Back
 GEP  Vol.8 No.8 , August 2020
Modelling the Potential Distribution of Vitellaria paradoxa subsp. nilotica (C.F. Gaertn) across the Kidepo Landscape of Uganda in the Face of Climate Change
Abstract: Climate change and human activities are increasingly linked with the extinction of species globally. In semi-arid regions, these pressures threaten the natural distribution and ecology of species. The threat that the shea butter tree (Vitellaria paradoxa subsp. nilotica) faces from human activity is well researched yet the sensitivity of its distribution to climate change remains barely known. We set out to assess the potential distribution of Vitellaria under different climate change scenarios using a MaxEnt. A current distribution model was first developed using only biophysical variables of soil type, temperature, precipitation, land use type, and elevation. This model was then projected onto two global warming scenarios (RCP 4.5 & RCP 8.5) for 2050 and 2070 using multi-model averages (BCC-CSM, CSM4, and MIROC5) derived from three general circulation models. Reductions are seen in distribution area across the landscape with soil type being the most important variable. These results draw useful implications for conservation of Vitellaria in that they show how it is vulnerable is to a changing climate as its natural range is mostly reduced. Since climate change is important in the distribution of the shea butter tree, the areas with highest suitability in this study can be used in establishing the Shea butter tree sustainable use zones/area within the Kidepo Critical Landscape (KCL), Uganda.

1. Introduction

Climate change has affected many species globally, causing them to migrate or redistribute to more habitable regions and thereby exposing them to the risk of extinction as their natural range is generally reduced ( Kelly & Goulden, 2008; Thomas et al., 2004; Thuiller, 2004). This reinforces suggestions by the Intergovernmental Panel on Climate Change (IPCC) that 20% - 30% and 40% - 70% of species globally are likely to face extinction under the moderate and extreme case scenarios of global warming respectively ( IPCC, 2007). The increasing risk of extinction is owed to prolonged and hard-to-predict variations in climate which cause unprecedented changes in natural limits of environmental cues like rainfall and temperature. Therefore, understanding species distributions is essential for safeguarding vulnerable and endangered species from further ecological and environmental stresses.

Species Distribution Models (SDMs) are essential in understanding species distributions in geographic and environmental space ( Elith & Leathwick, 2009 Peterson et al., 2012). They have been applied in ecological reserve design and conservation planning (Araújo et al., 2004; Ferrier et al., 2002; Kremen et al., 2008; Thorn et al., 2009), as well as assessing effects of climate change on species distributions (Qin et al., 2017; Thomas et al., 2004; Thuiller, 2004). MaxEnt (Phillips et al., 2004) is one of the many SDMs that use a combination of species presence points and environmental conditions to predict species distributions. The algorithm is often preferred to others because of its precise mathematical foundation, low sample size threshold, and simplicity in model output interpretation (Elith et al., 2011; Radosavljevic & Anderson, 2014).

In this study, we focused on Vitellaria paradoxa (hereafter Vitellaria), a semi-arid tree species that occurs only in Africa, stretching its range from Senegal to Uganda, where it is only present in the Northern part of the country (Allal et al., 2011; Hall et al., 1996). It has several human and ecological values (Ferris et al., 2001; Okullo et al., 2012). Shea butter trees grow on a variety of soils with high humus content and their natural stands are often found at an elevation between 650 - 1600 metres. The best temperature range for the species is between annual means of 25˚C and 29˚C. Precipitation ranges from 600 - 1400 mm/year (Hall et al., 1996).

Anthropogenic activity also bears impact on the distribution of the shea butter tree (Buyinza & Okullo, 2015). Since agriculture is the main human activity in the communities where the tree is common, it can be used as a measure of anthropogenic impact on the distribution of the tree (Platts et al., 2010). Although the impact of anthropogenic activities on its distribution is well documented, the climate change threat remains scarcely known ( Buyinza & Okullo, 2015 ; Okullo et al., 2012 ). If the climate change risk is left unaddressed, conservation efforts centred on offsetting human impact alone may not be sufficient in future conservation planning as global climate is expected to continually change. This study, thus, focussed on assessing the impact of climate change on distribution of Vitellaria paradoxa subsp. Nilotica (the shea butter tree) in the Kidepo Critical Landscape of Uganda.

