Perceptions of Ecosystem Services and Climate Change in the Communities Surrounding Mt. Kenya and Mt. Elgon, Kenya

Abstract

Local observations of climate change can be a critical resource for understanding the impacts of climate change, particularly in data-scarce areas. This study examines local observations of climate change in two montane areas of Kenya- Mt. Kenya and Mt. Elgon. Household questionnaires, focus group discussions, and interviews were used to explore local perceptions of ecosystem services and changes to those services. Results showed that communities had a strong appreciation for ecosystem services and had witnessed major changes in those services. Water provision was seen as the most important service and the one that had changed the most. Other observations of changes included shifts in species ranges, weather patterns, temperature, and soil properties. These changes are consistent with predictions from climate models, but they provide context-specific nuance that the models cannot provide. Spatial variables, such as distance to road and the alpine zone, played as large or larger role in affecting perceptions as demographics, which further points to the importance of context in understanding climate changes. Those that interacted with the mountains the most—the mountain guides—had particularly revealing observations of changes; these types of observations can be critical to understand and prepare for changes in the alpine areas of Kenya.

1. Introduction

Mountains around the world have an importance disproportionate to their global coverage. Although they only cover a fraction of the world’s surface, they provide over half of the world’s freshwater and biodiversity hotspots [1] due to orographic effects and steep altitudinal gradients [2,3]. Climate change is expected to impact mountain systems more than lowland areas. This is because the actual level of warming has been modelled to be higher in these areas (higher exposure), and because those areas are more susceptible to that warming (higher sensitivity) [1]. The Intergovernmental Panel on Climate Change (IPCC) predicts global climate to warm by at least 1.5 °C over the next century, but the warming in mountain areas—particularly in the tropics—is projected to be 2–3 times higher [4,5,6,7,8]. This warming will undoubtedly affect the biophysical properties of the mountains, which will in turn impact ecosystem services critical to the surrounding communities.

The concept of Ecosystem Services (ES) has emerged as a useful way of assessing climate change impacts on local communities [9,10]. While ES originally was understood primarily in financial terms [11,12], more recently the concept has been broadened to look at the full scope of nature’s contribution to livelihoods [13,14]. Local communities are much more likely to track closely those services that are of importance to them and therefore will be tuned into even minute changes in those services [15,16,17]. Across the globe, there is general agreement between local observations of climate change and model predictions [18,19]. Worldwide, the dominant change that has been observed at the local level is the alteration of precipitation patterns, but a variety of auxiliary changes have also been observed, such as changes in plant phenology and soil properties [19]. Local observations have enabled elucidation of local specificity that is currently lacking in model outputs [18]; this data is particularly important in rural parts of Africa where empirical data is limited [19], and yet the impacts of climate change are potentially greater due to higher vulnerabilities [20].

The isolated mountains of East Africa are of particular concern because biogeographically they function as islands [21]; as these ‘sky islands’ shrink due to climate change, their services may be lost forever. Mt. Kenya and Mt. Elgon are the two tallest mountains in Kenya and provide essential services to a large part of the country [22]. The impacts of climate change on the ecosystem services of these two critical mountain areas are still not fully understood, and climate model predictions may not be so useful given the heterogeneous and complex landscapes. Local observations can provide a critical role in this respect, acting as surrogate monitoring stations. In addition, how locals perceive their environment and changes to that environment can reveal something about their vulnerability and adaptive capacity. This study seeks to look at how communities surrounding these two mountains value ecosystem services and what changes they have perceived over the years. It seeks to analyze factors affecting these perceptions, how they vary according to spatial and demographic factors, and how they compare to local perceptions in other tropical mountain ecosystems.

2. Materials and Methods

2.1. Study Area

Mt. Kenya and Mt. Elgon are located in central and western Kenya, respectively (Figure 1). Mt. Kenya lies on the equator at 37.5° E longitude, and rises to 5199 m above sea level; Mt. Elgon lies at 1° N, 34.5° E, and straddles the boundary with Uganda. It tops out at 4320 m above sea level. The climate of both is characteristic of tropical mountains, with temperature and precipitation driven largely by elevation. The base of both mountains is warm and dry with savannah-like characteristic, while the peaks are alpine moorlands [23,24]. Rains on the mountain follow the bimodal seasonality of the lowlands with short rains in November –December, and long rains in April–May, although the onset and duration of these season also changes with elevation [25].

Mt. Kenya and Mt. Elgon (the Kenya side) are both managed as protected areas by the national government through the Kenya Wildlife Service (KWS) and the Kenya Forest Service (KFS) [26]. The forest reserve areas under KFS allow limited resource extraction whereas the national parks do not allow any resource extraction. Mt. Elgon also has a national reserve that is held in trust on behalf of the local communities. This national reserve covers much of the alpine zone and is inhabited by the pastoralists who traditionally lived there. Limited resource extraction, especially grazing, is permitted, but no farming is allowed [27].

Both mountains are surrounded by a high population density (300–400 persons/km²) and rich agricultural land [28]. This study focuses in particular on the communities to the southeast side of Mt. Elgon (Bungoma County) and the western side of Mt. Kenya (Nyeri County), which fall within a 20 km buffer of the alpine zone (Figure 1). These communities are primarily agriculturalists and the main ethnic groups are Kikuyu for Nyeri, and Luhya and Sabaots (Kalenjin) for Bungoma [29,30].

