Sunteți pe pagina 1din 19

Geoderma 90 1999.

321

A rudimentary mechanistic model for soil production and landscape development


Budiman Minasny ) , Alex. B. McBratney
Department of Agricultural Chemistry and Soil Science, The Uniersity of Sydney, NSW 2006, Australia Received 27 January 1998; accepted 8 October 1998

Abstract A rudimentary mechanistic model for soil production and landscape development is proposed. The continuity equation of the model assumes that the change in soil thickness over time depends on the production of soil from the weathering of bedrock and the transport of soil through natural surface erosion. The parameters for the model include the weathering rate, and erosive diffusivity. Weathering rate is expressed as an exponential decay function of soil thickness, which represents mechanical weathering. Erosive diffusivity can be estimated from soil erosion models. The model is solved numerically using the finite-difference approach and is applied to a numerical example. Simulation is performed for a hypothetical landscape with a series of hills and valleys. Results show the development rate of soil initially is very large, and slows down until it reaches a steady-state, defined by a constant change in soil thickness. The model also exhibits the characteristics of a nonlinear dynamic system: nonlinearity of soil thickness and curvature, and initial randomness appears to cause chaos instability. in the system. The results show promising progress in quantitative modelling of pedogenesis. The limitations and suggested improvements of the model are also presented. q 1999 Elsevier Science B.V. All rights reserved. Keywords: quantifying pedogenesis; mechanistic model; soil production; weathering rate; soil erosion; diffusional transport; nonlinear dynamic system

Corresponding author. Telefax: q61-2-935-13706; E-mail: budiman@acss.usyd.edu.au

0016-7061r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved. PII: S 0 0 1 6 - 7 0 6 1 9 8 . 0 0 1 1 5 - 3

B. Minasny, A.B. McBratney r Geoderma 90 (1999) 321

1. Introduction There is an increasing necessity to quantify the processes of pedogenesis. Hoosbeek and Bryant 1992. pointed out that quantification is a life-line to other environmental disciplines and also stressed the need for mechanistic models. Phillips 1998. has discussed the application of the nonlinear dynamic system NDS. approach to pedogenesis and suggested that this concept has been embedded in the classical pedological theory through the work of Jenny 1941. . In the earliest stage, quantitative models were mostly focused on an empirical approach, relating soil information to soil processes, such as erosion, soil organic matter production, mineral dissolution etc. The need for better understanding of the processes, has lead to the development of mechanistic models. The models developed, which are usually at the pedon or horizon scale, are the integration of sub-models, for example, water and solute movement, heat transport, soil organic matter decomposition, mineral dissolution, ion exchange, adsorption, speciation, complexation and precipitation. These sub-models arise directly from models developed in soil physics, chemistry and biology. Hoosbeek and Bryant 1994. termed this pedodynamics, which is defined as the quantitative integrated simulation of physical, chemical, and biological soil processes acting over short time increments in response to environmental factors. They analysed glacial outwash deposits and calculated the biogeochemical reactions and fluxes of major elements. The pedodynamics model successfully simulated the chemistry and movement of major elements in the podzolization process. Most of the models developed considered the chemical reactions and fluxes in the soil at the horizon scale. As an alternative, the model presented in this paper precedes the other models, and considers the soil production spatially at the catena scale and for time periods greater than 10 years. The following terms are used in describing the model in this paper: soil production: the formation of soil resulting from the weathering of bedrock in units of volume of soil per unit area mm3 mmy 2 s mm.; soil production rate: the rate of soil formation or increase in soil thickness over time, in units of volume of soil per unit area per unit time mm3 mmy2 yeary 1 s mm yeary 1 .; weathering rate: the rate of bedrock weathering or lowering of bedrock surface over time, in units of volume of rock per unit area per unit time mm3 mmy 2 yeary 1 s mm yeary 1 .. Little work has been done in soil science in order to formulate a general mechanistic model for the soil production and equilibrium in the landscape. In the fields of geology and geomorphology, integrated soil production and slope development formulations have been developed e.g., Anderson, 1988. , however. Heimsath et al. 1997. proposed a model for soil production in a landscape and have verified it with field data from Tennessee Valley. They also illustrate

B. Minasny, A.B. McBratney r Geoderma 90 (1999) 321

how this model can be used to determine whether a landscape is in equilibrium or not. The purposes of this paper are to introduce a simple mechanistic model for soil production at the catena scale and to illustrate the application of the model in quantifying pedogenesis which also show the application of the nonlinear dynamical system as proposed by Phillips 1998.. This paper also discusses the parameters of the model and attempts to bridge the hiatus of soil genetic modelling between geomorphology and soil science.

