Fig 1.
Mount Kilimanjaro’s main habitat types.
Most of the savanna1 zone (800–1100 m asl) is used for crop production (maize2, beans and sunflowers). Similarly to the lower montane3 forest zone (1100–2000 m asl) where many home gardens4 (traditional agroforestry systems), commercial coffee farms5 and grasslands6 occur. In this zone, only few remnants of the natural forest vegetation are left. In the montane Ocotea7 forest zone (2100 m to 2800 m asl), selective logging8 occurred until 30–60yr ago. In the upper montane Podocarpus9 forest zone (2800–3100 m asl), considerable areas were destroyed by fires several decades ago, the tree Erica excelsa (Ericacea) has colonized these previously burned10 areas. In the subalpine, Erica11 trimera zone (3500–4000 m asl), Africa’s highest forests can be found. Higher up we find the alpine Helichrysum12 zone (4000–4500 m asl). In our study, each habitat type was represented by five sites which were distributed over an east-west gradient along the southern mountain slope and elevation gradient within each vegetation zone.
Table 1.
Vertical and horizontal vegetation structure variables with description and method.
The correlation matrix presents Spearman’s rank correlations across the 30 sites in natural habitats at Mount Kilimanjaro. A correlation coefficient of > 0.7 of highly correlated variables is highlighted in bold. Sub-column names (1/6) in vertical and horizontal vegetation structure represent the variables in the respective rows of the first column, where they are named. This order is also consistent with other tables and figures.
Fig 2.
Relationship between elevation and vertical vegetation structure variables at Mt. Kilimanjaro, assessed as average of the structure variables among nine points in each of 30 sites.
The fitted quadratic functions indicate significant relationships (P < 0.05 level).
Fig 3.
Relationship between elevation and horizontal vegetation structure variables at Mt. Kilimanjaro, assessed as the coefficient of variation (CV) of the structure variables among nine points in each of 30 sites.
The fitted quadratic functions indicate significant relationships (P < 0.05 level).
Table 2.
ANOVA results comparing vertical and horizontal vegetation structure variables among the a) six natural habitat types along the elevation gradient and for modified habitat types in each elevation zone of Mt. Kilimanjaro, b) savanna vs. maize fields, c) lower montane forests vs. home gardens vs. coffee farms vs. grassland, d) natural vs. previously selectively logged Ocotea forests, and e) natural vs. previously burned Podocarpus forests.
Significant differences are shown in bold.
Fig 4.
Vertical vegetation structure for each of 12 habitat types (each represented by five study sites) at Mt. Kilimanjaro.
Bars indicate means ±SE for the six natural habitat types (white bars) and the six modified habitat types (grey bars). Dashed lines separate different elevation zones. Bars not sharing a character are significantly different from each other (P < 0.05 level).
Fig 5.
Horizontal vegetation structure (coefficients of variation (CV) of the structure variables per site) for each of 12 habitat types (each represented by five study sites) at Mt. Kilimanjaro.
Bars indicate means ±SE for the six natural habitat types (white bars) and the six modified habitat types (grey bars). Dashed lines separate different elevation zones. Bars not sharing a character are significantly different from each other (P < 0.05 level).