2.2. Quantitative Methods

Questionnaires were administered in March–April 2021 to households that fell within 20 km of the alpine zones of both mountains, using a cluster sampling strategy. A sample of 209 was chosen within this study area using the formula:

n = z² × (p × q)/d² × DEFF

(1)

with a desired precision (d) of 5%, an estimated prevalence (p) of 10%, a non-prevalence (q) of 90%, a normal distribution score (z) of 1.96, and a design factor (DEFF) of 1.5 [31]. The sub-location administration unit was used as the sample unit, and clusters were chosen with probability proportional to size based on the household population from the 2019 census [28]. The number of households to be sampled within each cluster was determined again according to the size of the cluster (see Tables S1 and S2). In each chosen sub-location, questionnaires were administered in villages that were closest to the mountains.

The primary goal of the questionnaires was to get a sense of what ecosystem services were recognized by locals, and what changes had been observed in these services over their lifetimes. First, respondents were asked about their overall level of interaction with the mountain, and then they were given a chance in open-ended questions to detail the benefits that they received from the mountain and the major changes they had witnessed. Then respondents were asked to rank the importance and amount of change for a suite of 13 different ecosystem services: climate regulation, cultural heritage, education, food gathering, livestock forage, medicines, nutrient cycling, raw materials, recreation, tourism, water purification, water supply, and wildlife habitat (see Table S3 for the questionnaire tool; Supplementary Materials). Questionnaires were administered using Kobo Toolbox, an open-source set of tools for collecting survey data digitally in the field [32].

2.3. Qualitative Methods

Two focus group discussions (FGDs) were conducted for each mountain with prominent stakeholder groups. For Mt. Elgon, this consisted of a Community Forest Association (CFA) and a group of youth from a high mountain community. For Mt. Kenya, the groups were a Water Users Association (WUA) and a group of mountain guides. The participants for the two community associations (CFA and WUA) were generally older (median age 60 and 56, respectively), and were respected members of the community whose work in the association meant they routinely dealt with environmental issues affecting the area at the base of the mountain. The other two groups were younger (median age 31 and 34), but people who interacted with the alpine environment on a daily basis, either through grazing sheep or acting as mountain guides. Each group consisted of 5–6 individuals chosen to represent a fairly homogeneous, but not totally uniform, cross-section of the group. The community associations also included a chairman; while having an authority figure in the group can interfere with the egalitarian nature of the discussion, it is nonetheless appropriate to include such leaders in these local settings (see for instance [33,34]). Another limitation was the gender imbalance, as there was only one woman present in the focus groups. This was also due to cultural norms (i.e., see [34]), but is a significant drawback nonetheless. The discussions were conducted in Kiswahili or English depending on the preference of the group, and were kept to around 1 h in length to keep the discussion focused and on topic.

Key informant interviews (KII) were also conducted at each study area with specific individuals who were identified as being able to provide more in-depth information on a one-on-one basis. There were five interviewees: two retired chiefs, a community elder, an older mountain guide, and a CFA chairman. These individuals knew the area very well and had a wealth of historical knowledge about the mountain and its resources. Interview protocols were written up and used as a guide for the interviews.

2.4. Data Analysis

Data were analysed using a mixed-methods approach, where the qualitative data supplemented the questionnaire data to provide context (Figure 2). Quantitative analysis and visualization of the questionnaire data were done using R software v. 3.6.3 [35]. The data were graphically displayed through bar charts to visualize the Likert scale responses for each ecosystem service. For discussion purposes, ecosystem services were divided into 3 categories: provisioning, regulating, and cultural, following the grouping of the Millennium Ecosystem Assessment report [10]. There is an additional category of support services, which refers to those services which are not directly used by people but which drive the other services. No true support services were examined in this study, although there is considerable overlap between regulating and support services [10,36].

Factors influencing perceptions were assessed through an Ordinal Logistic Regression. Covariates of interest were demographic characteristics—age, gender, ethnicity, education, and occupation—as well as spatial variables—elevation and distance to rivers, towns, roads, and the alpine zone of the mountain. Categories were simplified into binary categories to ensure adequate representation. For the spatial factors, the median distance to each feature was calculated, and those less than the median distance were termed ‘Near’ and those further were ‘Far’. Qualitative analysis of interviews and focus group discussions was also done using R- with the R Qualitative Data Analysis (RQDA) package [35], to extract themes and explore relationships in the data [37].

3. Results

3.1. Demographics

Respondents to the questionnaires were largely young to middle-aged male farmers with a primary school level of education. The dominant ethnicities were Kikuyu for Mt. Kenya and Kalenjin (Sabaot) for Mt. Elgon. The majority of the respondents were permanent residents, having been born and raised in the area (

Table 1. Questionnaire Demographics (n = 209).

Category	Percent	Category	Percent
Age		Occupation
<30	31%	Farmer/Agriculturalist	57%
30–60	56%	Business owner	17%
>60	13%	Day/Unskilled laborer	9%
		Professional	5%
Gender		Government employee	1%
Female	22%	Other	11%
Male	78%
		Ethnicity
Education		Kalenjin	39%
Primary School	51%	Kikuyu	45%
Secondary School	39%	Meru	8%
Diploma	10%	Luhya	6%
University Degree	1%	Other	1%
Immigrant status (Did you move here?)
No	60%
Yes	40%

). These characteristics largely reflect the underlying demographics of the area, except for the low turnout for women; which was mostly a function of limitations in access, given the survey was undertaken by a foreign male.

