18.1 Models of landscape evolution One of the most obvious questions we can ask about4ands€apes is &ow they came to be as they are. Indeed, the historical approach to landform analysis was the dominant perspective until the 1960s. Over the past two decades, however, the other obvious question - what areÍhe processes operating in the landscape today and how do they relate to the landforms we see - has become pre-eminent to the extent thai studies of landscape development through time have been rather neglected. The detailed work on surface processes over the past two decades has significantly increased our understanding of the relationships between process and form at the srrrall scale and over short periods of time. But the gap between our understanding of landform genesis at this scale and our knowledge of how whole landscapes function at long time scales has been widely acknowledged. A much better appreciation of the role of{ectonic and climatic contrglyover landscape development, coupled with the application of new dating techniques and*ajor theoretical advances, is beginning, once again, to bring the problem of long-term landscape development centre stage. This chapter draws extensively on topics introduced earlier in the book and tries to show how new concepts and data are beginning to shed a fiesh light on some long-standing problems in geomorphology. Geomorphology has seen various attempts to systematize the development of landscapes through time by isolating the key factors which determine the way in which landforms evglve. These models of landscape evolution, the most influential of which have been those proposed by W. M. Davis, W. Penck, L. C. King and J. Búdel, have had a profound effect on+ho*inds of problems that geomorphologists have considered and the ways they have attempted to tackle them. 1B Long-term landscape development 18.1.1 The Davisian cycle of erosion: peneplanation The model of landscape evolution usually known as the cycle of erosion was developed by W. M. Davis between 1884 and 1899 and owed much to the evolutionary thinking that had permeated both the natural and social sciences in Britain and North America during the latter half of the nineteenth century. Davis considered that in a similar way to life forms, landforms could be effectively analyzed in terms oftheř evolution. He regardedlandscapes as evolving through a progressive sequence of stages, each exhibiting charac+eristic landforms. In his view these sgg9_9ntial changes in form through time made it possible tb infer the temporal stage of development of a landscape from its form alone. A second key concept implicit in the cycle of erosion model (although not explicitly referred to by Davis) is that of thermodynamics. The development of the principles of thermodynamics had bee1 1mzu.9r achievement of nineteenthcentury science with repercussions just as profound as those of evolution. This aspect of the cycle of erosion model has been highlighted only relatively recently and relates closely to§ystems analysis in geomorphology. The second law of thermodynamics states that in an isolated system (that is, one which cannot give off or receive either mass or energy) entropy can never increase. The concept of entropy has been applied in many different contexts, but a general definition is that it is a measure of the energy in a system which is unable to perform work.-In a system in a state of Jow entropy there are large differences in the distribution of -€í}ergy- and the. flows from areas of high to areas of low ,energy allow work to be peďormed. Conversely, in a high entropy system energy is much more evenly distributed and the flow of energy and the petformance of work is correspondingly reduced, At the theoretical point of maximum 458 Endogenic+xogenicinteractions entropy the distribution of energy is entirely uniform and no work can be done in the system. In strict terms,, landscapes are neither isolated nor even closed systems since they are constantly importing and exporting mass and energy, Nevertheless, it has been argued that as the potential energy created by the uplift of a landsuďace is the major source of energy in the landscape system it is a justifiable simplification to regard landscapes as if they are isolated systems. Given this assumption, the cycle of erosion can be seen as representing a progressive decline in potential energy and increase in entropy as the landscape ,is eroded. IndeJ Davis saw thJstope angles andstream gradients at any particular point in the landscape as reflecting the distribution of potentia|'energy €xpressed as local differences in elevation. The total potential energy of the landscape, and hence its stage of evolution at-any given time, could be expressed in terms of its gp,ap,._ql,evation,above base..level (Fig. l 8. 