Estimating grazing-index values for Karoo plants



Grootfontein Agricultural Development Institute Private Bag X529. Middelburg 5900

JUNE 1995


At present, the Ecological Index Method (ElM), described by Vorster (1982) is used to calculate grazing capacities in the Karoo, The Karoo plant species are allocated index values on a group basis (Blom, unpublished report, 1981), giving rise to a disjunct 10, 7, 4, 1 ecological index value series (EN). The group classification is based on the ecological importance of the grass species, whilst the index values accorded the karoo-bush species are based on relative palatability ratings. These index values are used when the veld-condition scores are computed. The veld-condition score is an index of the state of health of that portion of veld, alluding to its ability to sustain animal production (Tainton, 1981).

Research by Botha (unpublished report, 1991), revealed that the index values currently used to calculate the veld-condition score, may lead to either over- or under-estimation of the current grazing capacity.

The construction of a single series of index values based on different variables for different plant groups is inappropriate. However, the greatest disadvantage lies in the fact that the series makes it impossible to distinguish between closely related species which have different grazing values. In order to correct these apparent shortcomings, the index values of the common species in the Karoo were subjectively adjusted by means of a species-by-species comparison.

The objectives of this paper are: * to illustrate the methods employed; * to debate the merits of the methodology; and * to supply a list of index values for various species.


A number of researchers and extension officers estimated the grazing values of those plant species encountered most frequently when conducting botanical surveys, according to the procedures carried out in the nominal-group technique. In order to ensure that the estimates of the species scores were comparable, each species was assessed on the basis of the following six variables:

1. The ability of the species to produce forage, i.e. the amount of grazable dry matter produced per year (low producers score 1, while high producers score 10).

2. The forage value during the growing season (species with a low feed value score 1, while species with a high feed value score 10).

3. The forage value during the dormant season (as in no. 2).

4. The relative ease with which the species can be grazed, i.e. the presence or absence of spines (sclerophyllous and very spiny plants score 1, while plants with soft leaves and plants without spines score 10).

5. Perenniality (annual plants score 1, while strong perennial plants score 10).

6. Ability of the plant to protect the soil against surface-soil erosion (upright karoo bushes have a low score while decumbent bushes score fairly high, annual tufted grasses score 1, tufted perennial grasses, depending on productivity and habit, have an intermediate score due to the erosive channelling effect they have on runoff water, while sod-forming grasses score 10).

The participants scored the species according to their knowledge and experience of a particular species. Each variable was scored out of ten. Variable scores were summed and the final score was calculated to lie between zero and ten.

The average of the scores submitted by the participants was calculated for each species. One standard deviation from the mean was calculated for each score. Values falling outside of one standard deviation were discarded and a new mean was calculated. The resultant score was vetted and either accepted or rejected. In the case of Tetrachne dregei, for instance, the participants' scores were 10, 10,9,7, 10, 10,8,6 and 10, respectively. Because 8,6 fell outside the one standard deviation, that score was ignored in the calculation of the new mean. The new mean is 9,95, or 10 when rounded off to one decimal place. This score is acceptable because Tetrachne is a dense sod-forming grass with a high production, exceptional soil-conservation characteristics and a high forage value. Consensus has it that Tetrachne represents the best available. This procedure was followed for all the other species. Rejected scores were scrutinized and adjusted on a consensus basis. The scores of related species were taken into consideration when the scores of individual species were adjusted. In the case of undesirable plants, such as poisonous and invader plants, etc., a factor was introduced in order to down-weight the grazing value of the particular species. The final scores are referred to as the grazing-index values (GIV).


Index-value series

The disjunct EN series of 10, 7, 4, 1, changed to a continuous series of grazing-index values ranging from 0.5 to 10. The grasses range in value from 0,8 to 10, while the index value of the karoo bushes range from 0.5 to 7,7. The climax grasses now range in score from 6 to 10, instead of the value of 10 used previously.

Comparisons of the index values of related Species

The series of grazing index-values (GN) enables fine distinctions in grazing value to be drawn between closely related species which have different grazing values (see Table 1).


The value of the adjustments can be seen when closely related species in the genus Pentzia are compared. In terms of Pentzia incana (good karoo/"ankerkaroo") EIV 4, the closely related Pentzia punctata is not as poor as the value of 1, as previously suggested, but it could obviously not be classified in the same category. Likewise, the inclusion of Pentzia sphaerocephala (EN 7) in a higher category than Pentzia incana is justified, but the difference in grazing value is not as large as its index value previously suggested.