Across demographics, there was a high level of interaction with the mountains. People attested that they visited the mountains frequently, that the mountains were very important in their lives, and that they had a high level of knowledge about them (Figure 3). Generally, those living further from the mountains did not visit as frequently and thus ranked their knowledge lower; however, the mountains were still seen as important in their lives.

3.2. Ecosystem Service Appreciation

In terms of services, people mostly mentioned water and livestock forage as their most valued ecosystem services, and to a lesser extent, medicines, food gathering, and agriculture. Almost all (90%, n = 209) of the respondents mentioned water as a primary benefit of the mountain. When prompted with the list of ecosystem services, most of the respondents agreed that these too were important, ranking almost all positively as ‘important’ or ‘very important’ (Figure 4). Provisioning services (food gathering, raw materials, water supply, and livestock forage) and regulating services (climate regulation, water purification, and nutrient cycling) all received high ratings—over 88% positive ratings. Cultural services generally scored lower: tourism and recreation had just over 70% positive ratings, whereas cultural heritage had only 46% of respondents giving it an ‘important’ or ‘very important’ score. Medicine and education were the two exceptions to this overall trend. Medicine, a provisioning service, scored just 69% positive responses; whereas education, a cultural service, scored 92% positive ratings.

In the discussions and interviews, respondents made note of all three types of ecosystem services, provisioning, cultural, and regulating (

Table 2. Ecosystem services identified in Focus Group Discussions and Interviews.

Ecosystem Service	Category	Examples
Food Gathering	PROVISIONING	natural honey from two varieties of bees bamboo and other local plants eaten as a vegetable fishing in mountain streams collection of fruits and beans
Livestock Forage		grazing in the forest and moorland; salt for livestock from the caves; also freshwater in the caves collection of fodder to bring down to settlements
Medicine		salt from caves has medicinal value for livestock milk from a certain tree used to kill intestinal worms bamboo used for stomach pains and malaria other medicinal trees and flowers
Raw Materials		collection of deadwood for firewood or construction toothpicks fashioned from twigs a certain tree face paints comes from leaves of a certain tree baskets or sacks from local straw gathering of stones and other building materials
Air	REGULATING	the forest acts as a carbon sink trees control the movement of lightening clean, cold, air with health benefits
Water		the mountain (forest) brings the rains, regulates the seasons, and cleans the water rivers comes from the mountain; there it rains all year round
Soil/Erosion control		trees hold in place the soil to prevent erosion and landslides
Cultural practices	CULTURAL	coming of age/isolation rites in the forest circumcision rites and healing rituals in the forest ancestral caves for various rites: circumcision, delivery of twins fortune telling and prophecy repository of culture and traditions
Education		forest as a repository of culture, a place to learn customs, traditions, and history
Spiritual Value		mountain as holy shrine: receive blessings for daily activities praying and fasting on the mountain conduct sacrifices on the mountain because God lives there the mountain is God’s store with provisions for the community it is a holy place, clean place, representing purity and sacredness myths and legends associated with the mountain specific places on the mountain with spiritual value-caves
Recreation and aesthetics		there are special places with flowers on the mountain, nice to visit runners use the mountain to train and exercise the mountain is home, you feel good there, you enjoy yourself tourists go in with bicycles to recreate
Tourism		animals are plentiful; monkeys, antelopes, elephants; squirrels, moles, lizards, snakes mountain is a good place for tourism; place of flowers, pleasant, safe

), and had some interesting examples of each one. Provisioning services mentioned included food gathering (honey, vegetables, fruits, and fish); livestock rearing (fodder and salts); medicine (various herbs and trees); and raw materials (firewood, timber, straw and stones). Cultural services included rites and rituals (healing ceremonies, circumcisions, and fortune-telling), education (coming-of-age ceremonies), spiritual value (prayers and fasting), and recreation/aesthetics (athletic training and a sense of home). Finally, regulating services included regulation of air, water, and soil. The role of the mountain in supporting wildlife and tourism was also recognized.

Overall, water was seen as the dominant service provided by the mountain. Many people used the term ‘water tower’ to refer to the mountain, and for many that was the principal role of the mountain in their everyday lives:

“Here in the mountain, the rains come according to the times that God planned, the forest itself regulates the seasons of the rain.”

(Mt. Elgon FGD #2)

Interestingly, although cultural services ranked lower in the questionnaires, in the discussions, respondents put more emphasis on these services. These elicited some very detailed and enthusiastic responses, although many were quick to point out that some of these services were no longer utilized:

“Traditions that we used to do in the forest were from a long time ago- those years ago, we would prepare children for circumcision- having been circumcised they would live in the forest for some time until they heal and then return. But these days life has changed, there is not this- it has left with the times, it is not there anymore. Also, there was the tradition of the old men, who would walk and live in the forest, because of the current life of today, this also is past- it’s not there anymore.”

