This action might not be possible to undo. Are you sure you want to continue?
Year 2 annual report DACA88-95-R-0020
Prepared by Greg Tucker Nicole Gasparini Stephen Lancaster Rafael Bras, PI
Department of Civil and Environmental Engineering Massachusetts Institute of Technology Cambridge, MA
An Integrated Hillslope and Channel Evolution Model as an Investigation and Prediction Tool
Year 2 Annual Report
Introduction ............................................................................................................................... 1-1 Overview of Model Components ............................................................................................. 2-1 Grid Elements ................................................................................................................. 2-3 Storm Size, Duration, and Frequency ............................................................................. 2-5 Flow Routing and Runoff ............................................................................................... 2-6 Bedrock Weathering (Regolith Production) ................................................................... 2-7 Hillslope Sediment Transport: Continuous Processes .................................................... 2-7 Hillslope Sediment Transport: Mass Movement ............................................................ 2-8 Stream Erosion and Deposition ...................................................................................... 2-8 Grain-Size Sorting ........................................................................................................ 2-10 Lateral Stream Channel Erosion (Meandering) ............................................................ 2-10 Vegetation ..................................................................................................................... 2-10 Modeling Landscape Evolution Using an Adaptive Irregular Simulation Mesh ................ 3-1 Grid Elements and Data Structures ................................................................................. 3-3 Drainage Networks on a Triangulated Irregular Mesh ................................................... 3-7 Numerical Algorithms .................................................................................................... 3-8 Variable Resolution for Modeling River Meandering .................................................. 3-10 A Stochastic Approach to Modeling Drainage Basin Evolution ........................................... 4-1 Model Description .......................................................................................................... 4-2 Sensitivity of Erosion Rate to Rainfall Variability ......................................................... 4-5 Morphologic Consequences of Rainfall Variability: Numerical Example ..................... 4-9 Dynamics of Vegetation and Runoff Erosion ......................................................................... 5-1 Model of Vegetation and Erosion ................................................................................... 5-2 Characteristic Form Profiles ........................................................................................... 5-6 Numerical Examples ..................................................................................................... 5-11 Understanding the Interactions of Multiple Grain Sizes ...................................................... 6-1 Meandering: A Simple Model ................................................................................................. A-1
Understanding the dynamics of landscape evolution is a challenging problem, for two reasons. First, the processes involved are inherently destructive, and therefore the geologic record of landscape development is often fragmentary. Second, the sculpture of terrain involves a fascinating but complex set of interacting nonlinear processes. While challenging, however, the problem is not intractable. Information on landscape history is still preserved in the form of topography itself, and often also in the form of associated sedimentary deposits such as alluvial valley ﬁlls. And despite the complexity of geomorphic processes and their interactions, the resultant landforms often exhibit an underlying similarity associated with different geologic and climatic settings. One of the important challenges in modern geomorphic research is to investigate landscape dynamics by developing simulation models that incorporate the process interactions we observe in nature. This report describes progress to date on the development of one such model. The Channel-Hillslope Integrated Landscape Development model (CHILD) is a descendant of earlier modeling efforts at MIT and at the Pennsylvania State University, and it incorporates many of the important processes not
Channel Hillslope Integrated Landscape Development Model
and ﬂoodplain evolution in non-glaciated drainage basins. This report focuses on the theory and implementation of the model itself. Implementation of the irregular mesh is described in Chapter 3. It simulates the interaction of two general types of process: “ﬂuvial” processes. The CHILD model is designed to simulate the evolution of ﬂuvially-dominated landscapes formed chieﬂy by physical erosion (thus.g. so that the user is free to formulate the problem in a way that is suitable to the problem at hand. and “hillslope” processes. Its primary near-term purpose is to serve as a quantitative tool for investigating the impact of Quaternary-scale climate change on hillslope. 1-2 Channel Hillslope Integrated Landscape Development Model . Use of an irregular simulation mesh allows for the incorporation of lateral stream erosion and makes it possible to represent different parts of the landscape at different spatial resolutions. To this end. and opens up the possibility of coupling the model with three-dimensional tectonic deformation models (e. Some of the important improvements over previous models include: • Adaptive. The computer code is written to maximize ﬂexibility and modularity. soil creep. Those shown in gray are planned components that have not yet been fully implemented as of this writing. for example). and on its application to several theoretical problems that are closely related to the problem of geomorphic responses to environmental change. channel. it does not include glacial erosion or karst development. rather than being dictated by the software itself. It also eliminates some of the gridding artifacts associated with ﬁxed-grid methods. irregular mesh: the model uses an irregular ﬁnite-difference gridding method based on the model of Braun and Sambridge (1997) to represent the landscape surface.. strikeslip or thrust faulting). prototypes of process modules have been developed within the framework of the ﬁxed-grid GOLEM model but not yet adapted to the variablemesh data structure used in CHILD.Introduction included in previous models. the model provides a number of different options for activating or deactivating different processes and/or different formulations of the same process. a category which encompasses erosion or deposition by runoff cascading across the landscape (including slope wash and channel and rill erosion). which includes weathering. In some cases. and other slope transport processes. The various components in the model are diagrammed in Figure 1-1.
FIGURE 1-1. Boxes shown in gray are still under development.y.t) Drainage area A(x.) FLUVIAL EROSION & DEPOSITION LANDSCAPE STATE VARIABLES: Elevation z(x. delete points) Schematic illustration of the state variables and process modules in the CHILD model. add.y.y.t) Regolith thickness C(x.y.t) WEATHERING SELECTIVE GRAIN SIZE TRANSPORT VEGETATION LATERAL EROSION (MEANDERING) MESH UPDATER (move.CLIMATE MODULE RUNOFF AND FLOW ROUTING STORM GENERATOR HILLSLOPE TRANSPORT (soil creep.t) Surface runoff Q(x.y.t) Vegetation cover V(x.t) Percent gravel D%(x. landsliding.y. etc. Channel Hillslope Integrated Landscape Development Model 1-3 .
The ﬁrst part describes the present version of the model and its implementation. The multi-size model is discussed in Chapter 6. • Vegetation: CHILD provides a module to simulate the interaction of vegetation growth and runoff erosion. it is discussed brieﬂy in Chapter 3. The vegetation model is discussion further in Chapter 5. Essentially. The second part of the report consists of a three short chapters that address the role of different processes incorporated in the model. This report is divided into two parts. • Multiple sediment sizes: A model of mixed-size sediment transport has been developed within the framework of the GOLEM model. vegetation removal due to runoff erosion. These chapters are built around the following set of guiding questions: 1-4 Channel Hillslope Integrated Landscape Development Model . The interface between CHILD and the meander model is currently under development. The CHILD code is written in C++ in order to exploit the advantages of an objectoriented programming language. and is currently being adapted for incorporation into CHILD. Meandering is simulated using the model developed by Lancaster and others (in prep) (a copy of which appears in Appendix A). The stochastic rainfall model is described further in Chapter 4. and vegetation regrowth during interstorm periods. each of these chapters focuses on the role of one process.Introduction • Stochastic rainfall forcing: Previous models have always modeled sediment trans- port by using a single “effective” rainfall or runoff rate that represents a geomorphic average. • Stream meandering: this is the ﬁrst landscape evolution model in which the processes of vertical stream erosion and lateral channel migration (meandering) are coupled. and includes both a description of the underlying theory and a sensitivity analysis. Chapter 2 presents an overview of the basic theory and a summary of the different model components. Chapter 3 gives a description of the algorithms and data structures used to implement the irregular mesh. This module simulates the effect of vegetation on surface resistance to overland ﬂow. CHILD relaxes that assumption by providing the option of stochastic rainfall input.
The sensitivity analyses described in these chapters are also important prerequisites to calibrating and applying the model to a speciﬁc ﬁeld site. What the model can do. is provide insights that can be further tested through ﬁeld studies. however. because they provide a basis for understanding the model’s behavior under idealized cases. Channel Hillslope Integrated Landscape Development Model 1-5 .• What is the role of rainfall variability in controlling drainage basin morphology and evolution? (Chapter 4) • What is the nature of the interaction between vegetation growth and runoff erosion in catchment evolution? (Chapter 5) • How do interactions between hillslope and channel processes and between different branches of a drainage network inﬂuence spatial variations in grain size? (Chapter 6) We do not expect the model to provide deﬁnitive answers to these questions.
Introduction 1-6 Channel Hillslope Integrated Landscape Development Model .
Channel Hillslope Integrated Landscape Development Model 2-1 . fractional vegetation cover. In addition. and composition of sediment layers previously deposited at a given point. highlighting the drainage networks that form naturally when converging ﬂow excavates valleys and leads to further ﬂow convergence. Figure 2-1 shows a typical simulation. thickness. as noted below. at the user’s option the model can track several additional state variables. including regolith thickness (depth to bedrock). and surface runoff (Figure 1-1). the module for selective transport of different sediment sizes is currently under development. drainage area. The model tracks four basic state variables that determine the depth of erosion or deposition at each point during a given iteration: elevation. percent gravel in the active surface sediment layer.1 In generic mathematical terms. slope. the model can be expressed by the following equation: 1.CHAPTER 2 Overview of Model Components The CHILD model simulates landscape evolution by tracking the passage of water and sediment across an irregular lattice of points that represents the landscape surface. and the number. At this writing.
