Projected Impact of Climate Change on Habitat Suitability of a Vulnerable Endemic Vachellia negrii (Pic.-Serm.) (Fabaceae) in Ethiopia

: Species tend to shift their suitable habitat both altitudinally and latitudinally under climate change. Range shift in plants brings about habitat contraction at rear edges, forcing leading edge populations to explore newly available suitable habitats. In order to detect these scenarios, modeling of the future geographical distribution of the species is widely used. Vachellia negrii (Pic.-Serm.) Kyal. & Boatwr. is endemic to Ethiopia and was assessed as vulnerable due to changes to its habitat by anthropogenic impacts. It occurs in upland wooded grassland from 2000–3100 m.a.s.l. The main objective of this study is to model the distribution of Vachellia negrii in Ethiopia by using Maxent under climate change. Nineteen bioclimatic variables were downloaded from an open source. Furthermore, topographic position index (tpi), solar radiation index (sri) and elevation were used. Two representative concentration pathways were selected (RCP 4.5 and RC P8.5) for the years 2050 and 2070 using the Community Climate System Model (CCSM 5). A correlation analysis of the bioclimatic variables has resulted in the retention of 10 bioclimatic variables for modeling. Forty-eight occurrence points were collected from herbarium specimens. The area under curve (AUC) is 0.94, indicating a high-performance level of the model. The distribution of the species is affected by elevation (26.4%), precipitation of the driest month (Bio 14, 21.7%), solar radiation (12.9%) and precipitation seasonality (Bio15, 12.2%). Whereas the RCP 8.5 has resulted in decrease of suitable areas of the species from the current 4,314,153.94 ha (3.80%) to 4,059,150.90 ha (3.58%) in 2050, this area will shrink to 3,555,828.71 ha in 2070 under the same scenario. As climate change severely affects the environment, highly suitable areas for the growth of the study subject will decrease by 758,325 ha. The study’s results shows that this vulnerable, endemic species is facing habitat contraction and requires interventions to ensure its long-term persistence.


Introduction
Climate change affects species by pushing them to their ecological limit, resulting in range shifts. Climate change is a key challenge for biodiversity and the functions of ecosystems [1][2][3]. Various studies elsewhere have shown that climate change results in the redistribution of species [4][5][6][7]. Species track suitable habitats at their leading edge [8] and remain in their warm range edge due to phenotypic plasticity [9,10]. In cases where these two options are not available to species, they are destined to extinction [11]. A recent review The 19 bioclimatic variables downloaded from Worldclim (www.worldclim.org, accessed on 20 May 2021) with 30s (~1 km 2 ) from the observation data version 2.1 and 3 topographic layers were used in this study (Table 1). To predict the future distribution of this species, the General Circulation Model (GCM) Community Climate System Model (CCSM ver.5) was used. The intermediate greenhouse gas scenario (RCP 4.5) and the highest emission scenario (RCP 8.5) from the fifth Report of the Intergovernmental Panel on Climate Change (IPCC5) were used. Furthermore, solar radiation data was downloaded from the www.worldclim.org (accessed on 20 May 2021) with the standardized tiff formats [42][43][44]. Topographic factors (altitude and slope) were downloaded from Aster DEM (space born thermal emission and reflection radiometer Digital Elevation Model) from www.gscloud.cn (accessed on 24 May 2021) [45].
Bioclimatic variables downloaded in TIFF file format and converted to Asci format by using SDM toolbox of ArcGIS version 10.5 [44]. In order to identify the multicollinearity among the 19bioclimatic and topographic variables, Pearson's correlation was used [43]. A correlation among the bioclimatic and topographic variables was assessed using the ENM tool in R package [46]. A cut-off point of <0.8 was used to exclude variables with high correlation from further analysis to minimize the effect of multicollinearity and model overfitting [47]. Out of the 22 environmental variables, 13 were considered for further analyses in this study.  (Table 1). To predict the future distribution of this species, the General Circulation Model (GCM) Community Climate System Model (CCSM ver.5) was used. The intermediate greenhouse gas scenario (RCP 4.5) and the highest emission scenario (RCP 8.5) from the fifth Report of the Intergovernmental Panel on Climate Change (IPCC5) were used. Furthermore, solar radiation data was downloaded from the www.worldclim.org (accessed on 20 May 2021) with the standardized tiff formats [42][43][44]. Topographic factors (altitude and slope) were downloaded from Aster DEM (space born thermal emission and reflection radiometer Digital Elevation Model) from www.gscloud.cn (accessed on 24 May 2021) [45]. Bioclimatic variables downloaded in TIFF file format and converted to Asci format by using SDM toolbox of ArcGIS version 10.5 [44]. In order to identify the multicollinearity among the 19bioclimatic and topographic variables, Pearson's correlation was used [43]. A correlation among the bioclimatic and topographic variables was assessed using the ENM tool in R package [46]. A cut-off point of <0.8 was used to exclude variables with high correlation from further analysis to minimize the effect of multicollinearity and model overfitting [47]. Out of the 22 environmental variables, 13 were considered for further analyses in this study.

