. SOME METHODS IN ASSESSING THE EFFECTS OF MIXED ...
.
SOME METHODS IN ASSESSING THE EFFECTS OF MIXED GRAZING LN HETEROGENOUS
ENVIRONMENTS
L
J . Connolly,’ T . Nolan2, L. M. Guillon3,
C.Sal14, K . Dieye’ and H. Guerin 3
Teagasc 19 Sandymount Ave. Dublin 4, Ireland.
2 Teagasc, Creagh, Ballinrobe, CO. Mayo, Ireland.
3 IEMVT, 10, Rue Pierre Curie, 94704 Maisons Alfort. France.
4 ISRA, LNERV, Dakar-Hann, Senegal.
SUMMARY
In extensive range conditions, with a heterogenous vegetation including
herbaceous and Woody species grazed by a number of animal species, animal
preferences for plant species cari be defined.
This. information cari be
combined with data on the relative abundance of the plant species using
linear programming approaches to help assess the effect of mixed grazing
on animal output and preservation of plant species. Limitations are
discussed. Three examples suggest that there may be major benef it s in
mixed grazing in range conditions.
INTRODUCTION
The aim of the programme of which this work is part is to investigate the
role of mixed grazing in preservation of range vegetation and sustaining
or improving animal production in the Senegalese Sahel. The objective of
the work reported here was to measure preferences of different livestock
for individual plant species to assess whether this could lead to a better
use of vegetation resources by mixed compared with mono grazing.
METHODS
At the ISRA research centre at Dahra in Senegal six experimental grazing
treatments were selected, mono cattle, mixed sheep and goats and mixed
c a t t l e , sheep and goats, each a t t w o stocking r a t e s . A n i m a l s w e r e stocked
in a mob at either a low or high stocking rate and introduced to the six
experimental plots for a fixed period each weekday for four weeks
according to a latin square type design.
The botanical composition of
the diet of each animal species was assessed in each experimental grazing
period using the ‘bergere method’ (Guerin et al, 1984). This basic
experiment was executed in November/December 1985 (runs 1 and 4) and 1986
and in Spring 1986 (run 3). For each of the six experimental plots
detailed vegetation surveys were carried out in early September and
November 1985 and in early September, October, November and December
1986. Freferences
(DeRancourt et al, 1980) of each animal species for
----.. _--_..
-
..-
- - - - - - “.._.
i
L. $. R. JL
i
-. _
i
!
t
1
L7"'! ';.'T, a:;:p,!;
:c..
‘-“L. , 2
f
. ..-
-:
1
> ;
,
." -
,
1
I
*
.
i
each plant species were calculated, using the survey and ‘bergere method’
data. These express the preference by a grazing species per unit of one
pasture fraction relative to a unit of another fraction.
Connolly (1974) showed how linear programming based on a knowledge of
differential selection by a number of grazing species for different sward
fractions or plant species could be used to get insight into the potential
benefits of mixed grazing in rangeland conditions. The method is based on
the assumption that the relative preferences of animals for different
plant species is constant over mixtures at a particular time. It cannot be
applied in its full form here since we do not have information on the
total intake for each animal species as well as the total quantity of feed
availble in each vegetation category. However, as outlined below, the
concepts cari be used to give a limited insight into the extent to which
differential selection cari lead to higher carrying capacity under mixed
compared with mono grazing. The high stocking rate dietary data is used as
the preferences seemed more stable between mono and mixed grazing there
(Nolan et al, 1988). The examples chosen are from runs 1, 3 and 4, with
details given for run 1 in Table 1.
Using the patterns of intake for different species as a basis, and taking
the intake of individual cattle sheep and goats as 100, 25 and 20 units
respectively, Table 1 gives, for run 1, estimated intakes per animal of
each of the top eight herbaceous species, the Woody species and then a
balancing ’ species’ , to make up for the remaining species. This is the
data for the ‘Al1 species’ case.
