Meat production on savannah-like grasslands ( dehesas ) in semi-arid zones of the province of Salamanca

Two savannah-like grassland systems (dehesas) in the province of Salamanca (Spain), developed on decomposing slate and granite soils and used for extensive meat production, were examined over a period of three years to determine the stocking rate they could support, their indices of pasture utilization, and the metabolisable energy necessary for meat production. The effective stocking rates were 0.36 and 0.30 cows ha year on the slate and granite systems respectively. The energy provided by supplementary feed was the 34% of the annual nutritional requirements of the herds, but reached more than 60% in the months of greatest need. Pasture utilization was 40% on the slate soil and 38% on the granite soil. The live weight obtained per reproductive cow was between 128 and 217 kg, equivalent to a weight per hectare of 57-90 kg year, of which 85% corresponded to calf production. The mean metabolisable energy required for producing one kilogram of meat (live weight) was 361 MJ on the slate soil and 369 MJ on the granite soil (range 215-499 MJ), depending on the year.


Introduction
In a previous paper (Martín Polo et al., 2003), estimations were reported regarding the amounts of available energy and protein in two savannah-like systems (dehesas) -one developed on decomposing slate soil, the other on decomposing granite soil, both typical of stock raising land in the province of Salamanca, Spain.Semi-arid conditions, such as those reigning in this part of the country, limit production.Feed supplementation periods therefore usually last longer than six months, although the greater part (80%) of the available feed is pasture.Prieto (1992) indicates that pasture utilisation is the main economic forte of the dehesa, and that pasture resources over the year condition stock raising production costs.
dehesa ecosystems in areas grazed by native Avileña, Morucha and Retinta cattle, Milán et al. (2001) estimated total stocking rates of 0.5 ± 0.1, 0.9 ± 0.1, and 0.4 ± 0.0 animals ha -1 of usable area (UA) respectively, 93%, 93% and 80% of which corresponded to cattle.Colson and Chatellier (1996) indicate the mean European stocking rate to be 1.45 cows ha -1 , with important differences between the northern and southern countries.Holechek (1988) estimated stocking rate as a function of the forage produced and utilised, bearing in mind the ingestion of the animals involved, the slope, distance to water, and the effects of life on the open range.
The present work estimated the current, theoretical and effective stocking rates, pasture utilisation, and meat production per reproductive cow per hectare and year for two dehesa systems of the province of Salamanca, one developed on decomposing slate soil, the other on decomposing granite soil, taking into account the energy requirements and productivity of each.

Animal production monitoring
The two dehesa systems (A = slate-derived soil, B = granite-derived soil) examined were used exclusively for the production of beef from native cattle (Morucha cow x Charolais bull crosses).Monthly stock censuses were taken, the weight of the «type» animal determined, and every two months the weight of 20 representative animals of the herd recorded.Since these animals are not easy to handle, visual weight estimations were made and the results contrasted with the true weights recorded for the eight most manageable adult cows.
The meat produced was considered to be the live weight of calves weaned at 6-7 months, the live weight of old animals, and that of stud bulls.
In 1989, no winter data were taken since work did not begin until the end of May.Data for the yearly census for this year were provided by the stock raiser.

Estimation of herd energy needs
The maintenance ration of a typical cow (MRC) was considered to be that which covered the metabo-lisable energy needs for the mean live weight of mother cows (MAFF, 1987).This was used as a unit of reference.The equivalents for herd animals, according to Bellido et al. (1986) are: 1.4 (i.e., 1.4 × maintenance ration of a mother cow) for lactating cows, 1.3 for cows in the last three months of gestation, 1.0 for non-pregnant cows (i.e., non-gestating or in the first six months of gestation), 1.5 × an activity factor (AF) of 1.15 for adult bulls, 1.25 × 1.15 AF for two year-old bulls, 0.9 for one year old cattle, 1.1 for two year-old females, and 0.9 for calves aged 4-7 months.Animal walking costs and the energy cost of life on the open range were taken into account, which amount to 30% of general maintenance needs according to Marchi (1978) but 20% on the Extremaduran dehesas according to Bellido et al. (1986).In this work, the figure of 20% was assumed because of the similarity between the dehesas of Extremadura and Salamanca.Since data on milk production in Morucha cows were lacking, those for Extremaduran Retinta cows living under similar conditions were used (1,130 kg over seven months of lactation) (Bellido, 1983).