Data Analysis
Maxent software version 3.4.4 was downloaded from http://www.cs.princeton.edu/ schapire/MaxEnt/ (accessed on 30 May 2021) and used for modeling the current and future distribution of the species. Maxent is not disposed to sample size and can produce species response curves. All environmental layers have been converted and overlaid onto the same pixel size of 30 m and projected as an ASCII raster grid format. The generated model was evaluated by calculating the AUC of ROC graph [48,49]. The data for species distribution classified as training and test data with proportion of 75% and 25% for the total occurrence record data.
AUC is an effective and efficient independent threshold index with the capacity of assessing the model's capacity to distinguish the presence and absence. AUC values are categorized in to five different classes based on performance [50,51]. The performance classes are failing (0.5 to 0.6), bad (0.6 to 0.7), reasonable (0.7 to 0.8), good (0.8 to 0.9) and great (0.9 to 1). Models with values less than 0.5 indicates that the occurrence in the real-life scenario is rare or can be considered as a guesstimate [52] Jackknife was run to systematically exclude each variable or evaluate the leading bioclimatic or topographic variables. Jack knife evaluates the leading variables in determining the potential distribution of species [53]. The relationship between the environmental and topographic factors and the potential habitat for the species is determined from the created response curve from the model [54,55].

Correlation among the Environmental Variables
The Pearson's correlation has resulted in the removal of Bio 5, Bio 6, Bio 8, Bio 10, Bio 11, Bio 13 and Bio 16 from further analysis due to their high multicollinearity ( Table 2).

Model Evaluation
Validation guarantees the reliability of modeling results. In this study, the AUC value showed that the Maxent model performed well (AUC = 0.940) under the current scenario ( Figure 2). The distribution of Vachellia negrii (Pic.-Serm.) Kyal. & Boatwr. under the current climatic conditions is highly influenced by elevation, Bio 14 (precipitation of the driest month), Bio 1 Annual Mean Temperature, Bio 15 (precipitation seasonality) and solar radiation (Table 3). Furthermore, the contributions of these variables are higher than the others.  Validation guarantees the reliability of modeling results. In this study, the AUC value showed that the MaxEnt model performed well (AUC = 0.940) under the current scenario ( Figure 2). The distribution of Vachellia negrii (Pic.-Serm.) Kyal. &Boatwr. under the current climatic conditions is highly influenced by elevation, Bio 14 (precipitation of the driest month), Bio 1 Annual Mean Temperature, Bio 15 (precipitation seasonality) and solar radiation (Table 3). Furthermore, the contributions of these variables are higher than the others.