2. Theory Consider a landscape with surface elevation z , soil with thickness h and soilbedrock interface e along a horizontal x-axis Fig. 1. . The change in soil thickness over time depends on two major processes: the production of soil from weathering of bedrock and the transport of soil by erosion process, or in simple mathematical form: Change in Soil Thickness s Weathering q Inflow y Outflow

1.

Soil production depends on the rate of breakdown or weathering of the underlying parent materials under physical, chemical and biological processes, which will result in the lowering of soilbedrock interface. Considering the simple nature of the model, these factors are not analyzed individually.

Fig. 1. Simplified model for soil production on a landscape Based on Carson and Kirkby, 1972; Heimsath et al., 1997..

B. Minasny, A.B. McBratney r Geoderma 90 (1999) 321

The rate of weathering or lowering of bedrock surface yE erE t . is usually represented as an exponential decline with thickening of soil Armstrong, 1980; Cox, 1980; Ahnert, 1988.: Ee s yP0 exp ybh . 2. Et where P0 wL Ty1 x is the potential weathering rate of bedrock at h s 0 and b wLy1 x is an empirical constant. Ahnert 1977. described that the reduction in mechanical weathering with thickening of soil corresponds to the exponential decrease of temperature range with increasing depth below soil surface. This can also be confirmed by looking at the equation for heat transport in soil. The analytical solution for heat transport for a uniform soil under annual periodic temperature changes is Jury et al., 1991.: T z , t . s TA q A 0 exp yzrd . sin v t q zrd . 3. where T z , t . is the temperature at depth z and time t , A 0 is the amplitude of temperature variation at soil surface, v is the angular frequency of temperature variation and d is called the damping depth, defined as:

4. v where D h is the thermal diffusivity wL2 Ty 1 x. The temperature at any depth of soil decreases at an amplitude of A 0 exp yzrd .. By comparing this relationship to Eq. 2., we are able to derive the relationship as follows: P0 should be proportional to A 0 , and b inversely proportional to d. Parameter P0 is mostly controlled by the climate, while b is mostly controlled by thermal properties of the rockrsoil. Thermal properties of various rocks can be found in Smith 1977., McGreevy 1985. and Warke and Smith 1998.. Kirkby 1985. argued that the exponential decrease in weathering is the result of very slow drainage of water. Temperature is the most important factor in mechanical breakdown of rock. It affects the breakdown indirectly through its control on moisture and processes such as freezingthawing, salt weathering and also controls chemical weathering Warke and Smith, 1998.. Despite the wide use of this relationship, the values for the parameters are rarely published. Table 1 shows weathering rates for different rock types under different climate conditions which can be used as a guide for the value of P0 . Note that the weathering rates represent both chemical and physical processes. Heimsath et al. 1997. obtained the production rate from calculated value of in situ produced cosmogenic nuclide concentration in bedrock samples. They found the value of P0 s 0.77 " 0.09 mm yeary1 and b s 0.0023 " 0.0003 mmy1 in the Tennessee Valley. Wakatsuki and Rasyidin 1992. predict the rate of rock weathering and soil formation from a simplified geochemical mass-balance calculation. Other methods for determining rock weathering can be found in Colman and Dethier 1986..

ds

2 Dh

B. Minasny, A.B. McBratney r Geoderma 90 (1999) 321 Table 1 Weathering rates for different parent materials in various climate regions Rock type Volcanic material Volcanic material Granitic rock Granitic rock Ultrabasic rock Rate mm yeary1 . 0.058 7.78 0.014 0.0135 0.0290.047 0.009 0.003 0.077 0.0003 0.011 0.003 0.0004 0.037 0.3 2.5 Climate Koppen. Af Af Aw Aw Aw Aw Aw Bw Ca Cb Db Db Db ET ET Location Papua New Guinea Krakatau, Indonesia Ivory coast Chad New Caledonia Uganda Senegal Tennessee Valley North Queensland Trnavka river basin, Chekoslovakia Pacific Northwest Nothern Virginia Southern Blue Ridge Longyeardalen, Spitsbergen Swiss Alps Author Ruxton, 1968. Jenny, 1941. Boulange, 1984. Gac, 1979. Trescases, 1973. Trendall, 1962. Fritz and Tardy, 1974. Heimsath et al., 1997. Pillans, 1997. Paces, 1986. Dethier, 1986. Pavich, 1986. Velbel, 1986. Jahn, 1976. Barsch, 1977.