Determination of the stocking rate
Based on the work of Bellido et al. (1986), three types of stocking rate were considered: -The current stocking rate (CSR): the result of dividing the total number of maintenance rations over one year required by the herd (MRCT) by the number of maintenance rations per year (i.e., 365) needed by the type animal (i.e., the MRC), and by the available pasture area (APA, ha), identified and estimated from aerial photographs: CSR = MRCT/ (MRC × 365 × APA).
-The theoretical stocking rate (TSR): this depends on feed availability at times of maximum production (spring pasture, autumn pasture, and acorn harvest) and assumes that full use is made of all resources.The TSR is estimated by dividing the total amount of metabolisable energy provided by the system in the form of pasture and acorns (TME) by the metabolisable energy of one maintenance ration for a type animal (MERT), the number of maintenance rations per year (365), and the APA: TSR = TME/(MERT × 365 × UA).
-The effective stocking rate (ESR) is that which the pasture can truly support.To calculate this, the total energy needs of the herd (MRCT × MERT) are subtrac-ted from 95% of the energy provided by supplementation (EPS) [there is a 5% loss according to Gaillard (1989)], and the difference divided by the MERT, 365, and the APA: ESR = (MRCT × MERT -EPS × 0.95)/ (MERT × 365 ×APA).
The pasture utilisation index (PUI) is the relationship between the energy actually used by the cattle and that which is available: PUI = (MRCT × MERT -EPS × 0.95)/ (TME × 100) (expressed as a percentage).

Changes in energy needs
Figure 1 shows the changes in adult cow live weight.The coefficient of variation, which is always below 15%, is relatively low and quite acceptable for a field study.It also shows that the herds were homogeneous.The live weight of animals in system A showed a spring maximum when the available pasture was at its peak (mean of 448 ± 8.5 kg), and a winter's end minimum (388 ± 9.0 kg).System B animals showed a similar trend (482 ± 11.7 kg in spring, 409 ± 2.3 kg at the end of winter).Weight losses were therefore 13.2 and 15.2%, respectively.According to Agabriel and Petit (1986) total winter weight loss should not surpass 8-9% by the time the cattle can graze once more, since this would increase the interval between births and reduce the production of milk.
Table 1 shows the nutritional needs of the herd.For the whole years of 1990 and 1991, supplementation reached 33.0% and 34.6% of the needs of the system A herd, and 31.6% and 36.5% of the system B herd.At certain moments (October and November of 1989) it actually reached as much as 63.7% and 63.3% of the two herds' needs respectively.Bearing in mind that weather conditions during the study period were worse than normal, and that these annual values do not exceed 50%, the systems can be characterised as 'extensive'.Calvo et al. (1997), who studied sheep raising on the Extremaduran dehesas, reported supplementation to reach 13.3% of total flock needs.Campos et al. (1996) reported general supplementation needs of 34.3% for Salamanca, 23.1% for the Alentejo (Portugal), and 46.7% for Badajoz.

Determination of the stocking rate
The census data of Table 2 were taken into account in order to express the total needs of the herds in terms of MRC.
The Table 2 shows that live weight remains practically constant over the study period in system A, but drops by 7.4% in system B. The mean number of calves born per reproductive cow was 0.77 in both systems, slightly greater than the 0.7 reported by Elena et al. (1986) for the Extremaduran dehesa.Table 2 shows the CSR, TSR and ESR and PUI values.CSR  and ESR showed little variation over the study period, e.g., mean CSR values were 0.55 ± 0.02 and 0.43 ± 0.01 in systems A and B respectively, and mean ESR was 0.36 ± 0.01 and 0.30 ± 0.02 cows ha -1 year -1 respectively.These values are higher than the 0.25 and 0.13 obtained respectively by Elena et al. (1986) and Miguel et al. (1989) for the Extremaduran dehesa.Milan et al. (2001) report a mean stocking rate for the Salamanca dehesa of 0.9 ± 0.1 cows ha -1 and 0.4 ± 0.0 for that of Extremadura.Coates and Mannetje (1990) determined an ESR of 0.17 cows ha -1 year -1 for natural pasture in Australia, and up to 0.68 for a combination of natural and sown pastures.The stocking rates of the present study are therefore high for semi-arid dehesa systems, but below the maximum 1.4 cows ha -1 demanded by the EU for extensification aid (BOE, 2000).
With respect to TSR, mean values of 0.90 ± 0.04 and 0.80 ± 0.05 cows ha -1 year -1 were obtained for systems A and B; system A would therefore appear to be more productive.
The two systems had more or less the same PUI: 40% for A and 38% for B, values recommended by Holecheck (1988) for arid (rainfall approx.400 mm year -1 ) areas.Winder et al. (2000), who studied arid areas (mean rainfall 235 mm year -1 ) stocked at 0.025 and 0.36 cows ha -1 year -1 , and with a dry matter production of 530-930 kg ha -1 year -1 , estimated pasture utilisations of 32% and 42% respectively.These values are high for the stocking rates but not so high with respect to production.For sheep grazing on the Extremaduran dehesa, Calvo et al. (1997) report a pasture utilisation of 61.7%.This figure is influenced by the stocking rate, supplementation, and the production of dry matter each year.It is also influenced by the natural degradation of pasture production, which, although it depends on the year, is more constant and beyond the control of the stock raiser.No data are available on this.