Response Curves of Bioclimatic Variables
The response curves show the relationships of probability of occurrence of Vachellianegrii (Pic.-Serm.) Kyal. & Boatwr. and each environmental variable ( Figure 3). The species response curve describes the relationship between the topography and bioclimatic factors and species occurrence probability ( Figure 3). Elevation, precipitation of the driest month (Bio 14), 26.4 percent and 21.7 percent, respectively, contributed the most to the spread of the species. The relevance of the bioclimatic and topographic variables of elevation and precipitation of the driest month (Bio 14) contribute the greatest percentages (30.4 and 27.2, respectively), followed by temperature seasonality (Bio 15), contributing13.9 in permutation ( Table 3). The response curves for the important variables show that they have a greater influence on the distribution of Vachellianegrii (Pic.-Serm.) Kyal. & Boatwr. (Figure 4). The suitability for growth increases as the altitudinal range increases, and ideally fits in the range of 2000 to 3000 m, where the rainfall of the driest month is adequate for the species' survival and growth ranges from 5 to 10m 3 during the driest season.

Potential Distribution of Vachellia negrii
Modelling the potential distribution Vachellia negrii (Pic.-Serm.) Kyal. & Boatwr. shows that it occurs on both sides of the Rift Valley ( Figure 5). It has a relatively higher range on the north and northwestern highlands of Ethiopia than on the southeastern highlands. In the latter, its highly suitable habitats are confined to the Bale Mountain range and Gara Muleta and its surroundings. In north and northwestern Ethiopia, highly suitable areas for this species are in moist and dry Afromontane forests. Moderately suitable habitats of this species occur at the fringes of the highly suitable areas and are much more extensive (Table 4).

Important Environmental Variables for Future Geographic Distribution of Vachellia negrii
The importance of environmental variables for the prediction of the distribution of Vachellia negrii (Pic.-Serm.) Kyal. & Boatwr for in the scenarios (RCP 4.5 and 8.5) are different for mid-century and for2070 ( Figure 6). There is also a within-scenarios variation. The distribution of this species is mainly determined by elevation, precipitation of the driest month (Bio 14) and precipitation of the warmest month (Bio 18) for RCP 4.

Predicted Distribution of Vachellia negrii (Pic.-Serm.) Kyal. & Boatwr.
The predicted distribution of highly suitable habitats for Vachellia negrii (Pic.-Serm.) Kyal. & Boatwr. decreases for both climate change scenarios (Figure 7). The species loses its northern range of highly suitable areas, and the southeastern range is more or less limited to Gara Muleta and its surroundings. Highly suitable habitats for this species decrease by 5.9% under RCP 4.5. in 2050 and 13.2% for RCP 8.5 in 2070 ( Table 4). The decline of The predicted distribution of highly suitable habitats for Vachellia negrii (Pic.-Serm.) Kyal. & Boatwr. decreases for both climate change scenarios (Figure 7). The species loses its northern range of highly suitable areas, and the southeastern range is more or less limited to Gara Muleta and its surroundings. Highly suitable habitats for this species decrease by 5.9% under RCP 4.5. in 2050 and 13.2% for RCP 8.5 in 2070 ( Table 4). The decline of highly suitable habitats for this species is dramatic for RCP 8.5 in 2050 and 2070, i.e., 17.6% in 2050 and 38% in 2070. A similar trend was observed for moderately and low suitable habitats. An exception is a slight increase in moderately suitable habitats for RCP 8.5 in 2070.

Discussion
Forecasting the future of climate change and its influence on species distribution is essential due to the wide range of climate simulation findings [56,57]. The distribution and well-being of forest ecosystems will be altered in either a positive or negative way by future global warming. In Ethiopia, the spread of Vachellia negrii (Pic.-Serm.) Kyal. & Boatwr. is now threatened. This vulnerable species is affected by both anthropogenic and climatic forces. The current impacts of climate change are affecting the whole world, especially in developing countries due to extreme population growth and illegal poaching [58]. Maxent modeling will have an effect on allocating the areas and the species which are endangered and in need of conservation [57].

Model Performance and Percent Contribution of Variables
According to Pearson et al., an AUC greater than 0.75 is a clear illustration of the strength and efficiency of a good species distribution (niche) model (2007). The model's AUC of 0.94 indicates that it is within the acceptable limits [57,59]. Because AUC values greater than 0.75 are considered informative [50], our results show that the model performed well in distinguishing between true and false positives [60]. The exclusion of variables owing to autocorrelation resulted in a flawless model outcome, according to references [61,62]. The topographic and climatic factors were shown to be autocorrelated, resulting in (Table 2) topographic variables (elevation and slope), precipitation (Bio 12, Bio 14, Bio 15, Bio 18 and Bio 19), temperate (Bio 1, Bio2, Bio3, Bio4 and Bio7) and solar radiation. Elevation has a significant role in determining the range of a species' habitat. As seen in Figure 4, the most suitable areas are found at elevations of 2500 m.a.s.l. When predicting the distribution of organisms, elevation is one of the most essential factors to consider [16].