Greenstone, sandstone Basalt Biotitic gneiss Sedimentary and metamorphic Granite Gneiss

Soil erosion depends on elevation z which acts as a potential driving the process. The first derivative of z with respect to space = z . is the slope or gradient, in which = s partial derivative vector E.r E x ., E.r E y .. Gradient is the maximum rate of change of elevation and is directly proportional to the rate of sediment flow. Odeh et al. 1991. concluded that gradient influences the distribution of surficial soil particles which is suggestive of selective removal and deposition of fine materials on mid to lower-slopes. The second derivative of z with respect to space = 2 z . is the curvature or profile curvature., defined as the rate of change in gradient. Curvature can affects flow acceleration and deceleration Odeh et al., 1991. . A positive curvature reflects a convergent slope valley., and negative one is a divergent slope hill. . The process of erosion or sediment flow may occur by overland flow and mass wasting. The transport rate q . of material soil. can be defined similarly to Darcys law for water transport in soil: q s yK e Ez Ex

5.

where q is the flux wLTy1 x or volume of material transported across unit area per unit time wL3 Ly 2 Ty 1 x and K e is the erosivity of the material wLTy1 x.

B. Minasny, A.B. McBratney r Geoderma 90 (1999) 321

The movement of materials in the landscape is usually expressed as diffusive transport Scheidegger, 1991. , where the flux is defined as: Ez qs s yD 6. Ex in which qs is the sediment flux wL2 Ty 1 x or volume of material that flows across a slope profile width per unit time wL3 Ly 1 Ty 1 x, and D is the erosive diffusivity of the material wL2 Ty 1 x. Carson and Kirkby 1972. recognize two limiting processes for material transport and weathering: weathering limited, where transport process is more rapid than weathering rate and, transport limited, where weathering rate is more rapid than transport processes. The erosive diffusivity depends on the erosion factors of the soil, e.g., vegetation cover, soil physical properties, and weather. For numerical modelling, Koons 1989. suggested that D should be replaced with DU , the effective erosional diffusivity, which is the diffusivity for the averaged time and space considered in the numerical modelling. He also found that D is largely a function of climatic parameters and can be spatially and temporally variable. Martin and Church 1997. suggested that D should be rewritten with two separate diffusion terms: diffusivity for slow, continuous mass movements e.g., creep. and diffusivity for rapid, episodic mass movement e.g., landslide. . Erosive diffusivity varies widely as shown in Table 2. In geomorphological studies, the value of D is usually estimated from Eq. 6. , by determining the sediment flux required to fill in unchannelled valleys determined by radiocarbon deposition rates. at particular gradient Reneau et al., 1989.. Another way to estimate the value for a certain kind of soil is by deriving it from the Universal Soil Loss Equation or USLE Wischmeier and Smith, 1978.. The USLE is given by: AsRPKPLPSPCPP 7. y2 y 1 x w Where A is the mass of soil lost from a unit area per year ML T , R is the rainfall erosivity, K is the soil erodibility factor, L is the slope length factor, S
Table 2 Comparison of erosive diffusivity values from various authors Authors Nash, 1980; Hanks et al., 1984. Colman and Watson, 1983. Koons, 1989. Flemings and Jordan, 1989. McKean et al., 1993. Kooi and Beaumont, 1996. Martin and Church, 1997. D mm2 yeary 1 . 1.1=10 1.6=10 9.0=10 2 1.5=10 5 1.5=10 7 1=10 8 5=10 9 4.9=10 3 "3.7=10 3 2=10 4 1=10 8 Landslide: 2=10 5; Creep: 2=10 2
3 4

Note hillslope erosion from arid regions derived from scarp study low rainfall region in Southern Alps sub-Andean foreland coastal mountains of California, Oregon and Washington numerical example numerical study based on available transport data

B. Minasny, A.B. McBratney r Geoderma 90 (1999) 321

is the slope factor, C is the cropping management factor and P is the erosion control practices factor. Using this relationship, K e and D can be derived: Kes A 1

rs = z

or

Ds

A Lx

rs = z

8.

where L x is the length of the slope, rs is the density of the soil wM Ly3 x. Values of D can be calculated as a function of K soil erodibility factor. and slopes. The results of such calculation by holding R constant at 2000 MJ mm hay1 hy1 . and C P P s 1 is shown in Fig. 2. We can formulate the continuity equation for the soil thickness over time at equilibrium as: the change in soil thickness over time plus the weathering of bedrock is equal to the transport of soil by erosion. The equation is: 1 ECs q 1 E Cr Eq sy Ex

rs E t

rr E t

9.

where Cs is the concentration of soil mass over the volume of soil. at h wMLy3 x, Cr is the concentration of rock at e. In diffusion terms,

rs

Eh Et

q rr

Ee sy Et

E Ex

rs qs .