1(A)). Although Davis acknowledged numerous complications that could affect the cycle of erosion, his detailed description of the anticipated sequence of forms (Fig. 18.2) was based on §evera] important assumptions: that denudation occurs under a'humid tepperate climate (which Davis regarded as"lnormal') on ď únifo.rm lithology, and that the cycle is initiated by the relativelý'brief and rapid uplift of a landsurface of minimal local relief which does not experience significant erosion during the uplift phase. Given these conditions, he described a series of stages in the cycle of erosion categorized by way of analogy to the stages of human life as youth, maturity and old age. Davis argued that there would be a-progressive decline in slope angles (Fig. '7.25(A)) and stream gradients through time which would ultimately result in the production of a landsurface close to base level with very šubdued relief. Such a suďace he termed a peneplain and consequently Davis's model of landscape development characterized by declining surface gradients through time is often referred to as peneplanation. It is important to note, however, that the term peneplain is used by some writers much more broadly to refer to any low relief surface however formed. The altemative FT YOUTH MATUR|TY oLD AGE Altitude divides Z__ oJ : Altitude oí main valley floors oíhighest PE N EPLAlN Altitude oí main valley floors Altitude of highest divides q) : E NDRUMPF Time YoUTH ..-.-".""...-.------------ MATURlTY oLD AGE ? o ! :: Altitude of main valley floors Altitude of highest divides PEDlPLAlN PEDlPLAlN.........-..---.---------- Time Fig, 18.1 Schematic representation of the key elements of the models of landscape evolution proposed by Davis (A), Penck (B) and Křns G). Note that for simplicity base level is assumed to be fixed through time and that the temporal scale is not necessarily cotip'arable between diagiams, In the Davisian scheme the sta7e of old age should be regarded as many times longer than Youth and matirity.(ModifiedfroiJ.B.ThornesandD,Brunsden,(1977)GeomorphologyandTime. Methuen,London,Fig.6.2, p.122.) /r"\ Long-term landscape development .l59 Fig. 18.2 The Davisian cycle of erosion under a humid climate. The assumed startinq-point is a landsurface with little local relief, either a peneplain developed during the previous cycle as shown in (A) or an emerged submarine surface, Uplift leads to rapid incision of the landsurface by rivers. In early youth (B) narrow river valleys separate broad areas of largely uneroded uplands and riyer gradients are irregular withwatetfalls, rapids and lakes formed in response to lithological yariations.These channel gradient irregularities haye been eliminated by the end of middle youth, and by the end of late youth (C) major rivers are graded and lateral erosion enables the development ofnarrowfloodplains in their lower courses.The flat uplands which have been steadily reduced in area during youth as the drainage network has grown are eliminated altogether by the beginning of maturity (D). This is the stage of maximum local relief and the drainage network becomes fully integrated and closely adjusted to structure. Hereafter .local relief begins to decline as the graded river channels,which by this stage have spreadfar up tributary valleys, are lowered progressively less rapidly th(tn intelíluves, Associated with this change is the reduction in average slope angles, as the steep slopes ofyouthwhich are close to the stability angle of the partially weathered debris are tranýormed into lower gradient slopes as the active basal remoyal of debris ceases. Throughout maturity , floodplains become gradually wider and major rivers develop meandering channels. By late maturity ( E) local relief has been significantly reduced and the landscape comprises qentle |alley-side slopes and extensive floodplains. As old age is reached (F ) the entire landscape is graded and floodplains are seyeral times broader than the active meandering belts within them. The mean elevation ofthe landsuface, already close to base level, is loweredfurther only very gradually. Note, howeyer, that in regions remote from the coastline to which rivers are flowing the developing peneplain will remain well above base level since river channels must haye a certain minimum gradient in order to transmit water. Low rates of erosion allow the accumulation of thick weatherinq mantles which, in progressively masking the underlying bedrock, gradually free river channels from structural controls. None the less, particularly resistant lithologies may allow erosional residuals, known as monadnocks, to suryiye into late old age. Finally, renewed uplift will initiate a new cycle of erosion (G). (From A. N. Strahler ( 1969) Phystcal Geography (3rd edn.) Wiley, New York, Fig. 27 .l , p. 166, drawn by E. Raisz.) spelling 'peneplane' which is used ocgasionally is certainly misleading as Davis in no sense enÝisaged the development of a planar surface as the ultimate product of the cycle of eroSion. As we have mentioned, Davis acknowledged the presence of factors that might complicate the stately progression of landscapes illustrated in Figure l8.2. The cycle might be intemrpted by renewed úplift]at any stage which would cause rejuvenation of the landscape through the development of youthful forms which would coexist with older forms and thereby create a polycyc|ic landscape. The simplest assumption was that such uplift would only manifest itself as a fall in base level at the downstream extremity of drainage basins (normally at the coastline), and would lead to the gradual enéiijá'c'hment of steeper river gradients and slopes upstream through the drainage systems of the uplifted landscape (Fig. 18.2(G)). A second complication Davis noted was climate: Davis 460 Endogenic+xogenic interactions effectively represented landscape development under the humid temperate morphoclimatic regime of his home