Current Distribution
The species' present distribution is limited to the highlands of Tigray, Amhara, Oromia and Dire Dawa. The spread of Vachellia negrii (Pic.-Serm.) Kyal. & Boatwr is constrained by elevation. Even a small number of bioclimatic factors [63,64] affects the distribution. Bioclimatic variables that were employed to anticipate the distribution of suitable land did not adequately determine microclimates [57,[65][66][67]. Although elevation is the most important factor in the distribution model, solar radiation also has a role in the species' future and current spread. Elevation influences the species' microclimate, which has a direct impact on distribution [67,68].
Growing may also be possible at various elevations around the country [68][69][70]. Other variables that influence distribution, such as Bio14 and Bio15, also have a role. The temperature during the dry season and the precipitation throughout the dry and rainy seasons have an impact on species survival and recruitment. Because water is involved in all of the physiological activities of the species, the availability of water throughout the dry season promotes flowering and seed dispersal [71].

Future Distribution
Climate change has a significant impact on the geographic range of Vachellia negrii (Pic.-Serm.) Kyal. & Boatwr. according to the IPCC climate change scenario. Highly suitable, moderately suitable, low suitable and unsuitable are the four categories of appropriate regions predicted by the model. Climate change would have a negative influence on the niche of the species [28,72]. As a result, the suitable habitat range decreases (Figure 8). References 43, 52 and 66 predict that the species will make a significant shift to a more appropriate habitat. Climate change is the most significant factor, although habitat degradation, invasive species invasions, biodiversity loss and land conversion also contribute to the species' decreasing land mass [9,10]. Sustainability 2021, 13, x FOR PEER REVIEW 13 of 17 Vachellia negrii (Pic.-Serm.)Kyal. & Boatwr. thrives in environments that are ideal for its distribution, which decrease by 73.86% in 2050 of RCP 4.5 and 7% in 2070of RCP 4.5. The species' habitat suitability decreases in 2050 based on RCP 8.5, but increases in 2070 based on RCP 8.5. The Ethiopian highlands, similar to all other highlands, are undergoing climate change, which has a beneficial impact on species growth [67]. If the temperature rises on a regular basis, it will alter the seasonal growth of species [73,74].
Highly suitable areas for distribution of Vachellia negrii (Pic.-Serm.) Kyal. &Boatwr. decreases by 73.86% in 2050 of RCP 4.5 and an increase in 7% in 2070of RCP 4.5. The projection of the species in 2050 based on RCP 8.5 shows a decrease in the moderately suitable areas, but in the case of 2070 RCP 8.5, there is an increase in the habitat suitability. Ethiopian highlands, similar to any other highlands, are facing change in temperature, which affects the species growth positively [67]. The increase in temperature would also affect the seasonal growth of species if it occur son a continuous basis [73,74].

Range Shift
The moderate, low and high potentials from the landscape and regions with adequate habitats for growth and survival showed a decline in both the mid-century (2050s) and the end-of-century (2070s) periods ( Figure 8). As a result, a rise in temperature may compel their habitat to move from a lower to a higher elevation, causing habitat contraction and expansion, in this case, contraction [57,75,76]. In addition, invading species tend to decrease and alter the distribution of native species in the region [21,71,77,78]. According to the findings, the species range decreases from the total land mass of 104,429,094.6 ha with no occupancy, whereas 5,753,201.233 ha is deemed as an appropriate habitat in both situations, but climate change has caused 2,915,668.9450000003 ha to be lost. Several new studies on climate change and global warming demonstrate that variations in temperature have an impact on biodiversity distribution [78,79].