10.

where rs is the density of soil and rr is the density of rock. It should be noted that this equation does not take into account chemical weathering.

Fig. 2. Relationship between slopes and K to values of D mm2 yeary 1 . at constant R of 2000 MJ mm hay1 hy1.

10

B. Minasny, A.B. McBratney r Geoderma 90 (1999) 321

If we combine Eqs. 5. and 9. , we obtain the general equation for one-dimensional soil production: 1 ECs

rs E t

1 E Cr

E sy Ex

rr E t

Ke

Ez Ex

11.

Likewise, by combining Eqs. 6. and 10. and assuming D and rs constant over space, then the soil production function is: Eh sy Et

rr E e rs E t

qD

E2 z Ex2

12.

The above equation destates that the production of soil thickening of h. over time depends on the weathering rate of rock lowering of e . and the transport of soil through erosion, which is dependent on the diffusivity D . and curvature of the landscape E 2 zrE x 2 .. The above equations are analogous to heat transport in soil. The soil surface z . is the analogue of temperature and D the analogue of soil thermal diffusivity D h .. Eq. 5. is analogous to Fouriers law, and Eq. 12. to the heat flow equation soil Jury et al., 1991. . Given appropriate initial and boundary conditions, the above partial differential equation Eq. 12.. can be solved numerically using the finite-difference approach. The formulation using explicit approximation on a rectangular grid over space x . and time t . is: h txq 1 y h tx Dt sy
tq 1 t rr e x yex

rs

Dt

qD

t t t zx q 1 y 2 z x q z xy 1

D x.

13.

under the following initial and boundary conditions: h s hi , z s zi Ee Et Eh Ex s yP0 exp ybh . s0 at t s 0, x G 0 at t ) 0, 0 - x - ` at t ) 0, x s 0 and x s `

The initial condition describes the soil thickness and elevation at h i and z i . The second condition states that the soil production rate is assumed to be an exponential decay function of the soil thickness. And the third condition defines the lower and upper boundary for the finite difference space grid. The numerical solution of Eq. 13. is stable under the condition where the stability ratio 2 D D tr D x . 2 should be F 1.0 Press et al., 1992. . Eq. 13. could be readily applied in a two-dimensional model landscape, using Digital Elevation Model data.

B. Minasny, A.B. McBratney r Geoderma 90 (1999) 321

11

3. Methods 3.1. Application of the model A numerical example is given to illustrate the application of Eq. 13. . A hypothetical landscape with initial elevation representing a series of hills and valleys are simulated with initial soil thickness h i . The density of rock and soil, and diffusivity D is assumed to be spatially and temporally constant. The weathering rate decreases exponentially with increasing soil thickness. The parameters used for the simulation are summarised in Table 3. The time-step used is 500 years. The calculation is implemented with a commercial spreadsheet program MS Excel. using rows and columns as time and space grids in the explicit finite-difference approach. 3.2. The effect of initial randomness on soil deelopment To illustrate the effect of irregularity and randomness on the stability of the solution and the soil development, uniformly distributed independent random numbers correlated with a moving average across the x-axis were added to the initial soil elevation which has the average height of 1000 mm. The parameters used are the same as in Table 3, and two time-steps are chosen D t s 250 and 500 years in order to satisfy the stability criterion of the numerical solution. For D t s 500 years the random number has mean of 5.2 mm and variance of 2.6 mm2 and for D t s 250 years the mean is 25.8 mm and variance is 81.6 mm2. 3.3. The effect of different rate of soil weathering on soil deelopment To illustrate the effect of different rates of soil weathering on the soil development in a landscape, two rock types are considered. On the hill to mid-slope the rock type is resistant with P0 s 0.10 mm yeary1, while on mid-slope to valley, the rock is more easily weatherable with P0 s 0.19 mm
Table 3 Parameters used in the simulation Parameters D mm2 yeary 1 . D x mm. rr g mmy3 . rs g mmy 3 . h i mm. P0 mm yeary 1 . b mmy 1 . Value 4000 500 2.6=10y 3 1.4=10y 3 20 0.19 0.005

12

B. Minasny, A.B. McBratney r Geoderma 90 (1999) 321

Table 4 Comparison of the square root and logarithmic function on describing the relationship between soil thickness and time Data Valley this paper. Hill this paper. Taranaki, New Zealand Trustum and De Rose, 1988. Ventura, California Harden and Taylor, 1983.
a

Eq. Eq. 14. Eq. 16. Eq. 14. Eq. 16. Eq. 14. Eq. 16. Eq. 14. Eq. 16.

