Around the world flammable shrublands have traditionally been managed by fire to reduce canopy cover, increase forage production, and to get rid of predators and pests (Keeley et al. 2012). This ancient management is well-known for grassy biomes (Bowman et al. 2011, Archibald et al. 2012), but was common also in heathlands (Hobbs and Gimingham 1987, Wesche et al. 2000) and flammable shrublands (Neumann et al. 2011, Keeley et al. 2012, Pausas and Fernández-Muñoz 2012). Without an understanding of the local fire ecology and management history, there is a risk that new land-management policies, such as climate mitigation projects like REDD+ (Reducing Emissions from Deforestation and forest Degradation, the role of conservation, sustainable management of forests and enhancement of forest carbon stocks in developing countries) may cause unintended negative consequences. Examples could be loss of biodiversity and reduced food security, or a shift in the fire regime, resulting in larger more destructive fires.
Natural and anthropogenic fire regimes can be difficult to separate in paleorecords, and prehistoric management is often unknown (Archibald 2016). Flammable shrublands can also be naturally ignited by lightning, and, e.g., some dry shrublands in Australia have likely been little influenced by prehistoric human ignition (Bradstock et al. 2012, Keith et al. 2014). A fire regime is defined as the frequency, season, intensity, and sizes of fires (Gill 1975). Anthropogenic fires are typically more frequent, early-season, low-intensity, and small-scale than wildfires (Archibald et al. 2013). In anthropogenic fire regimes, ungulate herbivores typically reduce standing fuels by consuming a large share of annual biomass production. Remaining surface fuels are then typically removed by frequent controlled early-season patch-burning (Laris and Wardell 2006, Archibald 2016). This traditional fire management practice typically creates an ignition-saturated, fuel-limited fire regime, in which fire size and intensity is regulated by fuel quantity, quality, and spatial distribution (Pausas and Fernández-Muñoz 2012, Pausas and Keeley 2014). Frequent anthropogenic ignition over long time periods often created a more fine-grained landscape mosaic structure, because recently burnt patches are nonflammable until surface fuels have rebuilt (Baeza et al. 2011, Johansson and Granström 2014). In shrublands this can take several years, and in such cases young stands act as fuel-breaks in the landscape, limiting the size of consecutive fires (Minnich and Chou 1997, Allen 2008, van Wilgen 2013, Johansson and Granström 2014).
Fire regime shifts have occurred several times in geological time scales, e.g., in response to increased atmospheric oxygen (Bowman et al. 2009) and the evolution of flammable grasses (Pausas and Keeley 2009, Strömberg 2011, Bond and Midgley 2012). When humans first arrived, dramatic increases in ignition frequency occurred, because of the need for open landscapes containing more food resources (Bowman et al. 2011). Also megaherbivore extinctions likely altered fire regimes globally (Johnson et al. 2018). Today most flammable biomes are experiencing the reverse trend; a rapid decrease in ignition rates (Bowman et al. 2011, Archibald et al. 2013). One example is fire suppression in Mediterranean-type forests, often causing fuel accumulation and increased fire intensity and size (Williams 2013, Pausas and Keeley 2014). Apart from burn bans, there are many socioeconomic changes that may cause fire-regime shifts, e.g., abandonment of traditional land management due to population growth (Butz 2009), agricultural intensification, rural depopulation, or nature conservation (Pausas and Fernández-Muñoz 2012, Pausas and Keeley 2014). Where traditional fire management has been terminated, fire regimes typically have shifted into ignition-limited fire regimes characterized by longer fire intervals, more surface fuels, higher fire intensities, and larger fires (Williams 2013). This has been directly observed mainly in the northern Mediterranean basin, where it happened quite recently (Pausas and Fernández-Muñoz 2012). Other shrub systems have also been traditionally managed by fire, e.g., the Californian Chaparral, but this prehistoric management terminated earlier and is less well understood (Keeley 2002, Anderson and Keeley 2018). The increased frequency of extreme fire weather due to climate change is also altering fire regimes globally, and this has large implications on current management options (Keith et al. 2014, Pausas and Keeley 2014, Dass et al. 2018).
