Human-climate induced drivers of mountain grassland over the last 40 years in Sidama, Ethiopia: perceptions versus empirical evidence

Mountain grasslands are globally important ecosystems. They are considered as heritage sites with “outstanding values” to ecological communities by adding a further “layer and support to the existing protection measures” (Thorsell 2002). Furthermore, mountain grasslands are regions of high diversity and species-rich ecosystems that co-evolved with centuries of livestock grazing (Spiegelberger et al. 2006). Climate change has the potential to alter these ecosystems and the services they provide to each other and to human society at large. Human societies depend on ecosystem provisioning, regulating, cultural, and supportive services for their own security. Thus, mountain grassland ecosystems provide many of the important goods and services for human beings including food, timber, fresh water, protection from natural hazards, carbon storage, and many other functions including recreation and tourism (Messerli and Ives 1997). They also provide forage for grazing animals and habitat for varieties of wild animals. The ecosystem goods and services from mountain grasslands are relied upon by some of the most impoverished people in the world who are often marginalized due to cultural factors and inaccessibility of most mountain areas (Gerlitz et al. 2012).

According to a recent study (Ward et al. 2014), mountain grasslands also store considerable amounts of carbon and these ecosystems should be accounted for climate policy. However, mountain grasslands are increasingly under threat from human activities (Körner et al. 2005) and impacts of climate change (Schroter et al. 2005). Besides, changes in land use such as increases in human settlements in the high mountain areas and the gradual degradation of remote sites threaten the unique biodiversity of mountain ecosystems (Spiegelberger et al. 2006). The floristic composition and species diversity of mountain grasslands are changing generally in response to human-induced factors (Spiegelberger et al. 2006). The magnitude of changes in mountain grasslands is linked to ecosystem processes and excessive interferences by human beings and over-exploitation of these ecosystems.

Human activities have modified mountain grassland ecosystems in the process of food production and other essentials for several thousands of years (Ellis and Pontius 2007). In recent years, the rates, extent, and intensity of human pressure on mountain grasslands are by far greater than before, affecting the functioning of these ecosystems. These affect consecutively the functional attributes of ecosystems, the provision of ecosystem services, and human welfare (Ellis and Pontius 2007).

Earlier evidences (Beniston and Rebetez 1996; Beniston et al. 1997; Theurillat and Guisan 2001; Nogue’s-Bravo et al. 2007) have shown that mountain grassland ecosystems are disproportionately exposed to the impact of climate change. This is of great concern in the face of climate change as mountain grassland ecosystems represent irreplaceable resources, both in terms of biodiversity conservation and the ecosystem services they provide (Körner et al. 2005; Viviroli and Weingartner 2004). Climate is an integral part of mountain ecosystems and organisms have adapted to their local climate over time. The goods and services provided by mountain ecosystems sustain almost half the human population worldwide, and also maintain the integrity of the Earth’s ecosystem through the provision of vital environmental regulation and rejuvenation functions (Karma et al. 2010).

Despite the importance of mountain grassland ecosystems worldwide, no previous studies have examined the mechanisms involved in land use/land cover changes in mountain grasslands of Sidama in South Ethiopia. In order to assess the potential impacts of human activities and climate change on trends of grassland ecosystems, a commonly employed method could be the use of a remotely sensed image on land use/land cover changes (Abate and Angassa 2016). This article presents human-climate induced drivers on trends of mountain grassland in Sidama Zone in South Ethiopia between 1973 and 2015 using earth-observing satellite imagery and perceptions of the local communities with the objectives of (1) determining the driving forces of land use/land cover changes, (2) understanding the perception of local communities on land use/land cover changes between 1973 and 2015, and (3) establishing and mapping the magnitude, rates, nature, and spatial distribution of land use/cover types that occurred in the study area.

The use of local perception and LULC changes in understanding the effects of anthropogenic factors linked to climate change on mountain grassland ecosystems is crucial for future management and conservation of these ecosystems. In the current study, using local perceptions and analyses of remotely sensed images, we generated evidence of human-climate induced patterns of LULC dynamics supported by GIS technology in the mountain grasslands of Sidama. In the mountain grasslands of Sidama, perceptions of respondents indicate that land use/cover changes are occurring as a result of multiple drivers with potential consequences of climate change. Observations of respondents and analyses of remotely sensed images further suggest the existence of extensive changes in land use cover dynamics with increasing drought, deforestation, and expansion of agricultural lands. These reflect rapid changes in human population growth, land tenure, and associated pressures on the environment with consequences of major shifts in land use, which also agree with Chase et al. (1999). Observations of respondents indicate that land use/cover dynamics in the mountain grasslands of Sidama in South Ethiopia is accelerated by drought and changes in the timing of rain. Among others, experiences of respondents show that the rapid growth of human population, expansion of settlements, and increased mountain side farming are major factors for the loss of forest landscape.

