Abstract:
Proc. Aust. Soc. Anim. Prod. (1972) 9: 16 A PRODUCTIVITY MODEL OF IMPROVED PASTURE GRAZED BY MERINO SHEEP P. J. VICKERY* and D. A. HEDGES* Summary A mathematical model which simulates the productivity of a Phalaris tuberosa/ white clover (Trifolium repens) pasture grazed with Merino sheep is described. Rainfall, pan evaporation and soil temperature recorded at Armidale, N.S.W., in 1967 and 1968 were used to predict herbage growth, liveweight changes and wool growth of wethers grazing at densities of 10, 20 and 30 animals per ha. These predictions were compared with experimental values obtained at Armidale in the same years, and the outputs from the model were in agreement with the productivities at the three stocking densities. I. INTRODUCTION Allden (1968) has drawn attention to the numerous stocking rate/sheep production experiments in which the production of several management treatments has been compared at a number of stocking densities. However, because of the cost of these experiments, it has not always been possible to fully examine the management/stocking density response. Thus when the results are used to give economic advice, it has been necessary to interpolate within experiments and to extrapolate beyond them, which are often uncertain and inaccurate procedures. I Arcus (1963) has suggested the use of a dynamic simulation model of the grazing management system, as an alternative basis for such economic analysis. The model described here was produced as an initial step in the development of this approach; it utilizes information gained from a number of grazing experiments by members `of the Pastoral Research Laboratory, Armidale. II. y THE MODEL An organizational flow chart of the model is given in Figure 1, and specific details of the predictive relationships are given in a memorandum available from the Librarian of the Pastoral Research Laboratory. The abbreviations. used in the following text are the same as those in Figure 1. Operation of the model was on the basis of weekly productivity and weekly rates of movement of material between & the pools. The exogenous variable inputs to the model were weekly totals of rainfall (P) and standard pan evaporation, and weekly mean soil temperature (T) from 3-hourly + readings taken 4 cm below a grassed area. Within the model actual evaporation (E) was calculated from pan evaporation using a site specific relationship with the current soil moisture status. The change in soil moisture (SM) for a defined effective root zone was then computed from actual evaporation and rainfall. The relative growth rate (RGR) of the Ph-ahris tuberosa/TrifoZium repens pasture (above and *CSIRO, Division of Animal Physiology, Pastoral Research Laboratory, Armidale, New South Wales, 2350. below ground) was derived from soil moisture, with a maximum RGR at 75 mm, and from soil temperature, with a maximum response at 22OC. Both pasture age, as a function of stocking density and time of year, and herbage pool sizes (> 2500 kg/ha) associated with canopy closure, further modified RGR. Estimated weekly net primary production (NPP) was calculated from RGR and the green herbage pool (GH) plus a proportion of root pool (R) that was assumed to be growing actively. This growth was divided between tops and roots (TR), with the proportion a function of season and stocking density. Losses from the green herbage pool, in addition to sheep intake (GI) and insect consumption (IC), occurred by a seasonal death rate (DR) to a dead material pool (DL). Losses from the dead material pool occurred by a seasonallitter disappearance rate due to the decomposers (DE), and by consumption due to sheep (DI), and the detritus feeders (DC). The root pool lost material through consumption by root-feeding insects (SC) and a seasonal disappearance rate due to the decomposers (DE'). Dry matter intake (TI) of the sheep was predicted from fleece-free liveweight and pasture digestibility, and it was adjusted for pasture availability and partitioned into green and dead components using relationships calculated at this Laboratory by Hamilton (personal communication). Twelve subpools (1 each for ages 1 to 12 months) were maintained for both green and dead herbage and these allowed for selection by the sheep for the younger and more digestible components of the pasture. Dry matter intake was converted into metabolizable energy intake, assuming that organic matter was 90 per cent of dry matter, the energy content of digestible organic matter was 19.25 MJ/kg, and that the metabolizability of digestible energy was 80 per cent. Using the metabolizable energy requirements for maintenance as determined by Young and Corbett (197 l), an energy balance was calculated and used to estimate liveweight change (LWC) assuming: (i) 80 per cent efficiency of utilization of catabolized energy (Marston 1948) or (ii) that the efficiency of conversion of surplus energy to body reserves was a function of the digestibility of the diet (Blaxter 1967). The energy content of liveweight was estimated from relationships based on the data of Langlands and Sutherland (1969) and Searle and Graham (1970), and wool growth (WG) was a function of the energy balance of the sheep. The energy losses of the above transactions were excreta (EX) and heat (H). Mortality (M) sin the sheep flock was a function of liveweight, and when mortalities occurred the number of sheep which died were replaced with new sheep (A) of the same liveweight. However, when 100 per cent mortality occurred in a particular week (liveweight < 25 kg), the replacement liveweight was 40 kg. III. EXPERIMENTAL DATA The model was run for Merino wethers grazing at 9.9, 19.8 and 29.7 animals per ha, using appropriate meterological data recorded at this Laboratory in 1967 and 1968. The output was compared with data obtained by Hutchinson (1969) for Merino wethers grazing P. tuber-ma/T. repens pastures at the same stocking densities in 1967 and 1968. At each stocking density there were 15 observations of the green and dead herbage pools, with each observation the mean of 16 measurements. For liveweight and fleece growth rate there were 16 observations and these were means of measurements on 8, 16 and 24 sheep