Within savanna environments, movements of elephant are influenced by changes in climate especially seasonal rainfall. In this study, we investigated the possible changes in elephant population based on projected rainfall changes using regional climate models (RCM) and Representative Concentration Pathways (RCPs). The relationship between elephant and rainfall was modelled against annual, wet season, dry season rainfall based on various time lags. Future relation between elephant and rainfall was projected based on three RCPs; 2.6, 4.5 and 8.5. There was a strong linear relationship between elephant and October-November-December (OND) rains with time lag of 13 years (Y = −4016.43 + 19.11x, r 2 = 0.459, P = 0.006). The rainfall trends for RCP 2.6 and 4.5 showed a slight increase in annual rainfall for the period 2006-2100 but driven by OND increases. Rainfall increase for RCP 8.5 was significant and was driven by increase in both March-April-May (MAM) and OND. These rainfall dynamics had influence on the projected elephant population in the Amboseli ecosystem. For RCP 2.6 and 4.5 the elephant population increase was 2455 and 2814 respectively. RCP 8.5 elephant population doubled to an average of 3348 elephants. In all the RCPs there are seasonal and yearly variations and absolute number varies from the average. The range of variation is small in RCPs 2.6 and 4.5 compared to RCP 8.5. Evidently, elephant population will increase based on projected rainfall projections surpassing park capacity. It therefore, requires that the Park authority put in place measures that could contain these numbers including opening of blocked wildlife corridors, maintain the cross border movement of Amboseli elephant with Tanzania in that case ensure there is no poaching. Lastly, work with local communities so that they can benefit from tourism through setting up conservancies through which they could minimize the human elephant conflicts based on the projected elephant population.
Global land surface warming and increasing temperature events are expected to occur more frequently and more extremely causing changes in biodiversity and altering movement and survival of large herbivores [
The various ecosystems in the world are expected to experience varying stress from climate change impacts. Species responses include toleration, habitat shift or extinction, which are outcomes of exceeding species tolerance level of water and temperature stresses [
Africa is extremely susceptible to the impacts of climate change [
While there is fairly inadequate evidence of current extinctions being caused by climate change, investigations insinuate that climate change could surpass habitat destruction as the greatest global threat to biodiversity over the next few decades [
Globally, the African elephant stands out as the largest mammal which was once widely distributed throughout Sub-Saharan Africa [
In this study we investigated the possible changes in elephant population based on projected rainfall changes based on regional climate models (RCM). The simulations used for the projections are the Representative Concentration Pathways (RCPs), which are based on radiative forcings (globally radiative energy imbalance) measured in Wm−2 by the year 2100 [
The Amboseli ecosystem is situated in the southwest of Kenya, bordering Tanzania. Geologically, the ecosystem covers part of a dry Pleistocene lake basin, which has a temporary lake that floods during years of heavy rainfall. Rainfall is bi-modal with short rains normally occurring in November and long rain period starts in March to May [
The ecosystem covers an area of approximately 5700 km2 stretching between Chyulu Hills and Tsavo West National Parks South to Mt Kilimanjaro in Tanzania (
Rainfall in Amboseli ecosystem is bi-modal with long rain period being observed in March-April-May (MAM) and short rains in October-November-December (OND). The short dry seasons are January-February (JF) and long-dry season occurs in June-July-August-September (JJAS). The average annual rainfall for Amboseli ecosystem is 582 mm (SE = 25) with MAM contributing about average 228 mm (SE = 14), OND contributing 262 mm (SE = 18), JJAS contributing 20 mm (SE = 3) and JF 67 mm (SE = 9). The Amboseli ecosystem is one of the few areas in Kenya where the short rains are heavier than the long rains that take place between March and May in many parts of Kenya. The Amboseli ecosystem falls in the rain-shadow of Mt Kilimanjaro placing it amongst the driest places in Kenya. However, water flowing underground from
Mt Kilimanjaro wells up here in a series of lush swamps that provide dry season water and forage for wildlife. This attracts high concentrations of migratory animals during the dry season.