(Mt. Elgon KII #2)

An overall ‘service’ that came out strongly was the sense of home that many felt living on the mountain. This was the feeling that this was their mountain, that it nurtured them and kept them healthy, and that it was where their culture was stored. For example:

“You know the forest in any part of the world, it conserves the culture of the people of that place. And it protects their language- the pronunciation and how it stays with the people- it is protected by the forest. Because the naturality [sic] of the forest- our brother here has said that it breathes- the trees breathe that natural air that’s here by almighty God.”

(Mt. Elgon FGD #2)

This preserving nature of the mountain can also be seen in other areas: not only does it preserve culture, it preserves the water, air, soil, and wildlife. One participant referred to the mountain as ‘God’s store’:

“You can see the importance of the mountain to the local people- indigenous people- and their belief was that...God does not reside on mountains, he is all over, but mountains are his store. If it is not his dwelling house- not his state-house- then it is his store. Because this is where you get supplies- water for life.”

(Mt. Kenya KII #3)

3.3. Climate Change Perceptions

With regard to climate changes, water again dominated, with 70% of the respondents listing rainfall pattern changes as the major change they had witnessed. Temperature and soil fertility changes were also featured prominently in the responses. Again, when prompted, most people agreed that other ecosystem services had changed as well, though there were also plenty of responses to the contrary (Figure 5). People reported ‘high’ or ‘very high’ changes in provisioning services, particularly food (79%) and raw materials (73%), and also to a lesser extent in cultural heritage (70%), climate regulation (59%), and tourism (57%). Medicine and cultural regulation received the highest amount of ‘very high change’ responses. The services perceived to have changed the least were water purification (34%) and livestock forage (41%). Many people reiterated that water was still clean and forage was still plentiful. The vast majority of the changes noted were negative changes (declines), though in Mt. Kenya there were a few instances of respondents indicating changes for the better, noting increases in rainfall and air quality.

Interview and focus group respondents generally agreed that there were changes in the environment, but they did not always associate these changes with global climate change, rather many attributed the changes to local destruction of the environment. Loss of various tree species, wildlife species and topsoil were often associated with deforestation and land degradation, as were changes in the air, water, and temperature. Other respondents, however, did associate observed changes in ecosystem services with global climate change. These people often had very specific examples of changes that had occurred, which they attributed to a global phenomenon. These included changes in vegetation, wildlife, soils, and weather patterns. Vegetation and wildlife changes noted included shifts in species ranges, loss of species, and changes in morphology/phenology. Hydrological changes included changes in rain intensity, storm cloud formation, predictability of seasons, and declines in water yields. Finally, temperature changes were seen in the loss of ice and snow on the mountain, and the in feeling that days were becoming warmer. These changes are consistent with IPCC observations and predictions (

Table 3. Environmental changes identified in Focus Group Discussions and Interviews.

Category	IPCC Observed Changes [38]	Observation in Alpine Areas of Kenya
Temperature	Glacier reduction (very high confidence) Decline in snow cover (high confidence)	snow line shifting up to 4200 m, and less snow overall ice disappearing, do not use ice axes anymore, routes becoming unstable warmer days—you feel you are getting burned colder nights—some of the coldest nights ever experienced recently Lewis glacier moved 30 m in just 10 years
Precipitation and Streamflow	Increase in natural disasters such as landslides (medium confidence) Changes in amount and timing of runoff and streamflow (very high confidence), both increases and decreases in water quantity Loss of aesthetics, recreation, and tourism due to glacier and snow decline (medium confidence)	daily rainfall pattern changing, less rains in the morning increase in hail rains coming from any and all directions increase in lightening rains becoming more unpredictable storms forming suddenly without warning increased rain intensity causing landslides rainy seasons changing planting season shorter than before total rain has decreased lake levels sinking; permanent rivers becoming seasonal or drying up
Vegetation and Wildlife	Increase in species in mountains due to upslope shifting (very high confidence), but also declines (extinctions) of cold adapted species on summits (high confidence) Increases in disturbances such as wildfire (high confidence) Temporary increases in plant productivity (medium confidence)	plants blooming/flowering more often moss and lichen growing at higher elevations pines tress growing at higher elevations giant rosettes plants starting to die at lower elevations new plants species appearing plants becoming smaller at higher elevations more plants at higher elevations, and growing faster and bigger loss of tree species: Elgon teak, cedar and olive animals rarer at higher elevations animal ranges expanding—birds going lower; lions going higher loss of various wildlife species: fish, snakes, elephants mosquitoes now present at elevations where before there were none
Soil and Air	Decrease stability of mountain slopes due to ice thaw (high confidence)	loss of solifluction [soil frost action] in the moorlands soil in the moorland more stable than before erosion, landslides and loss of topsoil decline in air quality

).

Nonetheless, while people acknowledged climate change existed and saw it as a serious challenge, few people saw it as an existential threat. An open-ended question asked respondents to identify the number one challenge facing the country: few listed climate change or environmental concerns, instead listing employment or government services as their chief concern.

The major change mentioned by most participants was changes in rain patterns and season:

“There used to be rain in the morning, there used to be rain in the evening, the water was in plenty. But now I don’t remember when I had rains in the morning.... And even when the rains come, they come in a form like the rains were not formed [previously]: in hailstones.”