General burn bans, deriving from the Central European antifire paradigm (Pyne 1997, 2016), were during colonial times implemented across Africa, based on the assumption that fire degrades ecosystems (Pyne 1997, Laris and Wardell 2006, Archibald et al. 2012, Neumann 2014). In African grassy biomes, burn bans from the 1970s typically resulted in pasture degradation, bush-encroachment, and biodiversity loss (Angassa and Oba 2008, Parr et al. 2014, Archibald 2016, Khatun et al. 2016). In South African fynbos, fire-suppression since the 1930s has threatened many species adapted to specific fire-return intervals (van Wilgen 2013). The East African ericaceous belt has likely been influenced by fire ever since it expanded after the last glacial retreat ~14 ka BP (Wesche et al. 2000, Umer et al. 2007, Schüler et al. 2012, Gil-Romera et al. 2019). Early land use history in the East African heathlands is largely unknown, as well as the effects of decades of fire-suppression. However, large wildfires have become increasingly problematic in protected heathland areas (e.g., Wesche et al. 2000, Hemp 2005, Belayneh et al. 2013). The importance of anthropogenic fire in shaping the East African ericaceous belt has been discussed (Hedberg 1964, Wesche et al. 2000, Jacob et al. 2015), but rarely from an ancient cultural landscape perspective.
Land use change and climate change are two major threats to ecosystem resilience and productivity today. Occurring simultaneously, they interact across landscapes and potentially reinforce each other. International policies for climate mitigation through forest restoration, such as REDD+, have many times become a threat to fire-adapted ecosystems (Parr et al. 2014, Khatun et al. 2016). Carbon forestry projects typically aim for total fire exclusion and grazing restrictions to promote tree regeneration, without assessing the associated risk of surface fuel accumulation (Barlow et al. 2012). There is a lack of ecological studies of still active anthropogenic fire regimes in flammable shrublands, because such systems are globally rare today. Bale Mountains is a unique social-ecological system in that the subalpine heathlands have been maintained by traditional fire management for thousands of years. Despite large fluctuations in livestock and human populations, nature conservation efforts, and a 50-year burn ban, the traditional patch-burning and free-range grazing has continued until present. However, since the creation of the Bale Mountains National Park in 1969, parts of the subalpine heathlands have been subjected to a stronger enforcement of the fire-exclusion policy, making the study area particularly suitable for studies of landscape effects of changes in land management.
The major aim of this study was to increase our understanding of how changes in policy and local land use has altered the fire regimes. Specifically we ask if and how the fire regimes have been affected by the creation of the Bale Mountains National Park, in comparison with similar, but less protected heathland areas outside the park. We compared remotely sensed images from inside and outside the park from 1968 to 2017 and analyzed interviews with 41 local agropastoralists. Our aim with the interviews was to obtain answers to the following questions: (1) How has fire management changed historically? (2) What are the differences in management between the park and outside? (3) What are the underlying causes of historical changes? (4) What is the traditional ecological knowledge regarding heathland fire ecology and fire management? Our hypothesis was that fires inside the park have become fewer but larger, because of the stronger fire-exclusion efforts, which leads to reduced numbers of nonflammable young stands acting as fuel-breaks in the landscape. In this paper we analyze drivers for change, and discuss how the traditional fire knowledge can be used as a base for developing a sounder, community-based fire-management program in order to conserve the unique ecosystem and at the same time adapt to climate change. Results from this project has implications for biodiversity conservation, local livelihoods and heathland management policy, given both climate change predictions and the new carbon sequestration ambitions.
The Bale Mountains in the southern highlands of Ethiopia harbors the largest area of subalpine ericaceous vegetation in Africa (~90 000 ha, out of which the heathlands compose ~70 000 ha), and harbors many threatened and endemic species (Miehe and Miehe 1994, Fetene et al. 2006). The Bale Mountains National Park (BMNP) was established in 1969 to protect the unique Afroalpine and montane forest habitats (Hillman 1988). This study focuses on the heathland zone (~3500–3900 m.a.s.l.) inside and outside the park in an area of ~95 x 40 km (centered at 6°53′42.65″N, 39°32′24.08″E; Fig. 1).
