Unravelling multidimensions of the flora of Kenya floristics, herbivory defences, and phylogenetic diversity

The wealth of plants in the vegetation of Kenya, owing to the country's wide range of climatic and ecological regions, can be complex and hence defies easy definition. This thesis, consisting of a literature review, four data chapters and a synthesis, addresses multiple aspects of that vegetation, including floristics, functional traits and phylogenetics, using a multidimensional approach, with a purpose to disentangle and explain drivers and patterns of the vegetation. The purpose of the study is achieved by focusing on the following four objectives: (1) assessing plant species composition and climate to present correlative support for the three African phytochoria divisions (Afromontane Archipelago, Somali-Masai and Zanzibar-Inhambane) prominent in the vegetation of Kenya; (2) investigate evidence for species filtering on the basis of soil properties maintaining broad-leaved non-spiny Combretum – Terminalia and narrow-leaved spiny Acacia – Commiphora woodlands mosaics in a mesic savanna ecosystem in south-eastern Kenya; (3) to evaluate structural and chemical plant defense manifestations between woody plants of the broad-leaved non-spiny Combretum – Terminalia and narrow-leaved spiny Acacia – Commiphora woodlands studied in Objective 2; and (4) to infer age, phylogenetic diversity and structure for the phytochoria of Africa predominantly represented in the vegetation of Kenya. Objective 1 (chapter 2) assessed community composition and climatic homogeneity phenomena for the three African phytochoria prominent in the vegetation Kenya. To test the phytochoria's taxonomic structure, occurrence data from 32 regional plant inventory checklists distributed across the phytogeographical regions of Kenya were grouped using multivariate analysis techniques. The checklists were intensively searched from published sources (i.e., journal articles and technical reports), which are considered representative of the vegetation of Kenya at species (70%) and genera (78%) levels, and at least comprise five samples for each phytochorion. To test climatic homogeneities for the studied phytochoria, 19 climatic variables of the BIOCLIM dataset extracted in each of the plant inventory areas were used, and a test of phytochoria-climate correlations executed in the canonical correspondence analysis (CCA). The study results show that vegetation types in Kenya were mainly partitioned by both precipitation and temperature vectors gradients, rather than on discontinuities in phytochoria units. For example, heterogeneity was observed in the Somali-Masai phytochorion, frequently interspersed with granitic hills, supporting montane-like vegetation at their summits. Precipitation was the most influential climatic variable in determining occurrence of vegetation type in the vegetation of Kenya, with its facets perhaps compounded by elevation and temperature. While phytochoria typically represent geographical areas with largely homogenous plant species composition that contrast with other phytochoria, in contrast, data presented in this study demonstrates that the vegetation of Kenya does not support the phytochoria classification. Instead, the vegetation is highly heterogeneous, partitioned by both precipitation and temperature vectors which change rapidly within a short distance particularly in response to elevation gradients. In chapter 3 (objective 2), I explored if soil characteristic is one of the micro-environmental filters driving and maintaining broad-leaved non-spiny Combretum – Terminalia and narrowleaved spiny Acacia – Commiphora woodlands as mosaics at the mesic savanna landscape of southern Kenya — herein testing evidence for species filtering on the basis of soil properties. I analysed empirical field data of woody species and soil characteristics from 57 plots that included 25 from Combretum – Terminalia and 32 from Acacia – Commiphora vegetation formations. I compared floristic composition and soil characteristics between the two systems and correlated edaphic statuses and vegetation states. I encountered 2,749 woody plant individuals, representing 115 species, 87 genera and 40 families. I found that the two vegetation types had discrete woody species composition and soil characteristics, and a significant speciesedaphic association. These results affirmed evidence for species filtering on the basis of soil properties (objective 2) and also a widely held view that soil characteristics are among the important