area of the north-east USA as 'normal'. But he accepted th'at the detailed nature of landform evolution under different prevailing climates would not be identical because of variations in the intensity of geomorphic processes under different molphoclimatic regimes. Consequently, he developed 'arid' and 'glacial' versions of the cycle of erosion while later disciples of his evolutionary approach added further variants. A thfud complication was provided by lithology and structure which Davis saw as exerting specific controls on landscape evolution largely through their influence on drainage pattems. He maintained, however, that such controls would become progressively less significant as the cycle of erosion proceeded. h the case of limestone terrains, later workers found it necessary to develop a specific karst cycle of erosion. Nevertheless, in spite of these complications, Davis maintained the value of regarding landscapes primarily in terms of their evolutionary stage in a unidirectional temporal sequence. Although his cyclic scheme never gained wide acceptance on the continent of Europe, it dominated Anglo-American geomorphology for several decades. Since the 1950s, however, both the theoretical utility and the empirical validity of the cycle of erosion have been increasingiy challenged. What, then, are the major criticisms of the model? Although contemporary critics have tended to focus on the rather vague understanding of surface processes evident in Davis's formulation of landform development in general, and slope development in particular, perhaps the most fundamental problem with the cycle of erosion arises from the assumptions conceming the rates and occurence of uplift. Presumably due to the lack of quantitative data when he was writing, Davis was never very specific about actual rates of uplift and denudation. The estimates he did give, such as the 20-200 Ma for the peneplanation of the faultblock mountains of Utah, indicates that he envisaged extensive time scales. Our current knowledge of upliítrates (see Chapter 15) suggests that few areas of the world remain stable for periods of tens of millions of years or more, and therefore it seems that polycyclic landscapes are likely to be the norm rather than the exception. Furthermore, jso_static uplift is an inevitable consequence of denudational unloading as the cycle runs its course. As a result, continuous crustal uplift, aibeit at a declining rate though time, will affect the entire duration of a cycle of erosion and greatly delay the attainment of full peneplanation. As we have seen in Chapters 3 arň 4, inter-plate and intra-plate tectonic mechanisms give rise to quite different temporal and spatial patterns of uplift, and in neither case does the elevation of the landsurface take the form of geologically brief, discrete episodes of rapid surface uplift. Epeirogenic movements characteristic of plate interiors usually involve slow, but prolonged sut{ace uplift, while the high crustal uplift rates characteristic of convergent and oblique-slip plate margins persist for as long as the plate interactions giving rise to them are sustained. Another major criticism of the Davisian model arises from its inability to accommodate the frequent and rapid climatic changes that have characterized the Quaternary. These have been of world-wide extent and, in conjunction with the frequent major changes in base level with which they have been associated through their effect on global sea level, they make it extremely unlikely that landscapes anywhere can be realistically viewed as representing a simple unidirectional sequence of forms. 18.1.2 The Penck model: uplift and denudation related As has already been pointed out, the Davisian model never gained universal support, and geomorphologists on the continent of Europe found its assumptions - especially those conceming the nature of uplift - drastically over-simplified, In spite of these criticisms the only coherent altemative scheme of landform development to emerge prior to the second world war was that of w. penck. penck's ideas have never been popular among English-speaking geomorphologists both because of his rather obscure writing style and terminology, and because the majority of geomorphologists unable to read German had to rely for several decades on misleading representations of his views by Davis and other writers. Penck's ideas on uplift differed significantly from those of Davis (Fig. 18.1(B)). Whereas the latter assumed brief episodes of rapid uplift punctuating prolonged periods of stability, Penck argued that, in orogenic belts at least, active trplift could continue for a considerable time and in such situations Davis's notion of evolutionary stages of landscape development would be of dubious value. On the basis of the evidence from sedimentary sequences flanking the Alps, Penck considered that rates of active uplift initially increased slowly before reaching a maximum ar-rd then declining gradually. In certain circumstances Penck thought that periods of increasing and decreasing rates of uplift might be reflected in slope forms. This link could arise from the effeci changing rates of crustal uplift could have on rates of river incision. High rates of crustal uplift, Penck argued, would raise +iver channels further above base level and thus jncrease their gradients. This would lead