Total consumption for the mixed grazing
treatment where cattle, sheep and goat stocking rates were (1,1.5,2) per
0.54 ha is shown in column 4, obtained by multiplying the first three
columns of the table by 1, 1.5 and 2 respectively and summing. We now
assume that the preferences as indicated by the consumption patterns are
constant for each animal species and that the total of feed consumed for
each plant species is the total available, a very narrow assumption. The
f ifth, sixth and seventh columns are obtained by dividing the fourth
column by the first, second and third columns in turn, and the lowest
value in each of them gives the maximum mono species stocking rate that
could be sustained subject to the assumptions.
For example, in the first
run these values for the ‘Al1 species’ case are 1.40, 2.19 and 2.00 for
cattle sheep and goats respectively.
The advantage to mixed grazing is calculated as indicated below. If the
c a t t l e , sheep and goat mono grazing maximum stocking rates are denoted by
(cl,o,oL (O,S,,O> and (O,O,Gl) respectively, then the equation of the
plane passing through these three points is
C
S
G
1 = --- + --_ + ---
(1)
5
s1
Gl
The line from the origin to the mixture (1,1.5,2), the actual mixed stocking
rate, is given by the equation
c = Sl1.5 = G/2
(2)
This line intersects the plane (1) at (k,1.5k,2k) where k is
k = l/(l/cl + 1.5/Sl + 2/Gl)
(3)
The relative advantage from mixing is estimated as l/k, being the ratio of
the actual mixed stocking rate achieved (1,1.5,2) to that expected if
there were no synergistic effects of mixing, (k,1,5k,2k).
RESULTS AND DISCUSSION
Values of l/k for the three runs for the ‘Al1 species’ and top f ive
species cases are
R u n
1
3
4
Top five species
1.30
1.76
2.60
Al1 species
2.40
2.42
2.63
Thus, in the first run, even with the reduced number of species, the
actual mixed stocking rate is 30% higher than that ‘expected’ assuming no
complementary ef f ect s. The figures for the benef it of mixing for the
other case and the other two runs are very much higher.
Mixed graz ing ,
under the assumptions would allow very much greater carrying capacity
than mono gr az ing .
Since the actual mixed grazing stocking rate was 1 cattle beast, 1.5
sheep and 2 goats, the mono grazing limits for sheep and goats are
unrealistically low in the ‘Al1 species’ case. If the analysis is
restricted to the top f ive species, the mono graz ing maxima are 1.40,
5.14 and 6.80 (the 6.80 being from the ‘balancing species’) for cattle,
sheep and goats respectively.
These conclusions rest crucially on two assumptions, neither of which cari
be expected to hold exactly in practice, SO the benefit may actually tend
to be lower than these figures would suggest.
The assumption of stable
relative preferences implies that in mono grazing the scarcity of one
component of feed Will proportionately limit the intake of the other
plant species.
In reality there may tend to be a greater level of
substitution between food sources, with switching as one source becomes
scarce and the results above confined to the five main species may be
more realistic in that the effects of small components may be unduly
distorting the estimation of carrying capacity of sheep and goats in the
'Al1 species' case. The second assumption, that the food actual consumed
in the mixture is a11 that is available is, of course, far too
simplistic,
assuming a matching of supply to demand that would rarely
occur in practice. However, it shows what could occur if feed supply
sources were well matched with requirements.
In linear programming
terms, what has been done by this assumption is to force a11 the
constraints to intersect at a common point, (1,1.5,2).
The actual
situation with some components in overabundance and with relative
preferences not SO rigidly defined may reduce this advantage, perhaps by
a considerable amount.
ACKNGWLEDGEMENTS
The authors wish to acknowledge the support of the Tropical Agriculture
Programme, DGXII, European Communities, in partially funding this work.
REFERENCES
Connolly, J. (1974). L' e
in ar programming and the optimum carrying
capacity of range under common use.
Journal of Agricultural Science
Cambridge, 83:259-265.
De Rancourt, M., Nolan, T. & Connolly, J. (1980). Measurement of
animal grazing preferences in mixed grazing.
