Understanding soil erosion and landscape evolution using computer based predictive models

Greg Hancock
The University of Newcastle, Australia
Author Profile

Shortcut URL: https://serc.carleton.edu/36192


Continent: global
UTM coordinates and datum: none


Climate Setting:
Tectonic setting:
Type: Process, Computation


The ability to measure and model soil erosion and resultant landscape evolution is important because soil erosion has a range of environmental impacts, including loss of organic matter and nutrients, reduction of landscape productivity and a reduction in downstream water quality. It is also necessary to understand how hillslope and catchment shape evolves through time (Figures 1 and 2) (Hancock et al., 2002). Soil erosion and landscape evolution models are also especially pertinent for landform design and the rehabilitation of mine sites as it is important for engineers to establish how resilient a proposed rehabilitated landform sculpted over the existing mine landform will be (Figure 1).

Numerous soil erosion models such as the Water Erosion Prediction Program (WEPP) (Laflen et al., 1991), Universal Soil Loss Equation (USLE), Modified Universal Soil Loss Equation (MUSLE) and the Revised Universal Soil Loss Equation (RUSLE) (Onstad and Foster, 1975; Wischmeier and Smith, 1978) have been developed, each with their own individual strengths and weaknesses. The models have been applied across a broad range of landscape types with varying degrees of success.

More recently, computer landscape evolution and soil erosion models that use digital elevation models to represent the landscape surface have been developed (Willgoose et al., 1991a,b; Braun and Sambridge, 1997; Tucker et al., 2001) with some models being applied to assessment of degraded and mining landscapes (Willgoose and Riley, 1998; Coulthard and Macklin, 2003). These models have considerable advantages over traditional modelling approaches, such as the RUSLE and WEPP, as they remove the need to manually determine slope length and angle. These models can also determine both erosion and deposition, something not possible with the RUSLE. A further advantage of using digital elevation based models is that they dynamically adjust the landscape to erosion and deposition, producing a better representation of slope and angle over the duration of the simulation.

In general, most models link hydrology and erosion sub-models. They also include a creep or diffusive erosion term. These mathematical equations route water and sediment through a digital elevation model using the D8 algorithm (water and sediment routed downslope in the steepest downslope direction) or the DInfinity algorithm (where water and sediment routed downslope in the proportionally). Models like SIBERIA (Willgoose et al., 1991a,b) use average annual hydrology and erosion parameters and produce annual output as a result while the CAESAR (Coulthard et al., 2002) model inputs hourly rainfall and produces hourly hydrology and sediment output (Figure 3).

These models all require some form of data input or calibration. For example, before the SIBERIA model can be employed the fluvial sediment transport equation and area-discharge relationship require calibration. The fluvial sediment transport equation is most reliably parameterised using input from field sediment transport and hydrology data (Evans et al., 1999, 2000; Hancock et al., 2000). While there are many factors, such as organic matter content, electrical conductivity, mineralogy and rock content that contribute to soil erodibility, one of the strongest influences on erodibility is soil particle size distribution. Where there was an absence of field data soil textural information such as that of Sheridan et al., (2000) provide a database. Calibration of the CAESAR model is considerable simpler than other models and requires a DEM of the study site or catchment, rainfall data (mm hr-1) and soil/sediment particle distribution data which can be input as 9 individual ranges.

In terms of ease of application all models have their strengths and weaknesses. The calibration of SIBERIA can be time consuming if using field hydrology and sediment transport data to derive model parameters in advance of using the model. The calibration of CAESAR is much simpler with only hourly rainfall data and soil particle size data required as input parameters. The difficulty with this is the point specific nature of the soil particle size data that can be used and the availability of hourly rainfall data. Soil particle size data varies considerably on the hillslope and catchment scale (both laterally and with depth) and we have demonstrated here that it can affect sediment transport rates. Therefore, it may be wise to evaluate the impacts of having areas with different grain size characteristics on the model performance. Long-term hourly rainfall data may not be available for study catchments especially in remote locations.

The SIBERIA model, when calibrated for use with average annual data is considerably quicker in terms of run time. For example, simulating a catchment with 5000 nodes takes approximately 1 hour per 1000 years for SIBERIA (Figure 2) while CAESAR takes approximately 100 hours for 1000 years (Figure 3) (with both models run using a computer with an AMD 2.4GHz processor with 2GB RAM). The reason for the time difference is that SIBERIA uses average annual hydrology and sediment transport data while CAESAR uses hourly rainfall data and has a more sophisticated representation of hydraulics and sediment transport which requires considerably more calculations. CAESAR also calculates the erosion and deposition of each soil size fraction. Consequently if average annual erosion rates are required then SIBERIA can provide the required output while if sediment transport rates down to the scale of individual storm events is required then CAESAR can provide this data on an hourly output.