The issue of unequal land distribution under the Feudal system prior to the 1970s was resolved by the 1975 land proclamation (Dessalegn 1984). Following the land reform decree in 1975, the local communities had ensured access to land. The 1975 land reform created an opportunity where landless peasants allocated farmlands that changed the form of ownership. Changes in land ownership from the landlords to peasantry lead to the exploitation of the land for agriculture that probably accelerated land use/cover changes. Likewise, the use of forest products for fuel and house construction greatly contributed to land cover changes. Overall, the demand for increased crop production for household consumption might exacerbate the processes of forest loss.

Perceptions of respondents regarding land use/cover dynamics are consistent with the current satellite imagery analyses. The results from satellite imagery analysis also confirm respondents’ perceptions of high forest cover in the past 40 to 50 years. Moreover, respondents’ reflections of vast grasslands in the past 40 to 50 years are generally consistent with the results of satellite imagery analysis. The reflections of respondents regarding the expansion of mountain side farming are disclosed by satellite imagery analyses, suggesting a dramatic increase in cropland between 1973 and 2015.

Observations of respondents reflect that they have a detailed knowledge of sharing their experiences about the impact of climate change using local indicators. These local indicators as indicated by respondents include increased temperature and livestock diseases, reduced milk production, decline in livestock feed resources and limited water availability, changes in the timing of rain, dried up of wet lands, and frequent droughts. The probable causes of land degradation and low productivity could be attributed to excessive human interference and over-exploitation of natural resources accelerating patterns of land use/cover changes.

The encroachment of cultivation into mountain grassland areas are major causes of land use/cover changes in the study areas. The results suggest that the encroachment of crop cultivation is increased by 646.74% between 1973 and 2015 with a considerable damage on the forest cover. The perceptions of respondents indicate that the demand for family’s food production is the major driver for the expansion of agricultural land in mountain areas. Gibbs et al. (2010) have reported a similar case where the high demand for food and fuel is a major factor for the expansion of crop cultivation into mountain areas in developing countries. Generally, the growth of human population and agrarian land transformation are major contributing factors.

Observations of respondents suggest that the processes of land use/cover changes are more manifested in terms of shifts in forest cover, shrinkage of grazing lands, and increase in crop fields. Respondents’ perceptions of deforestation, major shifts in grazing lands, and increased agricultural fields into mountain areas are also confirmed by the existing satellite image analyses. Respondents further believed that there are enormous ongoing changes with regard to the impact on forest and grassland resources. Generally, respondents stated that livestock number has exceeded the capacity of the land due to population growth and expansion of crop cultivation. Perceptions of reduced forest cover and available grazing lands are consistent with our results from the remotely sensed images. Our results indicate that although the forest cover initially showed a downward trend, there had been an improvement in forest cover of the study areas since 2000.

Quantitative results of the existing change analysis over the last 42 years during three different periods (1973–1986, 1986–2000, and 2000–2015) and change matrix between 1973 and 2015 revealed the extent of changes in terms of land use dynamics. In general, the trends in forest cover, grassland, and shrubland show incessant decline. Hence, the LULC change assessments indicate that 119.58% of the area underwent a transition between the natural and human-dominated activities. This increase in agricultural land reinforces the idea that human economic activities are the major driving factors behind LULC changes (Wang et al. 2008). Generally, crop land increased by 433.02 ha, which was gained from other land use types. This suggests that as human population grows, mountain side forests and grassland resources might be affected adversely with the expansion of crop cultivation. Furthermore, respondents’ observations suggest that as the growth of human population continues, the demand for food will rise. Zeleke and Hurni (2001) have also argued that changes in mountain side forests and grassland covers is a common phenomenon in other places with similar settings due to human interferences. Similarly, Shiferaw (2011) has reported the loss of forest cover due to the expansion of agriculture in South-Wollo highlands of Ethiopia.

