Comparison of the Floristic Composition and Diversity of Two Wetlands in Ethiopian

The main purpose of this study was to determine and compare the oristic composition and diversity indices of Washa and Borale Wetlands, located in Central Ethiopian. As revealed in the result, 74 species belonging to 26 families, and 57 genera were identied. Asteraceae and Poaceae were the most dominant families contributing 24.56% and 14.04% to the total genera, and 20.27% and 16.22% to the total species identied, respectively. Of the total, about 92% plant species were herbs, whereas 1% was climber, the least one. The alpha diversity of the Washa and Borale wetlands were 51 and 64, respectively. The average richness of the Washa and Borale wetlands were 12.3 ±0.91, and 15.35 ± 0.89, respectively. Likewise, the Shannon diversity (H’) and evenness (E) of Washa and Borale sites were 2.24 and 0.87, and 2.67 and 0.97 respectively. Accordingly, based on their average values, the diversity, evenness and richness indices were higher in Borale than Washa sites, and showed signicance difference between the two wetlands (P < 0.05). Likewise, in both sites, especially in Borale, the majority of the species were native, annual and upland, implying the suitability of the wetlands to these native, but to annual and upland invaders due to the ecological and hydrological modications of the wetlands, and competitive exclusion of the native aquatic plants by upland annual plants. Generally, many of the wetlands’ species were annual and upland invaders. Hence, in-situ and ex-situ strategic plans are required for restoring the wetlands via giving priority.


Introduction
Wetland ecosystems are essential in conserving wetland-dependent biodiversity and in delivering important and vital services for human well-being. For instance, the wetlands like ponds/small reservoirs provide sustainable solutions to key issues of water management and climate change such as nutrient retention, ood regulation, and carbon sequestration (Ce´re´ghino et al 2014), besides supporting various plant species. Many Ethiopian wetlands provide signi cant contribution to the local people chie y being the major source of water for irrigation, sh, livestock and human drink, and cleaning, as well as materials like grasses and fodders. Generally, they provide ecological, socio-economic and refreshment bene ts to humans (Moges et al 2018, Menbere and Menbere 2018; Wondie 2018). However, the continuous conversion of wetlands into crop and grazing land, over shing, drainage, eutrophication, and siltation threaten biodiversity, impair environmental functions of the wetlands and affect people's livelihoods (Gebresilassie 2014; Wondie 2018). Moreover, watershed degradation, overgrazing, invasion of alien species, urbanization and water diversion are other threats to wetland degradation (Soboka and Gemechu 2021). Similarly, agricultural practices such as farming, grazing, and extraction of water for irrigation are the major threat to the present study area. These all factors result in wetland degradation, which in turn has potential in uences on biodiversity, climate, ecological security and human health (Bai et al 2013). Thus, the most direct and effective way of evaluating the biological conditions of wetlands is to directly monitor the biological components of wetlands via using bioassessments, which help diagnose the causes of degradation, provide data to make informed management decisions, and/or prioritize wetlands being restored. Therefore, the main objectives of this study were to: 1) identify and document the plant species, 2) analyze the species composition and diversity indices of the study wetlands and 3) conduct a parametric t-test for a non-null hypothesis of mean diversity indices or variables between the two wetlands. reservoir including its surroundings is fed by seasonal ooding and three small streams draining its catchment, particularly in rainy season. It was designed to hold 280,000m 3 of water with 14m depth (Experts communication). Grazing, grass harvesting, farming and water extraction for irrigation were the main activities executed in and around the site.