Out of the total heathland area, 69% is located within the park, and 31% outside, partly in the Adaba-Dodola forest priority area to the west. All areas are since 2013 included in the Bale Mountains Eco-region REDD+ project. The project aims to conserve biodiversity and sequester carbon by increasing the enforcement of fire exclusion and grazing restrictions, including in the heathland zone (Watson et al. 2013, Oromia Forest and Wildlife Enterprise 2016). A brief background of geology, climate, vegetation, and land use history is important for understanding the social-ecological system (Appendix 1.1). The heathland vegetation comprises two codominating tree-heather species, Erica arborea and Erica trimera (Miehe and Miehe 1994), hereafter collectively referred to as Erica. Both species have potential to grow into tall trees, but when exposed to recurrent fire, they develop subterranean lignotubers and a multistemmed growth-form favorable for fire propagation (Johansson and Granström 2014). The heathlands are burnt on short rotation by local herders to improve pasture for free-ranging cattle (Johansson et al. 2012). They therefore consist of a mosaic of vegetatively regenerating Erica stands of different canopy age (time since last fire). This shifting patch mosaic favors a high biodiversity and creates extended habitat for Afroalpine plants (Johansson et al. 2018). Burning is done in the dry season (normally December–February, in drought-years extending into April). Erica stands rarely reach more than 2–3.5 m height (~16–30 years age) before being burnt again. Recently burnt stands appear black until the next rainy season because of the charred top soil (Fig. 2a). Young stands (1–3 years) are characterized by a grass/herb-dominated lawn in between resprouting large old Erica lignotubers (Fig. 2b). The grass lawn cures in the dry season and young stands appear yellow in remote-sensed imagery. Wild and domestic grazers/browsers prefer the young stands where they consume grass, herbs, and young Erica shoots (Evangelista et al. 2007, Gustafsson 2009). Stands younger than five years cannot burn because of lack of fine dead fuels, and therefore act as fuel breaks in the landscape (Johansson et al. 2012, Johansson and Granström 2014). From ~5–8 years, the evergreen Erica shrub canopy gradually closes (Fig. 2c) and flammability increases with age because of fuel accumulation and the flammable nature of the Erica shrubs. Mature, 8–15 year old stands are easily ignited and burn in high-intensity crown fires, also under moderate drought (Johansson et al. 2012). From ~15–16 years, stand flammability gradually decreases because the flammable shrub canopy starts to vertically separate from the surface fuels beneath (Johansson et al. 2012; Fig. 2d).
All land is state owned, but traditional user rights apply to grazing land (Bekele 2003, Tesfaye et al. 2012). Burning of all vegetation has been banned nationally since the mid-1970s according to the forest law (the law literally translates to “destruction of forests”; Bekele 2003, Angassa and Oba 2008). In the Bale Mountains the degree of enforcement of the burn ban and grazing restrictions has varied over time, both inside and outside the park (Abera and Kinahan 2011). Generally, burn-ban enforcement has been stronger inside the park, even though the exact location of the park border is unknown on the ground. The park is so large that enforcement has been efficient mainly near the roads (Abera and Kinahan 2011). There is large variation in the total area burnt in different fire years (Abera and Kinahan 2011, Johansson et al. 2012, Belayneh et al. 2013). This depends mainly on dry-season weather, the amount and spatial distribution of mature, burnable, stands, and the ignition frequency. Extreme fire years in Bale Mountains largely coincide with regional prolonged droughts due to ENSO-related short-rain failures in March-April (Jury 2016, Zeleke et al. 2017; see Table A1.1). These have been frequent the last two decades (Sass-Klaassen et al. 2008, Mokria et al. 2017). In extended drought years, the numbers of burnt patches accumulate throughout the prolonged dry season until April. Potential natural fuel-breaks like mires (diameters ~10–400 m), rocky outcrops, and patches of less flammable shrub species (Alchemilla and Helichrysum) are small and cover < 10% of the landscape (Fig. A1.1a).
