Evaluation of the upgraded spur model (spur2.4)

https://doi.org/10.1016/S0304-3800(99)00016-2Get rights and content

Abstract

spur (Simulation of Production and Utilization of Rangelands) is a grassland ecosystem simulation model. spur2.4 integrates all previous versions of spur. The changes from spur91 make the model more applicable to the southern Great Plains. The forage submodel from spur2 predicts forage intake and diet selection by grazing herbivores, and the cow-calf submodel from spur2 simulates all individuals in a herd for their full life cycle, based on defined genetic traits. The soil organic matter submodel from century was added to improve soil carbon and nitrogen cycling. spur2.4 output is compared with observed values from the Texas Experimental Ranch, Throckmorton, Texas. The changes made in creating spur2.4 have considerably improved the utility and accuracy of the spur model for north Texas. Soil moisture predictions are improved (spur91 is 137% and spur2.4 is 106% of observed, R=0.79–0.86), but run-off is still not adequate (R=0.65–0.76) and monthly evaporation is lower (spur91 is 94% and spur2.4 is 95%) than observed. Soil carbon level predictions by the century submodel are within range of the data, and output is stable over the simulation period. Predictions of individual plant species productivity compared with observed data are improved with spur2.4 (R=0.79–0.98). Between-season plant growth and long-term persistence are simulated well. The addition of the century soil organic matter submodel increased nitrogen level predictions resulting in more accurate predictions of animal weight gain. Steer weight gains are typical for north Texas. Simulation of cow-calf mass was good (R>0.98) as was production per hectare and per cow. The addition of the cow-calf model increases the utility of the model. The paper identifies portions of the model that need further validation and field research to improve model utility and credibility for use in natural resource management.

Introduction

Models have great potential as research tools to enhance our knowledge of ecosystem function and as decision aids for natural resource managers. spur (Simulation of Production and Utilization of Rangelands) was designed to simulate rangeland ecosystem function and predict ecosystem response to changing determinants and various management practices (Wight and Skiles, 1987, Baker et al., 1993, Carlson and Thurow, 1996). It has the potential to evaluate the environmental and economic impact of different management alternatives at the landscape or whole ranch scale. The output from spur can be selected to include rainfall run-off, soil loss, soil organic matter, plant production, forage selected and harvested by livestock and wildlife, animal weight and gain, and estimates of net economic return. Once spur is calibrated for a particular location, the model can be run to predict the long-term outcome of management strategies and weather sequences and to assess best management strategies or combinations of management practices (Baker et al., 1993).

Models must provide reliable results when measured against field data to be useful as a decision aid for resource managers. Modifications to improve the capability and accuracy of predictions of spur were made separately by Carlson and Thurow (1992) (spur91), Hanson et al. (1992) (spur2) and Foy (1993) (spur2.3). These improvements to spur have been combined to produce spur2.4. This paper compares the output from spur2.4 with data from the Texas Experimental Ranch (TER), Throckmorton, Texas to determine how well we can currently predict the outcome of climate, physical environment and different management, in this environment. Results on hydrology, soil organic carbon, herbaceous vegetation and cow-calf production are presented. Since spur91 has provided superior results when evaluated in the southern Great Plains against the original spur (Carlson and Thurow, 1996), differences between spur2.4 and spur91 are emphasized. spur2.4 was not tested against spur2 and spur2.3 since neither of these models include the changes to improve plant–water interactions made by Carlson and Thurow (1996) in spur91.

Section snippets

Model description

spur is composed of six basic submodels (Fig. 1) (see Carlson and Thurow, 1992, Hanson et al., 1992). The climate record provides daily inputs of precipitation, maximum and minimum temperature, solar radiation, and wind run. The hydrology component maintains daily water balance, calculates snow accumulation, snowmelt and sediment transport. The soil module tracks soil moisture by soil layer according to soil series characteristics and soil carbon and nitrogen levels. The plant module tracks

Hydrology

Monthly run-off was not adequately predicted by both models (R=0.66–0.76)(Table 1). Monthly wet year run-off (1986) was more highly correlated (R=0.76 in both models) than monthly dry year run-off (1988) (R=0.66 in both models). This is an inherent problem with the curve number technique used in the model (Carlson and Thurow, 1996). Total simulated run-off for 3 years was twice the observed run-off (Table 1). However, the amount of run-off as a percentage of precipitation in this area is <8% (

Conclusions

The changes made to spur in creating first spur91 and then spur2.4 have improved the utility and accuracy of the model considerably. The model is now able to do more than just predict general trends of management responses. The potential for aiding in the assessment of various management strategies and practices is now possible in limited areas, but much work remains to be done to expand this capability. The hydrology component, in particular, is not nearly accurate enough and this is one of