Sampling design and data collection procedures
Field survey is the appropriate approach for collecting the oristic data. To familiarize with, or to observe the changes in plants and ecological conditions in the study wetlands during the dry and wet seasons, and nally, determine the sampling design, reconnaissance survey was carried out for two weeks, in May and September 2019. Accordingly, systematic sampling design was employed for laying out the quadrats in both sites. For oristic survey, the quadrat method is used to take the plant specimens and estimate the percent-cover of plants for identifying their botanical names and determining the diversity indices (Shannon and Wiener, 1949). For this, wetlands boundaries were rstly delineated to the maximum extent of ooding or the edge of depressions (Flinn et al 2008; Moges 2016). Since almost all plants of the wetlands were herbaceous, 1 m x 2 m quadrat sizes (Mulatu et al 2014) was xed. Then, the quadrats were laid out in the peripheries of the reservoirs, and at downstream wetlands following the transect lines. The transect lines were laid out along the longest axis of each wetland (Du Toit et al 2017) using a 50 meters interval between the two consecutive quadrats. Yet, where wetland sites were wide enough, adjacent quadrats were aligned parallel to one another, 50 m apart. Hence, 40 total quadrats (20 per site) were the total sampled plots of the study area. Geographical coordinates and altitude of each quadrat were recorded in Global Positioning System (GPS) of Garmin and in a notebook for facilitating the second round of data gathering.
The specimens of all vascular plant species encountered (Alvarez et al 2012; Ruto et al 2012, Moges et al 2017) in each quadrat of the study area were collected. Additionally, the percent cover of every species met in each quadrat was estimated (Murillo-Pacheco et al\ 2016). Accordingly, the specimens were collected at the end of rainy and dry seasons. October is a month of the end of the rainy season, and of good owering time for most of the vascular plant species, whereas February is one of the dry months, in which the plants favoring the dry conditions grow up in Ethiopia. Accordingly, the rst round plant specimens' collection was made in October 2020, and the second round, in February 2021. The aerial ground cover-abundance value of each herbaceous species was determined in percent (%) using ocular estimation, and immediately converted into the Braun-Blanquette (BB) scales ranging from 1 to 9 (as modi ed by Van  ). However, for further identi cation being assisted by expert mounted specimens and microscope, the whole specimens were taken to the National Herbarium of Addis Ababa University (ETH). Finally, the voucher specimens were placed there.

Data analysis
For vegetative data analysis, statistical tools, diversity indices, and parametric test were made using SPSS version 21, R version 4 and Stata version 14 softwares, respectively.
The diversity of the wetland plants were analyzed and measured using species richness (S), diversity (H'), evenness (J) and Sorensen's similarity coe cient (SSC) indices of Shannon and Wiener (1949). Species richness is a biologically appropriate measure of alpha (α) diversity and is usually expressed as a number of species per sample unit (Whittaker 1972). The Shannon diversity (H ') and evenness (E') indices were also determined as a measure to incorporate both species richness and evenness (Magurran 1988;Kent 2012). The Shannon diversity index (H') was calculated from the equation: Where pi, is the proportion of individuals found in the i th species. The values of Shannon diversity index is usually found to fall between 1.5 and 3.5 and only rarely surpasses 4.5 (Magurran 1988, Kent 2012). The Shannon evenness index (J) is calculated from the ratio of observed diversity to maximum diversity using the equation: Where H max is the maximum level of diversity possible within a given population, which equals ln (number of species). J is normal between 0 and 1, and with 1 representing a situation in which all species are equally abundant (Magurran 1988).
Moreover, for comparison of community similarities between the two study sites, Sorensen similarity index was used. The Sorensen similarity index is preferred to that of the Jaccard as it gives more weight to the species that are common to both sites (quadrats) than to the unique species to either of the wetlands. The Sorensen similarity index is calculated from the equation: Where SSC = Sorensen similarity coe cient, a = the number of species common to both sites, b = the number of species present in one of the sites to be compared and c = the number of species present on the other site. Often, the SSC is multiplied by 100 to give a percentage similarity. The SSC values range from 0 (complete dissimilarity) to 1(total similarity). For hypothesis testing to the most ecological data, nonparametric test is recommended since the uniformity of the data distribution might not be achieved. Yet, as the data sources of the two wetlands are independent, and the extreme data of the wetlands were minimized via transforming the percentage data to the BB scales (1-9) for bring them into uniformity, the paired parametric (T-) test was preferred. Statistical signi cance was considered when the p-value was < 0.05.

Plant species composition and habits
Totally, 74 plant species, belonged to 26 families and 57 genera, were identi ed from both Washa and Borale Wetlands (Appendix A). Among the total, the most dominant family was Asteraceae followed by Poaceae (Table 2). Thus, Asteraceae and Poaceae contributed 14 and 8, and 12 and 15, genera and species, respectively ( Table 2). The next important families considered as co-dominant families were Commelinaceae, Lamiaceae and Apiaceae (Table 2), of which the family Lamiaceae contributed three and four genera and species, respectively. Still, the families including Cyperaceae, Fabaceae, Onagraceae, Polygonaceae and Schrophulariaceae consisted of two genera each with variable number of species. Of those families, Cyperaceae contributed the largest number of species (seven) next to Asteraceae and Poaceae. The remaining families (18) comprised one generum with species ranging from one to two. Further, three and nine families were found in only Washa and Borale sites, respectively; but 14 families, in both sites (Table 2). Table 2 The distribution of family, and number of general and species in the study area (n = number of genera or species, % = percentage of genera or species) While considering the plant habits, 68 species (91.89%) were herbs, followed by shrubs (6.75%) and climbers (1.35%), which had a very small share (Fig. 2).