Associated References

  • Braun J, Sambridge M. 1997. Modelling Landscape Evolution on Geological Time Scales: A New Method Based on Irregular Spatial Discetization, Basin Research, 9: 27-52.
  • Coulthard TJ. 2001. Landscape evolution models: a software review, Hydrological Processes, 15:165-173.
  • Coulthard TJ, Macklin MG. 2003. Modeling long-term contamination in river systems from historical metal mining, Geology, 31, no. 5, 451-454
  • Coulthard TJ, Macklin MG, Kirkby MJ. 2002. Simulating upland river catchment and alluvial fan evolution. Earth Surface Processes and Landforms. 27, 269-288.
  • Coulthard TJ, Van De Wiel MJ. 2006.A cellular model of river meandering, Earth Surface Processes and Landforms, 31, 123-132
  • Evans KG, Willgoose GR, Saynor MJ, Riley SJ. 2000. Post-mining landform evolution modelling. I. Derivation of sediment transport model and rainfall-runoff model parameters. Earth Surface Processes and Landforms. 25(7): 743-763.
  • Evans KG, Willgoose GR. 2000. Post-mining landform evolution modelling: 2. Effects of vegetation and surface ripping. Earth Surface Processes and Landforms, 25(8): 803-823.
  • Flanagan DC, Livingston SJ. 1995. Water Erosion Prediction Project (WEPP) Version 95.7 User Summary. In: (Flanagan DC, Livingston SJ) 'WEPP user summary', NSERL Report No 11, July 1995.
  • Hancock GR, Willgoose GR, Evans KG, Moliere DR, Saynor MJ. 2000. Medium term erosion simulation of a abandoned mine site using the SIBERIA landscape evolution model, Australian Journal of Soil Research, 38:249-263.
  • Hancock GR, Willgoose GR, Evans KG. 2002. Testing of the SIBERIA landscape evolution model using the Tin Camp Creek, Northern Territory, Australia, field catchment, Earth Surface Processes and Landforms, 27(2): 125-143.
  • Hancock GR, Crawter D, Fityus SG, Chandler J, Wells T. 2008. The measurement and modelling of rill erosion at angle of repose slopes in mine spoil, Earth Surface Processes and Landforms, 33, 1006-1020, DOI: 10.1002/esp.1585.
  • Hancock GR, Loughran RJ, Evans KG, Balog R. 2008. Estimation of soil erosion using field and modelling approaches in an undisturbed Arnhem Land catchment, Northern Territory, Australia, Geographical Research, in press.
  • Hancock GR, Lowry JBC, Coulthard TJ, Evans KG, Moliere DR. 2009. A catchment scale evaluation of the SIBERIA and CAESAR landscape evolution models, Earth Surface Processes and Landforms, DOI: 10.1002/esp.1863.
  • Laflen JM, Elliot WJ, Simanton J R, Holzhey CS, Kohl KD. 1991. WEPP soil erodibility experiments for rangeland and cropland soils. Journal of Soil and Water Conservation 39-34.
  • Onstad CA, Foster GR. 1975. Erosion modelling on a watershed. Transactions of the American Society of Agricultural Engineers, 26: 1102-1104.
  • Renard KG, Laflen JM, Foster GR, McCool DK. 1994. The revised universal soil loss equation. In Lal R, Soil Erosion Research Methods, Second Edition. Soil and Water Conservation Society, Ankeny IA.
  • Sheridan GJ, So HB, Loch RJ, Pocknee C, Walker CM. 2000. Use of laboratory scale rill and interrill erodibility measurements for the prediction of hillslope–scale erosion on rehabilitated coal mines soils and overburden. Australian Journal of Soil Research 38(2): 285 – 298.
  • Tucker G, Lancaster S, Gasparini N, Bras R. 2001. The Channel-Hillslope Integrated Landscape Development model (CHILD), Landscape erosion and evolution modeling, New York, NY, United States, Kluwer Academic/Plenum Publishers.
  • Willgoose GR, Bras RL, Rodriguez-Iturbe I. 1991a. A physically based coupled network growth and hillslope evolution model: 1 Theory, Water Resources Research, 27(7):1671-1684.
  • Willgoose GR, Bras RL, Rodriguez-Iturbe I. 1991b. A physically based coupled network growth and hillslope evolution model: 2 Applications, Water Resources Research, 27(7):1685- 1696.
  • Wischmeier WH, Smith DD. 1978. Predicting rainfall erosion losses-A guide to Conservation Planning. US Department of Agriculture, Agriculture Handbook No. 537.2/