Quantification and simulation of surface runoff from fescue grassland watersheds

https://doi.org/10.1016/S0378-3774(02)00124-5Get rights and content

Abstract

The topographic features of the foothills fescue grasslands in southern Alberta predispose them to runoff and soil loss via erosion. A study was conducted at Stavely Research Station, Alberta to determine the runoff from small grassland watersheds under three grazing intensities, viz. ungrazed (or control), heavy (2.4 animal unit months per hectare (AUM ha−1)) and very heavy (4.8 AUM ha−1) grazing. Total annual precipitation in 1998, 1999 and 2000 was 648, 399 and 263 mm, respectively. Surface runoff hydrographs indicated large summer storm runoff rates from heavy grazed compared to other watersheds, but large snow melt-induced runoff from very heavy grazed compared to other watersheds. Surface runoff rates measured from May and August ranged between 0 and 2.3 mm per day in 1998, 0–0.2 mm per day in 1999, and 0–0.07 mm per day in 2000. In all the years, the average rainfall runoff was <10% of average daily precipitation on all three watersheds. In 2000, snow melt-induced runoff was measured in March. Total surface runoff for this month was 0.07, 8.5 and 3.7 mm for ungrazed, heavy and very heavy watersheds, respectively. These accounted for 78, 96 and 92% of total annual runoff from ungrazed, heavy and very heavy watersheds, respectively. Surface runoff for 1999 and 2000 was simulated using Soil Water Assessment Tool (SWAT), a continuous time distributed parameter model developed for ungaged basins. Model calibration was conducted using data of 1998 and parameters adjusted until the predicted and observed results were visibly close. Evaluation of the model was conducted using statistical criteria that included calculations of average error (AE), residual mean square (RMS), coefficient of residual mass (CRM) and modeling efficiency (EF), and comparing these statistics against optimal values. The evaluation indicated that the model under-predicted surface runoff from the watersheds in both the years.

Introduction

Hydrologic research has traditionally focused across a broad range of climate zones. Hydrologic research on arid and semi-arid zones has not focused on the Canadian prairies, and as a result little information is available. Quantification of runoff, sediment yield and non-point pollution on an agricultural watershed are required in order to evaluate the effects of agricultural management systems on surface water quality. Studies in Alberta’s foothills area revealed that the majority of annual runoff occurred during snow melt and that few summer storms caused runoff (Naeth et al., 1991).

Impacts of grazing on hydrology result from alteration of plant species composition, density, ground cover and litter. Livestock trampling may result in soil compaction (Naeth et al., 1990, Mapfumo et al., 1999), increased runoff and change in soil water regimes (Twerdoff et al., 1999). A previous study conducted on foothills fescue grasslands indicated low values of runoff, sediment concentration, sediment yield and runoff coefficients under ungrazed compared with heavy or very heavy grazing intensities (Naeth and Chanasyk, 1996). However, in that study small plot areas (1 m2) were used for runoff data collection. Verification of these conclusions can only be possible from watershed-scale studies.

As hydrologic processes vary in space and time, the use of hydrologic models becomes essential to determine the best management practices that minimize adverse impacts on watersheds. The basic requirement of any watershed model is the capability to estimate surface runoff. The runoff process determines the transport of sediments and agricultural non-point source (NPS) pollutants such as pesticides, nitrates and phosphates (Manguerra and Engel, 1998). A distributed parameter modeling approach incorporates data concerning the areal distribution of parameter variations together with computational algorithms to evaluate the influence of this distribution on simulated behavior (Huggins and Burney, 1982). The Soil Water Assessment Tool (SWAT) is one such distributed parameter hydrologic model that provides opportunities to improve watershed modeling accuracy and better long-term prediction of hydrologic components such as surface runoff, streamflow and evapotranspiration (Arnold et al., 1998). It has been widely applied in various scenarios and watersheds, including simulating climate change effects in an agricultural watershed (Hanratty and Stefan, 1998).

The objective of this study was to quantify rainfall-induced surface runoff and runoff coefficients from foothills fescue watersheds under three levels of grazing intensities (ungrazed, heavy and very heavy). A further objective was to calibrate the SWAT model and evaluate its ability in predicting surface runoff from these watersheds.