To quantify patch size change, aerial photos from 1968 and 1984, Landsat images from 1973, 1986, 2000, 2015, and 2017, and SPOT images from 2000, 2006, 2008, and 2011 were used (Table A1.2). To have a clear definition of fire year, our search-window started 1 January (aerial photos from December 1967 were assigned to fire-season 1968). The end of the search window was set to April to include prolonged droughts, but no images were found after February in any year. We selected all available cloud-free images spaced out over the 50-year period, but for the 1990s no cloud-free images were available for the park area. Thus, the 1995 data is not included in analyses. Aerial photos were scanned and georeferenced using ground control points (GCPs) in ArcGIS because interior orientation parameters were unavailable. Stand areas in these images are therefore not corrected for slope, which means that stands on steeper slopes (max 35%) are slightly underestimated in size, but these are rare and assumed to not affect the results. For further information on data handling see Appendix 1.2. The study area was limited to the heathland zone between 3500–3900 m.a.s.l. (Fig. 3). One large area of atypical land-cover was excluded (an area with large lava-flows, restricting fire movement), grasslands and Alchemilla spp. vegetation in the northwest of the park, and a small area of alpine vegetation in the southwest). We also excluded a 4 km wide buffer zone around the park border because the exact border location is unknown on the ground. For ground-truthing of stand age classes (see Appendix 1.3), landscape transects with GPS points on stand borders were collected inside the park in June 2018. For outside the park we used previously collected (2016, 2008, 2007) GPS points. Sample points for patch size quantification were randomly selected for each year using a 500-m grid system (Fig. 3, Appendix 1.2).
Borders of each individual Erica stand patch (resulting from one separate fire event) were manually traced using ArcMap 10.0 (ESRI 2011; cf. Allen et al. 2016). Manual delineation of patches was necessary because automatization proved to be inaccurate for several reasons: (1) sedge-dominated mires optically resemble young grass-dominated Erica stands, (2) background knowledge of how fires are ignited and move in the landscape (normally better uphill) was necessary to correctly interpret stand borders, (3) sometimes it was necessary to consult predated images or Google Earth historical scenes, and follow individual stands over time to correctly interpret patterns of hill-shadows etc.
Because of the tedious work of manually delineating borders, we decided to do many point sample replicates instead of a full interpretation and delineation of the whole landscape. This limited our landscape metrics to a statistical sample of patch sizes from each area and time, but did not allow us to calculate other types of spatial statistics across the whole areas such as, for example, connectivity between patches of similar age. From repeated field trips to the area, as well as information from local land users and remote-sensed imagery, we knew beforehand that young stands were dispersed across the landscape, albeit varying in size. Therefore we focused the study on young stand patch sizes, and the proportion of young stands in the landscape, because these are the most important metrics for our main study question: consecutive fire size.
Image resolution differed between years, but in order to most correctly interpret border patterns, and to avoid losing information, we aimed at using the highest available resolution for each year, with minor resampling of most images to keep most images within the same spatial resolution. Images with lower resolution would by necessity have less accurate estimates of sizes, but we believe any systematic error from our resampling of the images used at two instances of the spatio-temporal analysis is kept to the minimum because the highest resolution images were not necessarily from just one end of the time frame as it is the case in most temporally varying imagery datasets. This is because the earliest images were also of high resolution that match the resolution of the images from later years (those accessed through SPOT; see Appendix 1.2).
In all images, the borders between recently burnt, black (0–1 year), or young yellow (1–3 years), and mature green (> 4–5 years) stands were clearly visible (Fig. 4). A few cases of intermediate color were classified as “uncertain age” because these could be either 4-year-old stands, or older stands on flat land with sparse lignotubers. In such cases we used predated images to follow stand history. “Uncertain age” stands were excluded from final analyses. Borders between mature stands of different age were less clearly visible, as confirmed by ground truthing, and therefore we might have a general overestimation of mature stand sizes. Mature stands are not included in statistical analyses. Initially, patch size data was also grouped into confidence intervals (high confidence = 64%), but preliminary analyses showed the same results when using all patches, so the presented statistics and graphs are done using all delineated burnt and young stands. Out of the total 1011 randomly sampled points, 65% hit Erica stands, out of these 48% were mature, 8% of uncertain age, 38% young, and 6% recently burnt stands. Proportion of points hitting Erica stands, other vegetation types, and different landscape features in each year is shown in Figure A1.1a. Total area cover of Erica stands in each successional stage is shown in Figure A1.1b.