S1 5.84 256.13 5.23 117.45 1.98 7.23 0.66 40.07

S2 y0.0062 18.40 0.01 5.09 0.02 0.08 2.6=10y5 0.18

hi 106.00 y2422.81 122.91 y717.60 0.00 5.30 93.33 y105.25

RMSRa 12.40 44.99 15.35 16.15 1.70 1.38 58.77 60.34

RMSR s Root mean squared of the residuals.

yeary1. The other parameters used are the same as in Table 4, and the time-step is 250 years.

4. Results and discussion 4.1. Application of the model The results on simulation of soil production in a landscape are shown in Figs. 3 and 4. The time-step used in this simulation is 500 years. The solution is found to be consistent with the smaller time-step of 250 years. Fig. 3 shows the change in soil surface with soil development over time. The relief of the soil

Fig. 3. Simulation of soil surface changes over time with a 500-year time-step. The numbers on each line represents the time in years.

B. Minasny, A.B. McBratney r Geoderma 90 (1999) 321

13

Fig. 4. Simulation of soil thickness development over time with a 500-year time-step.

surface tends to decrease with time. This is also shown for soil thickness development in Figs. 4 and 5, in which soil accumulation is greater for the valley than for the hill. With the same weathering rate, soil from the hill will be transported downslope by the erosion process and accumulated in the valley. Fig. 5 shows that with an initially thin soil, the rate of production is very high which corresponds to the exponential decline function in Eq. 2.. Soil thickness then increases gradually until it reaches a steady state production rate E hrE t s constant., in our example the production reached steady-state at around forty thousand years. The soil thickness increase with time can be fitted by a power series on the square root of time which is known to be the semi-analytical solution of the

Fig. 5. Soil thickness development over time in the valley, and on the mid-slope and hill.

14

B. Minasny, A.B. McBratney r Geoderma 90 (1999) 321

Fig. 6. Change in curvature with increasing soil thickness over time, each line in the vertical direction represents a 500-year time-step.

transport equation Jury et al., 1991.. In this example, the first three terms of the series were found to be adequate in representing the data: h s h i q S1't q S2 t

14.

where S1 wLTy1r2 x and S2 wLTy1 x are empirical constants. The rate of soil production can be obtained by differentiating Eq. 14. with respect to time: Eh s Et S1 2't q S2

15.

This function is also found to give a better fit and more realistic values than the logarithmic function, which is usually used in describing soil chronofunctions Mellor, 1985; Trustum and De Rose, 1988.: h s h1 q S1 ln S2 t .

16.

Table 4 shows the comparison between the square root and the logarithmic function for our model and also from published empirical data. At steady state, the rate of soil production on the divergent slope hill. reaches zero or E hrE t f 0, and Eq. 12. reduces to: Ee Et sD

rr E 2 z rs E x 2

17.

B. Minasny, A.B. McBratney r Geoderma 90 (1999) 321

15

Table 5 Summary of the response surface parameters of soil thickness mm. to curvature mmy1 . and time year. Term Intercept Curvature 'time Curvature 2 CurvatureP'time Time Estimate 92.8583 y44 099 205 5.5053 1.1163=10 13 702 842.42 y0.0082 Std. Error 1.2163 450 595.9 0.0144 1.86=10 11 2401.04 0.000042 t Ratio 76.35 y97.87 382.39 59.99 292.72 y196.60 Prob ) < t < 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

With soil thickening, the weathering rate is expected to attenuate as suggested by the exponential decline function of h or E erE t f 0, and Eq. 12. reduces to: Eh E2 z sD 2 18. Et Ex So under steady-state conditions soil production is only dependent on the erosive diffusivity and the slope curvature. Curvature initially changes linearly with increasing soil thickness until it reached around 500 mm then it starts to bend over, as shown in Fig. 6. Soil thickness beyond that point is mostly controlled by the profile curvature. Since the soil thickness is a function of the square root of time, a response surface of curvature and square root of time can be fitted to predict the soil thickness. The parameters of the fitted function are summarised in Table 5. The contour plot of soil thickness as function of time and curvature is given in Fig. 7, which shows that for a given soil thickness, it will take longer for a site at negative curvature to develop compared to one with a positive curvature.

Fig. 7. Contour plot of soil thickness mm. as a function of curvature and time.