2. Data and Methods

2.1. Description of the Study Area

This study was conducted in Uganda within the Kidepo Landscape located between latitudes 1.71˚N and 4.23˚N, and longitudes 33.81˚E and 34˚E (Figure 1). The landscape stretches over 25,000 km2 with 22% (1447.45 km2) of this area forming the Kidepo Valley National Park ( NEMA, 2011). It is bordered by South Sudan in the North and the whole of its Eastern borderline is shared with the Republic of Kenya.

The study area falls in the Karamoja climatic zone of Uganda which is characterized by a long dry season from October to March, with a single wet season from April to September. Average precipitation is between 300 mm and 600 mm

Figure 1. A map showing the Kidepo Critical Landscape.

declining from west to east of the landscape. The region is distinctively the driest and hottest part of the country. The dominant soils of the study area are ferralsols. The area is generally low lying and marshy with a few seasonal rivers. The common vegetation types within this area are Vitellaria, Acacia savannas, forest-savanna and open grasslands ( NEMA, 2011).

2.2. Methodology and Analysis

2.2.1. Research Data

70 species presence points were used in the modelling process after a rigorous process of data cleaning and spatial filtering. Five environmental variables were selected based on knowledge of ecological requirements for the species ( Hall et al., 1996; Naughton et al., 2015; Okullo et al., 2004; Orwa et al., 2009). As shown in Table 1, they included annual mean temperature (bio 1), annual precipitation (bio 12), elevation (dem), soil type (soil), and agricultural land use (land). Downscaled global climate data was obtained from Global Circulation Model (GCM) projections with resolution of 30 arc seconds (≈0.86 km2 at the equator) and corrected for bias. Climate data for the baseline period of 1950-2000 and two RCPs was used to create distribution maps for 2050 (average for 2041-2060) and 2070 (average for 2060-2080).

2.2.2. Data Processing and Analysis

Model validation and evaluation

Data was partitioned in a 70 - 30 ratio for training and testing respectively. Default MaxEnt calibration settings were used in running the model. Suitability was classified as; Highly suitable (>0.6); Suitable (0.4 - 0.6); Marginally suitable (0.2 - 0.4); Least suitable (0 - 0.2). The resulting model was evaluated using the Area Under the Receiver Operating Curve. AUC values were interpreted as AUC > 0.90 (Excellent); 0.80 > AUC ≤ 0.90 (Good); 0.70 > AUC ≤ 0.80 (Acceptable); 0.60 > AUC ≤ 0.70 (Bad); and 0.50 > AUC ≤ 0.60 (Invalid) (Araújo et al., 2005).

Modelling the distribution under climate change

Distribution models were created for 2050 and 2070 under two emission scenarios RCP 4.5 (moderate case) and RCP 8.5 (extreme case), and three different GCMs; BCC-CSM, CCSM4, and MIROC5. The GCMs used were selected because they had harmonized temporal resolutions. Distribution maps under climate change were developed after differencing current and future distribution

Table 1. Environmental variables used in the model.

models. Climate change impact was characterized as; High impact areas (where Vitellaria potentially occurs in the present climate but which will not be suitable anymore in the future); Low impact areas (where Vitellaria can potentially occur in both present and future climates); New suitable areas (where Vitellaria could potentially occur in the future, but which are not suitable for natural occurrence under present conditions). The areas for each suitability class were subsequently calculated in a GIS environment.

Analysis for variable importance

The jackknife test for variable importance was used to determine the most important variable in the modeling process. A model was first created by excluding each variable in turn while using the remaining variables, then using each variable in isolation, and lastly using all variables. Three plots were consequently generated. In addition to the jackknife analysis, two other tests were used to check for variable contribution in the modeling process, as well as analyze the sensitivity of the distribution to changes in environmental cues.

2.3. Results

Current Vitellaria distribution

The model for current distribution of Vitellaria showed variation in suitability across the landscape (Figure 2). Otuke and Kitgum visibly had the largest portion of highly suitable area (935.07 km2 and 632.32 km2 respectively. These were followed by Agago, Amuria, Abim and Pader (516.40 km2, 207.60 km2, 196.36 km2, and 135.81 km2 respectively). Kotido and Kaabong districts were the least suitable (Table 2). The average test AUC for the replicate runs was 0.923 (std. dev = 0.019).

Vitellaria distribution under different climate change scenarios

Figure 3 shows the distribution of Vitellaria paradoxa in future for the three GCMs. There was no major variation in Vitellaria distribution across the different GCMs and emission scenarios for 2050 and 2070. Majority of the highly suitable habitat was constrained to Otuke, Kitgum and Agago districts across all GCMs.