(Mt. Kenya FGD #2)

The general feeling was that the weather was becoming unpredictable:

“So, what I would say is I cannot plan what I’m doing on the mountain, looking at the seasons that used to be there. Totally destabilized. When you think it won’t rain, it rains, when you think it will rain it doesn’t rain. It’s no longer as easy as it used to be- people will think- I’m going to plant at this time, and I’ll benefit from that rain. These days, you just have to wait for it to happen.”

(Mt. Kenya KII #2)

The most detailed observations of climate change impacts, however, were those provided by the mountain guides. The guides were consistent visitors to the upper mountain areas, and their work requires them to be very observant of their environment:

“When you go to most of the rocks, you find that…the side that receives more sun obviously has more lichen than the side that doesn’t. And you can easily tell. Or if you look at the wind, you can almost decide this is where I’m going to camp.”

(Mt. Kenya KII #1)

Some of these guides had witnessed some very specific changes in vegetation and soil since the time they had started guiding:

“In the books, they say there are those alpine [species] like the giant Groundsels that will bloom like after a decade. And for me, I have seen them bloom in like months… since I started 11 years now, I have seen a couple of them. So that’s something [that has] happened. And also there are plants from up there that are now coming lower. I have noticed a heath- huge heath plant- here at below Chukuzilia…a huge one. Yeah, and it’s something like I couldn’t expect to be there.”

(Mt. Kenya FGD #1)

Another older guide could personally pull from a larger wealth of experience, without having to rely on guidebooks. His observations were also striking:

“I realized that we are starting to see more plants grow a little bit further up than it used to be. I have also realized that in the moorland after [the] treeline, when there’s fires, when things rejuvenate, they are coming at a greater force. So you’ll see more plants growing faster than we were noticing before. I am realizing that we have come to that point whereby plants like Senecios and Lobelias are now starting to die. The ones I remember, going up Mackinder sides, were huge- and I think must [be] going to the maximum. And now they’re starting to fall over.”

(Mt. Kenya KII #2)

The mountain guides seemed to have the best grasp on the full scope of climate change, having been able to see specific examples of changes up in the moorland that were obviously unrelated to local forest destruction. In contrast to most members of the surrounding community, the guides saw these climate changes as an existential threat.

3.4. Factors Influencing Perceptions

An ordinal logistic regression of demographic variables and perceptions shows that gender, immigrant status, ethnicity, education, and occupation all had some impact on how people said they interacted with the mountains. How frequently people visited the mountain was influenced by gender, occupation, and distance to the alpine area. The odds of a male rating themselves as a frequent visitor—very frequent or frequent—was greater than for females (log-odds of 0.86, p < 0.01). On the other hand, the odds were less for non-farmers (−0.54, p < 0.1) and, strangely, those living nearer to the alpine zone (−0.56, p < 0.05) (see discussion). Knowledge of the mountain was most influenced by immigrant status, education, and ethnicity, while importance was most affected by ethnicity and study site. Those with higher education were more likely to rate their knowledge of the mountain higher (0.69, p < 0.05), while immigrants were less likely (−0.52, p < 0.1). Kikuyu were more likely to rank both their knowledge and importance of the mountain as high compared to Kalenjin (1.14, p < 0.1 and 2.39, p < 0.01). However, the largest influence for importance came from the study site (−4.18, p < 0.01), which dwarfs the impact of ethnicity (

Table 4. Ordinal logistic regression between demographic and spatial factors and interaction with the mountains.

Factor	Frequency	Knowledge	Importance
Sex [Male]	0.86 ***	0.36	0.13
Move Here [Yes]	0.35	−0.52 *	0.00
Education [Secondary+]	0.31	0.69 **	0.09
Occupation [Other]	−0.54 *	−0.08	−0.35
Ethnicity [Kikuyu]	−0.18	1.14 *	2.39 ***
Ethnicity [Other]	−0.37	0.80	0.17
Distance Road [Near]	0.37	−0.69	0.37
Distance Alpine [Near]	−0.56 **	−0.27	0.26
Study Site [Mt Kenya}	−1.10	−0.41	−4.18 ***

* = p < 0.1, ** = p < 0.05, *** = p < 0.01. Dependent variables are the Likert scores regarding statements on frequency of visiting the mountain, importance of the mountain, and knowledge about the mountain. Regression coefficients show direction of the influence- positive numbers show higher odds, negative numbers show lower odds as compared to the reference.

).

With respect to the individual ecosystem services, there were also some significant factors (

Table 5. Ordinal Logistic Regression between demographic and spatial variables and perceptions of importance of individual ecosystem services.