C. and represents an area of about 11. is a function Channel Hillslope Integrated Landscape Development Model 2-2 . and terms in brackets  denote optional dependencies. FIGURE 2-1. t is time. Qs. ∂z = U ( [ x. The catchment outline is that of the Forsyth Creek watershed.5 km2. The second term. Q. surface slope in the direction of ﬂow. D. S. F(). [ Q s. (a) Perspective view of landscape. vegetation cover. which may vary in time and space. baselevel lowering).Overview of Model Components (a) (b) Drainage basin simulated using the CHILD model. Kansas. S.) (b) Plan view of drainage network and irregular mesh. U(). and percent gravel in the substrate. Q. The surface runoff rate. t ] ) – F ( Q. y. V. represents runoff erosion. represents tectonic uplift (or equivalently. the mesh has been interpolated onto a regular grid. C. Fort Riley. and is a function of the runoff rate. Visualization is from the SG3D module of GRASS. V . and optionally also of sediment ﬂux. D ] ) – H ( z ) ∂t (EQ 1) where z is surface elevation. regolith thickness. The ﬁrst term. No distinction is made between erosion by overland ﬂow and by fully channelized ﬂow. (For graphical display purposes.
1994. future versions may incorporate landsliding and its dependence on soil pore pressure. 1994. Surface runoff can be modeled using a series of random storm events. however. in which case the storm properties represent a geomorphically “average” event (the assumption made in most previous models [e. Howard. The model provides ﬂexibility in the way that each of the process terms are modeled. or it can be modeled as a constant runoff rate that represents an average effective geomorphic event. Ahnert. The third term. Each node Ni is associated with a surrounding Voronoi Region (or Voronoi Cell). the model iterates through a series of discrete storm events and interstorm periods. Grid Elements The irregular grid used in the model is described in detail in Chapter 3. but a brief overview now will help to clarify some of the concepts that follow. Elevation. 1987.. Willgoose et al. represents sediment transport by hillslope processes such as soil creep. and landsliding. drainage area. S (which can inﬂuence the likelihood of surface saturation and therefore of runoff production).Grid Elements of drainage area. A. 1991. the model uses a ﬁnite-difference rather than a ﬁnite-element approach). and optionally of slope. 1995]). or they may be uniform. Moglen and Bras. These storm events can be modeled stochastically. and other state variables are computed at the points rather than within the triangles (in other words. The model grid consists of a set of points N that are connected to form a mesh of triangles (Figure 23). hillslope transport depends solely on local topography. Tucker and Slingerland. In the present version of the model. with the intensity and duration of each chosen at random from a distribution. H(). Kirkby.g. raindrop impact. To compute the effects of these different processes. The component process models are summarized below. Implementation of the various processes within the model is diagrammed in the form of a ﬂow chart in Figure 2-2.. Further discussion of many of these components is provided in later chapters. 1976. which is the unique set of triangles that connect a given set of points in such a way that a circle passing through the three points in any triangle will contain no other points. The Voronoi Region for a node Ni is the region within Channel Hillslope Integrated Landscape Development Model 2-3 . Points (or nodes) are connected using the Delaunay triangulation.
or if the mesh has been updated. If the meandering option is selected and higher resolution is needed. 7. DONE? yes END Channel Hillslope Integrated Landscape Development Model 2-4 .Overview of Model Components INITIALIZE UPDATE MESH1 GENERATE NEW STORM2 COMPUTE EDGE SLOPES DRAINAGE DIRECTIONS DRAINAGE AREA3 SURFACE FLOW3 FIND OUTLETS FOR CLOSED DEPRESSIONS4 ADD NODES TO MEANDERING CHANNELS5 no STREAM EROSION/ DEPOSITION DURING STORM (INCL. EROSION OF VEG. otherwise. Flow chart showing the sequence of computations in the model. 3. otherwise. Only at selected output intervals. mean storm properties are used. ﬂow entering a closed depression is assumed to evaporate. Only on ﬁrst iteration or if mesh has changed during the previous iteration. If the meandering option is selected. 6. 4. 5. If the option for lakeﬁlling is selected. FIGURE 2-2. COVER) CHANNEL MEANDERING6 INTERSTORM HILLSLOPE TRANSPORT VEGETATION REGROWTH BASELEVEL CHANGE OUTPUT7 Notes: 1. 2. If option for stochastic storms is selected. If at least one ﬂow direction has changed.
The boundaries between Voronoi Cells are lines of equal distance between adjacent grid points. Duration. P. Storm Size. and Frequency Each model iteration represents a storm event. Duration. or they may be held constant throughout a simulation.exp – ----f ( T d ) = ----- Td T d (EQ 2) Storm duration (EQ 3) Channel Hillslope Integrated Landscape Development Model 2-5 . the storm properties are chosen at random from the following distributions: Rainfall (runoff) intensity 1 P . and Frequency Points (nodes) Voronoi Cell Edge FIGURE 2-3. In either case. a duration. These parameters may be chosen at random for each storm using the model of Eagleson (1978). which any arbitrary point Q would be closer to Ni than to any other node on the grid. Td.exp – -f ( P ) = - P P T d 1 . and the same assumption is applied to the resulting hydrographs (see below). Each storm event is associated with a rainfall (or runoff) intensity. Each Voronoi Cell has surface area Av. Schematic illustration of model grid components. and an interstorm period before the next event. Ti.Storm Size. storms are approximated as having constant rainfall intensity throughout their duration. For variable storms.
Surface discharge is computed from drainage area in one of the following ways: (1) Hortonian (inﬁltration-excess) runoff: Runoff production (rainfall minus inﬁltration) is assumed to be uniform across the landscape. or a lake-ﬁlling algorithm is invoked to ﬁnd an outlet for the closed depression (see Chapter 3). following the edge that has the steepest downhill slope. In the ﬁrst method. Flow Routing and Runoff Surface ﬂow collected at a point on the grid is routed downslope toward one of its adjacent neighbor nodes. Channel Hillslope Integrated Landscape Development Model 2-6 . ﬂow can be collected on each triangular element and routed downslope toward the lowest of the triangle’s three vertices. The drainage area. Alternatively. water is either assumed to evaporate at that point. the mean values are applied during each iteration (in which case the duration of each iteration is T s + T i ). the surface discharge at any point is equal to Q = (P – Ic) A . The local contribution from rainfall at a node can be computed in one of two ways. and interstorm period. duration. with no downhill route away from a given node. and T i are mean storm intensity.exp – ---f ( T i ) = ---- Ti T i (EQ 4) where P. for a node is the sum of the area of all Voronoi cells (or triangles. the rainfall associated with a node Ni is equal to the effective runoff rate times the node’s Voronoi area. respectively. If a pit occurs on the grid. Av. Assuming steady-state ﬂow. (EQ 5) where Ic is inﬁltration capacity (Q = 0 if P < Ic). T s. if the second method is used) that contribute ﬂow to that node. If stochastic storm generation is not used. A.Overview of Model Components Interstorm period T i 1 .
and elapsed time since the last storm.Bedrock Weathering (Regolith Production) (2) Saturation-excess runoff: An option is currently under development for computing direct runoff from saturated areas. and stream sediment. C 0 (EQ 6) weathering where t is time. depends on the thickness of the regolith mantle. with saturation dependent on position in the landscape. kw is the rate of bedrock-to-regolith conversion when C=0. Culling. Note that for purposes of computational efﬁciency. expressed in units of depth per unit time. ∂z ∂t ∂ z ∂ z = kd 2 + 2 . and C0 is a parameter that controls the rate of decrease in weathering rate with increasing regolith thickness. The rate of conversion.g. colluvium. Channel Hillslope Integrated Landscape Development Model 2-7 . bedrock and regolith. groundwater ﬂux rate. The term regolith is used to represent any disaggregated sediment material. Bedrock Weathering (Regolith Production) The model allows for two basic types of material.. Conversion of bedrock to regolith by weathering processes is assumed to take place at the bedrock-regolith contact. ∂x ∂y 2 2 (EQ 7) creep Numerical solution of equation (7) on an irregular mesh is discussed in Chapter 3. equation (6) is only computed when the model is run in “bedrock-alluvial” mode (see “Stream Erosion and Deposition” on page 8). 1960). and includes soil.. Hillslope Sediment Transport: Continuous Processes Sediment transport by “continuous” hillslope processes such as soil creep and raindrop impact is modeled using the well-known geomorphic diffusion equation (e. C. according to the exponential function ∂C ∂t C = k w exp – ----.
τcb = critical shear stress for erosion). Stream Erosion and Deposition The model distinguishes between detachment of material from a stream bed and transport of the detached material. mb. θcb is a threshold. kt. and pb are parameters. nb.” a term which includes both sediment weathered from bedrock and eroded from the channel bed) is C s = Wk f ( k t Q f S f – θ c ) . equation (9) can be expressed in terms of excess bed shear stress using suitably-chosen values for kf. Channel Hillslope Integrated Landscape Development Model 2-8 . The transport capacity for detached sediment material (referred to hereafter as “regolith. channel width is computed using the empirical relationship m W = k cw Q cw (the constant kcw is assumed to be absorbed into the transport coefﬁcient kf). and kb. As with equation (8). m n pf (EQ 9) where Cs is transport capacity. and will be incorporated into CHILD in a future update. and nf.Overview of Model Components Hillslope Sediment Transport: Mass Movement A module for landsliding presently exists within the GOLEM model. In the present version of the model. kt. m n pb (EQ 8) where Db is the detachment (erosion) rate. nf. mf. Note that with suitably chosen parameters. and pf are parameters. with D b = k b ( τ – τ cb ) b (τ = bed shear stress. and kf. equation (8) can reprep sent excess shear stress. The maximum detachment rate depends on local slope and discharge according to D b = k b ( k t Q b S b – θ cb ) . mf. W is channel width.