Impact of climate change on Vitellaria distribution

Maps showing how the distribution of Vitellaria is impacted by climate change are shown in Figure 4. There is an increase in the amount of habitat lost as a

Table 2. General suitability in the Kidepo Landscape.

Figure 2. Current potential distribution of Vitellaria in the Kidepo Landscape. Dark green = highly suitable; Light green = Suitable; Cream = marginally suitable; White = least suitable.

Figure 3. Future potential distribution of Vitellaria for 2050 and 2070 under different RCPs.

result of climate change. Under the moderate scenario (RCP 4.5), habitat averages of 786 km2 and 730 km2 become suitable. Similarly, for the extreme scenario (RCP 8.5), 1115 km2 and 884 km2 of previously suitable habitat became unsuitable in 2050 and 2070 (Table 3).

Environment response curves

Lower and upper limits for temperature, precipitation, and elevation are shown in Table 4. These limits were obtained from the response curves of Vitellaria distribution for changes in these parameters generated by MaxEnt. In 2050 the lower limit average for bio 1 (annual mean temperature) under both RCP 4.5 and RCP 8 was 21.50˚C, while the highest limit was 25.3˚C. Both the lower and upper limits were consistent across all future models (21.50˚C and 25.3˚C respectively). For RCP 4.5, the upper limit average was 1459 mm while that for RCP 8.5 was 1458.7 mm. With regard to elevation, the lowest limit varied across all future models, with an average of 726.83 m. The averaged upper limit on

the other hand was 2485.67 m.

2.4. Discussion

Model projections under climate change show that while the suitability reduces

Figure 4. Spatial variation of climate change impact on future potential distribution of Vitellaria for 2050 and 2070.

Table 3. Quantified loss in suitability for Vitellaria across all GCMs.

Table 4. Multi-model averages for environmental limits for Vitellaria.

in some areas of the landscape, it increases in other areas. The increase in the upper limit for precipitation for both the moderate and the extreme case scenarios in 2070 is an indication that Vitellaria is likely to adapt to higher precipitation. On the other hand, the substantial drop in the limits for elevation for both RCP 4.5 and RCP 8.5 of 2050 and 2070 suggests that the range of Vitellaria may shrink as these upper limits for these future years are above the idealized range for elevation of such a species. Such shifts are expected for most species globally ( Ayebare et al., 2013

All models predicted soil type as the most important variable in the modeling process for both current and future time periods. It is possible that differences in soil type between districts could have led to differences in suitability across the landscape as shown in the maps for current and future predicted distribution. This could be the case since the species is adapted to semi-arid conditions, where variations in temperatures and precipitation have been reported to have less influence in determining the species’ distribution ( Cahill et al., 2013). Similarly, the low importance attached to temperature could be explained by the ability of Vitellaria to live in hot areas, implying that temperature is not a very important limiting factor for the growth of Vitellaria. Indeed, temperature changes have been shown to have less of an effect on the suitability for the sheanut tree ( Cahill et al., 2013; Grossiord et al., 2017).

2.5. Conclusion and Recommendation

Predicting the impact of climate change on species distributions is a key step in developing actions for conservation management. Evident suitability losses in the Kidepo landscape give reason to further assert suggestions that climate change is likely to push vulnerable species such as Vitellaria to the brink of extinction in the long term. It should be noted, however, that while greater portions that are suitable become unsuitable due to climate change, coincidentally, patches of new suitable areas also emerge. This affirms that climate change will have a dual impact. Therefore, existing on-farm tree conservation (Okiror et al., 2012) should be upheld where existent, and is highly recommended where it is not being practiced.

Acknowledgements

This work was supported by the NEMA project “Conservation and Sustainable Use of the Threatened Savanna Woodland in the Kidepo Critical Landscape in North Eastern Uganda” (https://www.ug.undp.org/content/uganda/en/home/operations/projects/SustainableInclusiveEconomicDevelopmentProgramme/TheKidepoCriticalLandscapeProject.html). We are truly thankful to everyone who participated in this study tirelessly to this point.

