Predicting the value of lupins for sheep and cattle in cropping and pastoral farming systems.

Abstract

Proc. Aust. Soc. Anim. Prod. (1974) 10; 383 PREDICTING THE VALUE OF LUPINS FOR SHEEP AND CATTLE IN CROPPING AND PASTORAL FARMING SYSTEMS G.W. ARNOLD and K.A. GALBRAITH* Summary Data from seven years research on the growth, competition with other species and feeding value of lupins were used to produce a simulation model of crop and animal production from lupins. The model simulated quite well the known effects of time of planting on lupin yield and the feeding value of the whole crop or the stubble for dry sheep and yearling cattle. The model was coupled with a model of pasture growth and animal production used to predict the effects-on growth of weaner sheep of sowing varying proportions of the farm to lupins in Mediterranean environments. The sensitivity of annual pastures to the increased winter stocking rate necessary to allow lupins to be grown was shown. Further development will be necessary before optimum combinations of lupins and pasture can be predicted. I. INTRODUCTION Sheep and cattle lose weight in summer and autumn when grazing dry residues from annual pastures of this region. A reliable way of ensuring that sheep and cattle will gain in weight in summer in Western Australia is to graze them on dry standing crops of sweet lupins (Carbon et al,. 1972). However, the rate of liveweight gain is variable (Arnold -et al. 3974). Lupin stubbles are becoming an increasingly important crop in W.A., and their value may also vary. This paper reports the use of available data to produce simulation models of lupin growth and utilisation, and their use to predict likely animal production responses - both from lupin crops alone and within pastoral farming systems using a simulation model of pasture growth. II. THE SIMULATION NODELS The model of growth of the lupin crop (Lupinus angustifolius cv. Uniwhite) has driving forces of average screezrature, radiation (cals/mn2/day'1) rainfall and open pan evaporation. A sub-model WATSUB by Carbon and Galbraith (unpbl.) predicts actual evaporation, soil moisture and soil suction. Lupin seed is predicted to germinate following three consecutive days when the soil is wet (see Table 1). A fixed germination of 80% is assumed, and at emergence each seed is assumed to produce 0.054 g of photosynthetic mass. All nutrients except nitrogen are assumed to be non-limiting. Temperature controls the time to emergence, the time from emergence until nodulation occurs, and influences growth rate (Trinick, personal communication). A potential growth curve, derived from data of Farrington (personal communication) is used to predict growth. This is modified for soil moisture content by multiplying by E/E0 (actual/expected evapotranspiration) and by temperature until August 15th; field observations indicate no temperature effects after this date. There is senescence of leaves once yield exceeds 2900 kg/ha and after September lst, senescence of the whole crop occurs if there have been five days with E/E0 x0.1. Senescence rate is 10% per day. When all the crop is dry the mass is partitioned into stem, seed and 'other' (leaf + pod). The digestibilities of these fractions are predicted from data of Arnold et al. (1974); values for -m stem and 'other' decrease due to leaching with summer rain. *Division of Land Resources Management, CSIRO, Private Bag, P.OoI Wembley, Western Australia, 6014. 383 It is assumed that weaner and wether sheep select similar diets and prefer seed to 'other' to stem. Selection of seed takes time to reach its maximum of 100% in the diet and this maximum cannot be achieved if there is <SO0 kg/ha seed left on the ground. After selecting seed, the rest of the diet is assumed to be 'other' unless there is <225 kg/ha and then some stem is eaten. Once the diet of sheep has been predicted its digestion and utilisation is predicted using the ANPROD sub-routines (Arnold and Campbell, 1972). Cattle select differently from sheep, being unable to harvest seed as readily (Carbon et al. 1972). Like sheep, cattle prefer other to stem -m and eat no stem until the availability of 'other' is <225 kg/ha. The potential intake of cattle is predicted using a function derived from data of Hodgson and Wilkinson (1967) and Holmes et al. (1961). Potential intake is achieved only with abundant feed of a high digestibility. Thus the predicted potential is adjusted for available dry matter using a function derived from data of Bennett (1963) and Hodgson et al. (1971). It is further adjusted for digestibility of the diet using the grassTu;e of the model described by Arnold and Campbell (1972). Intake of metabolisable energy (ME) is then calculated. The maintenance requirement of the cattle