	Factor/ES	Climate Regulation	Cultural Heritage	Education	Food Gathering	Livestock	Medicines	Nutrient Cycling	Raw Materials	Recreation	Tourism	Water Purification	Water Supply #	Wildlife Habitat
Ecosystem Services	Sex [Male]	−0.16	0.51	0.12	0.01	0.35	0.22	0.05	0.16	−0.18	0.35	0.53	n/a	0.40
	Move Here [Yes]	−0.53	−0.68 **	−0.27	0.11	0.28	−0.83 **	−0.07	0.21	−0.28	−0.27	−0.38	n/a	0.30
	Education [Secondary+]	0.87 **	−0.17	−0.43	0.14	0.35	0.04	−0.30	−0.62	−0.22	0.26	0.27	n/a	0.11
	Occupation [Other]	0.13	0.10	1.05 **	0.57	−0.24	−0.15	0.78 *	0.30	0.16	0.13	0.83 **	n/a	−0.27
	Ethnicity [Kikuyu]	0.10	−0.13	0.09	−0.24	−0.04	0.04	−0.43	−0.86	0.67	0.43	−0.61	n/a	−0.91
	Ethnicity [Meru/Luhya]	0.54	−0.12	0.29	−0.42	0.35	0.25	0.40	−0.32	0.58	0.73	−0.48	n/a	−0.94
	Distance to Road [near]	0.48	1.58 **	−0.19	−0.05	−0.14	1.78 **	−0.09	1.43	0.56	1.87 ***	0.38	n/a	−0.10
	Distance to Alpine [near]	0.95 **	0.01	−0.49	0.07	0.10	0.46	0.15	−0.61	0.32	0.09	1.01 ***	n/a	0.43
	Study Area [Mt Kenya]	−0.59	−1.79 **	−1.11	−1.69	−1.31	−3.97 ***	−0.56	−1.70	−0.78	−0.71	−0.39	n/a	0.23
Ecosystem Service Changes	Sex [Male]	−0.75 **	−0.25	−0.06	−0.28	−0.51	−0.50	−0.41	0.12	−0.94 ***	−0.75 **	0.00	−0.51	0.11
	Move Here [Yes]	0.16	0.16	0.05	−0.30	0.10	0.65 **	0.09	0.10	0.19	0.30	0.13	−0.49 *	0.16
	Education [Secondary+]	−0.04	0.22	0.19	0.27	0.22	0.08	0.15	0.28	−0.01	−0.07	0.53 *	0.24	0.20
	Occupation [Other]	−0.32	−0.91 ***	−0.74 **	0.32	0.00	−0.91 ***	0.07	−0.32	−0.74 **	−0.35	−0.59	−0.25	−0.01
	Ethnicity [Kikuyu]	−0.13	−0.30	−0.03	−0.37	−0.22	0.02	0.05	0.53	−0.02	−0.29	0.19	−0.33	−0.37
	Ethnicity [Meru/Luhya]	0.10	−0.04	0.48	0.21	0.01	0.14	0.85	0.61	0.47	0.45	0.66	0.36	0.12
	Distance to Road [near]	−0.75	−0.44	0.99	−0.68	−0.12	−0.64	0.28	−1.47 *	0.74	−0.89	0.75	0.63	0.17
	Distance to Alpine [near]	−0.83 ***	−0.57 **	−0.55 *	−0.59 *	−0.07	−0.27	−0.50 *	−0.45	−0.98 ***	−1.02 ***	−0.36	−0.06	−0.39
	Study Area [Mt Kenya]	1.05	1.74 *	−0.60	0.22	0.37	2.41 **	−0.92	1.28	−0.80	0.00	−1.05	0.33	−0.53

* = p < 0.1, ** = p < 0.05, *** = p < 0.01. Logistic Regression Model: [Ecosystem Service] ~ Gender + Immigrant Status + Education + Occupation + Ethnicity + Distance to Road + Distance to Alpine Zone + Study Area. # The water supply ecosystem service had too few levels in the responses to fit the model.

). Gender had an influence on perceptions of change, with males less likely to rank changes highly as compared to females (climate −0.75, p < 0.05; recreation −0.94, p < 0.01; tourism, −0.75, p < 0.05). Immigrant status had an impact on perceptions of medicines and cultural services, with immigrants ranking these services lower than permanent residents (culture −0.68, p < 0.05; medicine −0.83, p < 0.05), but ranking changes higher (medicines 0.65, p < 0.05). Non-farmers viewed the importance of certain services higher (education 1.05, p < 0.05; water purification 0.83, p < 0.05), but generally viewed changes in services lower (culture −0.91, p < 0.01; education −0.74, p < 0.05; medicines −0.91, p < 0.01; recreation −0.74, p < 0.05). Ethnicity and education, on the other hand, did not have much of an influence on perceptions, with the exception of those with secondary education (or higher) ranking the importance of climate regulation higher (0.87, p < 0.05).

Location and spatial variables also had an influence on perceptions. For those living near roads, the odds of ranking various cultural services were higher than for those living far from roads (culture 1.58, p < 0.05; medicines 1.78, p < 0.05; tourism 1.87 p < 0.01), but did not significantly influence views of changes in those services. On the other hand, for those living near the alpine zone, the odds of ranking various regulating services were higher than for those living far (climate regulation 0.95, p < 0.05, water purification 1.01, p < 0.01), while the odds of ranking changes in these and other services were lower (climate regulation −0.83, p < 0.01; cultural heritage −0.57, p < 0.05; recreation −0.98, p < 0.01; tourism −1.02, p < 0.01). The study area only had a significant impact on views of medicines and culture, but the magnitude of the log-odds was much larger than for the other variables. The odds of someone in Mt. Kenya ranking culture and medicines as important was much lower than for Mt. Elgon (culture −1.79, p < 0.05, medicines −3.97, p < 0.01), while those in Mt. Kenya were much more likely to rank changes in medicine services highly (2.41, p < 0.05).

Logistic regression provides the relative importance for each factor, while holding all the other factors constant. Thus it is susceptible to multicollinearity if variables are too closely related.