Cite this paper: Okurut, I. , Okullo, J. , Waiswa, D. and Muyizzi, J. (2020) Modelling the Potential Distribution of Vitellaria paradoxa subsp. nilotica (C.F. Gaertn) across the Kidepo Landscape of Uganda in the Face of Climate Change. Journal of Geoscience and Environment Protection, 8, 14-24. doi: 10.4236/gep.2020.88002.
References

[1]   Allal, F., Sanou, H., Millet, L., Vaillant, A., Camus-Kulandaivelu, L., Logossa, Z. et al. (2011). Past Climate Changes Explain the Phylogeography of Vitellaria paradoxa over Africa. Heredity, 107, 174-186.
https://doi.org/10.1038/hdy.2011.5

[2]   Araújo, M. B., Cabeza, M., Thuiller, W., Hannah, L., & Williams, P. H. (2004). Would Climate Change Drive Species Out of Reserves? An Assessment of Existing Reserve- Selection Methods. Global Change Biology, 10, 1618-1626.
https://doi.org/10.1111/j.1365-2486.2004.00828.x

[3]   Araújo, M. B., Pearson, R. G., Thuiller, W., & Erhard, M. (2005). Validation of Species-Climate Impact Models under Climate Change. Global Change Biology, 11, 1504- 1513. https://doi.org/10.1111/j.1365-2486.2005.01000.x

[4]   Ayebare, S., Ponce-Reyes., Segan, D. B., Watson, J. E. M., Possingham, H. P., Seimon, A., & Plumptre, A. J. (2013). Identifying Climate Resilient Corridors for Conservation in the Albertine Rift. Unpublished Report by the Wildlife Conservation Society to MacArthur Foundation.

[5]   Buyinza, J., & Okullo, J. B. L. (2015). Threats to Conservation of Vitellaria paradoxa subsp. nilotica (Shea Butter) Tree in Nakasongola District, Central Uganda. International Research Journal of Environment Sciences, 4, 28-32.

[6]   Cahill, A. E., Aiello-Lammens, M. E., Fisher-Reid, M. C., Hua, X., Karanewsky, C. J., Ryu, H. Y. et al. (2013). Causes of Warm-Edge Range Limits: Systematic Review, Proximate Factors and Implications for Climate Change. Journal of Biogeography, 41, 429-442.
https://doi.org/10.1111/jbi.12231

[7]   Elith, J., & Leathwick, J. R. (2009). Species Distribution Models: Ecological Explanation and Prediction across Space and Time. Annual Review of Ecology, Evolution, and Systematics, 40, 677-697.
https://doi.org/10.1146/annurev.ecolsys.110308.120159

[8]   Elith, J., Phillips, S. J., Hastie, T., Dudík, M., Chee, Y. E., & Yates, C. J. (2011). A Statistical Explanation of MaxEnt for Ecologists. Diversity and Distributions, 17, 43-57.
https://doi.org/10.1111/j.1472-4642.2010.00725.x

[9]   Ferrier, S., Watson, G., Pearce, J., & Drielsma, M. (2002). Extended Statistical Approaches to Modelling Spatial Pattern in Biodiversity in Northeast New South Wales. I. Species-Level Modelling. Biodiversity & Conservation, 11, 2275-2307.
https://doi.org/10.1023/A:1021302930424

[10]   Grossiord, C., Sevanto, S., Adams, H. D., Collins, A. D., Dickman, L. T., McBranch, N. et al. (2017). Precipitation, Not Air Temperature, Drives Functional Responses of Trees in Semi-Arid Ecosystems. Journal of Ecology, 105, 163-175.
https://doi.org/10.1111/1365-2745.12662

[11]   Hall, J., Aebischer, P., Tomlinson, H., Osei-Amaning, E., & Hindle, J. (1996). Vitellaria paradoxa: A Monograph. Cardiff, Wales: School of Agricultural and Forest Sciences, University of Wales.

[12]   IPCC (2007). Summary for Policymakers. In: S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K. B. Averyt, M. Tignor, & H. L. Miller (Eds.), Climate Change 2007: The Physical Science Basis. Contribution of WorkingGroup I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge and New York: Cambridge University Press.