respectively for the three stocking densities. These data were not used in the development of the model, but because of the absence of precise information about the effect of stocking density on pasture growth, some adjustments of the pasture growth and mortality parameters were necessary to obtain agreement between observed and predicted values of all stocking densities. IV. RESULTS The agreement between predicted and observed green and dead herbage pool sizes at the low- and high stocking densities is shown in Figure 2. The regression equations of observed versus predicted values were significant at P < 0.05, for both green and dead herbage (Table 1). A similar level of agreement was obtained for TABLE 1 fleece free liveweight and fleece growth rate for the high and low stocking densities (Figure 2). Here the statistical parameters of the regression equations indicated that the predictions from the model were reasonable. However, the agreement was best at the two higher stocking densities for liveweight, while agreement was best at the middle stocking density for wool growth rate. V. DISCUSSION Considering the complexity of the model and the fact that some of the relationships used were based on incomplete data, agreement between observed data and the model predictions was reasonable for all stocking densities. The agreement at the three stocking densities was not consistent, for all model outputs, and it indicated a lack of detailed knowledge on the stocking density effects in the model. This was particularly important in estimating plant growth, and supports the con19 20 tention of Willoughby (1970) that growth performance data for grazed plants are scarce in the literature. The use of growth data from non-grazed plants was examined, but was abandoned because many of the rates differed widely from those which occur under grazed conditions. More detailed animal and plant data from grazing experiments with a wide range of stocking densities could improve the precision in this part of the model. However, the priority for such experiments depends on the intended use and subsequent development of the model, and it might be more desirable to expand the model to cover the productivity of a reproducing flock, as this would allow the testing of management strategies in economic analyses of livestock enterprizes. Development of a reproductive flock model is now in progress, but a major problem has been the lack of suitable equations to describe intake and energy utilization as the sheep progress from one state to another. To provide a continuum of growth, production or mortality requires experimental data on the transitional periods, and these do not appear to be readily available for the grazing sheep. Detailed sensitivity analyses have not been attempted with the model, but manipulations by the authors during the process of model development have shown several areas of high sensitivity. Relative growth rate was found to be a suitable means of relating growth to environmental factors, but it did not produce a stable model unless allowance was made for the `active' root biomass in the calculation of growth. Because pasture roots could have a variable proportion of their biomass in this `active' form, information on such variation should improve the model. In addition, a small change in the digestibility of the diet eaten had a substantial effect on the energy balance and liveweight of the sheep. But the model did not allow for situations where digestibility varied as a consequence of stocking density influencing the botanical composition of the pasture and diet. Thus increased agreement with field observations might be expected if the model was constructed on a multi-species basis. The model was limited in application because of the use of arbitrary assumptions, such as the use of a specified soil type and effective root zone for soil moisture calculations. In the pasture growth section, the limits on growth via environmental factors also represented arbitrary averages of the behaviour of a phalaris/white clover pasture in the Armidale area. In another climate or location many of these parameters would require adjustment. Limitations in the animal section occurred in the conversion and efficiency factors used, which represented average values which did not interact with the diet available. Another limitation in this model was the absence of plant competition between the phalaris and white clover, and the effect of stocking density on this competition. The addition of interaction between soil moisture balance and stocking density, because of the latter's effect on soil physical conditions, would also be a desirable refinement to the model. Despite these limitations, the results indicate the feasibility of modelling the productivity of a complete grazing system over a wide range of stocking densities. Further development to a reproducing flock situation will be profitable. 21 VI. ACKNOWLEDGMENTS We wish to thank Dr. K. J. Hutchinson for permission to use his experimental results for test data, and for his interest and encouragement in this project. We also wish to record the role of those members of the Pastoral Research Laboratory who participated in the initial modelling exercise which initiated the development of the present model; their co-operation and thoughts contributed greatly to the original concept of this model. VII. REFERENCES Allden, W. G. (1968). Proceedings of the Australian Grassland Conference, Perth, 2: 213. Arcus, P. L. (1963). Proceedings of the New Zealand Society of Animal Production, 23: 159. Blaxter, K. L. (1967). 'The Energy Metabolism of Ruminants'. (Hutchinson: London). Hutchinson, K. J. (1969). Ph.D. Thesis, University of New England, Armidale, N.S.W. Langlands, J. P., and Sutherland, H. A. M. (1969). British Journal of Nutrition, 23: 603. Marston, H. R. (1948). Australian Journal of Scientific Research, Series B., 1: 93. Searle, T. W., and Graham, N. McC. (1970). Proceedings of the Australian Society of Animal Willoughby, W. M. (1970). Proceedings of the Australian Society of Animal Production, 8: 415. Young, C. A., and Corbett, J. L. (1971). Australian Journal of Agricultural Research 23: 57. Production, 8: 472. 22