16

B. Minasny, A.B. McBratney r Geoderma 90 (1999) 321

Fig. 8. Simulation of soil surface development over time with initial randomness added, calculated with 500-year time-step.

4.2. The effect of initial randomness on soil deelopment Scheidegger 1991. in his book Theoretical Geomorphology pp. 260. described the nature of the diffusivity equation as follows: It is thus seen that, based upon entirely general principles, one must expect a decay of initial irregularities in a landscape. Correlated random numbers were added to the initial soil surface height to see the effect of the initial irregularities. The results of soil development are shown in Figs. 810. Using a 500-year time-step, where the stability ratio 2 D D tr D x . 2 is 1.0, the results show the instability of the numerical solution, where an initial small randomness can create chaotic behaviour in the system. An interesting point is shown in Fig. 10, where soil

Fig. 9. Simulation of soil thickness development over time with a 500-year time-step with initial randomness added.

B. Minasny, A.B. McBratney r Geoderma 90 (1999) 321

17

Fig. 10. Soil thickness development over time in the valley, and on the mid-slope and hill with a 500-year time-step with initial randomness added.

thickness exhibits periodic fluctuation, which is known as bifurcation Gleick, 1987.. Depending on the position on the slope, the fluctuation can become smaller and larger. Using a 250-year time-step with stability ratio 0.5, the system becomes more stable, initial randomness does not affect the soil surface but greatly affects soil thickness Fig. 11.. Initial irregularities do not diminish as foreseen by Scheidegger 1991. but tends to intensify. As the solution is based on the finite-difference of h and e, both show amplification of the initial irregularity. This result

Fig. 11. Simulation of soil thickness development over time with a 250-year time-step with initial randomness added.

18

B. Minasny, A.B. McBratney r Geoderma 90 (1999) 321

also exhibits one of the features of a nonlinear dynamic system characteristic: the recognition of the possibility for inherent stability, deterministic chaos Phillips, 1998.. 4.3. The effect of different rate of soil weathering on soil deelopment Numerical solutions have the advantage of flexibility for varying parameters at different positions on the space and time grid. Fig. 12 illustrate a condition in a landscape where the rock in the valley which has a faster weathering rate P0 s 0.19 mm yeary1 . is enclosed by the hill which has a lower weathering rate P0 s 0.10 mm yeary1 .. With initial elevation lower than previous simulations, results showed that after 40,000 years, the surface of the valley has almost reach the same level as the hill, and creates small ridges as the result of the sharp boundary or difference between the two rock types. Ollier 1984. noted that differential weathering of rocks give rise to numerous features such as trenches, walls and ridges on the landscape. 4.4. General discussion We have illustrated the application and results of the model and here we wish to indicate some limitations of, and suggest improvements to the model. The erosive diffusivity coefficient D is always assumed to be spatially and temporally constant in a landform or system and to be independent of slope or curvature. But looking at the results above, it can be seen that there is a nonlinear relationship between curvature and soil depth and a dependency of D on the gradient that is clearly affected by curvature. As the soil develops, the gradient and curvature will also change and consequently D which is assumed to be constant. can change. Martin and Church 1997. pointed out that the

Fig. 12. Soil surface development in a landscape with two different rock types with a 250-year time-step.

B. Minasny, A.B. McBratney r Geoderma 90 (1999) 321

19

relationship between transport rate and gradient may not be linear. Phillips 1998. indicated that one of the characteristics for the nonlinear dynamic system is the recognition of nonlinearities in the system. To make the model more realistic, a relationship or characteristic curve between the gradient and D should be incorporated. The weathering rate, expressed as an exponential decay of soil thickness, is dominantly the result of mechanical weathering of the rock Ahnert, 1977. . Chemical weathering also plays an important role, as it may decompose the bedrock and part of the material is lost. This requires a different continuity equation. The model also assumes that the landscape is a closed system where the material is only transported and there is no loss of material, which never exists in a real system. It needs an extraction term to take into account the amount of material lost from weathering and erosion. Our rudimentary model really only simulates physical weathering and creep as might be found in slopes in a very arid climates or on the moon. 5. Conclusions A rudimentary mechanistic model for soil production and landscape development has been presented and the possibility of the nonlinear dynamic system approach also has been studied. Although there are many assumptions and limitations, the results are promising in terms of quantitative modelling of pedogenesis. To utilise and improve the model, much field and laboratory work is needed to collect data to estimate the parameters of the model. For future research, the model needs to be extended to incorporate other important pedogenic processes and to take into account the nonlinearity of the soil system. Application of the model can include predicting the effect of erosion on soil development and suggesting management practices in a landscape. Acknowledgements We thank Dr. Marcel Hoosbeek of The Wagenigen Agricultural University and Dr. Richard W. Arnold of Soil Survey Division, USDA-NRS for their useful comments on the paper, Mr. Damien Field for his helpful suggestions, and Dr. Chris Moran of CSIRO Land and Water for stimulating this research. References
Ahnert, F., 1977. Some comments on the quantitative formulation of geomorphological process in a theoretical model. Earth Surface Processes 2, 191201. Ahnert, F., 1988. Modelling landform change. In: Anderson, M.G. Ed.., Modelling Geomorphological Systems. Wiley, New York, pp. 375400.

