Title: Predictive Spatial – TEMPORAL DISTRIBUTION of selected tree species and their CURRENT genetic STRUCTURE in East Africa
Predictive species distribution modeling is a valuable tool for decision-makers in biodiversity conservation, invasive species monitoring and other natural resource management fields. In order to conserve biodiversity, the genetic makeup of a species in question must also be fully understood. This study determined the potential distribution of twenty (20) tree species in East Africa, projected and assessed the distribution of the species in future (2050); determined the geographic structure of the genetic variation in selected three species, that is; Prunus africana, Lepidotrichilia volkensii and Syzygium guineense; anddetermined the phylogenetic affinities of Syzygium. The potential distribution of the tree species was determined through Maxent and GARP modeling techniques using data from different herbaria on the distribution of these tree species in East Africa. Species distribution was also used to assess the climate sensitivity, climate change risks and conservation implications for the studied trees. Predictive models were developed and used to project distribution of the twenty tree species for the 2050 climate scenario. For all species, the resulting distributions fit the provided occurrence data very well (AUC of test partition ≥ 0.763) and results are highly significant at the sensitivity-specificity-equality threshold (p < 0.001), with maxent using data at 1 km resolution. While some GARP distributions did not fit very well (AUC of test partition ranging from 0.44 to 0.905). Elevation was the most important predictor variable accounting for 50% of species distributions. Climate change is projected to decrease 45% of species potential suitable areas, increase of 45% of the species potential areas and have no effect on 10% of the species potential suitable areas.
A low gene diversity in Prunus africana (GD = 0.216, ND = 0.001) was obtained using matK sequences (GD = 0.216, ND = 0.001), as compared to a relatively high diversity using ISSR (%PPB = 61.25%). Low genetic differentiation of (0.417; 0.37) was obtained using both matK and ISSR. A high genetic diversity in Lepidotrichilia volkensii (Hd = 0.83244, π = 0.0243) and a low differentiation of 12.3% using ITS-4R sequences was revealed as compared to moderate diversity (PPB% = 61.29) and a low differentiation (Gst = 0.288) using ISSR. A low nucleotide diversity (π = 0.001199) and a high haplotype diversity (Hd =0.81699) in sequences using ITS-4R as compared to a high diversity using ISSR (PPB = 75.76%). A lack of genetic differentiation Fst = 0.08485 and Gst = 0.196, using both sequences and ISSR, respectively. Also, the ITS -4R region of Syzygium for 23 species resulted in to a consensus length of 650 to 662, using a matrix of 741 characters. The consistency index was 0.896396 (0.889952), the retention index was 0.990770 (0.990770), and the composite index was 0.888123 (0.881738) for all sites and parsimony-informative sites (in parentheses).
The study concludes with the observations that some output maps for the studied tree species have narrow distributions and can extend their occurrence beyond the high concentration areas (marked red) to the green and yellow shades. However with climate change and therefore many areas becoming drier, they are unlikely that they will extend to the potentially suitable areas. Therefore, the conservation of the current ecosystems where these occur is crucial for all stakeholders. At the East Africa regional level, environmental variables like soil, elevation and slope, at 1 km scale were more appropriate to explain species distribution as compared to climate (5 minutes) alone. Furthermore, the study recommends the use of land-cover or land use data at 1 km resolution for similar studies. Low diversity in all the three species studied that is Prunus africana, Syzygium guineense and Lepidotrichilia volkensii using chloroplast DNA call for special attention in the three countries to store germplasm of the species both In situ and ex- situ, as well as in gene banks. For Prunus africana, immediate attention should be paid to Mabira forest in Uganda which is under high deforestation yet the forest populations are isolated and quite unique compared to the other four.