A dissertation submitted to the Faculty of Science, University of the Witwatersrand, Johannesburg, for the degree of Master of Science. Johannesburg 1993. / The prediction that savanna communities with similar conditions of plant available moisture (PAM) and available nutrients (AN) have similar structural and functional features was tested for a range of southern African savanna types. This prediction forms the basis of an hypothesis that savannas can be classified functionally on the basis of PAM and AN. Nineteen South African savanna types were sampled over a rainfall gradient of 369 to 690 rnm pa and on different geological parent materials. Ecologically meaningful indices of PAM and AN were derived using climate and soil physical and chemical data. The floristic structural and functional characteristics of the woody plant and herbaceous components of the savanna communities were ordinated and classified in relation to the PAM and AN gradients. The functional classifications were assessed in the plane of PAM and AN. Plant available moisture was indexed as the mean number of growth days per annum derived form a water balance modelling approach. This index reflected the duration of the growing season in days and was suitably detailed for use at the community level. The AN indices ranged in complexity from a measure of the availability of individual nutrients in the soil to an index based on the results of a bioassay experiment. The usefulness of the AN indices was assessed on a number of criteria and the soil A horizon total nitrogen content in mg/kg was chosen as a suitable AN index. The species composition of the woody plant and graminoid communities was primarily related to the PAM gradient with AN having a significant but secondary effect. There was overall a high degree of similarity between woody plant and graminoid floristic types. The distribution of members of the woody plant families, Mimosoideae, Caesalpinoideae and Combretaceae were primarily related to the AN gradient. The higher taxonomic classifications of the graminoid communities, based on subfamilies and tribes, were primarily related to PAM. Structural trends of the woody plant and herbaceous communities were related to the PAM gradient but not to soil fertility. There were a range of structural types for similar conditions of PAM and AN. The functional trends of the woody plants were primarily related to the AN gradient with PAM having a secondary but significant effect while the graminoid community functional trends were primarily related to PAM with AN having a secondary but significant effect. The woody plant and graminoid communities were notgrouped into functional types in the same way. It was evident from this study that these South African savanna types were related floristically, structurally and functionally to either or both of PAM and AN, which are hypothesised to be the primary determinants of savanna structure and function. Although communities with similar conditions of PAM and AN were not consistently functionally similar, the functional classifications of both the woody plant ami graminoid communities mapped well onto the PAM-AN plane. This indicated that the PAM-AN plane does have potential foruse as an overall framework for the classification of savannas on a functional. basis. / AC2017

No description available.

No description available.

The majority of the vegetation types occurring on Malilangwe Estate, in the south-eastern lowveld of Zimbabwe, are dominated by Colophospermum mopane (mopane). Over the past 30-50 years the stand density of these mopane vegetation types has increased, and an investigation was undertaken to determine the effect of season of burning and type of fire on mopane woodlands. From this study the following was ascertained: 1) A single predictive equation cannot be used over all seasons to estimate standing crop (fuel load) using the standard disc pasture meter procedure. The calibration equations developed using this procedure accounted for between 39 and 72% of the variation in standing crop, illustrating the high variation in basal cover of the grass sward, as well as the variation between months. Although the revised procedure, developed for areas with low basal cover, accounts for a lot more of the variation in standing crop, this procedure was not used to estimate standing crop over the study period because the calibration equation covered a number of vegetation types, and was not specific to the mopane woodlands. 2) Standing crop tracks effective rainfall (monthly rainfall divided by monthly pan evaporation) closely, with a lag period of less than one month. Standing crop can be estimated using a predictive equation that utilizes effective rainfall from the previous month. There is a positive relationship between peak standing crop and rainfall. A predictive equation was developed to estimate peak standing crop, using annual rainfall. Standing crop declines through the dry season as effective rainfall decreases, and this 'decrease function' allows for the estimation of the standing crop for a particular month, after peak standing crop is reached. 3) Two leaf quantification equations were developed for mopane trees in the south-eastern lowveld of Zimbabwe, one for coppicing and for non-coppicing individuals. These allow for the estimation of leaf dry mass from measured canopy volume. 4) There was no significant difference between the fire intensities attained for the three seasons of burning. Over all seasons, head fires were significantly more intense than back fires. 5) Percentage topkill after late dry season burns was significantly higher than topkill after early dry season burns. There was no significant difference between mid and late dry season burns, and head fires led to significantly more topkill than back fires. Plants < 150 cm experienced significantly more topkill (80 %) than did individuals > 150 cm (44%). 6) Fire per se led to an increase in stand density over all seasons and types of fire, but this change was not significant. Fire did not influence the nett recruitment of new individuals. Height class one (0-50 cm) and three (151-350 cm) were impacted most by fire. This reflects a change in tree structure, with an increase in the amount of leaf material in height class three, and a subsequent decrease in the amount of material in height class one. 7) The effect of season of burning on the change in tree height was significant, whereas the effect of type of fire was not significant. All treatments, except early dry season back fires, led to a reduction in tree height, whereas trees in the no burn areas increased in height. 8) Burning in any season, and implementing either type of fire, led to an increase in the number of stems. Mid dry season burns led to the highest increase in number of stems. However, the more intense the fire the smaller the increase in number of stems. 9) All three seasons of burning (head and back fires) led to a significant decrease in maximum canopy diameter per tree, while the maximum canopy diameter of trees in the no burn areas increased. Mid dry season burns resulted in the greatest decrease in canopy diameter. 10) The effect of burning on the change in leaf dry mass per tree was highly significant. All three seasons of burning led to a decrease in leaf dry mass, while there was no difference between head and back fires. Leaf dry mass in the control areas increased however. High fire intensities led to the greatest decrease in leaf dry mass, late dry season head fires having the greatest decrease. This study suggests that mopane plants face a constraint due to fire and/or browsing, and a tradeoff occurs between canopy volume, canopy diameter, canopy area; and number of stems. Fire leads to an increase in the number of stems through coppicing, while canopy volume and leaf dry mass decreases. This decrease is either (i) a tradeoff in response to increasing stem number, or (ii) a reduction in canopy because additional leaves on the new stems contribute to photosynthesis. The most important response to season of burning is the altered phenophase (phenological stage) of the plant. Early dry season burns cause the trees to be leafless during the early dry season (when unburnt trees are carrying full leaf), and then to be in leaf at the end of the dry season (when unburnt trees are leafless). It would appear that fire disturbance initiates leaf senescence after burning, and then leaf expansion earlier than normal i.e the whole leaf senescence/growth process is brought forward. Trees in late dry season burn areas remain leafless at the start of the rains, while trees in unburnt areas are carrying leaf. Being leafless these trees do not photosynthesize during this time, and it is proposed that the grass sward is advantaged by the reduced competition from the tree component. The consequences of these two changes in phenophase could not be addressed in this study, but are pertinent questions that must be answered if mopane woodland dynamics are to be more fully understood. Management recommendations for (1) the removal of unacceptable moribund grass material, or (2) the reduction of encroachment by woody species on Malilangwe Estate are also given. In an attempt to combat the increase in stand density of mopane it is recommended that high intensity head fires be implemented, when standing crop (fuel load) is sufficient and climatic conditions are conducive to maintaining high intensity fires. These should be carried out at the end of the dry season, before the onset of the rains. / Thesis (M.Sc.)-University of Natal, Pietermaritzburg, 2000.