is predicted from data of ARC and subtracted from ME. Energy gain or loss is then converted to live weight using ARC (1965) standards. Trampling losses are not accounted for, but the coarse nature of the dry lupins means that little shattering occurs; some leaf material would be lost by trampling. Full details of the pasture model ANPAST will be published elsewhere (Arnold, Carbon and Galbraith, 1974), and can be obtained from the authors. ANPAST predicts the pools of green and dry matter, burrs and seeds for a 384 sub-clover pasture (cv. Geraldton) and their digestibilities. The driving forces are those used in the lupin model. The ANPROD model is used interactively with the ANPAST model. The model of lupin growth and animal performance was validated against data of Carbon et al. 1972 and Arnold et al. 1974, and the ANPAST model against -e data of Biddiscombe and Arnold (unpublGh=). III. USE OF THE MODELS Firstly, growth of crops of lupins grown on sandy-gravelly soil at the CSIRO 'Yalanbee' Experiment Station was simulated for six years (Table 2). These results show the known marked effect of time of sowing on yield, and agree reasonably with measured yields taken after senescensce and just prior to srazinq. TABLE 2 Simulation yields of lupins in six years with three times of * Secondly, the liveweight changes of 48 kg Merino wethers and 270 kg yearling cattle were simulated when grazed at three stocking rates on both whole standing crops and s,tubbles of the six simulated crops for 3967 and 1969 (Table 2). A selection of results is given in Table 3 and Figure 1. The results conform with known data and illustrate two points. At high stocking rates the available seed supply is eaten rapidly, and maximum live weight is lower and occurs sooner than at low stocking rates. The stubble has a higher feed value when crop yield is low, i.e. in 1969. Whilst these results accord with limited field data, further work on stubble values is needed to test whether lupin stubbles from crops grown in areas with short growing seasons have a better feed value than those from crops grown in high rainfall areas. 385 Thirdly, year long liveweight changes of weaner sheep were simulated for systems having various proportions of the feed area sown to lupins and grazed as whole standing crops for different periods. Three farm stocking rates were used (7.5, 10.0 and 12.5 per head). The systems envisaged 25 kg weaners enticing them the graze lupins from December 1st each year. In these tests, the system was run for a year to allow for carry-over effects on pasture yield and results (Table 4) are for the second years. Various sequences of lupin and pasture growth seasons were studied. In a year with high lupin yields followed by an early seasonal 'break', 25% of lupins gave the best weaner growth curves over the year at the highest stocking rate, but in years of good lupin growth preceded by and followed by a late seasonal 'break' for pasture growth, nearly all systems with lupins 'crashed' at 12.5/ha (Table 4) because clover seed production was reduced in the previous growing season and there was an insufficient initial mass of green material to sustain pasture growth the following year. However, at lO.O/ha, nearly all systems sustained themselves, but in some ; pasture yields in July were almost critically low. It seems that full nutritional advantage of lupins can only be achieved at moderate stocking rates, TABLE 4 Effects of stocking rate, proportion of farm sown to lupins and period of lupin grazing on the sheep and pastures Although the pasture and lupin growth models are 'first generation' ones and may well not simulate all situations with accuracy, they simulate conditions on gravelly soils quitetwell. The models allow a wide combination of seasonal conditions, type of livestock and stocking rates to be examined. It will probably be necessary to use them to develop criteria (an objective function) for assessing the optimum combinations of lupins and pasture and use this function in an optimisation model. V, REFERENCES AGRICULTURAL RESEARCH COUNCIL (1965). 'The Nutrient Requirements of Farm Livestock, No. 2 Ruminants.' (Agricultural Research Council: London). ARNOLD, G.W, and CAMPBELL, N.A. (1972). Proc. Aust. P-P - Soc. Anim. Prod, 2: 23; ARNOLD, G-W,, CARBON, B.A., and GALBRAITH, K.A. (1974). Proc. XII Int, Grassld -- Congr. Moscow. BENNETT, D. (1963). Proc. YP - Congr. - 3rd Aust. Grassld P CARBON, B.A. ARNOLD, G.W., and WALLACE, S.R. (1972). Prod. 9: 281. = HODGSON, J. WILKINSON, J.M. (1967). Proc, Aust. Soc- Anim. - Anim. Prod. = 365. 9: HODGSON, J., TAYLER, J.C., and LONSDALE, C.R. (1971). - Br. Grassld Soc. J. - P 26: 231. == HOLMES, W., JONES J.G.W., and DRAKE-BROCKMAN, R.M.' (1961). - Prod. 2: 251. Anim. 386