmination is primarily due to Apen-early and would have been avoided by the initial streams.nine structural deformation – perhaps to a sub-

Other culminations look instead to be late modificationssurface thrust. South-east of the saddle, the Majolicaof originally less steeply plunging anticlinal axes. Therises to a second culmination, crowned by Monte AcutoCandigliano River is barely deflected as it approaches the(a Jurassic fault-block high) and Monte Catria (mostlyFurlo Gorge, arguing that this striking culmination wasbasinal Jurassic), and the same reasoning implies thrust-originally less extreme, if present at all. Gorgo a Cerbararamp control at depth. The doubly plunging Monteis even more convincingly a late modification, becauseNerone – Monte Catria anticline, with a central saddle,the Candigliano follows a 7-km longitudinal tract beforemay reflect a pair of lateral ramps beneath the plungingentering Gorgo a Cerbara (Fig. 3). As discussed above,ends of the anticline, dipping toward the central saddlelongitudinal tracts apparently form behind obstacles while(Boyer & Elliott, 1982, fig. 25, section B–B∞).associated transverse tracts cross through lower saddles.If the culminations resulted from in-sequence thrust-This suggests that Gorgo a Cerbara was originally a lowing, they would have been present since the time of thepoint along the fold axis, whereas it is now at theinitial rise of the anticline, and would have been the firstculmination of the Montiego Anticline.parts to emerge from the coastal plain, or perhaps even

Rivers cutting through anticlines have often been takenfrom the sea. In either case they would have been avoidedas prima facie evidence for the presence of cross-cuttingby the incipient rivers. If the culminations were due tofaults, providing lines of weakness, but in the Apennine

late, out-of-sequence thrusting, there is no reason whycase there is commonly no other evidence for such faults.

they could not be transected by the drainage. TheAt Furlo, the Jurassic stratigraphy differs north-west and

fact that the modern gorges occupy the saddle, notsouth-east of the gorge, suggesting the presence of

the culminations, thus provides weak evidence for an original fault. In other gorges (Gorgo a Cerbara,in-sequence thrusting in this anticline. Acqualagna, Bosso, Burano, Fossombrone) there is no

Along the Umbria–Marche Ridge, the axial culmi- indication of fault control of the cross-cutting rivers.nations rise above a saddle which itself is 500 m above These observations suggest the possibility that thethe surrounding terrain. There is a possible analogue cross-cutting river controlled the position of the axialwith the same dimensions, representing an earlier stage culmination, perhaps by removing some of the weight ofof development, 20 km to the north-east: the easternmost the overburden. An analogous situation occurs in Utah,major anticlinal trend has axial culminations at Furlo and where the Colorado River in south-eastern CanyonlandsArcevia (Fig. 3), separated by an axial saddle at Pergola National Park (Huntoon et al., 1982) is followed by thewhich, because of lesser uplift, does not yet rise above axis of a long anticline apparently reflecting flow ofits surroundings or expose pre-Tertiary rocks. The underlying Pennsylvanian evaporites of the Paradox For-Cesano River and two smaller streams cross the saddle mation into the zone of reduced overburden along thein a pattern that should produce analogues to the Bosso river canyon. The transected axial culminations in theand Burano canyons after another few hundred metres Umbria–Marche Apennines may also be due to move-of uplift. By contrast to the culmination of Monte ment in evaporites. However, the underlying TriassicNerone, the Furlo culmination is transected by a deep Burano Formation is primarily composed of anhydrite,gorge. This would argue against control of the culmi- which should be more resistant to flow than the subsur-nation by in-sequence ramp thrusting. face salt of the Paradox Formation.

Monte Conero has about the same dimensions as the Alternatively, the transected culminations may rep-axial culminations just described, and may well represent resent easier uplift through motion on late, out-of-the earliest stage in the morphological expression of these sequence thrusts. This seems probable in the case ofstructural features. As noted above, drainage is diverted Gorgo a Cerbara, which is located directly in front ofaround Monte Conero in longitudinal valleys, and its the Monte Nerone culmination on the Umbria–Marcheappearance in the distant future should be much like that Ridge. The combination of structural form and drainage

suggests that the culmination at Gorgo a Cerbara startedof modern Monte Nerone.


W. Alvarez

out low and was later raised by indentation of the Monte It is not yet clear whether the Umbria–Marche fold-thrust belt formed through an orderly progression ofNerone Anticline. Another possibility is that more sedi-

ment was shed from the incipient Monte Nerone high, thrusts, followed by uniform uplift after passage of thethrust front, or whether out-of-sequence thrusting causedpushing the coastal plain out over an original culmination

in the Cerbara sector of the Montiego Anticline. the fold profiles to change as they were uplifted. Carefulstudy of the geomorphology may contribute to dis-tinguishing between these two possible histories.

