Study area and climate data
The study was conducted in Ethiopia and Eritrea, which are located in the Horn of Africa between 3° to 18° N and 33° to 48° E, covering a total area of ca. 1.251 million km2, and characterized with a wide variety of landscapes, diverse geological formations and topographic features with marked contrasts in relief where the altitude ranges from about 125 m below sea level to 4, 620 m above sea level [Fig. 1; 26, 27, 28].Due to these diverse topographic features and wide elevational ranges the region is also known to have very variable macro and micro-climatic conditions and experience large spatial variations in temperature and precipitations [25, 29]. The mean annual precipitations across the region vary from 500 mm to 2200 mm while the mean annual temperature ranges from below 10 o C to 30o C with very high local variability. The transition between lowlands and highlands is commonly very sharp, resulting in a variety of climates that vary from very arid to very humid typical of equatorial mountains. Moreover, precipitation varies with latitude, generally decreasing from south to north. As the result of these physiographic and climatic features the region is endowed with a complex mosaic of habitats and ecological zones ranging from desert scrubland vegetation to alpine vegetations in high mountain areas that are inhabited with rich diversity of plants, animals and microbial life forms . To analyze the spatial patterns of plant diversity in this region in terms of the proposed parameters, we used uniform grid cells of 0.5° latitudes × 0.5° longitudes as units of analyses (Fig. 1).
In order to examine the relationships between the climatic factors and the diversity and structures of plant communities across the geographical regions of Ethiopia and Eritrea, we considered the mean annual temperature and precipitation as important ecological drivers of plant taxa distributions. The mean annual temperature and precipitation data as climate variables for each of the grid cells were extracted from global climate model Worldclim [30, available: http://www.worldclim.org/] using ArcGIS 10.5 .
The plant data sets used in this study were extracted from the published Flora volumes of Ethiopia and Eritrea (FEE)  and from the global biodiversity information facility (GBIF, https://www. gbif.org/). Based on our sources, we compiled a comprehensive checklist of seed plants belonging to Ethiopia and Eritrea. The database consisted of species names, family names, life forms, altitudinal and geographical distribution information’s of each of the species. Then, to ensure a standard taxonomy in the analyses we adjusted the family and genera of these plants as per the Plant List version 1.1 (available at http://www.theplantlist.org), using the R package "plantlist" , where the circumscription of the angiosperm family is generally consistent with APG III 
We extracted the minimum and maximum elevations of each of these grids using ArcGIS, and then the plant data records in each of these cells were obtained based on the altitudinal and geographical distribution range information of each plant as described in the sources (FEE). To assess the spatial taxonomic diversity patterns of seed plant distributions for total, woody and herbaceous plant groups across the geographic regions of these countries we calculated and mapped the genus richness of these plants at the 0.5° grid cell levels. The uniform size grid cells were used to eliminate the effect of the differences in area of the spatial units of analysis .
We constructed a phylogenetic tree for all the seed plants we compiled for analysis and also for the woody and herbaceous plant groups separately at genus level using the online program Phylomatic version 3 . The Phylomatic tree version R20120829.new was used as a backbone of the super tree and the BLADJ algorithm with PHYLOCOM version 4.2 was used to obtain phylogeny including the branch lengths in millions of years (Ma) based on Wikstrom [37, 38].
To examine the variation in evolutionary diversity of plant communities across the space and with respect to climatic factors in Ethiopia and Eritrea, we quantified the standardized effect size phylogenetic diversity (SES_PD) as standardized PD metrics because PD is strongly and positively correlated with TD . For this PD was standardized to the observed taxa richness by using the null model randomization method by shuffling the taxa labels across the phylogeny 999 times. Then SES_PD was computed as equation below [5, 38]
SES_PD = (PDobserved – mean PDrandomized )/sdPDrandomized (1)
PDobserved is the metric value of the communities under the study; the PDrandomized is the mean metric value of the null communities and sdPDrandomized is the standard deviation for the metric value of the null communities.
To examine the phylogenetic structures of community assemblages that quantify the degree of phylogenetic relatedness among plant communities in each unit, we calculated the net relatedness index (NRI) and the nearest taxon index (NTI) using the following models :
NRI = −1×(MPDobserved −MPDrandomized)/sdMPDrandomized (2)
NTI = −1×(MNTDobserved −MNTDrandomized)/sdMNTDrandomized (3)
Where, MPD is a diversity metrics measuring mean phylogenetic distance among all pair of taxa in the assemblage, MPDobserved is the MPD metric values of the community under the study, MPDrandomized is the mean value of MPD for the null communities, and sdMPDrandomized is the standard deviation of the phylogenetic distances in the null communities . MNTD represents the mean phylogenetic relatedness between each taxon and its nearest relative in the assemblage. MNTDobserved is the MNTD metric values of the community under the study, MNTDrandomized is the metric value for the null communities, and sdMNTDrandomized is the standard deviation of phylogenetic distances in the null communities . To maintain the statistical significance of the observed patterns, the randomization process for each of the null communities were repeated 999 times. Positive values of NRI and NTI indicate phylogenetically clustering communities while their negative values reveal phylogenetically overdispersing communities.
The analysis of PD and phylogenetic structures were performed in R software, using ‘picante’ package . The spatial patterns of diversity and phylogenetic structure were analyzed and mapped with ArcGIS 10.5 . To assess the relationships among plant diversity indices, phylogenetic structure attributes and climatic factors, we performed simple linear regression analysis and fitted the model for each plant group in R software .