Search

The sustainable management of dry forests and woodlands of Sub-Saharan Africa (SSA) is crucial for the SSA economy and people’s livelihoods but it remains a key challenge to address. Maps of tree species distributions of economic value are lacking for the region; yet, this information is essential for supporting sustainable use. We capitalized on new nation-wide field survey data for the Republic of Zambia (Southern Africa) to generate the first suite of baseline species distribution models (SDMs) for 20 tree species of economic significance to support for sustainable management and conservation. We employed one regression- and two machine learning-based techniques to model and predict tree distribution. For each species, we compared the three model results for the relative importance of different predictive variables, revealing the most important predictors of each species niche and providing insight into how human activities influence current tree species distribution. Overall, environmental predictors that best explained tree species distribution were related to water availability, including mean potential evapotranspiration (PET), annual rainfall, and the variation in PET, as well as elevation and soil fertility. Human impact on distribution was notable for tree species used for charcoal and timber, including the proximity to roads for charcoal-favored species and the proximity to settlement for timber species. For all tree species, fire did not stand out as a variable of importance, contrary to expectations. The SDMs generated from this study will provide essential baseline information to support national conservation and management efforts, especially for preferred timber and charcoal species for which selective harvesting has had an impact on their distribution. Our results highlight the importance of rainfalls for the distribution of tree species in this seasonally dry region and calls for future research to forecast the impacts of climate change on habitat suitability.

Occurrence data for Julbernardia globiflora and J. paniculata at 617 sites in the miombo woodland region of central, eastern and southern Africa and forest inventory data for 512 woodland plots in Zambia were used to determine species distribution and dominance. Distribution of the two Julbernardia species overlaps in the central region of the miombo woodland range while the eastern and western range regions are exclusively for only one of the two species. In the region of co-occurrence, there is a clear spatial separation in the dominance of the two species. In old-growth woodland a significant proportion of the variation in the dominance of J. globiflora was explained by the dominance of J. paniculata while mean annual maximum temperature and tree species richness negatively affect the dominance of J. paniculata. Old-growth woodland clearing changes the local climatic conditions and alters the way Julbernardia species in re-growth stands respond to potential evapo-transpiration (PET). Climate change, especially global warming, may further reinforce the impacts of PET to differentially favour J. globiflora. Because of this altered response of Julbernardia species in re-growth miombo, preserving old-growth miombo and preventing present human disturbances in designated areas, such as forest reserves and national parks, may be a useful climate adaptation strategy for these species.

Aim: In tropical Africa, savannas cover huge areas, have high plant species richness and are considered as a major natural resource for most countries. There is, however, little information available on their floristics and biogeography at the continental scale, despite the importance of such information for our understanding of the drivers of species diversity at various scales and for effective conservation and management. Here, we collated and analysed floristic data from across the continent in order to propose a biogeographical regionalization for African savannas. Location: We collated floristic information (specifically woody species lists) for 298 samples of savanna vegetation across Africa, extending from 18° N to 33° S and from 17° W to 48° E. Taxa: We focused on native woody species. Methods: We used ordination and clustering to identify the floristic discontinuities and gradual transitions across African savannas. Floristic relationships, specificity and turnover, within and between floristic clusters, were analysed using a (dis-)similarity-based approach. Results: We identified eight floristic clusters across African savannas which in turn were grouped into two larger macro-units. Ordinations at species and genus levels showed a clear differentiation in woody species composition between the North/West macro-unit and the South/East macro-unit. This floristic discontinuity matches to the High (i.e. N&W) and Low (S&E) division of Africa previously proposed by White () and which tracks climatic and topographical variation. In the N&W savannas, the floristic gradient determined by rainfall was partitioned into the Sudanian (drier) and Guinean (wetter) clusters. Within the highly heterogeneous S&E savannas and woodlands, six clusters were identified: Ugandan, Ethiopian, Mozambican, Zambezian, Namibian and South African. Main conclusions: The proposed pan-African classification of savannas and woodlands might assist the development of coordinated management and conservation policies., SCOPUS: ar.j, info:eu-repo/semantics/published