This study of Apennine drainage evolution leads to aCONCLUSIONSview of the landscape as a delicately balanced system inwhich tectonic transport, generation of fold and thrustThe fold-thrust belt of the Umbria–Marche Apennines

displays deep gorges cutting through anticlinal ridges, structures, uplift, erosion of a stratigraphic sequence withinternal discontinuities, progradation of the coastal plain,allowing rivers to drain across the structural grain to the

Adriatic Sea. These gorges are well explained by the elongation of the distal tracts of the rivers and changesin sea level all interact in complicated ways. This viewmechanism of Mazzanti & Trevisan (1978), which is a

combination of antecedence and superposition, initiated raises the possibility that geomorphology may be able tohelp resolve difficult questions of the subsurface structureat the shoreline, where folds, rivers and coastal plain are

all forming simultaneously. Rivers are progressively elon- and evolution of fold-thrust belts.gated downstream as the coastal plain builds out, trappedbehind but also burying incipient anticlines, and allowing

ACKNOWLEDGEMENTSthe new river increments to cross the anticlines whilethese folds still have low relief. My indebtedness to many Italian colleagues over many

Mazzanti & Trevisan based their mechanism on strati- years of research in the Apennines will be obvious. Earlygraphic and structural evidence that the Apennine anti- drafts of this paper benefited from careful discussion andclinal ridges have emerged in sequence from south-west review by participants in a UC Berkeley graduate seminar,to north-east. Subsequent advances in understanding the especially John Stock and Daniel Karner, and fromstructure of fold-thrust belts show that this observation conversations with Alessandro Montanari. I thank Robertis a natural consequence of the forward advance of the Anderson, Douglas Burbank and Peter DeCelles for theirbasal detachment thrust, generating a sequence of thoughtful and detailed reviews for the journal. Myanticlines. research in the Apennines over the years has been

The sequence of anticlines across the fold-thrust belt, supported by the National Science Foundation and byfrom younger to older, provides a proxy for the topo- Chevron Overseas Petroleum, Inc. It has benefited fromgraphic history of a single anticline and shows all the the logistical support of the Osservatorio Geologico distages in the evolution of transverse gorges inferred by Coldigioco and the hospitality of the people of PiobbicoMazzanti & Trevisan. In some cases, however, rivers and Gubbio. This study is dedicated to the memory ofhave longitudinal tracts, parallel to the fold axes, between Giampaolo Pialli, one of my first Italian geological friends.transecting gorges. Longitudinal drainage is presently Paolo brought modern structural geology to the Umbria–forming near the coast, behind the high anticline of Marche Apennines in the 1970s and trained many of theMonte Conero. This allows the recognition that within leaders in the present generation. He was the drivingthe fold-thrust belt, longitudinal drainage reflects initial force behind the complex and highly successful CROP-03conditions dominated by rise of a new anticline, whereas crustal-profile project (Pialli et al., 1998). CROP-03transverse drainage reflects original coastal conditions should have been simply Paolo’s next step, but insteaddominated by buildout of the coastal sediments. The will be remembered as his greatest accomplishment.drainage patterns thus preserve a memory of the relativebalance between sedimentation and uplift at the advanc-

REFERENCESing coastline.Small turbidite basins on either side of the largest A, E., B, V., P, P. & S, M. (1970)

Introduction to the geology of the Northern Apennines.anticline, showing evidence of lateral sediment feed,Sediment. Geol., 4, 207–249.suggest that some drainage lines may be descended from

A, W. (1972) Recent Italian research. Geotimes, 17, 14–17.submarine precursors. Early emergent anticlines, rep-A, W. (1989a) Evolution of the Monte Nerone seamountresented in older structures by axial culminations on long

in the Umbria-Marche Apennines: 1. Jurassic-Tertiaryanticlines, are systematically avoided by drainage lines.stratigraphy. Soc. Geol. Ital. Boll., 108, 3–21.However, in anticlines that plunge out over short dis-

A, W. (1989b) Evolution of the Monte Nerone seamounttances in both directions, the culminations are in some in the Umbria-Marche Apennines: 2. Tectonic control of thecases precisely transected by the drainage, suggesting seamount-basin transition. Soc. Geol. Ital. Boll., 108, 23–39.that erosional removal of overburden may have allowed A, W., A, A.J., R, P.R., K, D.B.,the rise of the culmination, either through evaporite T, N. & M, A. (1996) Quaternary fluvial-mobility or through easier uplift on late, out-of- volcanic stratigraphy and geochronology of the Capitoline

Hill in Rome. Geology, 24, 751–754.sequence thrusts.