Stream Erosion and Deposition
Two end-member cases and one intermediate case arise from equations (8) and (9), and special subroutines are provided in the model to handle each of these cases separately: Detachment-limited: If the sediment transport capacity is everywhere greater than the sediment ﬂux, the rate of stream erosion is simply equal to the maximum detachment rate,
∂z b = –Dc , ∂t
where zb represents elevation of the channel bed above a datum within the underlying rock column. This formulation has been used in a number of studies to represent bedrock channel erosion (or more generally, detachment-limited erosion) (e.g., Seidl and Dietrich, 1992; Anderson, 1994; Howard et al., 1994; Seidl et al., 1994; Tucker and Slingerland, 1994; Moglen and Bras, 1995; Humphrey and Heller, 1995). It has the practical advantage of being simple and efﬁcient to integrate numerically. Transport-limited: If sufﬁcient sediment is always available for transport and/or the bed material is easily detached, streams can be assumed to be everywhere at their carrying capacity. Under this condition, continuity of mass gives the local rate of erosion or deposition as
∂z b 1 ∂C s = – ---, ρs ∂ x ∂t
where ρs is sediment bulk density and x is a vector oriented in the direction of ﬂow. These two different cases can be invoked in the model by selecting the “detachmentlimited” or “transport-limited” stream erosion option, respectively. Alternatively, the model may be run in “bedrock-alluvial” mode. This represents the most general case. In bedrock-alluvial mode, the rate of erosion is computed as the lesser of (1) the excess sediment carrying capacity or (2) the maximum detachment rate. Regolith is assumed to have an effectively inﬁnite detachment capacity; detachment capacity only becomes a limiting factor when bedrock is exposed in a channel.
Channel Hillslope Integrated Landscape Development Model
Overview of Model Components
A model of transport and sorting of multiple sediment sizes has been developed within the framework of the GOLEM ﬁxed-grid landscape evolution model by N. Gasparini. The sorting model uses a two-phase sand-gravel transport formula developed by P. Wilcock at Johns Hopkins University. The multi-size transport model includes a set of routines for recording the depth and composition of previously deposited sediment layers. The multi-size model is described further in Chapter 6. Adaptation of the multi-size model to the CHILD framework is currently in progress.
Lateral Stream Channel Erosion (Meandering)
An interface is currently under development that will allow the CHILD model to be coupled with a 2D model of river meandering in order to compute later river erosion and ﬂoodplain widening. The meander model, written by S. Lancaster, is based on topographic steering of channel ﬂow. The model parameterizes bank shear stress as a function of the cross-channel transfer of ﬂuid momentum due to topographic steering by point bars. The meander model is described in greater detail by Lancaster et al. (Appendix A).
Vegetation is modeled in terms of a percent vegetation cover on the surface, V. Vegetation increases the threshold shear stress that must be exceeded before runoff erosion can occur, according to
τ c ( V ) = τ cs + V τ cv ,
where τcs is the critical shear stress for an unvegetated surface (primarily a function of grain size) and τcv is the added critical shear stress under 100% vegetation cover.
Channel Hillslope Integrated Landscape Development Model
During storms, vegetation cover is eroded at a rate proportional to excess shear stress,
dV η (erosion) = – k vd V ( τ – τ c ) , dt
where kvd is a constant that controls the rate of vegetation removal. The rate of vegetation regrowth during interstorm periods is modeled as a linear function of the amount of vegetation cover present,
dV (growth) = k vg ( 1 – V ) . dt
The vegetation module is discussed further in Chapter 5.
Channel Hillslope Integrated Landscape Development Model
Overview of Model Components Channel Hillslope Integrated Landscape Development Model 2-12 .
Slingerland et al... Tarboton.g.. and (4) use of a ﬁxed grid makes it difﬁcult or impossible to model processes with a signiﬁcant horizontal component. 1996). Ahnert. 1992.. such as thrust propagation or river meandering. the same representation that is used in digital elevation models (e. Tucker and Slingerland. Sinclair and Ball. Kirkby..g. this limitation may be reduced by using multiple-ﬂow algorithms (e. 1994. Channel Hillslope Integrated Landscape Development Model 3-1 . 1994. Howard. (2) drainage directions are restricted to 45 degree increments (though for watershed-scale applications. Willgoose et al. 1994. 1997)). use of a regular grid introduces anisotropy that can lead to bias in drainage network patterns. Rigon et al. 1991. Most landscape evolution models represent three-dimensional terrain using a regular matrix of points. 1987. Chase. (3) under certain circumstances. Beaumont et al. Costa-Cabral and Burgess. 1992. 1993. 1994.CHAPTER 3 Modeling Landscape Evolution Using an Adaptive Irregular Simulation Mesh Continuing advances in computing technology have made three-dimensional modeling an attractive tool for investigating landscape evolution. which in practice means the highest resolution required by any feature or process of interest.. Although signiﬁcant insights have been gained from models using this type of regular spatial discretization. 1976. the technique suffers from several drawbacks: (1) landform elements must be represented at a uniform spatial resolution.
They demonstrated that use of a triangulated irregular network has the advantages of eliminating anisotropy in drainage directions and allowing for a variable-resolution representation of topography. Tucker and Slingerland. Triangulated irregular networks. 1994).g. Sugihara and Iri. 1994. Techniques for constructing Delaunay triangulations and their corresponding Voronoi (or Thiessen) diagrams are well established in the ﬁeld of computational geometry (e. and sedimentation promise to yield important insights into such issues as the relationship between deformation and the stratigraphic record. Knuth. Recently. Du.. One of the most important horizontal erosion processes is lateral stream erosion..Modeling Landscape Evolution Using an Adaptive Irregular Simulation Mesh The last of these constraints is especially signiﬁcant. which by widening a valley can signicantly alter the depositional geometry within a ﬂoodplain over geologic time. Braun and Sambridge (1997) adapted an existing large-scale landscape model to operate with a triangular irregular spatial discretization that overcomes many of the limitations associated with regular grids. Previous coupled models have either incorporated only the vertical component of deformation (e. Anderson. 1996). and brieﬂy discuss how the technique makes it possible to model three-dimensional valley formation by stream erosion. we describe the irregular mesh technique used in the CHILD model of hillslope and channel evolution. which are often based on the so-called Delaunay triangulation. One alternative to regular grids is the use of triangulated irregular networks (TINs) for representing topographic surfaces. but such models ultimately require the ability to model deformation in three dimensions. Kooi and Beaumont.. 1996) or have represented lateral translation by simply offsetting two ﬁxed grids (e.g.g. here we focus on implementation of the adaptive mesh. Channel Hillslope Integrated Landscape Development Model 3-2 . 1987. Coupled models of deformation. 1985. are commonly used for constructing ﬁnite element meshes and for representing surfaces within geographic information systems. erosion.” most crustal deformation processes involve a signiﬁcant amount of horizontal translation. Similarly. Sloan. 1992. The mechanics of the model are described elsewhere (see Chapters 2 and 4-7). erosional processes often have a signiﬁcant horizontal component that is neglected in current models. 1996.. In this chapter. Sambridge et al. 1995. Guibas and Stolﬁ. Although we conventially speak of “uplift. We describe an efﬁcient data structure for implementing the irregular mesh.
also known as a Thiessen polygon). Isometric view of a simulated 4000 catchment. in other words. The Voronoi Region for a node Ni is the region within which any arbitrary point Q would be closer to Ni than to any other node on the mesh. and are generally fairly straightforward 400 200 0 0 1000 2000 3000 FIGURE 3-1. drainage area. showing irregular mesh. Each node Ni is associated with a Voronoi Region (or Voronoi cell. 5000 Channel Hillslope Integrated Landscape Development Model 3-3 . and other state variables in the model (see Figure 1-1) are computed at the nodes rather than within the triangles. Topography is represented in the model by a set of nodes N that are connected to form a mesh of triangles using the Delaunay triangulation of N. The boundaries between Voronoi Cells are lines of equal distance between adjacent grid points. Elevation. the model uses a ﬁnite-difference rather than a ﬁnite-element approach. The Delaunay triangulation is a unique set of triangles that connect a set of points in such a way that a circle passing through the three points of any triangle will contain no other points.Grid Elements and Data Structures Grid Elements and Data Structures The irregular mesh used by the CHILD model is illustrated in Figure 3-1. A number of algorithms and data structures for representing topography and hydrology on a ﬁxed grid have been developed.
and right-hand Voronoi vertex b. Jenson and Domingue.nextedg = AC AB. (b) Complementary directed edge BA. The DualEdge structure is adapted from the QuadEdge data structure of Guibas and Stolﬁ (1985). Unlike regular matrices. and right-hand Voronoi vertex a.edg = BA BA.vvertex = b B A Voronoi Cell A A. (a) C b a D B (b) C b a D B. (a) Directed edge AB. 1988. and directed edges (Figure 3-2). The CHILD model uses a “DualEdge” data structure that provides an efﬁcient way to satisfy these requirements. and (2) is ﬂexible enough to handle dynamic changes in the mesh itself. showing triangular lattice (black) and corresponding Voronoi diagram (gray). a data structure should represent this variable connectivity in a way that (1) provides rapid access to adjacent mesh elements without demanding excessive storage space. the number of neighbors connected to a given node in a triangulated irregular network may be arbitrarily large.g. triangles. Illustration of the dual-edge data structure. Designing efﬁcient data structures for an irregular triangulated mesh is more complicated.edg = AB AB.Modeling Landscape Evolution Using an Adaptive Irregular Simulation Mesh (e. Channel Hillslope Integrated Landscape Development Model 3-4 .. counterclockwise nextedg AC. and consists of three geometric elements: nodes. The data structure is summarized in Figure 3-3.vvertex = a FIGURE 3-2.nextedg = BD BA. where each node is connected to either four or eight adjacent neighbors. 1991). counterclockwise nextedg BD. Tarboton et al.. Ideally.