20

B. Minasny, A.B. McBratney r Geoderma 90 (1999) 321

Anderson, M.G. Ed.., 1988. Modelling Geomorphological Systems. Wiley, New York. Armstrong, A.C., 1980. Soils and slopes in a humid temperate environment: a simulation study. Catena 7, 327338. Barsch, D., 1977. Eine Abschatzung von schuttproduktion und schuttransport im bereich activer blockgletscher der schweizer Alpen. Zeitschrift fur Geomorphologie Suppl. NS28, 148160. Boulange, de Cote B., 1984. Les formations bauxitiques lateritiques dIvoire. Les facies, leur transformation, leur distribution et levolution du modele. Trav. Docum. ORSTOM, Paris. Carson, M.A., Kirkby, M.J., 1972. Hillslope Form and Process. Cambridge Univ. Press, Cambridge. Colman, S.M., Dethier, D.P. Eds.., 1986. Rates of Chemical Weathering of Rocks and Minerals. Academic Press, Orlando. Colman, S.M., Watson, K., 1983. Ages estimated from a diffusion equation model for scarp degradation. Science 221, 263265. Cox, N.J., 1980. On the relationship between bedrock lowering and regolith thickness. Earth Surface Processes 5, 271274. Dethier, D.P., 1986. Weathering rates and the chemical flux from catchments in the Pacific Northwest, U.S.A. In: Colman, S.M., Dethier, D.P. Eds.., Rates of Chemical Weathering and Rocks and Minerals. Academic Press, Orlando, pp. 503530. Flemings, P.B., Jordan, T.E., 1989. A synthetic stratigraphic model of foreland basin development. Journal of Geophysical Research 94, 38513866. Fritz, B., Tardy, Y., 1974. Etude thermodynamique du systeme ` gibbsite, quartz, kaolinite, gaz carbonique. Application a Bulletin ` la genese ` des podzols et des bauxites. Sciences Geologiques ` 26, 339367. Gac, J.Y., 1979. Geochimie du bassin du lac Tchad. Thesis, Univ. of Strasbourg. Gleick, J., 1987. Chaos. Making A New Science. Sphere Books, London. Hanks, T.C., Buckman, R.C., Lajoie, K.R., Wallace, R.E., 1984. Modification of wave-cut and faulting-controlled landforms. Journal of Geophysical Research 89, 57715790. Harden, J.W., Taylor, E.M., 1983. A quantitative comparison of soil development in four climatic regimes. Quaternary Research 20, 342359. Heimsath, A.M., Dietrich, W.E., Nishiizumi, K., Finkel, R.C., 1997. The soil production function and landscape equilibrium. Nature 388, 358388. Hoosbeek, M.R., Bryant, R.B., 1992. Towards the quantitative modelling of pedogenesisa review. Geoderma 55, 183210. Hoosbeek, M.R., Bryant, R.B., 1994. Developing and adapting soil process submodels for use in the pedodynamic Orthod model. In: Bryant, R.B., Arnold, R.W. Eds.., Quantitative Modelling of Soil Forming Processes. SSSA Special Publication 39, Madison, WI, pp. 111128. Jahn, A., 1976. Geomorphological modelling and nature protection in Arctic and subarctic environments. Geoforum 7, 121137. Jenny, H., 1941. Factors of Soil Formation. A System of Quantitative Pedology. McGraw-Hill, New York. Jury, W.A., Gardner, W.R., Gardner, W.H., 1991. Soil Physics, 5th edn. Wiley, New York. Kirkby, M.J., 1985. A model for the evolution of regolith-mantled slopes. In: Woldenberg, M.J. Ed.., Models in Geomorphology, Allen and Unwin, Boston, pp. 213237. Kooi, H., Beaumont, C., 1996. Large scale geomorphology: classical concepts reconciled and integrated with contemporary ideas via a surface processes model. Journal of Geophysical Research 101, 33613386. Koons, P.O., 1989. The topographic evolution of collisional mountain belts: a numerical look at the Southern Alps, New Zealand. American Journal of Science 289, 10411069. Martin, Y., Church, M., 1997. Diffusion in landscape development models: on the nature of basic transport relations. Earth Surface and Processes and Landforms 22, 273279.