micro-environmental filters driving and maintaining woody vegetation mosaics in the tropics. Objective 3 (chapter 4), which builds on the results of objective 2, evaluates whether there is a dichotomy on defence traits and perhaps on leaf nutrients levels, where spines and higher leaf nutrients are associated with fine-leaved plants occurring in nutrient richer soils whereas broadleaved plants occurring in nutrient poor soils have low nutrients but high tannin and carbon levels in their leaves. I used data of 93 plants for eight species from eight sites — four sites from fertile Acacia – Commiphora vegetation (42 plants) and another four sites from infertile broadleaved non-spiny Combretum – Terminalia vegetation (39 plants). Nutrient levels, and chemical and morphological defenses results did not show a consistent divide between the broad-leaved non-spiny Combretum – Terminalia and narrow-leaved spiny Acacia – Commiphora woody communities. Instead, the study found a range of defense-trait and leaf nutrients levels combinations. The results, thus, do not support the study hypothesis and only partially upheld the generally view regarding existence of trade-off between structural and chemical defenses for the dry African savannas woody plants partitioned on soil statuses. Instead, the results corroborate the emerging notion that a section of spiny woody species growing in nutrient-rich soils in African savanna landscapes invest in chemical anti-herbivore defenses equally to non-spiny broad-leaved plants growing in nutrient-poor soils in African savannas. Objective 4 (Chapter 5) inferred age, phylogenetic diversity, as well as phylogenetic structure of three major African phytochoria dominant in the vegetation of Kenya—with the aim to test aspects of “museum and/or cradle” of plant radiations. This was achieved through applications of large-scale DNA barcoding, including novel sequencing of over 200 Kenyan species and adding these to a matrix of ~2000 vascular plants. Maximum Likelihood (ML) methods implemented in RAxML software was used to create study phylogeny, which was dated using fossil-derived calibrations using the penalised likelihood algorithm as implemented in treePL. Phytochoria spatial occurrence data of the sequenced species was obtained from the African Plant Database, and taxon occurrences mapped onto the phylogeny using the phylobase package. The constructed phylogeny and species spatial occurrence data were both analysed in the Picante R-library to compute age, phylogenetic diversity and phylogenetic structure indices, which include phylogenetic diversity (PD), Net Relatedness Index (NRI), Nearest Taxon Index (NTI) and evolutionary distinctiveness (ED) of lineages for each phytochorion. The study found that divergence times for the Somali-Masai lineages are within < 10 Ma, a range that is much younger than of the Afromontane Archipelago and Zanzibar-Inhambane phytochoria. Afromontane Archipelago and Zanzibar-Inhambane phytochoria, in addition to containing taxa of younger divergences (i.e., < 10 Ma), also have lineages older than 10 Ma age range. The Zanzibar-Inhambane phytochorion had relatively higher PD, NRI and NTI values compared to the Somali-Masai and Afromontane Archipelago, while the ED measure among the phytochoria was similar, where the upper and lower quantiles overlapped. However, Zanzibar-Inhambane compared to Afromontane Archipelago and Somali-Masai had a higher number of outlier ED scores. Investigated phylogenetic indices support the hypothesis that Zanzibar-Inhambane is a “museum” of plant radiations, while the other phytochoria are mostly cradles of evolution. Overall, contrary to the common standard style in science that relies on a single dimension, a holistic understanding is needed to disentangle drivers and patterns of biodiversity. The current study applied a multidimensional approach to unravel multiple dimensions of the vegetation of Kenya. For example, using novel multivariate techniques that enable a quantitative approach, I have been able to rigorously test the phytochoria classification in the vegetation of Kenya. Through multivariate techniques I have advanced knowledge on ecology of the broad-leaved non-spiny Combretum and fine-leaved spiny Acacia vegetation types in the mesic savannas of south-eastern Kenya and perhaps elsewhere in Africa. Lastly, through applications of DNA barcoding, the study has contributed a novel insights and data for the vegetation of Kenya, opening opportunities for further work in the country in the field of community phylogenetics