Future use of spur in natural resource management

spur in its upgraded form can evaluate climatic change consequences, stocking rate decisions, timed rotational grazing systems, enterprise-level economic consequences and the management impacts on water harvest, erosion, vegetative quality and quantity, and soil carbon and nitrogen dynamics. The impact of treatments can be simulated for 100 years or more. With this model, rangeland ecosystem and enterprise sustainability can be addressed at the ranch scale by assessing cumulative effects over

Acknowledgements

We would like to thank Mike Coughenour for providing the initial FORTRAN code for the century soil organic matter submodel. Bill Bailey and Jon Hanson facilitated transfer of the spur2.3 code from ARS so that the development of this model could proceed. Rick Bourdon assisted in initializing the cow-calf model. Steve Dowhower and Bill Pinchak provided data from the Texas Experimental Ranch. Dee Carlson and Tom Thurow kindly reviewed an earlier draft.

References (30)

  • Carlson, D.H., Thurow, T.L., 1992. spur-91: Workbook and user guide. Pub. No. MP-1743. Texas A&M University, Department...
  • D.H. Carlson et al.

    spur91: Simulation of Production and Utilization of Rangelands

  • D.H. Carlson et al.

    Effect of honey mesquite on the water balance of Texas Rolling Plains rangeland

    J. Range Manag.

    (1990)
  • Field, L.B., 1987. Simulation of beef-heifer production on rangeland. Thesis, Colorado State University, Fort Collins,...
  • Foy, J.K., 1993. Soil organic matter changes as a function of stocking rate and grazing system under different climate...
  • Cited by (37)

    • Application of grazing land models in ecosystem management: Current status and next frontiers

      2019, Advances in Agronomy
      Citation Excerpt :

      Interactions among model parameters were also an issue in model calibration (Ma et al., 2015). When a model is not adequately calibrated with a dataset, users should look for better modules to improve certain processes, rather than using unrealistic model parameters to match experimental observations (Foy et al., 1999). For some model inputs, such as weather data and soil properties, there is considerable temporal and spatial variability in measurements.

    • Process-based simulation of prairie growth

      2017, Ecological Modelling
      Citation Excerpt :

      Even if APEX included this feature, it would likely remain much more difficult to accurately simulate the individuals within a suite of noncultivated species growing in competition with one another, as compared to cultivated crops that have little genetic diversity and are grown in near monocultures. The challenge of simulating a suite of rangeland species was also noted in successive versions of the SPUR model; on several occasions, vegetative production of certain individual species was accurate while simultaneous simulation of particular competing species remained difficult (Carlson and Thurow, 1996; Foy et al., 1999; Pierson et al., 2001). Matching simulated and observed values in rangeland may be complicated by sampling error, which may be relatively high for individual rangeland species due to their uneven distribution, in contrast to monoculture crops.

    • Modelling rangeland productivity in response to degradation in a semi-arid climate

      2016, Ecological Modelling
      Citation Excerpt :

      Species composition change is not accounted for. The SPUR2.4 rangeland model (Foy et al., 1999) simulates grassland through 4 plant functional type (PFT) groups, calling them species, but seems to be capable of greater detail if enough parameters were to be available. Relative abundances of the PFT groups can vary freely.

    • Modeling the impact of rangeland management on forage production of sagebrush species in arid and semi-arid regions of Iran

      2013, Ecological Modelling
      Citation Excerpt :

      Different models have been applied in the literature to study processes of rangeland vegetation and forage production. Some examples of the process-based models are: GRASIM (Mohtar et al., 1997, 2000), which is a pasture and grassland model, but it cannot simulate watershed level, water and sediment routing; SPUR (Simulation of Production and Utilization of Rangeland) (Foy et al., 1999; Corson et al., 2006), which simulates plant growth, water balance, nutrient cycling, and grazing, but it cannot simulate different plant communities (multi-stories) and different growing conditions (Chen, 2000); and GRAZE (Parsch and Loewer, 1995), which is a beef forage grazing model used to estimate animal intake and growth, however, it lacks nutrient and soil components for the study of grazing impact on environment. Other relevant models are IFSM (Integrated Farm System Model) (Rotz et al., 2012), EPIC (Erosion Productivity Impact Calculator) (Williams, 1995), SVAT (Soil Vegetation Atmosphere Transfer) (Mo et al., 2005), and GEPIC (GIS-based EPIC) (Liu et al., 2007).

    View all citing articles on Scopus
    View full text