each directed edge data object includes a reference to the directed edge that lies immediately counter-clockwise relative to its origin node (Figure 3-2).nextedg END END Channel Hillslope Integrated Landscape Development Model 3-5 . each triangle is associated with a Voronoi vertex. Pseudo-code for the DirectedEdge data structure is shown in Figure 3-3. A Voronoi vertex is deﬁned as the intersection point of three Voronoi cells (Figure 3-2). In addition to the origin and destination nodes. Each directed edge object also includes the coordinates of the Voronoi vertex immediately clockwise. and a pointer to one of its directed edges (along with other information relevant to hydrologic routing and physical parameters. not shown) (Figure 3-3).x := current_edge. The format of the Node and DirectedEdge objects makes it possible to efﬁciently ﬁnd the Voronoi polygon associated with a node.edg FOR i=1. y.thenode. the number of adjacent nodes.vvertex_x voronoi_polygon(i). Directed edges are deﬁned by their origin and destination nodes.vvertex_y current_edge := current_edge. which share the same endpoints but point in opposite directions (Figure 3-2). each directed edge object includes a reference to the Voronoi vertex associated with the triangle on its right-hand side (clockwise).nnbrs) current_edge = thenode. In general.thenode.y := current_edge. This representation makes it possible to rapidly access all of the edges and neighboring nodes connected to a given node. using the following algorithm: FindVoronoiPolygon( Node thenode ) BEGIN XYPoint voronoi_polygon(1.Grid Elements and Data Structures Directed Edges Each edge of a triangle is associated with two directed edges. two directed edges that share the same endpoints are termed complementary edges. Nodes Each node object includes x.nnbrs DO voronoi_polygon(i).. z coordinates.
a list of neighboring nodes can be obtained by accessing current_edge..Modeling Landscape Evolution Using an Adaptive Irregular Simulation Mesh Class Node x y z edg nnbrs Class DirectedEdge org dest nextedg vvertex_x vvertex_y Class Triangle p(3) t(3) e(3) // x-coordinate // y-coordinate // z-coordinate (elevation) // ﬁrst connected edge // number of neighboring nodes // origin node // destination node // next directed edge counterclockwise // x-coordinate of right-hand Voronoi vertex // y-coordinate of right-hand Voronoi vertex // Vertex nodes // Adjacent triangles (t(1) is opposite p(1). the three neighboring triangles. and Triangle objects.) // Clockwise-oriented directed edges FIGURE 3-3.dest for each edge originating at a given node. and the three directed edges that are oriented clockwise with respect to the triangle. etc. their right-hand Voronoi vertex Channel Hillslope Integrated Landscape Development Model 3-6 . DirectedEdge. Pseudo-code summary of the DualEdge data structure. Because the three edges are clockwise-oriented. Similarly. The nodes and neighboring triangles are numbered in such a way that the nth neighboring triangle lies opposite the nth vertex (see Figure 3-4). Each triangle is associated with a Voronoi vertex. This vertex is the intersection of the three Voronoi cells associated with each of the triangle’s vertices (e. Figure 32).g. Triangles Triangle objects include pointers to the three nodes in the triangle. showing the data members belonging to Node.
adjacent triangles. the area of each triangle is assigned to the lowest of the triangle’s three nodes and to all nodes downstream of that point. The simplest method assumes that all water entering a “sink” evaporates at that point. In some cases a node may form a local depression. Drainage Networks on a Triangulated Irregular Mesh The surface discharge at each node is computed as a function of the upstream contributing area (see Chapters 2 and 4). In the ﬁrst method. Alternatively. In the second method. In general. and clockwise edges in Triangle data objects. In both methods. Illustration of numbering of triangle nodes. The model provides two alternative methods of ﬁnding the upstream contributing area at a node. The Lake Fill algorithm starts by creating a stack of contiguous ﬂooded nodes. an outlet can be found for each sink using the “Lake Fill” algorithm. these cases must be detected and handled differently). which initially Channel Hillslope Integrated Landscape Development Model 3-7 . This case can be handled in one of two ways in the model. is the Voronoi vertex associated with the triangle itself. the contributing area at a node i is equal to the sum of the Voronoi areas of all nodes that ﬂow to i (including i itself). each node is assigned a drainage direction along the steepest path (edge) toward one of its neighboring nodes. with no neighbors lower than itself. Voronoi vertices can be found by locating the intersection of the perpendicular bisectors of the triangle’s edges (note that there exist special cases in which this is not true.Drainage Networks on a Triangulated Irregular Mesh P0 T2 P1 e1 e2 T0 e0 T1 P2 FIGURE 3-4.
However. The perimeter of the ﬂooded region (“lake”) is then iteratively searched to identify the lowest node along the perimeter. and Avi is the Voronoi area of node i (the depositional area associated with each node is always its Voronoi area. it is ﬂagged as the outlet point for all nodes on the stack. If this node can drain downhill to a location other than the lake itself. in the more typical case of a few isolated sinks.Modeling Landscape Evolution Using an Adaptive Irregular Simulation Mesh contains just the sink itself. and is useful for modeling a rising baselevel or the damming of water and sediment behind an uplifting block. it is added to the stack. Numerical Algorithms Fluvial erosion and deposition at each node in the mesh is computed numerically from Q s j – Q si j = 1 dz i = ----------------------------------dt Av i ∑ n (EQ 1) where zi is the elevation at node i. regardless of the method used to compute contributing drainage area). The sediment outﬂux Qsi depends on discharge. n is the number of nodes that ﬂow directly to i. Qs is sediment ﬂux. If a node is encountered that is part of a pre-existing lake (one initiated at a different sink). An example of a lake computed using the Lake Fill algorithm is shown in Figure 3-5. whereas the cascade algorithm requires a number of iterations equal to the maximum number of segments along any continuous stream regardless of the number or depth of sinks. For a mesh with numerous sinks. it is also added to the stack. as discussed in Chapter 2. (Note that this equation only applies to the case in which sediment ﬂux is Channel Hillslope Integrated Landscape Development Model 3-8 . the lake-ﬁlling algorithm is probably slower than the “cascade” algorithm of Braun and Sambridge (1997). slope. and (possibly) sediment inﬂux ΣQsj. The algorithm is robust enough to handle any arbitrary initial condition. If not. the lake-ﬁlling algorithm is likely to be faster than the cascade algorithm. The number of iterations needed by the lake-ﬁll algorithm depends on the number of ﬂooded points. t is time.
tracked. The downslope sediment ﬂux per unit slope width is given by qs = k d ∂z . ∂x (EQ 2) where the vector symbol denotes orientation in the downslope direction.. Erosion or deposition at a node due to diffusion is then approximated numerically by Channel Hillslope Integrated Landscape Development Model 3-9 . The system of ordinary differential equations described by equation (1) is solved using a predicted-corrector method (e. 1970). the erosion rate dz/dt is assumed to be independent of Qs. The slope width between two adjacent nodes is approximated by the length of their shared Voronoi cell edge.g. and is computed directly as a function of discharge and slope). Λij. Example of lake formation in the model. The lake has formed in response to a “digital dam” that was created by artiﬁcially raising the elevation of nodes near the catchment outlet. in the “detachment-limited” end-member case. equation (7)) is solved in the following way. Acton.Numerical Algorithms FIGURE 3-5. The green asterisks denote lake nodes. The hillslope diffusion equation (Chapter 2. The lake outlet is indicated by the thick blue line at the right-hand edge.
and (3) move.f.y) points. Use of an irregular. An interface between CHILD and the meander model is currently being tested and debugged. deﬁned as positive downwards. The interface includes routines to (1) identify meandering channels on the basis of a threshold drainage area. and/or delete nodes in response to channel migration. the equation can be solved efﬁciently by ﬁrst computing the mass exchange along each physical edge (or equivalently. Channel Hillslope Integrated Landscape Development Model 3-10 . Coupling the two models will make it possible to investigate the dynamics of valley and ﬂoodplain widening in an eroding landscape. add. Note that because diffusive mass exchange takes place along mesh edges. (2) collect the nodes within a meandering channel reach and arrange them in data structures that can be passed to the FORTRAN program meander. then updating the node elevations accordingly. This system of equations is solved using a simple forward-difference method. (Appendix A) describe a physically-based model of stream meandering that represents a meandering stream as a series of (x. across each Voronoi polygon face).Modeling Landscape Evolution Using an Adaptive Irregular Simulation Mesh Λ ij S ij dz i j=1 = – k d ---------------------Av i dt ∑ n (EQ 3) where Sij is the slope from node i to node j. Variable Resolution for Modeling River Meandering Lancaster et al. deformable mesh allows the meander model to be incorporated into the threedimensional landscape evolution model.
Andrews.g. Scant attention has been paid. landscape evolution is in fact driven by discrete events. with “effectiveness” deﬁned either on the basis of denudation rate (e. 1982. Baker. Wolman and Miller. We do so by developing a stochastic theory for erosion and sedimentation in a drainage basin. 1988) or landform genesis (e. earthquakes. 1977. Harvey. 1977. The topography of a typical mountain range. Channel Hillslope Integrated Landscape Development Model 4-1 . For example. Much of the previous research on this problem has focussed on deﬁning the recurrence interval of the most effective geomorphic event. is shaped by a quasi-random sequence of ﬂoods.g. to the question of how variability in geomorphic forces impacts the morphology and rate of evolution of landforms. we address the problem of rainfall variability and its impact on catchment geomorphology. The importance of the frequency spectra of geomorphic events is a fundamental problem in geomorphic research. 1980. 1960.CHAPTER 4 A Stochastic Approach to Modeling Drainage Basin Evolution Although it is often modeled as a continuum process.. and with each frequency distribution perhaps varying in time and space as well. for example. In this chapter. and landslides. with each process having its own characteristic frequency distribution.. Ashmore and Day. Webb and Walling. the relative geomorphic signiﬁcance of climate variability as opposed to mean climate has been widely debated and is not well understood. Wolman and Gerson. however. 1978).