Proceedings of Workshop
on Mixed Grazing, Galway, p.127-139. Published by An Foras Taluntais,
19, Sandymount Ave., Dublin 4, Ireland.
Guerin, H., Friot, D., MBaye, nD., 1983-1984 - Methodologie d'etude
de la valeur alimentaire des parcours naturels a faible productivite: 1
- Approche bibliographique no. 103/LNFRV, 1983-31 p.; II - Protocoles
et premiers resultats no. 13/LNERV, 1984-33 p.
Nolan, T., Connolly, J., Sali, C., Guillon, L.M., Dieye, K. and Guerin, H.
(1988). Mixed species in range grazing and preservation Ireland/
Senegal/ France. Report of project TSD/A/412 funded in part under the
STD subprogramme 'Tropical Agriculture', of DGXII of the European
Communities.
Table 1: Data from Run 1 for a partial linear programming analysis of
the impact of differential selection in mixed grazing.
_______--_------------------
-i--------------------------------------------
Intake per animal
Total intake/
Total
intake per animal
Plant species
Cattle Sheep
Goats intake Cattle Sheep Goats
-------------------------------------------------------------------------
Zornia glochidiata
41.0 5.3 4.2
57.3
1.40 10.91 13.64
Alysicarpus ovalifolius 14.0 3.5 3.4
26.1
1.86
7.44
7.66
Portulaca foliosa
8.0 2.8 1.0
14.1
1.77
5.14 14.13
Corchorus tridens
7.0 1.3 1.8
12.5
1.78
9.98
6.93
Ipomea pestigridis
6.0 2.0 1.4
11.8
1.97
5.90
8.43
Ceratotheca sesamoides 0.0 1.0 1.2 3.9
3.90
3.25
Cassia mimosoides
0.0 1.8 0.6 3.8
-
2.19
6.38
Ipomea vagans
3.0 0.8 0.4 4.9
1.64
6.57 12.31
Woody species
0.0 0.0 2.4 4.8
-
-
-
2.00
Balance of species
21.0 6.8 3.6
38.3
1.83
5.68 10.65
-----_-------------------------------------------------------------------
I?
.
SOME METHODS IN ASSESSING THE EFFECTS OF MIXED GRAZING LN HETEROGENOUS
ENVIRONMENTS
L
J . Connolly,’ T . Nolan2, L. M. Guillon3,
C.Sal14, K . Dieye’ and H. Guerin 3
Teagasc 19 Sandymount Ave. Dublin 4, Ireland.
2 Teagasc, Creagh, Ballinrobe, CO. Mayo, Ireland.
3 IEMVT, 10, Rue Pierre Curie, 94704 Maisons Alfort. France.
4 ISRA, LNERV, Dakar-Hann, Senegal.
SUMMARY
In extensive range conditions, with a heterogenous vegetation including
herbaceous and Woody species grazed by a number of animal species, animal
preferences for plant species cari be defined.
This. information cari be
combined with data on the relative abundance of the plant species using
linear programming approaches to help assess the effect of mixed grazing
on animal output and preservation of plant species. Limitations are
discussed. Three examples suggest that there may be major benef it s in
mixed grazing in range conditions.
INTRODUCTION
The aim of the programme of which this work is part is to investigate the
role of mixed grazing in preservation of range vegetation and sustaining
or improving animal production in the Senegalese Sahel. The objective of
the work reported here was to measure preferences of different livestock
for individual plant species to assess whether this could lead to a better
use of vegetation resources by mixed compared with mono grazing.
METHODS
At the ISRA research centre at Dahra in Senegal six experimental grazing
treatments were selected, mono cattle, mixed sheep and goats and mixed
c a t t l e , sheep and goats, each a t t w o stocking r a t e s . A n i m a l s w e r e stocked
in a mob at either a low or high stocking rate and introduced to the six
experimental plots for a fixed period each weekday for four weeks
according to a latin square type design.