In the present study we used elephant counts from 16 censuses conducted in the ecosystem between 1977 and 2014. Eight censuses were conducted in March-April-May (long rains), six in October-November-December (short rains) and two conducted in January-February (short dry season). The data was collected by the Directorate of Resource Surveys and Remote sensing (DRSRS) using a Systematic Reconnaissance Flight (SRF) [
The rainfall and temperature data were extracted from the Geospatial Climate (GeoCLIM) software. GeoCLIM is a regional gridded climate data set tool that interpolates time-series grids of precipitation and temperature values from station observations and associated satellite imagery, elevation data, and other spatially continuous fields [
The IPCC (2013) [
In this study we used the simulated data from the Ross by Center Regional Atmospheric Model (RCA4) driven by the Earth system version of the Max Planck Institute for Meteorology (MPI-ESM-LR) coupled global climate model from the on-going Coordinated Regional Downscaling Experiment (CORDEX) project. The goal of the RCM program is to advance the predictive understanding of Earth’s climate by focusing on scientific analysis of the dominant sets of governing processes that describe climate change on regional scales. The model was integrated into the CORDEX-Africa domain, with a horizontal grid spacing of 0.44 degrees―translates to a 50 by 50 km grid (refer to [
The total animal population size, its variance and standard error are calculated using jolly’s method 2 for aerial transects of unequal length [
V a r ( Y ^ ) = N ( N − n ) n ( s y 2 − 2 R ^ s z y + R ^ 2 s z 2 ) and standard deviation S E ( Y ^ ) = √ ( V a r ( Y ^ ) ) . Z is the area of the census zone (e.g. county) and R ^ = ∑ y ∑ z is the sample population density calculated as the total number of all
animals counted in each sampling unit y divided by the area of each sampling unit z summed over all the units included in the survey sample. N is the population of all the sampling units in the census zone whereas n is the number of sampling units included in the survey sample. s y 2 is the sample variance of the number of animals counted in all the sampled units while s z 2 is the variance of the area of all the sampling units included in the survey sample. s z y is the covariance between the number of animals counted and the area of each sampling unit.
We mapped elephant distribution based on aerial counts. The maps were based on grids cells of 5 by 5 km. For any of the observed elephant we calculated its population in a grid and mapped those using ArcGIS.
The monthly rainfall for all the three RCPs (2.6, 4.5 and 8.5) was analysed using linear and quadratic models for the period 2006-2100. The corrected Akaike Information Criterion (AICc) was used to choose between the supporting models. The model with the least AIC was selected as the supporting model [
We related annual elephant population elephant to rainfall components. The rainfall components defined as annual (October-September), short rains (October-December), long rains (March-May), long dry season (June-September), and short dry season (January-February) (see examples used in Lake Nakuru― [
We tested the differences between projected elephant population between RCPs: 2.5, 4.5 and 8.5 based on two-sample t-test. The null hypothesis tends that there is no difference between the two RCPs elephant population or more formally, that the difference is zero. Formally expressed as H0: p1 − p2 = 0, where p1 is the first population and p2 is the second population.
In this study we mapped the distribution of elephants between 1977 and 2014. The elephant in Amboseli ecosystem disperse with the onset of rains, but there was little indication of evidence of large scale elephant migration in Amboseli. It is reported that in Amboseli the elephant seasonal range is influenced to some extent by the intensity of poaching, rainfall and the development of artificial water resources within the park and surrounding areas. In the last 2 decades the elephant population has increased and also they are moving out of the park more often.
Our analysis indicates two patterns of elephant distribution in the Amboseli ecosystem. These two patterns indicate that the elephant can be confined to the park or at times move out. Most of the elephants were confined to the park in the 1970s and to late 1980s. The confinement of elephants was mainly because of the heavy poaching in the 1970s and early 1980s. However after 1991 we have observed large dispersal of elephant beyond the park and especially between the
periods 2000 to 2014. The total elephant range increased significantly and with many elephants sighted in the park and in adjacent group ranches of Lengesim, Imbirikani, Kuku Olgulului and Kimana (refer to
The historical trends of monthly rainfall in the Amboseli ecosystem shows slight increases rainfall in March, August and December, slight decline in rainfall in January, February, April, May, September, October and November (
The relationship between elephant population and temperature were not significant as tested with moving average from one to twelve months (refer to
The projected seasonal rainfall in Amboseli ecosystem varies across RCPs 2.6, 4.5 and 8.5.