Species richness, diversity and evenness of the study sites
The alpha diversity (richness) of the study area, drawn from 40 total plots, was 74. The species richness distribution in 40 plots of the study area was ranging from eight species in each plot of 20 and 21 to 20 species in each plot of 7 and 12 ( Table 3). The beta diversity (average richness) via pooling the whole sampling sites (40 plots) of the study area was 14.55. Similarly, the richness of Washa and Borale sites were 51 and 64 species, respectively ( Table 2). The minimum, maximum and average richness of the two study sites were different. The minimum and maximum richness of the plant species in Washa Wetland were eight (in plot 8), and 17 (in each plot of 25 and 28), respectively ( Table 3). The beta diversity through pooling the richness of the 20 plots of Washa Wetland was 13.2. Likewise, the minimum and maximum plant richness in Borale Wetland were eight (in plot 20) and 20 (in each plot of 7 and 12), respectively, but the beta diversity was 15.9 (Table 3).

SSC and Paired T-test between the two study wetlands
The SSC between Washa and Borale wetlands was 0.71, which means the two wetlands were similar by 71% in their plant species composition. However, the paired (parametric) t-test was made between the mean values of richness, Shannon diversity (H') and evenness of Washa and Borale wetlands (Table 4). Accordingly, based their average richness, Shannon diversity and evenness values, there were high (p = 0.0113) and very high (p = 0000) signi cant differences between the two wetlands, respectively (Table 4).  (Table 5). However, the wetland species, consisting of obligate wetland (18.18%) and facultative wetland (21.21%) species, were 39.39%, and were less than the uplands. Still, the facultative species was very few (1.3%) ( Table 5).  Table 6). The common wetland (OB + FW) and upland (FU + OU) species to the two study wetlands were 10 (24.39%) and 21 (51.22%), respectively (Table 6).  The present study was carried out in Washa and Borale reservoir-based wetlands, located in highlands of DBT of North Shewa Zone, Central Ethiopia. Thus, the plant species (and other related data) were sampled from both the surroundings of the reservoirs, and their downstream natural wetlands so that the study sites were mixed types of both man-made lacustrine reservoirs and ooded natural wetlands. Accordingly, the discussion on comparison of the oristic composition, richness, Shannon diversity, characteristic and common species to the two wetlands, the SSC, and paired T-test was made via considering those two mixed wetlands types. ). This might also be due to the hydrological and ecological modi cations of the study wetlands, resulting in creating suitable microhabitats for growing many upland and native, but less abundant in their coverage, might be owing to the shortage of time for their invading the study area dominantly.

Plant composition
While comparing the two study sites in terms of their dominant families, Borale Wetland supported a larger number of families than that of Washa Wetland (Fig. 3). This might be due to the variations of their ecological status, i.e., Borale site was more disturbed than Washa site. Mulatu et al (2014) and Moges et al (2017) also reported from the southwestern highlands of Ethiopia that the disturbed wetlands support more plant species than the non-disturbed ones. Despite unlike in their extent of their ecological disturbances, the two wetlands had common physiognomies in their climate (moisture and temperature) and agro-ecological zones, resulting in having more than half common families growing in the two sites (~54%) (Table 1, Fig. 3). This might be because climate and altitude are the most determinant factors for such same plant species growth.
Regarding the plant habits, there were only three habits (herbs, shrubs and climbers) identi ed in the study area. Of those, the majority of the species (~92%) was herb, followed by shrub (~7%). Contrarily, climber/liana (~1%) was found to be the least number of species. This report also agrees with the ndings of Mulatu et al (2014) and Moges et al (2017). This might be due to that the wetlands having surface or subsurface water mostly support more herbs than the other plant habits.