Our results suggest that the growth of human population, changes in land use policy, and expansion of agriculture are major reasons for the loss of mountain grasslands in the study areas. The growth of human population and modification of the natural environment to drive livelihood needs most likely caused damage to forests that further facilitated bush encroachment as a result of secondary succession following disturbances. Our results indicate that major forest loss occurred between 1973 and 2000. The probable explanation could be the reallocation of lands to peasants following the 1975 land proclamation in Ethiopia (Dessalegn 1984).

Our results are in agreement with the report by Braimoh (2006), suggesting that a gp ratio of greater than 1 is an indication of the possibility for a landscape category to expand beyond its original size. Related to our finding, the probable explanation for the greater than 1 values of gp could be a result of the exposure of a given landscape toward transition (Braimoh, 2006). This suggests a considerable change in terms of the different land use types as oppose to their original sizes. In contrast, the less than 1 gp value for the forest cover suggests that the forest resource is the most vulnerable land use type.

In connection with LULC persistence, the negative values for the net-change in the case of forestland, grassland, and shrubland covers reflect the net loss in the different land use classes. The probable reason for higher ratio of lp (i.e., >  1) denotes the tendency for a class of land use type toward transition, which is also consistent with the report by Braimoh (2006). Throughout the study period, the forestland displayed a lower amount of loss. This is an indicator of a better chance of expansion in terms of forestland with a minimal tendency for loss. The values of lp ratio for the grassland, shrubland, bushland, and agricultural land were greater than one, suggesting their exposure to transition.

We found that all land use/cover classes had greater than 1 lp values with exception of forestland. Overall, the agricultural land type experienced the highest lp ratio (22.56) followed by the bushland (14.24) and shrubland (8.03) cover types, suggesting that mountain side cultivation as the most driving force for the loss of forests and grassland resources. Likewise, previous studies (Lemenih et al. 2004; Hailu et al. 2015) have shown that the expansion of agriculture is a major factor for the loss of forest cover in Ethiopia. According to Foody (2001), land use land cover changes are strongly related to the patterns of rainfall.

The existing analysis indicates that the net change within the landscape between 1973 and 2015 is accounted for 48%. On the other hand, the observed total change within the landscape is 119.58, suggesting that both swap and net changes are vital in recognizing the total conversions within a landscape. Generally, analyses of land use conversion that only consider the net change might weaken the existing total changes of the ecosystem. Thus, our results confirm that the swap changes that had taken place between 1973 and 2015 are attributed to the gains and losses in the different landscapes.

Generally, our results are justified using the actual LULC maps of 1973, 1986, 2000, and 2015 with an overall accuracies of 86.5% and a Kappa statistic of 0.803, which is within the range of accuracies reported in previous studies (Echeverria et al. 2008; Kamusoko et al. 2011). The LULC patterns and changes in the mountain grasslands of Sidama in South Ethiopia are reflective of the various prospects and restraints for the use and conservation of mountain grassland ecosystems in the region.

The significant amount of spatial changes among the different land use/cover types over the last four to five decades mainly contributed to the ongoing degradation of mountain grassland ecosystems in Sidama, South Ethiopia. In 1973, the forestland cover was the dominant land use/cover type that accounted for 58% of the studied landscapes, followed by grassland (22%), bushland (16%), and Shrubland (12%). Four decades later in 2015, the proportions of forest, grassland, shrubland, and bushland covers were 50%, 11%, 7%, and 16%, respectively. On the other hand, the agricultural landscape was increased from 2% in 1973 to 16% in 2015 with a net gain of 14%, while the net gain for the forest cover was 8.2%. The replacement of natural forest by agricultural landscape exposed top soils to erosion with major consequences on this ecosystem.

Based on the results of GIS and remotely sensed analyses versus local communities’ perceptions, it can be concluded that the mountain grassland ecosystems of Sidama had already altered in terms of land use/cover dynamics over the last four to five decades. The LULC dynamics in the study area was evident by the declines in terms of forest cover (15.9%), grassland (53.57%), and shrubland (40.96%) at the end of 2015. Communities’ perceptions confirmed that the major drivers of LULC dynamics were attributed to changes in land tenure, increased population pressure, excessive exploitation, increased agricultural landscapes, overgrazing, impacts of climate change, and the high demand for wood products that resulted in deforestation. The results indicate a gradual transition in land use/cover changes caused by human-climate induced factors that calls for proper management and conservation of mountain grassland ecosystem in the study areas.