RCP 4.5 annual rainfall is projected to increase marginally and is mainly driven by increases in rainfall for the OND season (
Rainfall component | Constant | Slope | SE | F-ratio | P value | r-squared |
---|---|---|---|---|---|---|
OND8 | −1428.83 | 9.4658 | 4.7722 | 3.93 | 0.069 | 0.232 |
OND10 | −1874.40 | 10.977 | 5.172 | 4.50 | 0.054 | 0.257 |
OND12 | −2757.43 | 14.744 | 5.107 | 8.33 | 0.012 | 0.390 |
OND13 | −4016.43 | 19.911 | 6.001 | 11.01 | 0.006 | 0.459 |
OND14 | −4476.79 | 21.325 | 6.881 | 9.61 | 0.008 | 0.425 |
OND15 | −3509.13 | 17.558 | 8.934 | 3.86 | 0.071 | 0.229 |
Annual12 | 6082.95 | −7.774 | 3.569 | 4.74 | 0.048 | 0.267 |
Annual13 | 6588.23 | −8.519 | 3.824 | 4.96 | 0.044 | 0.276 |
Annual14 | 7528.14 | −9.875 | 3.009 | 10.77 | 0.006 | 0.453 |
Annual15 | 8425.17 | −11.233 | 2.460 | 20.56 | 0.001 | 0.616 |
Month | Equation | F-Ratio | P-Value |
---|---|---|---|
Annual | Y = 763.587 − 0.0640x | 0.0154 | 0.9014 |
MAM | Y = 346.714 − 0.0615x | 0.0434 | 0.8354 |
JJAS | Y = 94.4536 − 0.0295x | 0.1308 | 0.7183 |
OND | Y = 485.237 − 0.0799x | 0.0445 | 0.8333 |
Month | Equation | F-Ratio | P-Value |
---|---|---|---|
Annual | Y = 287.983 + 0.1968x | 0.0990 | 0.7536 |
MAM | Y = 465.295 − 0.0996x | 0.0681 | 0.7946 |
JJAS | Y = 81.2521 − 0.0236x | 0.1082 | 0.7428 |
OND | Y = −25.1041 + 0.2879x | 0.6140 | 0.4352 |
Month | Equation | F-Ratio | P-Value |
---|---|---|---|
Annual | Y = −400.653 + 2.3022x | 15.2661 | 0.0002 |
MAM | Y = −344.632 + 0.2937x | 0.7950 | 0.3748 |
JJAS | Y = 390.734 − 0.1767x | 6.4588 | 0.0126 |
OND | Y = −3017.99 + 1.6509x | 17.251 | 0.0001 |
Amongst the three RCPs the RCP 8.5 is projected will have an increase in rainfall significantly for the period 2006-2100 (P = 0.0002; refer
RCPs. The OND maximum rainfall is projected at 782 mm and minimum 167 mm while the maximum in MAM is 523 mm and minimum is 51 mm (
The three RCPs project an increase in elephant population in the Amboseli Ecosystem (
2056 after this decline to about 2160 elephants by 2070. There will be another phase of increase of elephants and by 2080 it is projected there will be 2980 elephant but after this period it will decline to about 1870 animals by 2100 (
The trend for RCP 4.5 shows three phases―the first phase is between 2019 and 2050 where the projected population of elephants is likely to increase from 2300 to 3020 elephants by 2050 (
As indicated in the analysis of rainfall that RCP 8.5 both the annual and OND is projected will increase significantly. The population of elephant is mirrored to the rainfall pattern. The population of elephant under RCP 4.5 has seven different trends or broken stick (
It is broadly recognized that, within savanna environments, movements of elephant are influenced by changes in seasonal rainfall [
Maximum elephant population | Minimum elephant population | |||||
---|---|---|---|---|---|---|
RCP | Mean | N | Year | Number | Year | Number |
2.6 | 2457 (SD 340) | 82 | 2056 | 3510 | 2100 | 1870 |
4.5 | 2815 (SD 322) | 82 | 2075 | 3725 | 2039 | 2216 |
8.5 | 3349 (SD 1226) | 82 | 2088 | 5820 | 2036 | 1272 |
in Amboseli ecosystem as indicated in the distribution elephants between 1977 and 2014. The distribution was mainly governed by rainfall season and security. In the 1970s, 1980s and early 1990s the elephant were mostly found in the park and later years the elephant moved more often outside the park because of better protection [
Therefore, in this paper we focused on analysing the relationships between elephant population and rainfall and temperature in the Amboseli ecosystem. Analysis of historical rainfall data of the study area from 1960-2014 showed a bimodal pattern of rainfall with two rainy seasons and two dry seasons [
In this study our analysis has established a strong positive relationship between elephant population and OND season with a lag of 13 years. Previous studies in the savannah ecosystem such as in the Tsavo [
Other drivers of changes to elephant population are extreme events such as droughts and high rainfall as experienced during the El Niño and increasing temperatures. A number of studies have also shown a positive or negative correlation between extreme events such as El Niño or droughts and elephant population. In Addo National Park the highest elephant population growth rate coincided with El Niño rains [