to an:acceleration in river dcwncutting until the rate of incision matched the rate of crustal uplift. The converse situation would apply during a decline in the rate of crustal uplift, with rates of+iver incision decreashg .as Cowncutting reduced channď gradients- Penck considered that a uniform rate of river incision would give rise to {§ stFaight slopes whieh would retreat at a constant angle, If the rate of downcutting were to increase, however, a phase of waxing development would ensue and slopes would steepen progressively from the base upwards to produce a convex profile. Conversely, a decrease in river downcutting could create a phase of waning development and slopes could become progressively less steep from the base upwards, creating concave profi les. Penck's model of landscape evolution can thus be summarized as follows. An initial gradual increase in the rate of crustal uplift of a primary surface (Primiirnrmpf) leads to the widespread development of convex slopes. Further acceleration in the rate of uplift results in the formation of a series of benches (Piedmottreppen) around the margins of the primary uplifted suďace. As the rate of uplift begins to decline there is a transition from waxing development, characterized by rapid downcutting, to waning development where the rate of stream incision is reduced and valley widening through the parallel retreat of individual slope elements gradually bécomes dominant (Fig. 7.25(C)). As noted in Chapter 7 (see Section 1.6.2), this form of slope evolution is perhaps best described as slope replacement to distinguish it from the version of whole-slope parallel retreat advocated by King and misattributed by Davis to Penck (Fig. 7.25(8)). The steepest slopQ elements forming free faces retreat most rapidly leaving behind basal series of lower angle debris slope segments. The retreat of free faces eventually leads to the formation at drainage divides of large residual hills, or inselbergs. which are flanked by pediments. The eventual elimination of inselbergs leaves a landscape termed by Penck an endrumpf consisting entirely of slowly retreating, low angle concave slopes. Although Penck's emphasis on the response of drainage systems to changing rates of uplift provides useful pointers as to how we might attempt to integrate tectonics into models of ,long-term landform development, his scheme as a whole is'untenable as it pays insufficient attention to other factors affecting landform development. In particular it fails to acknowledge the importance of changes in river discharge which might arise as a result of climatic change, and it also underplays the role oflithology and the nature of Long.term |andscapedeve|opment 4ó1 weathering, both of which can significantly affect relationships between stream activity and slope form. 18.1.3 The King model: pediplanation L. C. King's model of landscape evolution resembles Davis's in.assuming that uplift is episodic and rapid in comparison with rates of denudation, and that the overall morphology of a landscape at any point in time is diagnostic of its evolutionary stage of development (Fig. l8.1(C)). The essential, and significant, difference in King's scheme lies in the mode of slope development he proposed. King initially developed his model to account for the -landscapes of ,southern Ařica. These are characterized by extensive, gently inclined suďaces dotted with inselbergs and separated by escaípments, and have developed under predominantly arid to tropical wet-dry morphoclimatic regimes. King's notion of slope development appears to owe much to Davis's misrepresentation of Penck's ideas (compare Figure 7.25(8), (D)), Rather than the sequential replacement of parallel retreating slope segments by lower angle elements, King envisaged the parallel retreat of a single free-face slope unit leaving a broad, concave pediment sloping at an angle of 6-Jo or less at its base. Gradually over time, pediments coalesce to form pediplains and this mode of landscape development is therefore called pediplanation. King considered that once pediment suďaces have been formed they persist with little change until the next phase of surface uplift promotes a new cycle of river incision and escarpment retreat which consumes existing pediplains and creates new ones. As in the Davisian model, the dating of such denudational episodes can be described in terms of the timing of the fall in base level initiating each new landscape cycle. None the less, the landsurface itselfis diachronous because in King's model .landscapes essentially develop {hrough backwearing as escarpments experience paraliel .retreat; landsurfaces, therefore, are progressively older away -from escarpments (Fig. 18.3). Consequently, it is possible 19 16|[ gfthe local age ofa landsurface, and even to refer to a terminal age determined by the final removal of a pediplain remnant. Terminal age ofD lnitial age ofA AtLocal age ofA Fig, 18,3 Dffirent criteria for defining the ages of erosion suríaces according to the model of landscape development proposed by L. C. King. The surfaces labelled A-D were initiated during three episodes of base level fall. Each is diachronous and deposits on the surface may be capable of yielding a minimum local age at that point. The final elimination of the last remnant of a particular sutface (D) gives its terminal age. 462 Endogenic