B, A.W., B, L., C, C. & G, R. (1986) Giurese umbro-marchigiano ed ipotesi per un suo inquadri-mento regionale. Soc. Geol. Ital. Mem., 9, 839–874.Balanced sections and seismic reflection profiles across the

C, M.P., D D, M., M, S., P, W.Central Apennines. Soc. Geol. Ital. Mem., 35, 257–310.& W, F.C. (1999) Frontal part of the northern ApenninesB, A.W., G, P.L. & S, G.A. (1966) Structure,fold and thrust belt in the Romagna-Marche area (Italy):seismic data, and orogenic evolution of southern CanadianShallow and deep structural styles. Tectonics, 18, 559–574.Rocky Mountains. Bull. Can Petroleum Geol, 14, 337–381.

C, S., M, S., P, G., B, A. & R,B, M., G, F., L, G., L, O. B-V. (1989) Stratigrafia del Mesozoico e Cenozoico nell’area, R. (1988) Sezioni geologiche bilanciate attraverso ilUmbro–Marchigiana/Mesozoic–Cenozoic stratigraphy in thesistema a pieghe umbro-marchigiano: 1. la sezione Trevi-Umbria–Marche area [in Italian and English]. Mem. DecrittiveValle dell’Ambro. BSGI, 107, 109–130.della Carta Geologica d’Italia, 39.B, G. (1891) Il territorio di Gubbio. Notizie geologiche.

D, F.A., S, J. & D, D. (1984) Mechanics of fold-Tipografia Economica, Roma, 38.and-thrust belts and accretionary wedges; cohesive CoulombB, V., P, P., S, M. & S, G. (1970)theory. J. Geophys. Res., 89, 10 087–10 101.Development of the Northern Apennines geosyncline – the

D, C.D.A. (1969) Balanced cross sections. Can.miogeosynclinal sequences. Sediment. Geol., 4, 341–444.J. Earth Sci., 6, 743–757.B, S.E. & E, D. (1982) Thrust systems. Am. Ass.

D C, R. & S, U. (1987) Analisi ambientale quantita-Petrol. Geol. Bull., 66, 1196–1230.tiva dei litorali marchigiani fra Gabicce e Ancona. LivelloB, D., M, A. & B, N. (1996) Interactions ofdel rischio naturale e del degrado, distribuzione dei sedimentigrowing folds and coeval depositional systems. Basin Res.,e loro possibile impiego per ripascimento artificiale. Soc.8, 199–223.Geol. Ital. Boll., 106, 377–423.C, D., D Donatis, M., Mazzoli, S. & Wezel, F.C.

D, A.V., P, L. & P, G. (1983) Osservazioni(1995) The Colbordolo-Pesaro area of the Romagna-Marchegeologiche nelle aree comprese fra i massicci perugini ed ifoothill zone (Northern Apennines, Italy). Giornale Geol.,rilievi di Gubbio. Giornale Geol., 45, 127–150.57, 277–295.

D, D., S, J. & D, F.A. (1983) Mechanics of fold-C, A., C, R., P, A. & C,and-thrust belts and accretionary wedges. J. Geophys. Res.,C. (1982) The Jurassic-Lower Pliocene history of the Anzio-88, 1153–1172.Ancona Line (Central Italy). Soc. Geol. Ital. Mem., 24,

D D, M., I, C., L, A., M, S. &325–336.P, M. (1998) CROP 03: Structure of the MontecalvoC, A. & S, K.G. (1989) Exotic clasts in ain Foglia-Adriatic Sea segment. Soc. Geol. Ital. Mem., 52,Pliocene conglomerate near Pesaro have an Alpine source.617–630.Soc. Geol. Ital. Boll., 108, 607–618.