The probability density functions for storm intensity. and interstorm interval are given by Rainfall (runoff) intensity 1 P . Model Description Stochastic Rainfall Model Rainfall is modeled as a series of discrete random storm events. using the model of Eagleson (1978). Combining the stochastic rainfall model with the landscape evolution model enables us to simulate the long-term geomorphic impact of natural variability in storm size. duration.exp – ----f ( T d ) = ----- Td T d Ti 1 .A Stochastic Approach to Modeling Drainage Basin Evolution and exploring the consequences of that theory in the framework of the CHILD model. and frequency. The stochastic model is based on the rainfall model of Eagleson (1978) and describes the probability distribution of storm depth. What are the morphologic consequences of rainfall variability? 1.exp – -f ( P ) = - P P T d 1 . We focus in particular on two questions: What is the predicted sensitivity of long-term average sediment transport rates to the degree of variability in rainfall intensity? To what degree does the importance of variability depend on the presence of thresholds in the landscape. duration.exp – ---f ( T i ) = ---- Ti Ti (EQ 1) Storm duration (EQ 2) Interstorm period (EQ 3) Channel Hillslope Integrated Landscape Development Model 4-2 . such as a threshold for sediment entrainment? 2. Each storm event is treated as having a constant rainfall rate R that lasts for a duration Td and is separated from the next event by an interstorm interval Ti.
1978]). evaporation. Runoff rate. R > 0.g. and/or canopy interception: R = P – I. τc is critical shear stress for sediment entrainment. the formulation could be modiﬁed to account for saturation-excess runoff production [Dunne. dR (EQ 5) 1 ( R + I ) -.. (However. is deﬁned as precipitation minus losses to inﬁltration.Model Description In order to derive a distribution of runoff rates f(R). 1996) and higher for suspended load Channel Hillslope Integrated Landscape Development Model 4-3 . (EQ 6) The mean runoff rate is therefore I R = P exp – -. I = Ic + Il .5-3 for bedload (e.. W is channel width. (EQ 4) where Ic = inﬁltration capacity and Il = evaporation and canopy interception losses. Qs = k f W ( τ – τc ) . τ is average bed shear stress.exp – ---------------. and p is an exponent typically on the order of 1. P (EQ 7) Sediment Transport by Runoff The instantaneous rate of sediment transport by runoff is modeled as a function of excess shear stress. Yang.. The derived distribution for R is obtained from f (R) = f (P) dP . P f (R) = P 0 otherwise. R. we start by assuming runoff is Hortonian and uniform across the landscape. p (EQ 8) where Qs is the sediment transport rate integrated across the width of a channel or rill. kf is a transport coefﬁcient.
1997). Q α n . τ = k t ---W (EQ 9) with kt. ω –α m n p (EQ 10) where m = α(1-ω). uniform ﬂow and adopting an empirical bed friction relationship (such as the Manning-Strickler or Darcy-Weisbach relation). as we assume in the analysis below. this becomes Q = RA .. These effects can in principle be accounted for by relating hydrograph duration and peak attenuation to basin geometric parameters such as total stream length. Substituting the empirical width-discharge relationship W = k w Q ω . shear stress can be expressed as a power function of discharge.A Stochastic Approach to Modeling Drainage Basin Evolution (Whipple et al. α. A (EQ 11) where R is runoff per unit area and A is drainage area. and slope. we use the simple steady-state relationship Q = ∫ R ( a ) da . (Derivations are given by Willgoose et al. Channel Hillslope Integrated Landscape Development Model 4-4 . and n as parameters. S. 1994. 1991. though for the sake of simplicity we do not do so here. Q. If R is uniform across the basin. To write Q in terms of runoff.S ... Howard et al. Assuming steady. (EQ 12) This is clearly a simpliﬁcation. Tucker and Slingerland. Qs = k f k w Q ( k t k w Q S – τc ) . because it neglects hydrograph attenuation by equating hydrograph duration with storm duration. in review).
We start by considering only sediment transported during storm events. 〈 P〉 . is equal to mean storm rainfall times storm frequency times mean storm duration. It remains now to determine the mean storm sediment transport rate. and Qs is the mean transport rate produced by a storm. (EQ 14) Combining equations (13) and (14). Note that 〈 Q s〉 can be related to mean annual rainfall. or 〈 Q s〉 = N T d Q s . 〈 P〉 = PN T d . We make the following assumptions: 1. 〈 P〉 . Td is mean storm duration. (EQ 13) where 〈 Q s〉 is the long-term mean transport rate (mean annual. which for most rivers constitutes the bulk of sediment carried. N is the average number of storms per year.Sensitivity of Erosion Rate to Rainfall Variability Sensitivity of Erosion Rate to Rainfall Variability Fluvial sediment transport can be viewed as a random process in time which in turn is a function of another random process. Q s . We can combine the models derived above for rainfall and sediment transport to analyze the sensitivity of long-term mean sediment transport rate to the degree of temporal variability in rainfall. runoff. Under this condition. the mean annual sediment ﬂux is equal to the mean transport rate produced by a storm times the mean storm duration times the number of storms per year. equation (15) implies that the long-term mean sediment transport rate should be linearly related to mean annual rainfall.Qs . Each storm can be approximated as having a constant rainfall rate throughout its duration. Mean annual rainfall. Channel Hillslope Integrated Landscape Development Model 4-5 . 〈 Q s〉 = -------P (EQ 15) All else being equal. if Td is in years).
(EQ 16) Substituting the sediment transport formula (equations (10) and (12)) and the runoff distribution (equation (6)). γ np (EQ 20) Assuming that γ ranges from ~1-2. Combining with equation (15) gives the expected mean annual transport rate: 〈 Q s 〉 = 〈 P〉 P γ –1 ktk f kw (1 – α) A S Γ(γ + 1) . γ np γ (EQ 19) where Γ() is the gamma function. The integral term can be solved by the substitution u = R ⁄ P . 2.R dR . steady discharge equal to the runoff rate times contributing area. I=0. In that case. P (EQ 18) where γ = mp+ω. this equation predicts that long-term average sediment transport rates should generally be more sensitive to mean annual Channel Hillslope Integrated Landscape Development Model 4-6 .R ( k t k w A R S – τ c ) dR . The steady discharge assumption allows us to write the average storm sediment transport rate as Qs = ∫ ∞ 0 f ( R ) Q s ( R ) dR . which gives Qs = k t k f k w (1 – α) A S P γ np γ ∞ ∫ 0 exp ( – u ) u du = k t k f k w γ (1 – α) A S P Γ(γ + 1) . P (EQ 17) We do not know of an analytical solution to this equation. ktk f kw A S Q s = --------------------------------------P (1 – α) γ np ∫ ∞ 0 R γ exp – -. k f kw A Q s = -----------------P ω ∞ ∫ 0 p ( R + I ) ω –α m m n exp – ---------------. Runoff rate per unit area has the exponential distribution given in equation (6). as is typical. but an analytical solution can be found for the special case τc=0.A Stochastic Approach to Modeling Drainage Basin Evolution At each point in the landscape. runoff (if nonzero) produces a constant. 3.
as long as the threshold and inﬁltration terms are negligible. and either P’ or R’ (in addition to the exponent terms). = R' τ' = ---Q p W p Q p Q p τp τc τ c' = ---τp I I ' = -------〈 P〉 W ( τ – τc ) p ϖ m Q s' = ------------------------. Wp refers to W(<P>). Nondimensional rainfall intensity Runoff rate Channel width P' = P ⁄ 〈 P〉 R' = R ⁄ 〈 P〉 Q ϖ W ' = K w -----. τc’. Q p Q α W –α Q α Q –αϖ τ m .exp – ------------. and similarly for Qp and τp. Nondimensionalization To facilitate numerical analysis of equation (19).Sensitivity of Erosion Rate to Rainfall Variability rainfall than to mean rainfall intensity. -----.= -----.= R' ( R' – τ c' ) W pτ p 1 I' 〈 Q s〉 ' = ---. the equation is nondimensionalized as follows.Q' I ' + R' s R' p (EQ 21) (EQ 22) (EQ 23) Shear stress (EQ 24) Critical shear stress (EQ 25) Inﬁltration rate (EQ 26) Storm transport rate (EQ 27) Mean annual transport rate (EQ 28) Here. and shear stress and critical shear stress are normalized by the shear stress that would be produced by the mean annual rainfall (at a given slope and contributing area). Rainfall and runoff intensity are normalized by mean annual rainfall. Parameter m=α(1-ω). Channel Hillslope Integrated Landscape Development Model 4-7 . ------= -----. The nondimensionalization reduces the equation to three dimensionless parameters: I’.