[13]   Kelly, A. E., & Goulden, M. L. (2008). Rapid Shifts in Plant Distribution with Recent Climate Change. Proceedings of the National Academy of Sciences of the United States of America, 105, 11823-11826.
https://doi.org/10.1073/pnas.0802891105

[14]   Kremen, C., Cameron, A., Moilanen, A., Phillips, S. J., Thomas, C. D., Beentje, H. et al. (2008). Aligning across Taxa Conservation Priorities with Tools in Madagascar Planning High-Resolution Planning Tools. Science, 320, 222-226.
https://doi.org/10.1126/science.1155193

[15]   McClean, C. J., Lovett, J. C., Kuper, W., Hannah, L., Sommer, J. H., Barthlott, W. et al. (2005). African Plant Diversity and Climate Change. Annals of the Missouri Botanical Garden, 92, 139-152.
http://www.jstor.org/tc/accept?origin=/stable/pdf/3298511.pdf

[16]   Naughton, C. C., Lovett, P. N., & Mihelcic, J. R. (2015). Land Suitability Modeling of Shea (Vitellaria paradoxa) Distribution across Sub-Saharan Africa. Applied Geography, 58, 217-227. https://doi.org/10.1016/j.apgeog.2015.02.007

[17]   NEMA (2011). Conservation and Sustainable Use of the Threatened Savanna Woodland in the Kidepo Critical Landscape in North Eastern Uganda. Kampala, Uganda.

[18]   Okiror, P., Agea, J. G., Okia, C. A., & Okullo, J. B. L. (2012). On-Farm Management of Vitellaria paradoxa C. F. Gaertn. in Amuria District, Eastern Uganda. International Journal of Forestry Research, 2012, Article ID: 768946.
https://doi.org/10.1155/2012/768946

[19]   Okullo, J. B. L., Hall, J. B., & Obua, J. (2004). Leafing, Flowering and Fruiting of Vitellaria paradoxa subsp. nilotica in Savanna Parklands in Uganda. Agroforestry Systems, 60, 77-91. https://doi.org/10.1023/B:AGFO.0000009407.63892.99

[20]   Okullo, J. B. L., Ongom, B., Waiswa, D., Alum, W., & Obua, J. (2012). Utilisation and Conservation of Shea Butter Trees (Vitellaria paradoxa) in and around Aloi Internally Displaced Persons’ Camps in Lira District, Northern Uganda. In Third RUFORUM Biennial Meeting (pp. 22-48).

[21]   Orwa, C., Mutua, A., Kindt, R., & Jamnadass, R. S. A. (2009). Agroforestree Database: A Tree Reference and Selection Guide Version 4.0. Nairobi: World Agroforestry Centre ICRAF.

[22]   Pearson, R. G., Dawson, T. P., Pearson, R. G., & Dawson, T. P. (2003). Predicting the Impacts of Climate Change on the Distribution of Species: Are Bioclimate Envelope Models Useful? Global Ecology and Biogeography, 12, 361-371.
https://doi.org/10.1046/j.1466-822X.2003.00042.x

[23]   Peterson, T., Peterson, A. T., & Soberón, J. (2012). Species Distribution Modeling and Ecological Niche Modeling: Getting the Concepts Right. Natureza & Conservação, 10, 102-107. https://doi.org/10.4322/natcon.2012.019

[24]   Phillips, S. J., Dudík, M., & Schapire, R. E. (2004). A Maximum Entropy Approach to Species Distribution Modeling. Proceedings of the 21st International Conference on Machine Learning, New York, July 2004, 83.
https://doi.org/10.1145/1015330.1015412

[25]   Qin, A., Liu, B., Guo, Q., Bussmann, R. W., Ma, F., Jian, Z. et al. (2017). Maxent Modeling for Predicting Impacts of Climate Change on the Potential Distribution of Thuja sutchuenensis Franch., an Extremely Endangered Conifer from Southwestern China. Global Ecology and Conservation, 10, 139-146.
https://doi.org/10.1016/j.gecco.2017.02.004

[26]   Radosavljevic, A., & Anderson, R. P. (2014). Making Better Maxent Models of Species Distributions: Complexity, Overfitting and Evaluation. Journal of Biogeography, 41, 629-643. https://doi.org/10.1111/jbi.12227

[27]   Thomas, C. D., Cameron, A., Green, R. E., Bakkenes, M., Beaumont, L. J., Collingham, Y. C. et al. (2004). Extinction Risk from Climate Change. Nature, 427, 145-148.
http://eprints.whiterose.ac.uk/83/
https://doi.org/10.1038/nature02121


[28]   Thuiller, W. (2004). Patterns and Uncertainties of Species’ Range Shifts under Climate Change. Global Change Biology, 10, 2020-2027.
https://doi.org/10.1111/j.1365-2486.2004.00859.x

 
 
Top