Grazing pressure is generally lower inside the park (Vial et al. 2011). Therefore Erica postfire growth might be slightly higher. Also, the southern slope dry season is slightly shorter and Erica postfire growth rates higher (~1 cm higher per year, personal observation). Therefore Erica stands could potentially become flammable earlier than the minimum 5 year age previously shown for outside the park on the northern aspect of the massif (Johansson et al. 2012). To confirm that the 5-year minimum age for flammability applies also inside the park and on the southern aspect, we analyzed pairs of satellite images, dated 1–2 years apart, to determine the prefire age-class of burnt stands. We used three pairs of images, 1999/2000, 2006/2008, 2015/2017, and for the later image we selected the delineated burnt-stand layer, and overlaid it on the 1–2 year predated image, and recorded the age-class of the same patches. Using the time-difference between images, we calculated the preburn age of each burnt stand at the time of fire. None of the burnt stands were younger than 5 years old when burnt (Appendix 1.4, Table A1.3).
First, to test if the proportion burnt area had changed over time, and if this differed between inside and outside the park, we performed a linear model test, with year and site (inside/outside park) as explanatory variables and the proportion of points hitting burnt/young stands, compared to all points hitting Erica stands of all age-classes. This variable was arcsin transformed to meet assumptions of normality. All further analyses were done on stand patch sizes (log-transformed to meet test assumptions) of pooled burnt plus young stands, (excluding mature stands) to represent successive fire events over a three-year period before the image date. Four main tests were conducted: (1) to test if patch size changed as a consequence of the national park establishment, we run a linear model with patch size as the dependent variable and site (inside/outside park), year, and their interaction, as explanatory variables; (2) patch size change over time was tested for the park and outside separately by linear models with patch size as the dependent and year as the explanatory variable; (3) number of large patches (> 100 ha) was tested separately with the same linear model, with patch size as the dependent variable and site, year, and their interaction as explanatory variables; and (4) minimum stand age at time of fire (Appendix 1.4) was tested by the probability for young and mature stands to burn by Pearson’s Chi-squared test with Yates’ continuity correction data. For all analyses we used the statistical software R3.1 (R Core Team 2014) and studied residuals plots to evaluate if the models met the statistical assumptions.
Between May and June 2016 we conducted 6 individual, and 10 group interviews (with 2–4 persons) involving a total of 41 respondents, who are all members of local agropastoral communities and utilize the heathlands inside and outside Bale Mountains National Park. Respondents were selected with the help of village heads and the local interpreter. Interviews were semistructured and photo-aided in order to collect detailed information on forest and heathland utilization and management and traditional ecological knowledge on fire ecology and pasture management. Interviews were 1–2 hours long, directly interpreted from Oromo language to English, and also tape-recorded and transcribed by independent translators. We asked about historical and current policy changes through the establishment of the national park, the joint forest management project outside the park, and the new REDD+ project. Interview questions and photos are presented in Appendix 2. To understand the social-ecological mechanisms behind heathland ecology and structure, and variation in burn sizes we specifically asked about ecological relationships, management goals and strategies, and historical changes in fire management. We asked for what they understand are the reasons behind the burn ban, their opinion on it, and if and how burn-ban enforcement has varied over time and space, and the effects of this. Specifically we sought answers to the following questions: What kind of fire management was practiced historically and today? What are the differences between inside the park and outside? What are the underlying causes of changes over the last decades? What is the status of the local traditional ecological knowledge regarding fire management? Questions were based on our previous knowledge about the system, mainly from outside the park (Johansson et al. 2012). There was high coherence between direct translation and audio transcripts. Respondents talked frankly about fire management, despite that heathland burning is illegal. Because of the qualitative nature of the questions asked, no quantitative analyses were done, but patterns of typical answers and exceptions were sought in relation to respondents’ wealth class, age group, or residency (Table A1.5). The same Erica species dominate both the old tree heather cloud forest below the tree line, and the fire-managed heathlands (which would become forest, if unburnt for > 50 years), and both zones are locally called Sato (meaning Erica). Therefore the heathlands were often referred to as “forest” in the responses.
