Relationships between native tree species and soil properties in the indigenous forest fragments of the Eastern Arc Mountains of the Taita Hills, Kenya

The relationship between soil properties and spatial distribution of native woody species was studied in three Taita Hills forest fragments which, although degraded, are ranked among 34 biodiversity hotspots of the world due to their high biodiversity of both plant and animal species. This relationship was assessed by using Spearman correlation and principal component analyses (PCA). The results of these analyses should be useful in instituting forest restoration programs that are crucial for the forests. Both the soil and vegetation studied were sampled from 17 subplots in the natural forest fragments of Ngangao (120 ha), Chawia (86 ha) and Mbololo (185 ha). The soil variables measured were: pH, texture, soil nutrients of C, N, Ca, P, K, Mg and Na. In total 36 native tree species from 13 families were identified from the three forest fragments. Ordination results show that axis 1 accounted for 35% and axis 2 for 25% of the total variation in species composition, indicating that the structure of vegetation is related to two major environmental gradients. The correlation analyses of species and soil properties showed that Na and clay particles were the most important determinants of species distribution; pH and soil variables such as C, N, Ca and P also played minor roles. Unexpectedly, some species (e.g. Psychotria petitii
) showed positive relationships with Na attributed to possible substitution for K. Relationships with P were both positive (e.g. Craibia zimmermannii
) and negative (e.g. Albizia gummifera
) with some species, attributable to pH levels. An ANOVA for soil variables showed that there were differences in the Ca content in Mbololo (due to the parent material) and P in Ngangao where a special relationship was observed between some of the species. The presence of gaps accounted for the distribution of seedlings but not for the saplings, whose distribution responded more to factors similar to those to which mature trees respond. Soil-species relationships that were established may be utilized along with soil analyses when choosing native species for restoration.