D, F. & B, C. (1986) Aspetti geomorfologici delC, B. (1934) Studi morfologici nell’Italia centrale.territorio marchigiano. In: La Geologia delle Marche (specialBoll. R. Soc. Geogr. It., Ser. 6, 11, 22–30.volume of Studi Geologici Camerti) (Ed. by E. Centamore and

C, C. (1976) Correlazione tra piani carsici ipogei eG. Deiana), pp. 99–103. Universita di Camerino, Camerino.

terrazzi fluviali nella valle del F. Esino (Marche). Soc. Geol. F, A.J. & M, M. (1986) Back thrusting in

Ital. Boll., 95, 313–326. forelimbs of rootless anticlines, with examples from theC, C., C, C. & G, L. (1989) Geo- Umbro-Marchean Apennines (Italy). Soc. Geol. Ital. Mem.,

morphology of north-east Umbria and the Marche region [in 35, 357–370.English and Italian]. Mem. Descrittive Della Carta Geol.

F, A.J., M, N., P, G., M, M.Italia, 39, 37–44. & V, F. (1990) Palaeotectonic significance of gravity

C, C. (1988) Evoluzione del reticolo idrografico in displacement structures in the Miocene turbidite series ofun tratto umbro-marchigiano dello spartiacque principale the M. Pollo syncline (Umbro-Marchean Apennines, Italy).dell’Appenninico. Geografia Fisica e Dinamica Quaternaria, Geol. Mijnb., 69, 69–86.11, 11–24. E, P., G, G., T, M. & T, L. (1975)

C, E., C, M., D, G., M, A. Tensional and compressional areas in the recent (Tortonian& P, U. (1971) Contributo alla conoscenza del to present) evolution of the Northern Apennines. Boll. Geofis.Giurassico dell’Appennino Umbro-Marchigiano. Studi Geo- Teorica Applicata, 17, 3–18.logici Camerti (Camerino), 1, 7–89. G, R. (1958) Spostamento dello spartiacque dell’Ap-

C, E., C, U. & M, A. (1977) Analisi pennino Settentrionale in conseguenza di catture idrografiche.dell’evoluzione tettonico-sedimentaria dei ‘bacini minori’ tor- Atti della Societa Toscana di Scienze Naturali, Memorie, Ser.biditici del Miocene medio-superiore nell’Appennino A, 65, 25–38.Umbro-Marchigiano e Laziale-Abruzzese: (3) Le arenarie G, R. (1962) Evoluzione e spostamento dello sparti-di M. Vicino, un modello di conoide sottomarina acque appenninico tra il Monte Fumaiolo e Gualdo Tadino.affogata (Marche Settentrionale). Studi Geologici Camerti Atti Della Societa Toscana Di Scienze Naturali, Memorie, Ser.(Camerino), 3, 7–56. A, 68, 57–67.

C, E., C, U., R Lucchi, F. & Salvati, G, F., B, M., B, A.W., M, I. & V,L. (1976) La sedimentazione clastica del Miocene medio- L. (1993) Conflicting balanced structural sections across thesuperiore nel bacino marchigiano interno tra il T. Tarugo ed Central Apennines (Italy): problems and implications. In:Arcevia. Studi Geologici Camerti (Camerino), 2, 73–106. Generation, Accumulation and Production of Europe’s hydro-

C, M.B. & R, W.B.F. (1978) Messinian erosional surfaces carbons III, European Association of Petroleum Geologistsin the Mediterranean. Mar. Geol., 27, 193–363. Special Publication, v. 3 (Ed. by A.M. Spencer), pp. 219–231.

Springer-Verlag, Berlin.C, R., P, L. & P, G. (1970) Nuovi dati sul


W. Alvarez

G, E. & P, L. (1949) Considerazioni sullo volume of Studi Geologici Camerti) (Ed. by E. Centamore andG. Deiana), pp. 105–133. Camerino.sviluppo dell’idrografia in relazione alle piu recenti teorie

sull’orogenesi appenninica. Atti della Societa Toscana di N, T., P, E. & R, M.L. (1986) Il bacinopleistocenico marchigiano. In: Atti della riunione-escursioneScienze Naturali, Memorie, Ser. A, 56, 144–147.

G, L. & P, G. (1969) Alcune idee sulla evoluzione del Gruppo di Sedimentologia del CNR sul Pleistocene marchigi-ano (Ed. by T. Nanni), pp. 13–43. Dipartimento di Scienzeoro-idrografica dell’Appennino settentrionale. L’Ateneo Par-

mense, Acta Naturalia, 1, 29–45. dei Materiale e della Terra, Universita degli Studi di Ancona,Ancona (Italy).H, N. (1996) Regular spacing of drainage outlets from

linear mountain belts. Basin Res., 8, 29–44. O, T.M. (1965) The Zagros streams: a new interpret-ation of transverse drainage in an orogenic zone. SyracuseH, C.B. (1956) Cenozoic geology of the Colorado Plateau.