The botanical composition of
the diet of each animal species was assessed in each experimental grazing
period using the ‘bergere method’ (Guerin et al, 1984). This basic
experiment was executed in November/December 1985 (runs 1 and 4) and 1986
and in Spring 1986 (run 3). For each of the six experimental plots
detailed vegetation surveys were carried out in early September and
November 1985 and in early September, October, November and December
1986. Freferences
(DeRancourt et al, 1980) of each animal species for
----.. _--_..
-
..-
- - - - - - “.._.
i
L. $. R. JL
i
-. _
i
!
t
1
L7"'! ';.'T, a:;:p,!;
:c..
‘-“L. , 2
f
. ..-
-:
1
> ;
,
." -
,
1
I
*
.
i
each plant species were calculated, using the survey and ‘bergere method’
data. These express the preference by a grazing species per unit of one
pasture fraction relative to a unit of another fraction.
Connolly (1974) showed how linear programming based on a knowledge of
differential selection by a number of grazing species for different sward
fractions or plant species could be used to get insight into the potential
benefits of mixed grazing in rangeland conditions. The method is based on
the assumption that the relative preferences of animals for different
plant species is constant over mixtures at a particular time. It cannot be
applied in its full form here since we do not have information on the
total intake for each animal species as well as the total quantity of feed
availble in each vegetation category. However, as outlined below, the
concepts cari be used to give a limited insight into the extent to which
differential selection cari lead to higher carrying capacity under mixed
compared with mono grazing. The high stocking rate dietary data is used as
the preferences seemed more stable between mono and mixed grazing there
(Nolan et al, 1988). The examples chosen are from runs 1, 3 and 4, with
details given for run 1 in Table 1.
Using the patterns of intake for different species as a basis, and taking
the intake of individual cattle sheep and goats as 100, 25 and 20 units
respectively, Table 1 gives, for run 1, estimated intakes per animal of
each of the top eight herbaceous species, the Woody species and then a
balancing ’ species’ , to make up for the remaining species. This is the
data for the ‘Al1 species’ case.
Total consumption for the mixed grazing
treatment where cattle, sheep and goat stocking rates were (1,1.5,2) per
0.54 ha is shown in column 4, obtained by multiplying the first three
columns of the table by 1, 1.5 and 2 respectively and summing. We now
assume that the preferences as indicated by the consumption patterns are
constant for each animal species and that the total of feed consumed for
each plant species is the total available, a very narrow assumption. The
f ifth, sixth and seventh columns are obtained by dividing the fourth
column by the first, second and third columns in turn, and the lowest
value in each of them gives the maximum mono species stocking rate that
could be sustained subject to the assumptions.
For example, in the first
run these values for the ‘Al1 species’ case are 1.40, 2.19 and 2.00 for
cattle sheep and goats respectively.
The advantage to mixed grazing is calculated as indicated below. If the
c a t t l e , sheep and goat mono grazing maximum stocking rates are denoted by
(cl,o,oL (O,S,,O> and (O,O,Gl) respectively, then the equation of the
plane passing through these three points is
C
S
G
1 = --- + --_ + ---
(1)
5
s1
Gl
The line from the origin to the mixture (1,1.5,2), the actual mixed stocking
rate, is given by the equation
c = Sl1.5 = G/2
(2)
This line intersects the plane (1) at (k,1.5k,2k) where k is
k = l/(l/cl + 1.5/Sl + 2/Gl)
(3)
The relative advantage from mixing is estimated as l/k, being the ratio of
the actual mixed stocking rate achieved (1,1.5,2) to that expected if
there were no synergistic effects of mixing, (k,1,5k,2k).
RESULTS AND DISCUSSION
Values of l/k for the three runs for the ‘Al1 species’ and top f ive
species cases are
R u n
1
3
4
Top five species
1.30
1.76
2.60
Al1 species
2.40
2.42
2.63
Thus, in the first run, even with the reduced number of species, the
actual mixed stocking rate is 30% higher than that ‘expected’ assuming no
complementary ef f ect s. The figures for the benef it of mixing for the
other case and the other two runs are very much higher.
Mixed graz ing ,
under the assumptions would allow very much greater carrying capacity
than mono gr az ing .