Figure 4-1a illustrates how the inﬁltration parameter I’ inﬂuences the relationship between mean sediment discharge and rainfall variability. The results are plotted in Figure 4-1.000 realizations of Q s' . mean sediment discharge increases as the square root of rainfall intensity.5 1. inﬁltration capacity.5. Plot of normalized mean sediment ﬂux 〈 Q s〉 ' versus the rainfall variability parameter P' .A Stochastic Approach to Modeling Drainage Basin Evolution Sensitivity to Runoff Variability: Numerical Solutions In order to analyze the relationship between mean transport rate and the three parameters describing rainfall intensity.5 (a) Normalized <Qs>prime 3 (b) Tc‘ = 0 Tc‘ = 1 Tc‘ = 2 Tc‘ = 4 2. so that for low I’ and τc’. equation (20) predicts a linear relationship between mean sediment discharge and mean rainfall. and τ c' . m = 1 ⁄ 3. each curve is normalized by the mean value of 〈 Q s〉 ' . Channel Hillslope Integrated Landscape Development Model 4-8 .5 Normalized <Qs>prime 2 2 1. however. showing how the relationship changes as a function of (a) inﬁltration rate I ' and (b) critical shear stress τ c' . The random Qs’ values were then averaged and multiplied by 1 ⁄ P to obtain 〈 Q s〉 ' (see equation (28)). By comparison.000 random values of P’ were chosen and the corresponding Qs’ was computed for each. sediment discharge is more sensitive to the total amount of rainfall than to rainfall variability. 10. As I’ increases.5 0. and p = 3 . I '.5 0 0 0 0 10 20 30 40 50 Pprime 60 70 80 90 100 FIGURE 4-1. For each set of parameters P'. sensitivity to rainfall variability rises as increasingly larger rainfall events are required to produce signiﬁcant runoff. To facilitate comparison between the curves. and critical shear stress. Exponent parameters are ϖ = 0. Each plotted point represents an average of 10. as predicted by equation (20) for γ=ω+mp=3/2.5 1 1 I=0 I = 10 I = 100 10 20 30 40 50 Pprime 60 70 80 90 100 0. With I’=0 and τc’=0. the nondimensional form of equation (17) is solved using a Monte Carlo approach.5 Monte carlo solution to Qs=f(P) 2. Monte carlo solution to Qs=f(P) 3 3.
Given a typical mean annual rainfall of ~1 meter. Morphologic Consequences of Rainfall Variability: Numerical Example The consequences of rainfall variability for catchment evolution are next explored through simulations with the CHILD model. Figure 4-2 shows a synthetic catchment formed by a combination of steady tectonic uplift relative to the outlet at a rate of 2 x 10-5 m/yr and erosion by a sequence of random storm events. there exists the possibility for high sensitivity to climate variability in the case of boulder-bed or bedrock channels. this analysis does not account for direct runoff from saturated areas. these values correspond to I’ ~ 20 . Clearly. and inﬁltration capacity Ic = 10 m/yr.4 mm/day). Channel Hillslope Integrated Landscape Development Model 4-9 . In natural catchments. The topography consists of a main valley axis ﬂanked by a series of hollows (ﬁrst-order valleys). and is therefore also a potentially important variable. The analysis also does not account for antecedent soil moisture. In this simulation.6 cm/hr for midwestern agricultural silt-loam soils. An additional complication is the ﬁnding by Parker (1978) that channel geometry in streams with mobile bed and banks tends to adjust such that the ratio τ/ τc ~ Kch = constant. implying that vegetated soils may have the effect of amplifying the importance of rainfall variability. underscoring the point made by Baker (1977) that extreme events become increasingly important in bedrock channels and/ or channels bearing very coarse bedload material.2 . which would tend to increase the importance of smaller storms. The relationship between mean sediment discharge and rainfall variability is quite sensitive to the shear stress threshold τc’. Because τc’ depends on the meanﬂow shear stress of the stream in question. the distribution of interstorm periods inﬂuences the likelihood of precipitation falling on already-saturated soils. mean annual rainfall is 1 meter. mean rainfall intensity is 10 m/yr (~27. it is difﬁcult to judge what typical values might be.Morphologic Consequences of Rainfall Variability: Numerical Example Which of these curves is most appropriate to natural catchments? A rough estimate can be obtained from published ﬁgures.700. Dunne (1978) reports measurements of inﬁltration capacity on the order of 0. On the other hand. and on the order of 8 cm/hr for vegetated forest soils. however.
with mean annual rainfall held constant. The increased efﬁciency of stream erosion (predicted by equation (20)) leads to accelerated erosion of the channel network (Figure 4-3b). the slope-area relationship is close to the equilibrium value predicted by equation (20) (the small difference is due to the fact that the simulation has not yet reached complete equilibrium between uplift and erosion) (Figure 4-5a). Initially. When the mean rainfall intensity later reverts to its original value (Figure 4-4). which extend headward to produce a signiﬁcant increase in drainage density (Figure 4-3a). Simulated drainage 4000of basin formed by a combination steady tectonic uplift and storms of variable intensity and duration.A Stochastic Approach to Modeling Drainage Basin Evolution 400 200 0 0 1000 2000 3000 FIGURE 4-2. The morphologic changes illustrated in Figures 4-2–4-4 are apparent on a plot of local slope versus contributing area (Figure 4-5). 5000 2000 6000 Figure 4-3 illustrates what happens when the catchment in Figure 4-2 is subjected to a tenfold increase in mean rainfall intensity.000 model years. Erosion is greatest within the ﬁrst-order valleys. for a total duration of 100. The increase in rainfall intensity (Figure 4-5b) Channel Hillslope Integrated Landscape Development Model 4-10 . the pattern reverses as the valley network begins to ﬁll in with sediment.
(Spike-like features in (b) are distorted Voronoi cells created by boundary 4000 effects in the triangulation). Effect of an increase in rainfall intensity on the synthetic 2000 catchment shown in Figure 4-2. (b) Pattern of erosion: red = erosion. 5000 2000 6000 7000 0 1000 Channel Hillslope Integrated Landscape Development Model 4-11 . green = little or no change. blue 3000 (not shown) = deposition or uplift.Morphologic Consequences of Rainfall Variability: Numerical Example 400 200 0 0 1000 FIGURE 4-3. (a) Topography.
This numerical example illustrates the predicted relationship between rainfall intensity. higher drainage density. Initial condition is the topography in Figure 4-3. Catchment is shown 10. Channel Hillslope Integrated Landscape Development Model 4-12 . 0 leads to a decrease in gradient along the main valley network as the catchment adjusts toward a new equilibrium (dashed line). 1997) and vegetation cover (Chapter 5). higher rainfall intensity should be correlated with lower overall relief. Blue = net deposition.000 model years after decrease in mean rainfall intensity. gradients along the ﬁrst-order valleys and hillslopes increase in response to headward valley erosion and a consequent steepening of side-slopes. One implication of the model is that all else being equal. catchment relief. and/or higher sediment ﬂux. At the same time. Erosion and deposition pattern following a decrease in mean rainfall intensity (to its original value in the simulation show in Figure 4-2). though the nature and pattern of response may vary depending on the magnitude of erosion thresholds (Tucker and Slingerland. and drainage density. and green = little or no change. red = net erosion. The model also implies a pattern valley erosion and valley inﬁlling. in response to variations in rainfall intensity. however. respectively.A Stochastic Approach to Modeling Drainage Basin Evolution 000 000 000 000 FIGURE 4-4. These predictions thus provide a basis for comparison with morphologic and sediment ﬂux data.
Channel Hillslope Integrated Landscape Development Model 4-13 . a variable that also can clearly have important geomorphic consequences. Lines show the equilibrium slope-area relationship calculated by solving equation (20) for slope.Morphologic Consequences of Rainfall Variability: Numerical Example 10 0 (a) 10 Numerical simulation Theory 0 (b) Numerical simulation Previous equilibrium New equilibrium Slope 10 −1 Slope 3 4 5 6 7 8 10 −1 10 −2 10 −2 10 10 10 10 Drainage area (m2) 10 10 10 3 10 4 10 10 Drainage area (m2) 5 6 10 7 10 8 FIGURE 4-5. In the next chapter. Plot of local slope versus contributing area for the numerical simulations shown in (a) Figure 4-2 and (b) Figure 4-3. under the condition <Qs> = UA (uplift rate times drainage area). we develop a model for the interaction of vegetation and erosion and use the model to explore some of those consequences. Climatic factors such as mean rainfall and rainfall variability are often closely correlated with vegetation cover.
A Stochastic Approach to Modeling Drainage Basin Evolution Channel Hillslope Integrated Landscape Development Model 4-14 .
the data of Melton (1958) show an inverse correlation between drainage density and humidity. Vegetation inﬂuences physical erosion both directly.. by increasing surface resistance to wash erosion. vegetation is potentially important in governing landscape responses to environmental change (e. vegetation is widely believed to contribute to the observed relationship between climate and sediment yield. and indirectly.g. 1967. yet the biosphere. 1994). Horton. 1996).g. 1989. 1973. Wilson. 1958. Montgomery and Dietrich. in press). By imposing a signiﬁcant threshold for runoff erosion. Douglas. A number of data sets show a decrease in sediment yield with increasing mean annual rainfall in semiarid to humid climates. clearly plays an important role in landscape evolution. and evapotranspiration.g. a ﬁnding that has been attributed to the stabilizing effects of vegetation (e. 1997. Moglen et al. Channel Hillslope Integrated Landscape Development Model 5-1 .. Similarly. 1993. Kirkby. in press). by inﬂuencing inﬁltration. Moglen et al. 1945. For example.CHAPTER 5 Dynamics of Vegetation and Runoff Erosion Most models of landscape evolution emphasize physical rather than biological processes.. and vegetation in particular. Dietrich et al... Howard. Finally. vegetation cover may effectively impose an upper limit to channel network extent (e. runoff. Tucker and Slingerland.. despite the presumed increase in erosive energy associated with more humid climates (Langbein and Schumm.