Animal production
Table 3 shows that meat production per reproductive cow per ha and year tended to fall after 1989.In this kind of system, the weather conditions of one year  APA: area of pasture available (ha).MERT: metabolisable energy of one maintenance ration for a typical animal (MJ day -1 ).CSR: current stocking rate (cows ha -1 year -1 ).TSR: theoretical stocking rate (cows ha -1 year -1 ).ESR: effective stocking rate (cows ha -1 year -1 ).
have an effect on events in the next.This is mainly manifested in the number of calves born, which significantly affects total meat production.Some 82% and 87% of meat production in systems A and B respectively corresponded to calf production (although in system B in 1991 this figure fell to 69% as a consequence of an abnormally large number of old animals at the same time as meat production rose with respect to 1990).Mean values of 138 and 148 kg calf per cow and year were obtained in systems A and B respectively, giving a total production per reproductive cow of 166 and 187 kg respectively, and a mean production per ha of 82 and 64 kg (live weight) respectively [large and significant variations (P > 0.05) were seen between years].Herbel et al. (1984) reported 142 kg of calf per reproductive cow per year as a low figure for arid areas of New Mexico.In the same area, Winder et al. (2000), with very low stocking rates of 0.025 and 0.036 cows ha -1 , obtained 178 and 135 kg calf per reproductive cow respectively.
Guimaraes et al. (1999) obtained a calf production per ha and year of 25.5 kg, and 109.5 kg cow -1 year -1 , when weaning was at 205 days.
The metabolisable energy that must be supplied (as pasture, acorn and supplements) to produce 1 kg of meat (live weight) varies between 220 and 449 MJ in system A and form 215 to 480 MJ in system B. The differences (P < 0.05) between years are influenced by the one year phase shift between pasture production and meat production, i.e., in a year of high production there will be more matings and more calves the following year.Calf meat makes up some 85% of total meat production in these systems.In an assay with calves produced by different crosses that received different food rations, Meissner et al. (1995) showed it was necessary to offer between 56.0 and 67.9 MJ to increase body weight by 1 kg.In Cantabrian pastures (northern Spain) during the pasture period of March to June, Zea and Diaz (1996) found it was necessary to supply between 87.4 and 98.9 MJ to increase calf weight by 1 kg, depending on pasture availability.Using different forages, McDonald et al. (1986), Dumont et al. (1989) and Steen (1990) required between 82 and 92 MJ to increase calf weight by 1 kg during fattening.No references on the transformation of metabolisable energy into meat were found for systems similar to those of the present work.
In this study, over 4 or even 5 times more energy was required to produce 1 kg of meat than the other systems mentioned above.This might be explained by the following: a) open range grazing in extensive systems increases maintenance requirements by some 20-30% (Marchi, 1978;Bellido et al., 1986); b) since the spring pasture cannot all be utilised at its optimum moment it loses quality due to adverse environmental conditions and trampling; c) there are losses of around 5% in supplementation (Gaillard, 1989); d) some of the available feed is consumed by wild animals; e) in the management of these exploitations a number of unproductive animals have to be maintained every year, such as non-pregnant cows, young animals for eventual breeding, and stud bulls.
As conclusions: -The mean supplementation requirement was 34% of the needs of the herd in both systems.Despite the adverse conditions of the study period, this figure characterises the systems as extensive.
-T h e s t o c k i n g r a t e s s u p p o r t e d a r e h i g h for the type of pasture and the semi-arid nature of the area.-The PUI for system A (slate-derived soils) was 40%, while that of system B (granite-derived soil) was 38%.
-A calf live weight per reproductive cow of 111-162 kg was achieved, as well as a total live weight of 128-217 kg, corresponding to a production per ha and year of 57-90 kg.Some 85% of total meat production was obtained as calves.
-The mean energy necessary to produce 1 kg of meat (live weight) was 361MJ and 369 MJ for systems A and B respectively.

Figure 1 .
Figure 1.Changes in mean live weight of the «type» animal (adult cow) over 1989, 1990 and 1991 in the slate-derived (A) and granite-derived (B) systems.

Table 2 .
Mean annual livestock census by system and year A: slate-derived system.B: granite-derived system.LW: live weight of «type» animal (adult cow) in kg.N: number of calves per reproductive cow.H: proportion of cows that did not give birth (expressed over 1).

Table 1 .
Maintenance rations per mother cow (MRC) and totals for the herd (MRCT), stocking rates and pasture utilisation indices (PUI)

Table 3 .
Meat production per reproductive cow, hectare and year, and MJ of metabolisable energy (ME)