Aim Population pressure and communal land ownership are often perceived as serious threats to forest conservation in savanna woodlands of central and southern Africa. I aimed at testing the hypothesis that the rate of miombo woodland recovery after clearing and re-growth structure are determined by land tenure and use. Location Miombo woodland under customary, leasehold, forest reserve and national park on ten permanent and temporary sites was studied in central Zambia. Two sites were in mature woodland and eight sites were in re-growth miombo ranging in age from 1 to 30 years. Methods I enumerated and measured girth at breast height (1.3 m above ground) of trees/stems in sixty-four 20 × 10 m plots in 1982, 1986 and 2000 at six sites and annually from 1990 to 2001 at four sites to determine stem density and status (live, dead or cut) and wood biomass. A total of 239 trees were cut, wood biomass measured and the data used to develop equations for estimating wood biomass on study plots. Distance between each study site and the nearest human settlement was estimated during each sampling period using aerial photographs, topographical maps and the global positioning system. Results Land tenure was responsible for significant differences in stem density, wood biomass and rate of biomass accumulation in re-growth following clearing of mature miombo woodland. Although stem density was highest on customary land, wood biomass and accumulation rate were lowest. The highest biomass was on plots in forest reserves, with intermediate values for leasehold and national park. Fire was responsible for tree mortality at all the study sites and its impact was highest at a site in a national park. Sites close to human settlements had the highest density of cut stems but this activity did not significantly reduce wood biomass. Rate of woodland recovery was higher on sites cleared in the 1970s than on sites cleared in the 1990s, irrespective of age of re-growth. The development of the first, second and third re-growths following successive woodland clearing in 1972, 1981 and 1990, respectively, was not significantly different, except for stem density which was highest in the second re-growth. Analysis of interactions between five land tenure and use factors (independent variables) and re-growth structure revealed that 52% (P=0.0000) of the variation in stem density was because of re-growth age and decade in which the woodland was cleared while distance to human settlements and age of re-growth explained 42% (P=0.0000) of the variation in wood biomass. Individually, distance to human settlements explained 25% (P=0.0000) of the variation in wood biomass accumulation rate. Conclusion The results supported the hypothesis that rate of miombo woodland recovery and structure were influenced by land tenure and use. However, analysis of interactions between factors revealed that use related factors (i.e. decade in which woodland was cleared and distance to human settlements) and re-growth related factors (age and type of re-growth) were more important than land tenure per se in explaining variation in miombo recovery. The conclusion from these results is that regulation of land use is more important than change in land tenure to the proper management of miombo woodland.

A survey on customary land and an adjacent forest reserve was conducted in the Kopa area of northern Zambia in November 2000. The objectives of the survey were to describe habitats of edible caterpillars and assess temporal and spatial variations in caterpillar abundance, the impact of tree cutting during caterpillar harvesting and the potential of edible caterpillars as sources of income for rural households. The results showed that edible caterpillars have the potential, even in years of moderate abundance, to generate incomes of over US$60 per household that are comparable to, or even higher than, incomes from the sale of agricultural crops. The most favourable habitats for the production of the two most popular edible caterpillars, Gynanisa maja Strand and Gonimbrasia zambesina Walker, were in young re-growth Brachystegia-Julbernardia (miombo) woodland regenerating after abandonment of chitemene shifting cultivation, probably because of the increase in abundance of their main host tree, Julbe...

Between 1937 and 1983 a total of 391 400 ha of woodland, representing 51% of the 768 900 ha in the Zambia Copperbelt, had been deforested for industrial and household woodfuel. This has occurred in spite of a forest reservation programme that increased the area in forest reserves from 31 707 ha in 1942 to 252 715 ha in 1965. Between 1965 and 1980 the reserved forests were reduced to 246 930 ha and since then the area has remained static. The rapidly growing demand for charcoal by the increasing urban population, and the growing problem of acid dust and rain pollution from copper smelters, pose serious threats to forest conservation in this industrial region of Zambia. Current forest conservation is restricted to the establishment of exotic forest plantations and recently deforested areas are left to regenerate naturally. Forest destruction and conservation in the Copperbelt area were studied and the structure of old-growth and coppice woodland stands at 9 and 14 sites, respectively, investigated during 1982–1984. The Copperbelt vegetation is dominated by Brachystegia-Julbernardia woodland. Mean species diversity of 24·40·1 ha −1 in coppice stands was higher than that of 18·9 found in old-growth stands. Stem density in coppice was 3·3 times that observed in old-growth woodlands. The majority of species (over 50%) in both old-growth and coppice were represented by less than 11 stems per sampling plot. Mean basal area at breast height in coppice aged 19–23 years was 14·33 m 2 , which was 95% of the mean basal area of 15·15 m 2 found in old-growth woodlands.

The purpose of natural fallow in bush-fallow cultivaton systems is to improve soil fertility following a phase of cultivation and to provide useful forest products, including livestock feed. When natural fallow fails to serve these purposes, it can be supplemented or replaced by planted trees. This paper describes the development and function ofAcacia fallow in the Soli tribal land of central Zambia and examines the supplementary role of planted trees.