Conservation Strategy
In forest and land management, the ultimate goal of species distribution modeling (SDM) is to create future information about species distribution. The (SDM's) findings will encourage conservationists to develop and use coping mechanisms. The suitable region of the species' geographical range has shifted, according to the current research. Environmental plans must be based on research findings and recommendations. The species Vachellia negrii (Pic.-Serm.) Kyal. & Boatwr. thrives in environments that are ideal for its distribution, which decrease by 73.86% in 2050 of RCP 4.5 and 7% in 2070of RCP 4.5. The species' habitat suitability decreases in 2050 based on RCP 8.5, but increases in 2070 based on RCP 8.5. The Ethiopian highlands, similar to all other highlands, are undergoing climate change, which has a beneficial impact on species growth [67]. If the temperature rises on a regular basis, it will alter the seasonal growth of species [73,74].
Highly suitable areas for distribution of Vachellia negrii (Pic.-Serm.) Kyal. & Boatwr. decreases by 73.86% in 2050 of RCP 4.5 and an increase in 7% in 2070of RCP 4.5. The projection of the species in 2050 based on RCP 8.5 shows a decrease in the moderately suitable areas, but in the case of 2070 RCP 8.5, there is an increase in the habitat suitability. Ethiopian highlands, similar to any other highlands, are facing change in temperature, which affects the species growth positively [67]. The increase in temperature would also affect the seasonal growth of species if it occur son a continuous basis [73,74].

Range Shift
The moderate, low and high potentials from the landscape and regions with adequate habitats for growth and survival showed a decline in both the mid-century (2050s) and the end-of-century (2070s) periods ( Figure 8). As a result, a rise in temperature may compel their habitat to move from a lower to a higher elevation, causing habitat contraction and expansion, in this case, contraction [57,75,76]. In addition, invading species tend to decrease and alter the distribution of native species in the region [21,71,77,78]. According to the findings, the species range decreases from the total land mass of 104,429,094.6 ha with no occupancy, whereas 5,753,201.233 ha is deemed as an appropriate habitat in both situations, but climate change has caused 2,915,668.9450000003 ha to be lost. Several new studies on climate change and global warming demonstrate that variations in temperature have an impact on biodiversity distribution [78,79].

Conservation Strategy
In forest and land management, the ultimate goal of species distribution modeling (SDM) is to create future information about species distribution. The (SDM's) findings will encourage conservationists to develop and use coping mechanisms. The suitable region of the species' geographical range has shifted, according to the current research. Environmental plans must be based on research findings and recommendations. The species should be included in a list that may be planted during the country's yearly green movement, increasing the availability of the species in the country's high lands [55]. During the country's yearly greening, these designated appropriate regions would be used for planting. The government and stakeholders must focus on planting rare and endemic endangered species since the allocated sites for yearly plantation are ideal habitats for the species. In addition to the Vachellia negrii (Pic.-Serm.) Kyal. & Boatwr. plantation, the closure area would have an effect on safeguarding other treated plants and facilitating the establishment of additional species from seed banks [80,81].

Conclusions
Climate change is affecting the geographic distribution of Vachellia negrii (Pic.-Serm.) Kyal. & Boatwr. according to the present research. The model clearly demonstrates that the range of appropriate habitat for the species is shrinking. This has long-term consequences for the survival of the species. As a result, yearly tree-planting activities in Ethiopia should take into account predicted climatic conditions as well as the planting of Vachellia negrii (Pic.-Serm.) Kyal. & Boatwr. in appropriate habitats. For the Ethiopian government's annual plantation, the central highlands are now the most favored planting regions. Because these areas have been recognized as appropriate, it is suggested that these species be introduced. It is advised that these plants be planted on the hilltops since these areas have been designated as suitable. Efforts to conserve these severely endangered species should be made in general. Other endangered plant species, in addition to Vachellia negrii (Pic.-Serm.) Kyal. & Boatwr., should be conserved, therefore seed banks, botanical gardens and other insitu conservation techniques should be prioritized. To assist in the rescue of these species from extinction in the wild, further physiological, ecological, and microclimate research should be conducted.