Model of Vegetation and Erosion Conceptual model In the context of drainage network development. intermittent vegetation (clumps of grass. and higherorder channels with year-round ﬂow are free of vegetation except along the banks and bars. Below we attempt to quantify these three interacting processes. In the ﬁrst part. In deriving a model to describe vegetation-erosion dynamics. and often also on sediment yield. and use that theory to explore the nature of that interaction on “geomorphic” time scales (by which we mean time scales relevant to morphologic development as opposed to. we present the theory and explore some simple outcomes for slope proﬁle development. there is a dynamic competition between vegetation growth and the disruption of vegetation by runoff erosion. In humid or semi-humid landscapes. while low-order ephemeral channels contain sparse. bushes. Within well-established channels. Channel Hillslope Integrated Landscape Development Model 5-2 . In this chapter.). In the second part. the time scale for signiﬁcant erosion of an agricultural ﬁeld). runoff erosion clearly “wins. hillslopes are typically vegetation-covered. we develop a simple theory to describe the dynamic interaction between vegetation and erosion.” while near drainage divides it is the vegetation that “wins.” The interesting part is what happens in between these two extremes. we present several different numerical examples using the CHILD model. say. we focus on the physical interaction of vegetation and runoff erosion within rills and channels. The most important aspects of that interaction are (1) the effect of vegetation cover on soil/sediment erodibility. (2) the disruption of vegetation due to erosive overland ﬂow. etc. it may be the determining factor). The outcome of the competition between erosion and vegetation growth in the vicinity of ﬁrst order channels may have a signiﬁcant effect on drainage density (in some cases. and (3) the rate at which vegetation regrows after being disrupted. one of the most important issues is the nature of the hillslope-channel transition.Dynamics of Vegetation and Runoff Erosion In fully vegetated landscapes.
for example). Horton. based on DEM analysis).. we assume a linear relationship. τ is ﬂuid shear stress. By binding the soil with roots and leaf cover (grasses. which ranges from zero to one (or more generally from zero to V max ≤ 1 . 1990. and τc is a threshold for particle entrainment. an idea consistent with the threshold channel initiation hypothesis (e.. vegetation effectively increases τc. Dietrich et al. 1989) and also with Foster’s (1982) soil erosion model. Willgoose et al.Model of Vegetation and Erosion Modeling approach Vegetation reduces erodibility There are a number of ways in which the effects of vegetation on surface erodibility might be modeled. Vegetation can be considered in terms of a fractional ground cover.. Note that this approach assumes. where Vmax is the maximum percent cover that can be supported in a given environment).. Montgomery and Dietrich. which forms part of the WEPP agricultural model (Foster et al. 1995). We account for changes in the total shear stress due to increased roughness and for changes in the fraction of shear stress applied to the soil surface using the approach Channel Hillslope Integrated Landscape Development Model 5-3 . Many sediment transport equations have the form qs = k ( τ – τc ) . 1945.g. via equation (1). that transport capacity depends on vegetation cover. (EQ 2) where τcs is the critical shear stress for an unvegetated surface (primarily a function of grain size) and τcv is the added critical shear stress under 100% vegetation cover. V. A simple but physically plausible approach is to assume that vegetation increases the both the effective shear stress for runoff erosion and the critical shear stress for particle entrainment. p (EQ 1) where qs is sediment transport capacity per unit ﬂow width. recognizing that this may be an oversimpliﬁcation: τ c ( V ) = τ cs + V τ cv . 1993 give estimates of τc at channel heads. Few data sets exist with which to constrain the relationship between τc and V (Foster 1982 provides some ﬁeld-calibrated values. In the absence of better information.
R f = k rv V . again assuming a linear relationship. 1995). The “humped” curves reﬂect the dual inﬂuence of vegetation cover on critical shear stress and on vegetation erodibility. 1990. there is little vegetation to erode. but only recently have there been attempts to model the process quantitatively (Thornes. This has the potential for a self-enhancing feedback: removal of vegetation decreases τc. Kirkby. (EQ 4) Runoff erosion disrupts vegetation Surface runoff clearly disrupts vegetation when it becomes strong enough. The total shear stress and effective shear stress τf applied to the soil are assumed to be related by fs -τ .Dynamics of Vegetation and Runoff Erosion discussed by Foster (1982). the erosion potential is reduced due to the increase in τc. Channel Hillslope Integrated Landscape Development Model 5-4 . 1995. We speculate that the rate of vegetation destruction by rill or channel runoff is proportional to excess shear stress and to the fractional vegetation cover remaining: dV η (erosion) = – k vd V ( τ – τ c ) . k rv ≤ 1 . dt (EQ 5) with kvd being the rate of vegetation loss per unit excess shear stress (or stress to a power) at 100% vegetation cover. The ratio Rf=fs/ft is parameterized as a function of V. when cover is sparse. Solutions to (5) are sketched as a function of the dimensionless parameter G = τ/τcv (ratio of applied shear stress to vegetation-added critical shear stress) in Figure 5-1. Kirkby and Cox. τ f = ---ft (EQ 3) with fs/ft being the ratio of the friction factor produced by the soil alone to the total friction factor (including vegetation cover). which in turn increases the rate of vegetation loss (but which is also compensated for by reduction in V). but when cover is dense.
3 0. For the present purposes.. The rate of regrowth of vegetation biomass Vb is sometimes modeled using a logistic growth equation.2 0.9 Nondimensional vegetation destruction rate 0. it is reasonable to assume that there will always be a ready supply of seeds and colonizing roots near a devegetated area (recall that V represents vegetation cover within a rill or channel.9 1 FIGURE 5-1.Model of Vegetation and Erosion 1 0. the surrounding hillsides are assumed to have V=Vmax). nutrients.1 0. Vegetation destruction rate as a function of percent vegetation cover.5 0. Thornes.1 0 0 0.6 % vegetation cover (V) 0.g. soil moisture. however. which assumes a linear rate of biomass growth and an exponential rela- Channel Hillslope Integrated Landscape Development Model 5-5 . and also approaches zero if the biomass is very small (in which case there are few or no organisms to reproduce) (e. Vegetation regrowth After vegetation is disrupted (for example.5 1 2 0. dt (EQ 6) The parameter kvg is the rate of regrowth on a bare surface. Note that this approach is consistent with the crop growth model used in the WEPP agriculture model (Arnold et al.6 0. for different values of G (nondimensional shear stress).4 0. in which the rate of growth approaches zero as the biomass becomes large.5 0.3 0. 1995). with the rate of regrowth being proportional to the existing cover: dV (growth) = k vg ( 1 – V ) . during a storm) it will begin to regrow.. and so on.4 0.2 0. 1990).7 0.7 0. and would be a function of solar radiation. a simpler linear approach seems warranted. In view of this.8 0.8 0.
dV = k vg ( 1 – V ) – k vd V ( R f τ – τ c ) . we explore analytical solutions to this equation.. and τcs=0. η=1. Characteristic Form Proﬁles Nondimensionalization It is interesting to consider what a vegetation proﬁle across a steady-state (characteristic form) hillslope might look like. We can obtain a steady-state vegetation proﬁle by solving Channel Hillslope Integrated Landscape Development Model 5-6 . Combining the terms for vegetation growth and erosion.e. dt dt (EQ 8) which implies that k dV -(1 – V ) . vegetation does not affect total shear stress). and how the presence of vegetation inﬂuences the shape of the hillslope. The cover-biomass relationship in WEPP is V = 1 – exp ( – V b ⁄ B ) (EQ 7) and the biomass growth rate is dV b d = ( – B ln ( 1 – V ) ) = k . = -B dt (EQ 9) Equations (8) and (9) say that although the biomass can continue to grow indeﬁnitely. we make the further simpliﬁcations Rf=1 (i. it eventually reaches a point where additional biomass growth does not significantly increase the fractional vegetation cover on the surface.Dynamics of Vegetation and Runoff Erosion tionship between plant cover and biomass. In doing so. dt (EQ 10) In the next section.
Wash Proﬁle The equilibrium vegetation proﬁle depends in part on the rate at which it is being disrupted by runoff erosion. Introducing these dimensionless numbers into equation (10). at a uniform rate E such that qs = Ex’. 1 -p 1 -p LE N E = --------k τ cv (EQ 17) Substituting this into equation (16) yields Channel Hillslope Integrated Landscape Development Model 5-7 . First we consider the case of a hillslope proﬁle that is eroded solely by wash (equation (1)).Characteristic Form Proﬁles equation (10) for the case dV/dt=0. where E is an erosion rate that is constant along the proﬁle. we have dV 2 = N v V – ( 1 + N τ' ) V + 1 . we introduce the following nondimensionalization: Vegetation growth time scale Nondimensional time Nondimensional shear stress Nondimensional distance along slope proﬁle Vegetation number T v = 1 ⁄ k vg t' = t ⁄ T v τ' = τ ⁄ τ cv x' = x ⁄ L. and vice-versa. the equilibrium shear stress is τ' = N E x' + V . low values represent fast-growing and/or destruction-resistant vegetation. k vd τ cv N v = ------------k vg L = slope length (EQ 11) (EQ 12) (EQ 13) (EQ 14) (EQ 15) The vegetation number represents the efﬁciency of vegetation growth relative to destruction. d t' (EQ 16) which is a quadratic equation that can be solved if dV/dt is constant or zero. which may vary along a slope. To facilitate the analysis. and solving equation (1) for the case qs = Ex. Under that condition.
