Ontario - Canada Logos
SWEEP Banner

SWEEP Report #75

Watershed Modelling
(PWS Report #7)

Beak Consultants Ltd., Guelph, Ont.

View/Download Report  [3490 KB pdf]  (with Appendices A, C)



List By Number | List By Sub-Program

Completed: September, 1994

Executive Summary


The objectives of the modelling component of the Pilot Watershed Study (PWS) were as follows:

  • develop spatially descriptive erosion and runoff models of the study sub-watersheds;

  • apply the site specific models to the evaluation of future soil and environmental benefits relating to various agricultural land management practices; and

  • use the models as tools to aid in the assessment of the erosion and phosphorus runoff processes and the effects of land management upon the runoff processes.

Watershed Models

Two different watershed models were used in the PWS. The Guelph model for evaluating efforts of Agricultural Management systems on Erosion and Sedimentation (GAMES) with the Phosphorus component (GAMESP) was set up and calibrated for the purpose of acting as a planning and communication aid. In fact, a modified version of GAMES was taken to farm level planning meetings to assist in illustrating to farm managers the long term benefit of various localized and field level changes to farming practice. GAMES also was used to highlight problematic areas in the sub-watersheds, that is, where erosion prone factors combined to increase the potential for high erosion. Further, GAMES was applied to estimating losses from hypothetical land management systems across entire sub-watersheds. These predictions represent potential long term benefit, in terms of soil conservation and improved runoff water quality, in the long term.

A second more complex model was set up and applied to a different set of tasks. The Hydrological Simulation Program-Fortran (HSPF) is a dynamic predictive model with runoff and erosion prediction capabilities as well as instream flow and material transport and transformation capabilities. This model predicts and accounts for the complete hydrological cycle and various storage compartments within the watershed (i.e., groundwater, instream, interception etc.). This model was applied to aiding in the assessment of watershed processes which affect runoff and erosion. In effect, this model was used to provide estimates of rates and fluxes within the sub-watersheds for processes which can be easily measured as well as for those that cannot. HSPF was rigorously calibrated using the continuous runoff and loading information documented in PWS Reports #5, Hydrology (SWEEP Report #73) and #6, Water Quality (SWEEP Report #74).

Both models have excellent documentation (Cook et al. 1985, Johanson, 1980). Some of the more relevant components of the models are described in more detail along with model input requirements, data sources and the set up process.

Model Setup

In terms of setup, both models require that land areas and their characteristics be represented as model input. Both models use a system of irregular polygons with homogenous attributes. That is, each polygon has one characteristic associated with soil type, land use and slope and differs from its neighbouring polygons in at least one of the attributes. Polygons were determined using a SPANS Geographical Information System and digital information on sub-watershed soils, land use and slopes. Other model input information was determined from PWS surveys and monitoring programs.

Model Testing

HSPF was calibrated using PWS sub-watershed outflow and loading estimates for 1989. Verification was achieved by comparing predicted and observed results through 1990.

GAMESP was calibrated using 1989 to 1991 PWS soil and phosphorus losses at the sub-watershed mouths.

HSPF was shown to be capable of accurately predicting outflow rates, soil loss and total phosphorus loss on a sub-watershed scale over monthly and annual periods. Daily total flows and loads were generally reliable although snowmelt periods were not modelled accurately in the time frame of days. The model was able to respond to very short term changes in moisture conditions in the sub-watersheds which could result in increases in runoff of several orders of magnitude. The small headwater PWS study areas are very responsive to rainfall or snowmelt events. In general, annual flows and loads were predicted to within 5% of the observed flows and loads while daily predicted values were accurate to within 50%.

GAMESP is calibrated by adjusting two parameters which affect soil delivery until correspondence is achieved at the sub-watershed mouth. A similar process is followed for phosphorus. Actual erosion is determined using a derivation of the Universal Soil Loss Equation and is based upon long term conditions for rainfall. Soil delivery to open watershed channels is based upon overland flow path hydraulic parameters which relate to the transport capacity of overland flow.

Model Applications

HSPF and GAMESP inputs were adjusted to reflect two extreme land management schemes. A continuous fallow system was chosen to represent the worst case wherein no soil cover protection is provided throughout the year. A no-till case was used to reflect a best case in which all tilled fields would be managed for high crop residue levels.

GAMESP predicted that fallow conditions would result in up to ten times more soil delivery to the Essex sub-watershed mouths depending upon the simulation year rainfall, over the existing conditions (1989 to 1991 were used).

No-till conditions were predicted to result in a reduction in soil loss of up to 85%, with the highest value in the wettest year.

A similar range of relative increases and decreases were predicted for Kettle Creek and Pittock.

GAMESP predicts the changes in soil delivery are primarily due to increased crop cover and reduced overland delivery.

HSPF predictions were somewhat less optimistic. HSPF predicted affects of land management upon runoff and water balance as well as soil and phosphorus losses. Land management practice adjustments had an overall affect on runoff due to modified land surface conditions. These changes in runoff over a two-year period ranged up to 19%. Soil loss increases due to fallow conditions were predicted to range up to 87% or about double from the existing while no-till was predicted to save 60% of the soil at best. Phosphorus losses under fallow and no-till were predicted to be slightly reduced from the soil loss figures since phosphorus pathways include subsurface and other soluble and non adsorbed forms. HSPF predicts that phosphorus transport is not as controllable as soil loss.


Conclusions and Recommendations

The following conclusions have been presented from the modelling activities undertaken during the PWS:

  • GAMESP is an effective screening level soil and phosphorus management tool. It lends itself to interactive planning at the farm field level up to the small watershed scale. The model is easy to apply and interpret using readily available information.

  • HSPF is suited to assessment as it clearly links cause and effect at the primary process level. Provided with detailed input regarding watershed characteristics, management practices and watershed runoff, the model can assist researchers and planners to identify the effects of conservation tillage in the short and long term through runoff and erosion process alteration. HSPF provides real time estimates of runoff as well as erosion and can simulate subsurface as well as surface transport processes for phosphorus. In this regard, HSPF is more complete than GAMES and can be used to more fully assess and plan farm management at field and watershed scales.

  • Both models predict that significant benefit can be realized through the implementation of farm conservation management systems. The GAMES predictions are somewhat optimistic and do not account for the complex interaction of processes governing runoff. The HSPF predictions indicate that factors of two to five may be relevant between worst case and best case erosion and phosphorus delivery rates. These benefits are significant in light of water quality targets for the Great Lakes and on-farm conservation expectations.

The following recommendations have been made:

  • Screening level models such as GAMESP should be an integral part of the farm level planning process to achieve full implementation of conservation farming.

  • Detailed deterministic models such as HSPF should be used to improve the estimation of conservation system benefits in the short and long term and as an aid to researchers in accounting for the complex system of processes governing runoff and erosion.

  • More research is required to define the cause and effect relationships involved in erosion and runoff from farm fields. The currently available models are only as accurate in predicting benefit as the accuracy of input parameter adjustments. At present, several processes are poorly quantified in this regard.

  • Pilot watershed studies must be continued over longer periods of time to allow for effective adoption followed by acclimatization. The PWS had achieved a reasonable level of adoption after five years. However the program did not allow for a new equilibrium to be achieved with respect to soil structure and quality, farmer practices, and crop performance.




List By Number | List By Sub-Program | LSP Report List


Created: 05-28-1996
Last Revised: Thursday, May 19, 2011 07:49:13 PM