Veld-condition scores

The value of the adjustments to the index values can be seen when veld-condition scores using the EN are compared with veld-condition scores computed using the GIV (Table 2). Veld-condition scores were computed for 510 sites. These sites represent 14 districts falling into eight of the nine agro-ecological areas into which the Karoo is sub-divided. One mean veld-condition score was calculated for each of the 14 districts. These scores were regressed on the median rainfall and the applied stocking rates noted down for these districts (Botha, unpublished report, 1991). The fit of the regression of veld-condition score on rainfall is better where veld-condition scores were computed using the GIV (r2 = 0,8653 and the index of agreement, d = 0,9285) (Willmot, 1982), than where the veld-condition scores were computed using the EN (r2 = 0,7738 and the index of agreement. d = 0,8468-cf. Fig. 1 and 2). The same applies to regressions of veld-condition scores on the applied stocking rates. Although the index of agreement suggests convergence between the GN and the EIV, the index of agreement, d = 0.9632 and r2 = 0,9455, small adjustments to the grazing values significantly affect the fit of the regressions of veld-condition scores on the applied stocking rates. In the case of the EN, d = 0,0439 and the r2 = 0,6874, while for the GN, d = 0,0481 and the r2 = 0,8193. This is an improvement of 13 % in the fit of the regression of GN veld-condition score on applied stocking rate over the EN.




In areas where the botanical composition is dominated by grasses, the ElM apparently over-estimates the current grazing capacity, owing to the high index value of the climax grasses. Areas dominated by these grasses have high veld-condition scores and hence high grazing capacities are calculated. It was found that the scores for these areas lie relatively far above the regression line describing the rainfall to veld-condition score relation. Fig. 1 illustrates the regression of veld-condition score on rainfall arrived at by means of the ElM. It is known that not all the climax grasses are of equal grazing value. When the grazing-index values of the species are used in the computations of the veld-condition score, a "better fit" is obtained when regressing veld-condition score on rainfall, as illustrated in Fig. 2 (Table 2).


The marked influence that the climax grasses have on the calculated grazing capacity is clear. In climax-grassveld areas, the grazing capacity is high when the EIV are used in the computation of the veld-condition score. This led to the assumption that a heavy stocking rate could be applied. When the GIV are used in the computation of the veld-condition score, the grazing capacity is lower, indicating the necessity of applying a lighter stocking rate.


The veld-condition scores of areas which are being dominated by karoo bushes, lie close to the regression line describing the rainfall to veld-condition-score relation. In these areas, use of the GN in the computation of the veld-condition score, yields a higher grazing-capacity figure, i.e. a heavier stocking rate can be applied than when the EN is used.

Species list with proposed grazing-index values

A list containing the proposed grazing values of some 714 of the common species occurring in the Karoo is appended as Table 3. Nomenclature follows Gibbs Russell, Reid, Van Rooy & Smook (1985), Gibbs Russell, WeIman, Retief, Immelman, Germishuisen, Pienaar, Van Wyk & Nicholas (1987) and Gibbs Russell, Watson, Koekemoer, Smook, Barker, Anderson & Dallwitz (1990). In the current nomenclature, Pentzia lanata is regarded as a synonym of Pentzia spinescens, but according to experience in the Karoo, these two taxa represent two different species. In the case of Pentzia lanata. the value of 5.7 is higher than the value of 4.8 recorded for Pentzia spinescens. The same reasoning applies to several species. i.e. Eriocephalus ericoides and Rosenia humilis.


Veld-condition scores for which the GIV of the species in the computations were used, result in a better fit when regressed on rainfall, than when the EIV of the species are used in the computations (see Fig. 1 and 2).

Veld-condition scores, calculated by use of the two series of index values, were compared. Condition scores for four sample sites from three different homogeneous areas are presented in Table 1. The scores for climax-grassveld areas are lower than previously and suggest the use of a lower stocking rate, whereas the scores for Karoo-bush veld are slightly higher. suggesting a higher stocking rate. The method of calculating the GIV can be applied successfully to species not included in the attached list. By taking into account the GIV of related species. the relative grazing-index value of any species can be calculated.



GIBBS RUSSELL. G.E., REID. C., VAN ROOY. J. & SMOOK. L. 1985. List of species of southern African plants. 2nd ed. Recent literature and synonyms. 1. Cryptogams. Gymno-sperms. Monocotyledons. Pretoria: Government Printer. (Memoirs of the Botanical Survey of South Africa: no. 51).

GIBBS RUSSELL. G.E., WELMAN. W.G., RETIEF. E., IMMELMAN. K.L. GERMISHUISEN. G., PIENAAR. B.J., VAN WYK. M. & NICHOLAS. A. 1987. List of species of southern African plants. 2nd ed. Recent literature and synonyms. 2. Dicotyledons. Pretoria: Government Printer. (Memoirs of the Botanical Survey of South Africa: no. 56).

GIBBS RUSSELL. G.E., WATSON. L., KOEKEMOER. M., SMOOK. L., BARKER. N.P., ANDERSON. H.M. & DALLWITZ. M.J. 1990. Grasses of Southern Africa. Pretoria: Government Printer. (Memoirs of the Botanical Survey of South Africa: no. 58).

TAINTON. N.M. 1981. The ecology of the main grazing lands of South Africa. in Veld and Pasture Management in South Africa. edited by N.M.Tainton. Pietermaritzburg: Shuter & Shooter.

VORSTER. M. 1982. The development of the Ecological Index Method for assessing veld condition in the Karoo. Proceedings of the Grassland Society of southern Africa. 17: 84-89.

WILLMOT. C.J. 1982. Some comments on the evaluation of model performance. Bulletin American Meteorological Society. 63: 84-89.




Technical Communication nr 239 p 1 - 15