Richness, Shannon diversity, evenness and similarity between Washa and Borale sites
Richness is an apt measure of diversity, and the richness of a range of habitats (of both wetlands) is also termed as gamma diversity (Whittaker 1972), but the average richness is called beta diversity (Whittaker 1972, Kent 2012). However, the total number of species per wetland site is called alpha diversity (Du Toit et al 2021). Thus, the species richness (S) is the most repeatedly used index (Magurran 2004) for comparing the diversity between sites (wetlands) (Woldu 1985) and may be used gamma, beta and alpha diversity interchangeable and accordingly for this paper.
Therefore, the gamma diversity of the present study area was 74. The richness per plot (2m 2 ) in the study area was also ranged from eight (8) Connell (1987) and Moges et al (2017) also reported that environmental heterogeneity, disturbance and competitive exclusive processes are the determinant factors for either increasing or decreasing the diversity of an ecosystem. As also reported during the ecological assessment of the two study sites, Borale site was highly impacted compared to Washa site due to the anthropogenic activities taking place within and around the two study sites (Supplementary le 1). Generally, diversity patterns seem to be driven by high landscape heterogeneity and wetland management.
Concerning the dominant species found in the two study sites, the species E. marginatum, G. dissectum, C. scherianus, P. thunbergii, Alchemilla abyssinicum, and L. stolonifera, were the most dominant species in Borale study site. Similarly, the most dominant species in Washa site were E. marginatum, L. stolonifera, G. dissectum, C. scherianus, P. thunbergii and Eleusine occifolia. Thus, the species E. marginatum, C. scherianus, L. stolonifera, P. thunbergii and G. dissectum were the most common dominant species to the two study wetlands. This result indicated that the dominant species in the two wetlands are almost similar, and more than half of them were obligate wetland species. Of course, the SSC between the two sites was about 71% similarity, and con rmed this nding. Barbour et al., (1987) also reported that any two plant communities those have more than 50% similarity shows the same association. Additionally, Kent (2012) stated that when calculated between all pairs of quadrats of a study area, or between two sites, a similarity or dissimilarity matrix is produced. Thus, based on this general rule, the Washa and Borale sites had similarity in their plant composition. However, Murillo-Pacheco et al (2016) from Colombia, and Moges et al (2017) from Ethiopia reported the opposite result that the similarity in composition of aquatic plants among the study sites was low. These also imply that although the SSC showed similarity between the present study sites due to their similar agro-ecological and climatic factors, the two study sites could have signi cance difference owing to the variations of the extent of ecological disturbances, resulting in creating heterogeneous microhabitats, which in turn, leading to support different species.

Characteristic and common species of the study area
The characteristic species are species that characterize the speci c site or habitat being unique to only that of other sites of the study area. Thus, persicaria decipens, Cyperus pauper, Cyperus brevifolius, Commelina diffusa, Anthemis tigreansis, Emilia leptocephala, Galium acrophyum, Senecio myriocephalus, and Taraxacum o cinale were the characteristic species to Washa Wetland since they were found only from Washa site. Particularly, the rst four plant species were common wetland plants in Washa and other less or non-impacted wetlands despite having many uplands (Table 5). Contrarily, the plant species including T. latifolia, L. tomentosa, Guizotia scrab, Plectranthus punctatus, Rumex abysinicus, Sida schimperiana and some others listed (Table 5) were unique to only Borale Wetland. Many characteristic species of Borale wetland were also uplands, and are considered as good indicators of disturbed sites. Even, some aquatic plants such as T. latifolia and Emilia leptocephala (Mattf.) C.Jeffrey are important characteristic plants being indicators of water pollution or ecological degradation in Borale Wetland. As also reported from Jimma Highlands, Boye Wetland was highly impaired and invaded by T. latifolia (and Emilia leptocephala) ( Additionally, from the total characteristic species (33), 90.9% and 54.54% were native and perennials, respectively. Contrarily, of the total (74), 41 (55.4%) species were common to the two wetlands. Of these common species, 92.68% and 56.1% were native and annual species, respectively (Appendix B). Seventeen species (41.46%) were perennials, and only one (2.44%) was biannual. Regarding these wetland indicator species of the common plants to the two sites, 10 (24.39%), 10 (24.39%), and 19 (51.22%) were wetland species (OB + FW), facultative (F), and upland species (FU + UP), respectively. This also indicates that the majority of the wetland indicator species growing in both wetlands were uplands. This nding is in line with Thompson et al (2007). These all imply that the invasion of alien species in the study sites was very low; however, due to their ecological disturbances, many of the aquatic plants were replaced by native and annual upland invaders. Still, those OB, FW and F characteristic species together represented about 41%, which grow and easily adapt the wetland ecosystems, and which indicate the restoration potential of the study sites.

Conclusion
Seventy four plant species were identi ed and documented from the two wetlands, located at Central Ethiopia. Asteraceae and Poaceae were dominant families, and contributed the largest number to the total species. The species diversity, richness and evenness of Borale were higher than Wash site, and showed signi cant differences between the two sites since Borale site was highly disturbed due to anthropogenic activities. Additionally, the study sites largely supported native and annual species, but many of them were uplands, considered as native invaders due to the hydrological and ecological changes in the study sites, creating a good environment for widely growth of upland species by excluding wetland species. Thus, if apt measures are not taken soon, both wetlands, especially Borale site would reach unrestored conditions. Therefore, to ensure a long-term conservation and sustainable use of wetlands, it is essential to develop and implement in-situ restoration and management strategic plans that take both reservoirs and natural wetlands with their catchments into consideration.

Declarations
Competing Interest: The diagram illustrating the habits of the species in the study area Figure 3