Temporal and spatial changes in land use, vegetation cover, deforestation and reforestation in the Zambian Copperbelt were studied using a combination of aerial photograph analysis, literature review and inquiries among relevant government institutions. The study showed that between 1937 and 1984 loss of natural woodlands in the Copperbelt of Zambia (total area of Copperbelt: 9,615 km2) amounted to 41 per cent woodland area (8,419 km2). The major causes of this deforestation are woodfuel collection (for firewood and charcoal), cultivation and replacement of natural woodland with forest plantations. These accounted for 38 per cent, 37 per cent and 15 per cent of the total deforestation, respectively. Before 1962 the copper mining industry used large quantities of firewood to generate electricity and this resulted in the loss of 150,413 ha of woodland between 1937 and 1961. When the mining industry switched to hydroelectricity, urban households became the major users of woodfuel. The urban population in the Copperbelt increased from 0.412 million in 1960 to 1.400 million in 1984 and its consumption of woodfuel led to the deforestation of at least 89,436 ha between 1962 and 1984. Deforested areas are left to regenerate naturally. However, due to poor management about 34 per cent of the area deforested before 1962 has failed to regenerate. Although 50,200 ha have been reforested with exotic trees, only 5,020 ha of this is on land previously deforested. Thus only about 10 per cent of reforestation efforts have contributed to the rehabilitation of damaged land. Ninety per cent of the reforestation has actually replaced standing indigenous woodland. Uncontrolled bushfires have destroyed forest plantations and have also caused the failure of woodland regeneration in deforested areas. These observations indicate that if forest resources in the Copperbelt of Zambia are to be properly managed it will be necessary: 1to effectively regulate land use changes and; 2to control bushfires.

The status of chitemene shifting cultivation, in northern Zambia was investigated and reviewed. This paper presents results concerning (a) changes in soil nutrients under chitemene shifting cultivation, (b) the relationship between population density and deforestation and (c) how the chitemene shifting cultivation system has survived under conditions of overpopulation relative to the critical population density.

A survey of 94 miombo woodland stands was conducted in different parts of Zambia, and stackwood yield for charcoal production was assessed at 90 sites. The basal area of miombo woodland stands was significantly correlated to the aridity ratio in the study areas. Mean basal area at breast height per ha differed significantly among the three aridity zones in which study areas were located. In the aridity ratio 0.7–1.0 zone, mean basal area at breast height was 11.66 m 2 per ha, and this increased to 13.02 m 2 in the aridity ratio 1.0–1.3 zone and to 16.52 m 2 in the aridity ratio 1.3–1.6 zone. But mean basal area of woodland stands in different topographic zones was not significantly different. The volume of stackwood yield per m 2 basal area at stump height did not differ significantly among study areas: mean stackwood yield was 14.8 m 3 per m 2 basal area. Correlationship and regression analysis of the data led to the development of regression models for estimating standing wood stocks and yield for charcoal production in miombo woodlands in different parts of Zambia. The results presented in the paper are useful for managing natural woodlands for fuelwood production.

Miombo woodland in Zambia is divided into wetter and drier types by the 1100 mm mean rainfall isohyet. In the present study two wetter (Northern, Northwestern) and three drier (Central, Eastern, Western) miombo subtypes were recognized.A survey of the woody flora of 94 old-growth and 58 coppice stands in the five miombo subtypes revealed some significant differences in species diversity among them. The floristic differentiantion of miombo subtypes was best achieved by the use of the Shannon species diversity index (H′) and species density (number of species 0.1 ha−1 or 0.4 ha−1). H′ in miombo decreased from 2.247 in Northwestern to 2.136 in Northern, 2.052 in Eastern, 1.918 in Central and 1.824 in Western subtypes. Although species density was not significantly different between wetter miombo subtypes, this decreased in a similar fashion from 18.78 0.1 ha−1 in Northwestern to 18.30 in Northern, 16.65 in Eastern, 14.79 in Central and 12.29 in Western subtypes. The low species diversity in Western drier miombo was attributed to the combined effects of low soil fertility and rainfall.Although the mean number of species per 0.1 ha plot was higher in coppice samples, species richness (d) and index of diversity (H′) were lower than in old-growth samples. These results indicate lower species diversity in regenerating miombo. The dynamic aspects of species structure should therefore be considered in the future conservation of biological diversity in miombo woodland.