U.S. Geol. Surv. Prof. Pap., 279, 1–99. University Geographical Series, 1.O, T.M. (1985) Origin of drainage transverse toH, P.W., B, G.H. J & B, W.J. (1982)

Geologic map of Canyonlands National Park and vicinity, structure in orogens. In: Tectonic geomorphology (Ed. by M.Morisawa and J.T. Hack), pp. 155–182. Allen & Unwin,Utah (1: 62,500). Canyonlands Natural History Association,

Moab, Utah. Boston.O, G.G. & F, P.F. (1984) Sedimentary basins formedK, D.B. & M, F. (1998) Correlation of fluviodeltaic

aggradational sections with glacial climate history: a revision and carried piggyback on active thrust sheets. Geology,12, 475–478.of the Pleistocene stratigraphy of Rome. Geol Soc. Am. Bull.,

110, 748–758. P, G. & A, W. (1997) Tectonic setting of theMiocene Northern Apennines: the problem of contempor-L, G., B, F., B, M., M, M. &

K, J.V.A. (1994) Seismotectonic zoning of east-central aneous compression and extension. In: Miocene Stratigraphy:an Integrated Approach (Ed. by A. Montanari, G.S. Odin andItaly deduced from an analysis of the Neogene to present

deformations and related stress fields. Geol. Soc. Am. Bull., R. Coccioni), pp. 167–185. Elsevier, Amsterdam.P, G., B, M. & M, G. (1998) Results of the106, 1107–1120.

M, A. & R, W.B.F. (1986) Extension in the CROP03 deep seismic reflection profile. Soc. Geol. Ital.Mem., 52, 657.Tyrrhenian Sea and shortening in the Apennines as a result

of arc migration driven by sinking of the lithosphere. Tecton- P C (ca. 560) History of the Wars, v. 3 (TheGothic War). Harvard University Press (1919, 1968),ics, 5, 227–245.

M, O. (1926) La maggiore discordanza tra orografia Cambridge, MA.R, K.J., G, P. & C, H. (1980) Lithospheric splite idrografia nell’Appennino. Riv. Geograf. Ital., 33,

65–74. in the descending plate: observations from the NorthernApennines. Tectonophysics, 64, T1–T9.M, R. & T, L. (1978) Evoluzione della rete

idrografica nell’Appennino centro-settentrionale. Geograf. R L, F. & P, G. (1973) Apporti secondari nellaMarnoso-arenacea: 1. Torbiditi di conoide e di pianuraFisica e Dinamica Quaternaria, 1, 55–62.

M, D.A. (1992) Geometry and kinematics of an active, sottomarina a est-nord-est di Perugia. Soc. Geol. Ital. Boll.,92, 669–712.laterally propagating wedge thrust, Wheeler Ridge, Cali-

fornia. In: Structural Geology of Fold and Thrust Belts (Ed. R L, F. & V, E. (1980) Basin-wide turbiditesin a Miocene, over-supplied deep-sea plain: a geometricalby S. Mitra and G.W. Fisher), pp. 3–28. Johns Hopkins

University Press, Baltimore, MD. analysis. Sedimentology, 27, 241–270.S, R. (1952) Il bacino del Metauro. Giornale Geol., 24,M, G. (1938) Il Tevere. Monografia idrologica. v. 1, parte

2. Geologia e permeabilita dei terreni del bacino. Servizio 1–268.S, A. (1950) Sull’origine della rete idrografica e dei baciniIdrografico (Roma), Pubblicazioni, 22.

M, G. (1951) Geologia dell’Appennino settentrionale. Soc. intermontani nell’Appennino centro-settentrionale. Riv. Geo-graf. Ital., 57, 249–256.Geol. Ital. Boll., 70, 95–382.

M, C.I. (1948) I cunei composti nell’orogenesi. Soc. S, A.N. (1945) Hypotheses of stream development inthe folded Appalachians of Pennsylvania. Geol. Soc. Am.Geol. Ital. Boll., 67.

M, C.K. (1988) Out-of-sequence thrusts. Tectonics, 7, Bull., 56, 45–88.T, W.D. (1965) Regional Geomorphology of the United539–561.

N, T. & V, P. (1986) Inquadramento idrogeologico States. John Wiley and Sons, New York, 609.ed influenza della tettonica sugli acquiferi di subalveo dellepianure marchigiane. In: La Geologia delle Marche (special Received 31 August 1998; revision accepted 30 July 1999.


(1999) 11, 267–284 drainage on evolving fold-thrust belts...

Documents