Table 6. Chi-Square test of independence among demographic and spatial variables.

Factor	Sex	Ethnicity	Education	Move Here	Occupation	Study Site	Distance Road	Distance Alpine
Sex	X	0.73	1.66	1.16	0.93	0.20	0.02	0.04
Ethnicity		X	17.84	23.78	35.07	169.33	140.96	33.79
Education			X	1.92	38.37	13.66	9.70	4.59
Move Here				X	0.01	26.43	23.93	0.58
Occupation					X	31.14	32.22	12.65
Study Site						X	179.26	24.34
Distance Road							X	14.48
Distance Alpine								X

Bold values show significance at the p < 0.05 level.

presents the chi-square relationships between the different factors. Ethnicity, study site, and distance to roads, in particular, are highly associated with each other: (Ethnicity/Study Site- χ² = 169.33, df = 2, p < 0.01; Ethnicity/Distance to Roads- χ² = 140.96, df = 2, p < 0.01; Study site/Distance to Roads- χ² = 179.26, df = 1, p < 0.01). The study site, therefore, may be confounding the influence of these other two variables.

4. Discussion

Respondents overall indicated a high level of appreciation for mountain ecosystem services and also were cognizant of changes to those services. All types of services— provisioning, regulating, and cultural services—were recognized and rated highly (Table 2; Figure 4). However, the discussions highlighted the fact that the standard ES categories [10] were not always adequate to characterize the services; some services spanned multiple categories in peoples’ minds. For example, medicines were often referred to more as a cultural service than a provisioning service as the use of medicines was closely tied to traditional knowledge. Similarly, wildlife was seen for some as a source of food, for others as a means of attracting tourists, while others saw it as a source of biodiversity. It is for this reason that the concept of Nature’s Contributions to People (NCP) can sometimes be more useful than ecosystem services, as the categories are more fluid and the emphasis is on how nature contributes to quality of life [13].

Overall, water was seen as the primary service of the mountains and also the one that had changed the most in peoples’ lives. This is consistent with what other studies have found around the globe. Local observations have reported changes in precipitation quantity [9,39,40], seasonality [15,41], and pattern [15,42] (see [18,19] for a more complete list). There was also a general sense that weather patterns had become more unpredictable. Indigenous communities in the Arctic and subarctic regions, another hotspot in climate change, have made similar observations regarding climate changes. Such communities have pointed out the unpredictability in weather patterns which affects migration routes, forage and prey abundance, and therefore entire ways of living [43,44,45,46].

These changes, as well as the other major changes noted—upward movements of plants and animals, melting glaciers, rising temperatures, and declining water yields (Table 3)—match well with climate model predictions [47]. They also have been reported before in East Africa, where model and empirical data have demonstrated shifts in species ranges [48,49,50], changes in hydrologic patterns [25,51,52,53], temperature changes [51,54,55], and glacier reductions [56,57,58]. Some of the other changes observed, however, are less often reported in the literature although still consistent with global climate change. For example, the observation of plants flowering more often and growing bigger and faster at higher elevations is an intriguing observation that merits further study. Also, the changes observed in rainfall characteristics—more hail and lightning, and rain coming from different directions—are interesting details. These are details that are not readily captured in meteorological records, and yet can have a direct impact on human livelihoods [59].

Nonetheless, local observations are rarely uniform, and in fact, the same community, with common experiences, can have different perceptions of changes [60]. Even in this study, local knowledge was not a homogeneous quantity, and there were interesting differences in perceptions according to demographic and spatial variables (Table 4 and Table 5). This can sometimes lead to observations that at first appear to be conflicting: for example, some reported expansion of wildlife ranges whereas others reported contracting ranges; some reported warmer temperatures others colder; and some saw a decrease in soil stability while others observed an increase (Table 3). However, these observations may not be as inconsistent as they first appear: climate change is predicted to have variable effects even within small geographic areas, and different organisms respond differently to that change [47,61]. On the other hand, some perceptions may indeed be faulty or incomplete.

As a general rule, in each study site, those that visited the mountain more also had higher levels of knowledge of the mountain and ranked its importance higher. These people typically valued individual ecosystem services higher and were more likely to be able to identify specific changes that had occurred. These were the people that derived a livelihood from the mountain—generally lower educated farmers who were long-term residents, living far from roads but closer to the alpine area. Higher educated people, non-farmers, and immigrants had less cultural connection to the landscape, and so overall did not value services as highly, particularly cultural services. The services that they did value were the more intangible services: climate regulation, water purification, and education. Their views on changes were mixed: immigrants, with no baseline to go by, were more likely to rank changes highly, while non-farmers with less reason to track closely changes that had occurred, were less likely to rank changes highly (Table 5).

Other studies around the globe have variously shown that ethnicity, gender, or economic status are the major factors influencing perceptions [16,33,34,62,63,64,65]. In this study, socio-economic status, which is a function of education, occupation, and migrant status, was the biggest driver of perceptions. This makes sense: the degree to which one’s livelihood is directly dependent on ecosystem services will naturally determine how those services are valued. There is also a connection between socio-economic status and access to information, which may explain this finding [66,67]. Ethnicity did not influence perceptions much in this study, while gender only had an impact on views of change, with males overall viewing less changes in ES as compared to females. However, given the low sample size for females, this result should be treated with caution. Men and women often have overlapping but complementary views regarding ecosystem services and climate change [46,68,69]. While some studies find women valuing certain services more than men, other studies find the opposite or no significant difference [33,44,63,65,70,71].