B. Minasny, A.B. McBratney r Geoderma 90 (1999) 321

21

McGreevy, J.P., 1985. Thermal properties as controls on rock surface temperature maxima, and possible implications for rock weathering. Earth Surface Processes and Landforms 10, 125136. McKean, J.A., Dietrich, W.E., Finkel, R.C., Southon, J.R., Caffee, M.W., 1993. Quantification of soil production and downslope creep rates from cosmogenic 10 Be accumulations on a hillslope profile. Geology 21, 343346. Mellor, A., 1985. Soil chronosequences on Neoglacial moraine ridges, Jostedalsbreen and Jotunheimen, southern Norway: a quantitative pedogenic approach. In: Richards, K.S., Arnett, R.R., Ellis, S. Eds.., Geomorphology and Soils. Allen and Unwin, Boston. Nash, D.B., 1980. Forms of bluffs degraded for different lengths of time in Emmet County, Michigan, U.S.A. Earth Surface Processes and Landforms 5, 331345. Odeh, I.O.A., Chittleborough, D.J., McBratney, A.B., 1991. Evaluation of soillandform interrelationship by canonical ordination analysis. Geoderma 49, 132. Ollier, C., 1984. Weathering, 2nd edn. Longman, London. Paces, T., 1986. Rates of weathering and erosion derived from mass balance in small drainage basins. In: Colman, S.M., Dethier, D.P. Eds.., Rates of Chemical Weathering and Rocks and Minerals. Academic Press, Orlando, pp. 531550. Pavich, M.J., 1986. Processes and rates of saprolite production and erosion on a foliated granitic rock of the Virginia Piedmont. In: Colman, S.M., Dethier, D.P. Eds.., Rates of Chemical Weathering and Rocks and Minerals. Academic Press, Orlando, pp. 552590. Phillips, J.D., 1998. On the relation between complex systems and the factorial model of soil formation with Discussion.. Geoderma 86, 142. Pillans, B., 1997. Soil development at snails pace: evidence from a 6 Ma soil chronosequence on basalt in North Queensland, Australia. Geoderma 80, 117128. Press, W.H., Flannery, B.P., Teukolsky, S.A., Vetterling, W.A., 1992. Numerical Recipes: The Art of Scientific Computing. Cambridge Univ. Press, Cambridge. Reneau, S.L., Dietrich, W.E., Rubin, M., Donahue, D.J., Jull, A.J.T., 1989. Analysis of hillslope erosion rates using dated colluvial deposits. Journal of Geology 97, 4563. Ruxton, B.P., 1968. Measures of the degree of chemical weathering of rocks. Journal of Geology 76, 518527. Scheidegger, A.E., 1991. Theoretical Geomorphology, 3rd edn. Springer-Verlag, Berlin. Smith, B.J., 1977. Rock temperature from the northwest Sahara and their implications for rock weathering. Catena 4, 4163. Trendall, A.F., 1962. The formation of apparent peneplains by a process of combined laterisation and surface wash. Zeitschrift fur Geomorphologie NF6, 183197. Trescases, J.J., 1973. Weathering and geochemical behaviour of the elements of ultramafic rocks in New Caledonia. Bureau of Mineral Resources, Geology and Geophysics, Canberra 141, 149161. Trustum, N.A., De Rose, R.C., 1988. Soil depthage relationship of landslides on deforested hillslopes, Taranaki, New Zealand. Geomorphology 1, 143160. Velbel, M.A., 1986. The mathematical basis for determining rates of geochemical and geomorphic processes in small forested watersheds by mass balance: examples and implications. In: Colman, S.M., Dethier, D.P. Eds.., Rates of Chemical Weathering and Rocks and Minerals. Academic Press, Orlando. Wakatsuki, T., Rasyidin, A., 1992. Rates of weathering and soil formation. Geoderma 52, 251263. Warke, P.A., Smith, B.J., 1998. Effects of direct and indirect heating on the validity of rock weathering simulation studies and durability tests. Geomorphology 22, 347357. Wischmeier, W.H., Smith, D.D., 1978. Predicting Rainfall Erosion LossesA Guide to Conservation Planning. Agric. Handbook No. 537, USDA, Washington, DC.

S-ar putea să vă placă și