Since the actual mixed grazing stocking rate was 1 cattle beast, 1.5
sheep and 2 goats, the mono grazing limits for sheep and goats are
unrealistically low in the ‘Al1 species’ case. If the analysis is
restricted to the top f ive species, the mono graz ing maxima are 1.40,
5.14 and 6.80 (the 6.80 being from the ‘balancing species’) for cattle,
sheep and goats respectively.
These conclusions rest crucially on two assumptions, neither of which cari
be expected to hold exactly in practice, SO the benefit may actually tend
to be lower than these figures would suggest.
The assumption of stable
relative preferences implies that in mono grazing the scarcity of one
component of feed Will proportionately limit the intake of the other
plant species.
In reality there may tend to be a greater level of
substitution between food sources, with switching as one source becomes
scarce and the results above confined to the five main species may be
more realistic in that the effects of small components may be unduly
distorting the estimation of carrying capacity of sheep and goats in the
'Al1 species' case. The second assumption, that the food actual consumed
in the mixture is a11 that is available is, of course, far too
simplistic,
assuming a matching of supply to demand that would rarely
occur in practice. However, it shows what could occur if feed supply
sources were well matched with requirements.
In linear programming
terms, what has been done by this assumption is to force a11 the
constraints to intersect at a common point, (1,1.5,2).
The actual
situation with some components in overabundance and with relative
preferences not SO rigidly defined may reduce this advantage, perhaps by
a considerable amount.
ACKNGWLEDGEMENTS
The authors wish to acknowledge the support of the Tropical Agriculture
Programme, DGXII, European Communities, in partially funding this work.
REFERENCES
Connolly, J. (1974). L' e
in ar programming and the optimum carrying
capacity of range under common use.
Journal of Agricultural Science
Cambridge, 83:259-265.
De Rancourt, M., Nolan, T. & Connolly, J. (1980). Measurement of
animal grazing preferences in mixed grazing.
Proceedings of Workshop
on Mixed Grazing, Galway, p.127-139. Published by An Foras Taluntais,
19, Sandymount Ave., Dublin 4, Ireland.
Guerin, H., Friot, D., MBaye, nD., 1983-1984 - Methodologie d'etude
de la valeur alimentaire des parcours naturels a faible productivite: 1
- Approche bibliographique no. 103/LNFRV, 1983-31 p.; II - Protocoles
et premiers resultats no. 13/LNERV, 1984-33 p.
Nolan, T., Connolly, J., Sali, C., Guillon, L.M., Dieye, K. and Guerin, H.
(1988). Mixed species in range grazing and preservation Ireland/
Senegal/ France. Report of project TSD/A/412 funded in part under the
STD subprogramme 'Tropical Agriculture', of DGXII of the European
Communities.
Table 1: Data from Run 1 for a partial linear programming analysis of
the impact of differential selection in mixed grazing.
_______--_------------------
-i--------------------------------------------
Intake per animal
Total intake/
Total
intake per animal
Plant species
Cattle Sheep
Goats intake Cattle Sheep Goats
-------------------------------------------------------------------------
Zornia glochidiata
41.0 5.3 4.2
57.3
1.40 10.91 13.64
Alysicarpus ovalifolius 14.0 3.5 3.4
26.1
1.86
7.44
7.66
Portulaca foliosa
8.0 2.8 1.0
14.1
1.77
5.14 14.13
Corchorus tridens
7.0 1.3 1.8
12.5
1.78
9.98
6.93
Ipomea pestigridis
6.0 2.0 1.4
11.8
1.97
5.90
8.43
Ceratotheca sesamoides 0.0 1.0 1.2 3.9
3.90
3.25
Cassia mimosoides
0.0 1.8 0.6 3.8
-
2.19
6.38
Ipomea vagans
3.0 0.8 0.4 4.9
1.64
6.57 12.31
Woody species
0.0 0.0 2.4 4.8
-
-
-
2.00
Balance of species
21.0 6.8 3.6
38.3
1.83
5.68 10.65
-----_-------------------------------------------------------------------
I?
.