A study on potential impacts of projected temperature based on RCP 2.5, 4.5 and 8.5 changes on the elephant range showed minimal impacts on elephant whilst for other large herbivore the impacts on their range will significant [
Therefore, in the study we focused on analysing the potential projected impacts of rainfall on the elephant population in the Amboseli ecosystem. It is important to project the potential changes of climate on both animals and plants. There have been efforts to project rainfall and temperatures based on Global Circulation Models (GCMs) in the last six decades [
The three climate models selected for this study had variations in projections for rainfall between 2006 and 2100 for Amboseli ecosystem. Projections for RCP 2.6 indicate an overall decline in rainfall driven by projected declines of rainfall in MAM, JJAS and OND. RCP 4.5 predicts that there will be overall marginal increase in annual rainfall which is driven by increase in OND. The mean annual rainfall for RCP 4.5 is 692 mm, for OND is 340 mm, for MAM is 261 mm and for JJAS is 33 mm. Projections for RCP 8.5 indicate a statistically significant increase in annual rainfall which is mainly driven by significant projected increase in OND rainfall. The OND maximum rainfall is projected at 782 mm and minimum 167 mm while the maximum in MAM is 523 mm and minimum is 51 mm. [
Increase in intensity of extreme precipitation events has been found to be related to increases in temperatures [
Since the rainfall pattern for RCP 2.6 and RCP 4.5 were almost similar the elephant population projection and pattern are much similar. RCP 2.6 projects an average elephant population at 2457 with a peak population being realized in 2056 at a population density of 3510 elephants while RCP 4.5 projects an average population density of 2815 elephants in the long term with the highest population being realized in 2075 at 3725 elephants) and lowest population is projected by 2100 at a population density of 1870 elephants. The population of elephant under RCP 8.5 reveals a different trend; initially the population is expected to drop from a population of 2715 elephants to 1270 elephants by 2036 then an increase to a maximum of 5820 elephants by 2088. After this period it is projected the elephants will decline to a population of about 4130 by 2100. In between 2045 and 2060 the average population is projected around 2966 elephants and another stable state of population will be between 2065 and 2080 where is projected the elephant average population will be 4039 elephants. Recent study by [
In terms of breeding and genetics a change in the intensity or duration of the rainy versus drought seasons could change relative breeding rates and, hence, genetic structures in these populations [
As indicated in this study that it is projected there is likely possibilities of increase in elephant population based on all the three RCPs 2.6, 4.5 and 8.5 in the Amboseli ecosystem which could boost tourism activities in the area. However, sustainability of the population in the limited park space will be a big challenge because the Amboseli Park alone will not be able to sustain the numbers and therefore most elephants will have to move into the dispersal areas outside the park. It is therefore likely that there will be an increase in human elephant conflicts as the human population in rangelands is expected to continue increasing in the coming years [
In addition encroachment on elephant habitats that is blocking migration routes [
In most savanna ecosystems land fragmentation is recognized to be a major threat to wildlife distribution [
Further, lack of direct benefits from wildlife based-tourism has prompted the local people to shift to horticultural farming and leasing of farms to commercial farmers [
The projected elephant population based on the three RCPs indicates moderate increases in elephant for RCP 2.6 and RCP 4.5 and high increase for RCP 8.5. On average the elephant population for RCP 2.6 is projected at 2457 while for RCP 4.5 is about 2815 elephant and RCP 8.5 are at 3349 elephants. The projected increase can be managed through engaging the local communities, since elephants disperse to the community and private lands and their survival depends on the goodwill of the local community. This implies that any conservation action should have a human socio-economic dimension and not only focus on elephant welfare, but that of the local community who bear the cost of elephant conservation.