Section snippets

Model description

The Soil Water Assessment Tool is a river basin scale, continuous time model that operates on a daily time step. Using a routing command language the model can simulate a basin subdivided into grid cells or sub-basins. The sub-basin components of SWAT include hydrology, weather, sediment yield, nutrients, pesticides, soil temperature, crop growth, tillage and residue and agricultural management practices (Arnold et al., 1995). The hydrology component is based on the water balance equation,

Model input

The model input data required consisted of climate, soils, land use and vegetation. Precipitation and weather data of daily average, maximum and minimum temperatures, solar radiation were obtained from the nearby Stavely Weather Station. Default values of overland flow N, channel N and USLE erosion control practice factor generated by SWAT were used in running the model. Antecedent soil water was measured using a neutron probe throughout the 2 years of study. Results of this study can be found

Study site, meteorological conditions and hydrologic measurements

The study was conducted at the Agriculture Canada Stavely Range Substation, in the fescue grasslands in the Porcupine Hills of southwestern Alberta, approximately 100 km south of Calgary, Canada. The site was fenced to create paddocks in 1949. This site is characterized by hilly topography with slopes ranging from 18 to 37%. Annual average precipitation is 550 mm with 40% of it occurring as snow. The most prevalent soils on the site are Typic Haplustolls of loam to clay loam texture. The plant

Surface runoff depths and coefficients

Total annual precipitation amounts were 648, 399 and 263 mm, in 1998, 1999 and 2000, respectively. In all the 3 years, monthly precipitation amounts were greatest in June and more than 50% of total annual precipitation occurred in the months of June–August (Table 1).

The total number of days between May and August, inclusive with surface runoff were 27, 23 and 16 for 1998, 1999 and 2000, respectively (Table 2). In general, surface runoff amounts were very small over the entire study period,

Conclusions

Rainfall-induced runoff was a very small fraction of total rainfall. Runoff volume and peak discharge were generally higher for the heavy than the very heavy treatment. These parameters for the ungrazed treatment were very low, unexpectedly so for snow melt. Simulation of surface runoff by the SWAT model under-predicted the amount of surface runoff in 1999 and 2000, after being calibrated in 1998. This was evident in the statistics calculated as well as the overall fitting of the model to the

Acknowledgements

We gratefully acknowledge funding received from the Alberta Agricultural Research Institute (AARI) and the Natural Sciences and Engineering Research Council (NSERC). Thanks to Brian Henderek and Kelly Ostermann for assistance with field work and data collection.

References (27)

  • St Elmaloglou et al.

    Simulation of soil moisture content of a prairie field with SWAP93

    Agric. Water Manage

    (2000)
  • K Loague et al.

    Statistical and graphical methods for evaluating solute transport models: overview and application

    J. Contam. Hydrol.

    (1991)
  • J.E Nash et al.

    River flow forecasting through conceptual models. Part 1. A discussion of principles

    J. Hydrol.

    (1970)
  • V.Z Antonopoulos et al.

    Modeling of water and nitrogen dynamics on an undisturbed soil and a restored soil after open-cast mining

    Agric. Water Manage

    (1998)
  • Arnold, J.G., Williams, J.R., Srinivasan, R., King, K.W., 1995. Soil water assessment tool manual. USDA-ARS. 94...
  • J.G Arnold et al.

    Large area hydrologic modeling and assessment. Part I: Model development

    J. Am. Water Resource Assoc.

    (1998)
  • D.E Carlson et al.

    Effect of honey mesquite on the water balance of Texas rolling plains rangeland

    J. Range Manage

    (1990)
  • Chanasyk, D.S., Mapfumo, E., Willms, W.D., Naeth, M.A., 2002. Quantification and simulation of soil water on grazed...
  • S.I Gill et al.

    Runoff and sediment yield from snow melt and rainfall as influenced by forage type and grazing intensity

    Can. J. Soil Sci.

    (1998)
  • M.P Hanratty et al.

    Simulating climate change effects in a Minnesota Agricultural Watershed

    J. Environ. Qual.

    (1998)
  • T.E Harms et al.

    Runoff response from two reclaimed watersheds

    J. Am. Water Resource Assoc.

    (1998)
  • G.H Hargreaves et al.

    Reference crop evapotranspiration from temperature

    Appl. Eng. Agric.

    (1985)
  • Huggins, L.F., Burney, J.R., 1982. Surface runoff, storage and routing. In: Haan, C.T., Johnson, H.P., Brakensiek, D.L....
  • Cited by (0)

    View full text