When Nv is large (Figure 5-2. NE q – m 1 . Channel Hillslope Integrated Landscape Development Model 5-8 . q=1/p=2/3 (e. p (EQ 18) Similarly. assuming typical values of m=2/3. Figure 5-2 depicts solutions to (18) and (20) for different values of the dimensionless parameters Nv. we can solve for equilibrium slope along the proﬁle using the shear stress relationship τ = k t L x' S . Vegetation cover decreases downslope in response to the increasing shear stress. NE. 1909). we assume that wash controls the amount of vegetation but does not contribute signiﬁcantly to shaping the topography. Gilbert.Dynamics of Vegetation and Runoff Erosion 1 V = -----------------------------. there is an abrupt reduction in vegetation cover in the upper part of the proﬁle that corresponds to a sharp concavity in the topography. This vegetation gradient has the effect of increasing the concavity of the hillslope proﬁle. top and bottom). asymptotically approaching zero as x’ becomes large. middle). Diffusive Proﬁle The wash model ignores the effect of creep-related (slope-dependent) processes that would tend to produce convex-upward slopes near a drainage divide (e. Essentially. (EQ 20) where Nt = ktLm/τcv.. q q 1 + N v N E x' 1 q = --. and Nt.g. The proﬁle shape under the action of both creep and wash processes cannot be solved analytically. the Meyer-Peter and Mueller relation) or q=1/p=1/3 (the Einstein-Brown relation).g.. n=2/3. The vegetation number Nv controls the downslope rate of reduction in vegetation cover. particularly when the erosion number NE is small relative to the other parameters (compare Figure 5-2. m m n (EQ 19) Solving for the equilibrium condition. but we can obtain some idea of what a vegetation proﬁle might look like on a diffusion-dominated slope by solving for the characteristic-form slope as if diffusion were the sole process.x' S = -----+ -----------------------------------------------q q m Nt N x' ( 1 + N N x' ) t v E q 1⁄n .
6 0.4 0.8 0.2 0.8 0.4 0.2 0.2 0.3 0. p=3) Slope profile Vegetation profile Slope profile w/o veg 1 1 0.4 0.2 0.7 0.7 0.9 Nondimensional elevation / % vegetation 0.1 0.5 0.9 Nondimensional elevation / % vegetation 0.2 0. Solutions to equations (18) and (20) for wash-dominated hillslope proﬁles.9 1 Characteristic form wash profile with vegetation (Ne=10) 1 0.Characteristic Form Proﬁles Characteristic form wash profile with vegetation 1 0.1 0.3 0.4 0.3 0.7 0.7 0.5 0.7 0.5 0.2 0.4 0.1 0 0 Slope profile Vegetation profile Slope profile w/o veg Nondimensional elevation / % vegetation 1 0.1 0 0 Slope profile Vegetation profile Slope profile w/o veg 0.8 0.8 0.7 0.2 0.3 0.9 Nondimensional distance Characteristic form wash profile with vegetation (Nv=4.7 0.9 0.6 0.3 0.6 0.4 0.3 0.8 0.5 0.1 0.6 Nondimensional distance 0.5 0.1 0.2 0.9 Nondimensional elevation / % vegetation 0. p=3) Slope profile Vegetation profile Slope profile w/o veg 0.1 0 0 0.3 0.4 0.8 0.7 0.7 0.6 0.8 0.8 0. showing the effect of vegetation on proﬁle shape.1 Characteristic form wash profile with vegetation (Nv=10) 1 0.5 0.6 Nondimensional distance 0.3 0.6 0.5 0.7 0.1 0.3 0.2 0.9 1 0 0 Characteristic form wash profile with vegetation (p=3) Slope profile Vegetation profile Slope profile w/o veg 0.6 0.3 0.6 Nondimensional distance 0. Channel Hillslope Integrated Landscape Development Model 5-9 .4 0.1 0.7 0.3 0.7 0.5 0.9 1 0.5 0.9 0.1 0 0 Characteristic form wash profile with vegetation (Ne=10.2 0.4 0.1 0 0 Slope profile Vegetation profile Slope profile w/o veg Nondimensional elevation / % vegetation 1 0.2 0.8 0.5 0.8 0.5 0.9 Nondimensional elevation / % vegetation 0.8 0.6 0.4 0.5 0.3 0.9 1 0.2 0.4 0.6 Nondimensional distance 0.4 0.8 0.9 1 FIGURE 5-2.6 Nondimensional distance 0.
Depending on the initial conditions. while the other implies decreasing cover.x' τ' = ----------------= N dt x' .8 0.x' = N d x' .6 0. In the example shown in the ﬁgure.9 1 Channel Hillslope Integrated Landscape Development Model 5-10 .1 0.2 Hillslope profile Vegetation (solution 1) Vegetation (solution 2) 0 0 0.3 0. 0.4 0.6 Nondimensional distance 0.5 0. Solutions to equation (23) for equilibrium vegetation cover on a diffusion-dominated hillslope on which vegetation growth is limited by wash erosion.Dynamics of Vegetation and Runoff Erosion Assuming that the transport rate by creep is given by qs = kd S. This example of bi-stability is similar to the bistable landscape states explored by Howard (1996). the characteristic form proﬁle is UL . as well as the multiple phase-states in the soil erosion model of Thornes (1990).7 0. τ cv m n (EQ 22) Combining with (16) (for dV/dt’=0) gives the vegetation proﬁle N v V – ( 1 + N v N dt x' 2 m+n )V + 1 = 0 . This analytical example of a diffusion-dominated proﬁle is of course rather artiﬁcial because it does not incorporate the linkage between slope form and Nondimensional elevation / Vegetation cover % 1 0.4 FIGURE 5-3.2 0. Note that equation (23) has two roots. S = ------kd (EQ 21) Substituting into equation (19). one of the roots implies full vegetation cover along the proﬁle. (EQ 23) Example solutions to (21) and (23) are shown in Figure 5-3. the shear stress is given by kt L N d m + n m+n . the system may evolve toward a different equilibrium state.8 0.
The basin is close to a state of dynamic equilibrium between uplift and erosion. In this particular example. which reduces slopes and thus shear stresses (as in the extreme case of Figure 5-3). the downstream transition from full to sparse vegetation cover is a function of two factors: (1) the threshold imposed by the vegetation itself. which retards runoff erosion on upper slopes (as seen in Figure 5-2). and vegetation growth. (a) 400 200 0 0 1000 2000 3000 4000 5000 2000 6000 7000 0 1000 (b) FIGURE 5-4. (The presence of a small vegetation cover within the channels is due to the fact that vegetation always regrows during interstorm periods. and (2) the rounding of hillslope proﬁles due to diffusion. Numerical Examples Figure 5-4 shows the topography and percent vegetation cover in a synthetic catchment simulated using the CHILD model. Simulated catchment with variable vegetation cover. numerical simulations are needed. the hillslope diffusivity constant kd is sufﬁciently high that the latter effect is more important. (a) Topography. White = 100% cover. black = 0% cover.Numerical Examples wash erosion. regardless of position within the landscape). In general. Vegetation cover ranges from 100% on the hillslopes to very low values in the main channels. Channel Hillslope Integrated Landscape Development Model 5-11 . To explore the complete interaction of wash. (b) Percent vegetation cover. diffusive transport.
but also the soil inﬁltration capacity and hence runoff production. This result is somewhat contrary to intuition. where the initial vegetation cover (and thus the threshold) was greatest (this type of behavior was observed by Tucker and Slingerland (1997) in experiments with a constant-threshold model). Short term-effect of a decrease in τcv on percent vegetation cover in the synthetic catchment shown in Figure 1000 5-4. and little or no change (green). Figure 5-5 illustrates the effect of a hypothetical decrease in the maximum sustainable vegetation cover. as it did in the constant-threshold experiments of Tucker and Slingerland Channel Hillslope Integrated Landscape Development Model 5-12 . the valley network extent in the simulation is limited by hillslope diffusion rather than by the threshold imposed by vegetation cover. The ﬁgure depicts areas of erosion (red). reducing the threshold does not lead to a rapid expansion of the valley network. Thus. which leads to an upslope extension of the sparsely-vegetated tributaries.Dynamics of Vegetation and Runoff Erosion 4000 3000 2000 FIGURE 5-5. deposition (not shown). The explanation appears to lie in the topography of the catchment and the nature of the hillslope-valley transition. Figure 5-6 shows a hypothetical example in which the synthetic catchment shown in Figure 5-4 is subjected to a complete loss in vegetation accompanied by a reduction in inﬁltration capacity (Ic = 0). The net effect of vegetation loss in this example is an increase in the rate of scour along the main channels. Vegetation cover typically inﬂuences not only surface resistance to erosion. which would suggest that a loss of vegetation should lead to accelerated erosion on hillslopes. The decrease in τcv is accompanied by an increase in the effectiveness of runoff erosion. The decrease in sustainable vegetation maximum is modeled by a tenfold decrease in the threshold parameter τcv. With a relatively high diffusivity constant kd.
by extension. green=no change) following a complete loss of vegetation and a corresponding 1000 reduction in inﬁltration capacity in the synthetic catchment shown in Figure 5-4. erosion. There is still much to be learned about what controls the type of response in the model and. We anticipate that this question can be fruitfully addressed through a program of systematic model sensitivity experiments and comparison of the model predictions with the Holocene stratigraphic record observed at Fort Riley. Patterns of erosion (red=max. Kansas. and in particular on whether that transition is governed by a process transition or by a vegetationimposed erosion threshold.Numerical Examples 4000 3000 2000 FIGURE 5-6. 0 (1997). Channel Hillslope Integrated Landscape Development Model 5-13 . in natural catchments. An implication of this result is that the nature of catchment responses may depend to a large extent on the nature of hillslope-valley transitions.
Dynamics of Vegetation and Runoff Erosion Channel Hillslope Integrated Landscape Development Model 5-14 .
Chapter 6 Understanding the Interactions of Multiple Grain Sizes 6-1 .
Appendix A Meandering: A Simple Model A-1 .
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue reading from where you left off, or restart the preview.