Population predictions or projections based on time series models, linear models, deterministic or stochastic structured models and simulations play an important role in population management [
As indicated in the rainfall analysis, for RCP 8.5, both the annual and OND are projected to increase significantly. The population of elephant is mirrored to the rainfall pattern. Research analysis established a strong positive relationship between elephant population and OND season with a lag of 13 years. Elephant is one of the key wildlife species that has a long life span and this study projects an increase in elephant population based on all the three RCPs 2.6, 4.5 and 8.5 in the Amboseli ecosystem. The increase could boost tourism activities in the area. However, sustainability of the population in the limited park space will be a big challenge because the Amboseli Park alone will not be able to sustain the numbers and therefore most elephants will have to move into the dispersal areas outside the park. It is therefore likely that there will be an increase in human elephant conflicts as the human population in rangelands is expected to continue increasing in the coming years. Therefore, there is need to have national and county policies that will allow translocation of animals to other locations to ease the pressure on the diminishing resources. The County of Kajiado should also come up with a strategic plan that focuses on encouraging local community not to shift to crop cultivation in entirety, instead support the community to come together and form conservancies where the main land use is wildlife conservation so as to secure the wildlife corridors. This can be enhanced by Economic incentives to communities that promote wildlife conservation as an important resolution to encourage involvement of local communities in biodiversity conservation efforts. Finally, there is need to enhance the trans-boundary agreement and control of poaching between Kenya and Tanzania for the population to grow and benefit the tourism sectors of both counties.
The authors gratefully acknowledge funding support provided through the National commission for Science, Technology and Innovation (NACOSTI) under the Research Endowment Fund 6th call postgraduate 2014/2015 grants. The authors also wish to thank anonymous reviewers for their suggestions, which resulted in the improvement of the manuscript. We thank the Directorate of Resource Surveys and Remote Sensing of Kenya (DRSRS) for permission to use the aerial survey data. We are grateful to Hussen Seid and Herbert Misiani of ICPAC for providing the projected rainfall data. Mohamed Said was supported by the Pathways to Resilience in Semi-Arid Economies (PRISE) Project 5―Property regimes, investments and economic development in the context of climate change in semi-arid lands of East Africa.
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
The authors declare that they have no conflict of interest regarding publication of this article.
Aduma, M.M., Said, M.Y., Ouma, G., Wayumba, G. and Njino, L.W. (2018) Projection of Future Changes in Elephant Population in Amboseli under Representative Concentration Pathways. American Journal of Climate Change, 7, 649-679. https://doi.org/10.4236/ajcc.2018.74040