The proportion of points hitting 0–3-year-old stands (burnt plus young stands pooled, Fig. A1.1.a) out of all Erica stands, had a tendency to increase over time, both inside and outside the park (P = 0.051), but there was no consistent difference in the proportion 0–3-year-old stands inside vs. outside the park, and no significant interaction. The main analysis showed that the variation in patch size of 0–3 year old stands was explained by an interaction effect between year and site (inside/outside park; P = 0.010), indicating that the trajectories of patch size change since 1969 have been different inside vs outside the park (Fig. 5). In 1968 and 1973 the average patch size (back-transformed from means of logged values) of 0–3-year-old stands in the two areas were in the same order of magnitude (8 vs. 10 ha in 1968, and 29 vs. 27 ha in 1973, inside/outside park, respectively). However, in 1984 and 1987 average patch sizes were larger inside the park (21 vs. 5 ha in 1984, and 27 vs. 5 ha in 1987, inside/outside park, respectively). In the 2000 Landsat image and the 2008 SPOT image (Fig. A1.3b) few, but large, black burns are visible inside the park, while outside the park burns are many and small. The statistical models showed a tendency for patch sizes to increase inside the park (P = 0.052) and a clear decrease outside the park (P < 0.001; Figs. 5 and A1.2).
Specifically, the largest fires (> 100 ha) have increased over time inside the park, but not outside (P = 0.019; Figs. 5 and 6). Prior to the 1990s, large fires did not occur every sampled year, and when occurring numbers were 2 or less. After 2000, large fires occurred every sampled year, in numbers of 3–5 inside the park and 0–1 outside the park. Ground photos of typical burn patterns are given in Figure 7a-b, and an enlarged section of the 2008 SPOT image is shown in Fig 7c.
Responses regarding livelihood strategies and heathland management were similar inside and outside the park. Older respondents gave more details on historical trends. Population growth, cessation of seasonal migration, agricultural expansion, and increased droughts were the major changes the last 50 years. Wealth-class, age, gender, or proportion livestock income (Table A1.5) had little influence on fire management knowledge or objectives, and responses were highly consistent. Respondents lived their whole life in the area, and their grandparents also lived there. Respondents from road-less villages on the southern aspect historically had less interaction with authorities and experienced less conflict regarding fire and grazing.
Old respondents said that 40 years ago all income came from cattle and honey, and they reared larger herds (some > 100 cows) comprising cattle and horses only. Today respondents reported to own a maximum of 30 cows, 10 horses, and 30 sheep. Few respondents reported keeping goats or donkeys. All respondents depend on the heathlands for pasture, especially during the dry season when the forest grass cures.
All respondents said that without frequent burning, the heathlands would soon turn into a continuous tall shrub without any grazing value. Reasons for burning the heathlands were found to be the same inside the park as outside. The major objectives were to increase grass/herb production, to get rid of a harmful urticating moth caterpillar, and to minimize livestock loss to hyenas and leopards (cf. Johansson et al. 2012). The heathlands are also important for honey production. Local wild bees are reared in log beehives, placed in trees at the tree line, and many forest tree species also contribute nectar. Erica arborea was said to produce more nectar than Erica trimera, and short fire cycles favors E. arborea because it flowers and sets seed earlier (from ~4 years, compared to E. trimera that starts flowering at ~11 years).
Agriculture accounted for up to 75% of household income and was dominated by barley, garlic, spring onions and cabbage. Livestock accounted for 10–90% of household income and this proportion was higher for older respondents. A few respondents earned a substantial share of their income from honey, bamboo, ecotourism, and park employment (Table A1.5).
The objectives for burning were the same today as prior to park establishment. Because it’s necessary they still burn, despite risking imprisonment. They said that eventually, if there was no fire for more than 50 years, the shrubs would grow into trees, but that this is unlikely to happen because complete ignition exclusion is impossible. Before the 1970s the fire management practice was different (Tables 1 and 2). Burning was done openly, and often collectively, by line ignition along the contour lines. Normally it was started earlier in the dry season: “As soon as the moss was dry, a group of men went out in the afternoon and ignited a mature stand in a somehow controlled way.” First they made sure that no cattle or children were present by shouting Guba! (Fire!). Then several men simultaneously ignited many shrubs along a line at the down-slope border of a “mature” (~8–14 years old) stand, in order to make an upslope high-intensity crown fire. Such, so-called guba qulquulo (clean fires), regular-shaped fires, consuming all shrub canopy within its borders, were preferred because they are perceived to produce better pasture and reduce the size of the remaining stumps, which restrict cattle access. Early-season fires normally produce white smoke (due to high water content of the foliage). Late-season fires produce black smoke (Fig. 7b) and these fires were referred to as “black-smoke fires.” These are more common today, according to respondents, and were seen as destructive fires, mainly because they burn the soil. Today, according to respondents, ignition is not made in the best way. Now, especially inside the park because of the risk of being caught, ignition is often made later in the dry season. They explained that the reason for this is because it is easier to succeed with a quick-point ignition on the sly when the vegetation is very dry. This is often done by a single man passing on horseback, “...just throwing a cigarette, or quickly igniting a single shrub when no one sees him.” Some ignitions are today also purposely made late in the dry season because of conflict with the park, according to respondents.