Many studies do not explicitly look at location or spatial characteristics as factors, and it is likely that in many cases demographic and socio-economic variables are confounded by location to some degree, as was found in this study. Studies that have considered location have often found it to be the biggest factor determining appreciation of ecosystem services, as location plays a large role in shaping underlying community values [33,34,70,72]. Ethnicity in particular is closely tied to location, and it can be difficult to separate the two [34]. In this study, location (study area) was too closely related to ethnicity and distance to roads to separate out their individual effects (Table 6). Mt. Kenya has a very different physical and socio-economic setting than Mt. Elgon. The differences in perceptions could therefore be a function of different land-use histories, different value systems and beliefs, or different levels of access to ecosystem services.

Interestingly, those living near the alpine zone ranked changes in many services less than those living further away. Those nearest to the alpine zone were largely from the mountain-dwelling Sabaot communities in Mt. Elgon, and even in the focus group discussion with this community, there were lower views of change. It may be that they have a more sanguine view of climate change as they live in more pristine areas untouched by human disturbance, or it could be that they see the changes on a much shorter time interval so they habituate to the changes without realizing it. Those living near the alpine zone also ranked their frequency of visitation lower, but this may just have to do with confusion in terms: for those living closer to the alpine zone, the mountain is considered the peak itself, whereas those living further away consider the whole land mass to be the mountain.

The cultural services elicited the most differing opinions (Figure 4). These were the services most significantly associated with specific groups of people. The biggest factor appeared to be location, with those in Mt. Elgon ranking the importance of cultural services higher and the amount of change lower. In Mt. Kenya, the lessened importance of these services was attributed to a loss of traditional knowledge. This squares with what has been seen in other parts of the world, where a cultural transition has been noted, with the loss of traditional knowledge in some sectors, particularly the younger generation [43,73]. On the other hand, the Mt. Kenya guides did indicate a sort of indigenous knowledge that they relied upon when guiding. This can be thought of as a hybrid knowledge, combining traditional knowledge with modern technology. Even the Inuit, who continue to rely on traditional knowledge for their livelihoods, also use GPS, radio and cellphones [43].

Explanations for changes in ecosystem services vary, with some attributing them to climate change while others point to local causes, and still, others say that the changes are caused by God. This too is in line with what other studies have found [15,41,42,43]. Perhaps because of this uncertainty, climate change did not rank as the number one concern for most respondents, which has been seen elsewhere as well [74]. Attribution of change is tricky anywhere, and in Kenya in particular, as Kenya has seen drastic changes over the past half-century in political, economic, and social spheres. Separating which changes are due to a global phenomenon and which are local changes is difficult.

Mountain guides are in a unique position at the front line of climate change, as they experience it every day and are trained to stay attuned to the slightest changes in their environment. They can therefore play an important role as ‘citizen scientists’, providing detailed observations to complement scientific investigations [75,76]. The guides in Mt. Kenya had very detailed observations of changes in cloud formation, plant phenology, and species range. They also often had concrete examples of temperature changes, noting specific locations where ice no longer formed or where soil properties had changed, or where species had migrated. These details are similar to observations by mountain guides in other parts of the globe, who have noted the increasing hazards associated with their routes, because of ice melting and unpredictable weather [77,78,79].

This study used a mixed-methods approach to try to quantify local perceptions of the environment and changes to that environment, while also qualitatively providing some context to these perceptions. There are limitations to this approach, as grouping data collected in different ways limits the statistical robustness; nonetheless, it enables an increased depth and breadth than either approach by itself can allow [80]. This was particularly important for this study, as the questionnaires were limited by challenges of access, which prevented perfect balance in demographics, and also limitations in the level of engagement and comprehension of questions themselves. Nonetheless, further research will be needed to tease out how prevalent certain perceptions are across a wider geographic area as well as perceptions of climate change versus climate variability.

5. Conclusions

Local communities in climate change-vulnerable areas such as Mt. Kenya and Mt. Elgon can be invaluable sources of information for understanding how climate change is impacting different parts of the globe. Local observations can provide a level of specificity that is lacking in global climate change models. Overall, these communities have a strong appreciation of the services provided by the mountains as well as an understanding of the changes occurring to those services. The biggest change they have noted is regarding water availability which is the service from the mountain that they appreciate the most. Weather patterns have become increasingly unpredictable which has become a source of concern for many, although generally not their number one concern. Cultural services of the mountain are also clearly incredibly important, but a cultural transition appears to be underway whereby the mountain is not used as much as in the past for cultural activities. Factors influencing perceptions appear to have as much or more to do with location rather than inherent demographic characteristics. Those nearest to and interacting the most with the mountain have the strongest appreciation for ecosystem services, but also often have lower perceptions of changes in those services. On the other hand, the guides that use the mountain on a daily basis for their livelihoods have very detailed observations of changes on the mountain. These guides can play an important role in citizen science in documenting the impacts of climate change in these vulnerable areas. These observations must necessarily play a role in management planning to provide context-specific insights on vulnerability and adaptive capacity in the face of climate change.