Month | Equation | F-Ratio | P-Value |
---|---|---|---|
January | Y = 337.476 − 0.148x | 0.120 | 0.729 |
February | Y = 329.801 − 0.151x | 0.312 | 0.578 |
March | Y = −200.929 + 0.145x | 0.077 | 0.782 |
April | Y = 853.925 − 0.381x | 0.481 | 0.491 |
May | Y = 787.946 − 0.374x | 1.676 | 0.201 |
June | Y = 244.554 − 0.121x | 3.979 | 0.051 |
July | Y = 254.688 − 0.127x | 5.672 | 0.021 |
August | Y = −6.734 + 0.006x | 0.012 | 0.914 |
September | Y = 211.626 − 0.104x | 1.573 | 0.215 |
October | Y = 538.203 − 0.254x | 0.709 | 0.403 |
November | Y = 1399.651 − 0.646x | 1.156 | 0.287 |
December | Y = −123.509 + 0.119x | 0.066 | 0.798 |
Month | Equation | F-Ratio | P-Value |
---|---|---|---|
January | Y = −142.717 + 0.082x | 0.602 | 0.440 |
February | Y = −20.183 + 0.025x | 0.046 | 0.831 |
March | Y = 54.053 + 0.001x | 0.000 | 0.997 |
April | Y = -211.375 + 0.157x | 0.506 | 0.478 |
May | Y = 504.036 − 0.220x | 2.546 | 0.114 |
June | Y = 11.645 − 0.001x | 0.001 | 0.982 |
July | Y = 10.758 − 0.004x | 0.049 | 0.826 |
August | Y = 5.334 + 0.000x | 0.000 | 0.985 |
September | Y = 66.299 − 0.025x | 0.273 | 0.602 |
October | Y = −131.986 + 0.107x | 0.422 | 0.517 |
November | Y = 1098.179 − 0.451x | 3.382 | 0.069 |
December | Y = −480.955 + 0.264x | 2.851 | 0.095 |
Month | Equation | F-Ratio | P-Value |
---|---|---|---|
January | Y = −22.115 + 0.026x | 0.053 | 0.819 |
February | Y = −14.593 + 0.006x | 0.004 | 0.952 |
March | Y = −86.153 + 0.069x | 0.210 | 0.648 |
April | Y = 133.989 + 0.004x | 0.000 | 0.989 |
May | Y = 417.550 − 0.172x | 1.430 | 0.235 |
June | Y = 46.587 − 0.019x | 0.393 | 0.532 |
July | Y = 3.364 − 0.001x | 0.000 | 0.983 |
August | Y = 5.594 − 0.001x | 0.001 | 0.978 |
September | Y = 25.706 − 0.005x | 0.007 | 0.933 |
October | Y = 35.804 + 0.030x | 0.028 | 0.867 |
November | Y = −235.315 + 0.202x | 0.577 | 0.450 |
December | Y = −51.530 + 0.056x | 0.187 | 0.667 |
Month | Equation | F-Ratio | P-Value |
---|---|---|---|
January | Y = −560.252 + 0.288x | 6.674 | 0.011 |
February | Y = −474.387 + 0.246x | 4.650 | 0.034 |
March | Y = −787.628 + 0.413x | 5.883 | 0.017 |
April | Y = 141.288 − 0.003x | 0.000 | 0.900 |
May | Y = 301.708 − 0.117x | 0.570 | 0.452 |
June | Y = 14.681 − 0.003x | 0.007 | 0.935 |
July | Y = 30.474 − 0.014x | 0.746 | 0.390 |
August | Y = 74.154 − 0.034x | 2.705 | 0.103 |
September | Y = 271-425 − 0.126x | 8.538 | 0.004 |
October | Y = -402.387 + 0.239x | 1.653 | 0.202 |
November | Y = −1461.509 + 0.810x | 8.719 | 0.004 |
December | Y = −1154.100 + 0.602x | 10.94 | 0.001 |
Temperature | Equation | F-Ratio | P-Value |
---|---|---|---|
MonthT1 | Y = −2121.957 + 103.151x | 1.018 | 0.333 |
MonthT2 | Y = −2110.941 + 102.035x | 0.889 | 0.364 |
MonthT3 | Y = −2855.950 + 127.978x | 1.855 | 0.198 |
MonthT4 | Y = −1946.839 + 98.369x | 1.067 | 0.322 |
MonthT5 | Y = −1743.923 + 92.378x | 0.831 | 0.374 |
MonthT6 | Y = −1232.805 + 74.864x | 0.465 | 0.508 |
MonthT7 | Y = −1253.273 + 75.607x | 0.341 | 0.569 |
MonthT8 | Y = −1743.729 + 92.873x | 0.283 | 0.604 |
MonthT9 | Y = −3852.064 + 166.924x | 0.469 | 0.506 |
MonthT10 | Y = −7301.423 + 288.319x | 1.038 | 0.328 |
MonthT11 | Y = −8802.127 + 341.049x | 1.379 | 0.262 |
MonthT12 | Y = −7837.906 + 306.831x | 1.125 | 0.309 |