**2.1 Sampling of vegetation**

Generally for biodiversity assessments vegetation is sampled in designated plots whose designs vary. In the Eastern Arc Mountains, a Y design developed by USDA35 (Fig.2) and modified for Forest Health Monitoring (FHM) in the Eastern Arc Mountains of Kenya and Tanzania36,37 have been used. Once the study plots are identified, sampling for vegetation entails enumerating the individual tree found in each plot by species. In order to assess the regeneration dynamics, it is important that during sampling, seedlings and saplings are enumerates as well. In this study, all live juvenile trees of > 5cm in diameter at breast height (dbh) were identified and recorded from each of the two randomly selected subplots. The seedlings were categorised based on height up to 1.3 m in height while sapling were categorized to be higher than 1.3m and with diameters less than 5cm dbh

#### **2.2 Diversity calculations**

There are several methods for analyzing plant diversity and its elements38,39, and these methods provide indices which provide bases for comparison. Lou40 has gone a step further by deriving Effective Numbers from these indices. These are whole numbers which make comparisons between different diversities easier since the numbers show the magnitude of the differences which makes it easy to perceive. Once the data collected for the diversity analysis is sorted, the diversities can be derived based on several formulae depending on which index is selected. An example is shown for Shannon-Weaver index that has been used in this analysis as Box 1. For practical use however, a Biodiversity Calculator software39 and can be used for the calculations. This calculator is freely available online (http://www.columbia.edu/itc/cerc/dunoffburg/MBD/LIMK.html). In this software, values for species derived from the data are fed in the calculator which then provides a range of diversity indices (e.g. Shannon-Weaver, Simpson and Broken stick) which one can select to use. In many cases such as in the Eastern Arc Mountain forest, the Shannon-Weaver index (H') has been commonly used for these calculations25,33. This is because it provides an account of both the abundance and evenness38 and does not disproportionately favour any species as it counts all the species according to their frequency39,40 unlike indices such as Simpson's which disproportionately highlights common species (in terms of abundance) instead of showing their frequencies (richness) in samples. One documented disadvantage of the Shannon-Weaver index is however, that it requires a large sample size in order to minimize biases41. Shannon-Weaver's indices ranges are typically from 1.5 to 3.5 and rarely reach 4.542. Other parameters such as species richness (S) and species evenness (H'E) are also generated from the same calculator.

#### **Box 1. Shannon-Weaver biodiversity Index**

Shannon's index, (H') is defined by

244 Biodiversity Loss in a Changing Planet

their ecosystem functions. A study by Himberg17 shows how the communities living around these forests decry the losses of the services such as reliable water flow and some of the

Several forest degradation types have been documented, at the global level25,26,27 as well as at the local level28, and most severe one consists of a total loss of forest cover due to disturbances. The consequences of this loss not only effects the biodiversity but also ecosystem functions28, including pest control and pollination in agricultural crops29, seed dispersal 30, and the regulation of water resources29,31. While in many instances losses in forest cover have negative impacts, in other circumstances, positive impacts have been noted in which there have been increases in plant species diversity. Such improvement in species diversity has been attributed to resilience which allow the certain plants to regenerate profusely after disturbance or to changed forest conditions such as increased light to forest floor 32 which favour the establishment of pioneer and early successional species whose seeds may have been stored in the soil seed bank. In the Taita Hills, the consequences of the disturbances documented have shown marked changes in tree species compositions25,33. The most disturbed forest fragment had a higher degree of changes in tree species composition in which 58 different tree species regenerated with stem densities varying between 10 and 2000 trees per hectare33. Differences in species composition occurred between the forest stand types, with native forests showing higher species diversities than the exotic plantations. The loss of biodiversity had also affected the animal

Generally for biodiversity assessments vegetation is sampled in designated plots whose designs vary. In the Eastern Arc Mountains, a Y design developed by USDA35 (Fig.2) and modified for Forest Health Monitoring (FHM) in the Eastern Arc Mountains of Kenya and Tanzania36,37 have been used. Once the study plots are identified, sampling for vegetation entails enumerating the individual tree found in each plot by species. In order to assess the regeneration dynamics, it is important that during sampling, seedlings and saplings are enumerates as well. In this study, all live juvenile trees of > 5cm in diameter at breast height (dbh) were identified and recorded from each of the two randomly selected subplots. The seedlings were categorised based on height up to 1.3 m in height while sapling were

There are several methods for analyzing plant diversity and its elements38,39, and these methods provide indices which provide bases for comparison. Lou40 has gone a step further by deriving Effective Numbers from these indices. These are whole numbers which make comparisons between different diversities easier since the numbers show the magnitude of the differences which makes it easy to perceive. Once the data collected for the diversity analysis is sorted, the diversities can be derived based on several formulae depending on which index is selected. An example is shown for Shannon-Weaver index that has been used in this analysis as Box 1. For practical use however, a Biodiversity Calculator software39

categorized to be higher than 1.3m and with diameters less than 5cm dbh

**1.2 Forest disturbance and its consequences on biodiversity** 

medicinal plants.

species composition34.

**2. Biodiversity assessment 2.1 Sampling of vegetation** 

**2.2 Diversity calculations** 

$$H' = -\sum\_{i=1}^{S} p\_i \ln p\_i \tag{1}$$

Where i, is the proportion of the species relative to the total number of species (pi) multiplied by the natural logarithm of this proportion (ln pi) and the final product multiplied by -1.

Species richness, the number of species present in an ecosystem (S) was defined by

$$S = \sum n \tag{2}$$

Where *n* is number of species in a community.

Species evenness, the proportion of individuals among species in an ecosystem is often assessed by Shannon's equitability index (H'E) which is calculated by

$$\mathbf{H'E} = \mathbf{H'} \;/\; H\_{\text{max}} \tag{3}$$

Where Hmax is defined as ln S. H'E values range from 0 to 1 and 1 indicates complete evenness.

The Shannon-Weaver's indices obtained in the study were converted to *effective numbers* using a method by Lou40; this can be done to obtain values that can be used to compare the differences in species diversities. The *effective numbers* are calculated as an exponential of the Shannon's index as:

$$N\_{\text{Effect}} \text{ of species (pi)} = \exp(-\sum\_{i=1}^{S} p\_i \ln p\_i) \tag{4}$$

To obtain similarities in species composition between forest fragments and forest stands, the Jaccard's index43 was used. Jaccard's index (Cj) is defined by

Cj = a/a+b+c (5)

where a, is the number of species present in both forest types or locations compared, b is the number of species in only one forest type or location; and c is the number of species present in the other forest type or location.

When establishing diversities and studying the regeneration dynamics, densities of native tree species in the study sites and forest stands (e.g., in exotic and native forests) have to be calculated on per area basis as well. In order to show the differences both in diversity and regeneration levels, further statistical analyses have to be done for the indices and densities derived. There are several statistical methods (SAS, STATA, SPSS) which can be used to assess such data sets. In the case of Taita hills, a one way ANOVA was used and SPSS 15 for windows software method adopted for the statistical analyses. The means were separated by applying Tukey's test to test the differences in the diversities and densities between the forest sites and types.