Respondents said that fires inside the park have recently become larger and more dangerous, and exemplified this with the March 2015 fire in which one park staff was killed. They said that casualties never happened before. Though the park has some resources to patrol the heathlands, the area is too large to enforce the law everywhere and communities do not report ignition: “Because everyone needs pasture land, no one reports a burning incidence to the park.” The risk of being caught igniting is higher near routes frequently patrolled by park rangers. The penalty for igniting is imprisonment with large penalty fees, both inside and outside the park. The park has more resources to control grazing and new settlements, but grazing pressure increases also inside the park, and new homesteads have been established above the tree line (Table A1.4). Respondents inside the park had in the past experienced more conflict with authorities, especially regarding burning and grazing. Some respondents also mentioned positive outcomes of the park, such as improved environmental protection and ecotourism. All respondents had an overall positive attitude toward the national park and the REDD+ project because they see the need for conservation and anticipate benefits from carbon trading and ecotourism. Relations between communities and the park has improved, with new benefits-sharing projects like schools, water, and electricity, but crop-raiding by wildlife from the park is also an increasing problem for the communities. Respondents also expected that they will be allowed to stay in their villages in the future. One respondent inside the park stressed the need to stay as follows: “If communities get money, they will manage the forest [including heathland] ... but if they get displaced from their village they would not want to receive even a billion birr, but they would choose to stay.” Young respondents, especially outside the park, were aware of the concept of carbon storage, and said that they had been promised payments for planting trees (which some respondents had done, and were asking when payments will be made). Only one respondent, outside the park, received payments for forest protection. There was varied understanding of how REDD+ payments will work and whether it will benefit the communities or not, as indicated in the following quotes: “I didn’t agree with carbon money because they may take our good forest and give us bad climate ... but if it doesn’t have negative side-effects it might be a good idea.” “Money paid for tree protection is good, but if cutting forest is totally forbidden it will be problem because then we will not be allowed even the stick we hold in our hand.” “It is possible to protect our forest to store carbon ... it is good if they pay 800–1000 birr [~US$27–34] per ha of pasture land.” “It is a good idea to get money for protecting the forest because the reason for deforestation is also to get money.” Outside the park, ignition frequency has increased since the 1970s, despite everyone knowing that it is illegal. Therefore, according to respondents, stand patches have become smaller. Grazing pressure has increased and is generally higher than in the park. Attempting to increase pasture production, men, and especially children, now ignite increasingly frequently, even under poor burning conditions. Children are often igniting because they cannot be put in prison, and ignitions increase during school breaks, according to respondents and observations. The new ignition practice with quick-point ignitions on the sly now also applies outside the park because burn-ban enforcement has recently increased on account of the REDD+ project.
Respondents said that lately the extended droughts occurred every 3–4 years, and that these make the fires more destructive, especially in the park. Some respondents explained that the reason behind increased fire sizes is primarily the lack of young stands that stop the fires. It appears that most respondents clearly understand that when few stands are burnt early-season or in poor fire years, the fires in drought years become larger. Another explanation according to the respondents, could be that during extremely dry and windy conditions, fire might travel even in 5-year-old stands, and in old “overgrown” stands (Fig. 2d), which normally do not burn. Respondents said that young stands are impossible to burn, and old respondents said this was the case also when grazing pressure was lower when they were young. Respondents inside the park were worried that fires have become more dangerous and out of control, like the April 2008 fires when the park ordered helicopters, military servants, and school children to stop fires, at the same time as local youngsters were igniting down slope. Discussing a photo of this event, respondents said, “It is very dangerous, stopping fire in mature Erica is impossible” and “It is because of the conflict that someone was still igniting below.” Respondents also emphasized the problem with soil fires, and that this was uncommon earlier. They said that late-season burning, when the humus is dry, should be banned because soil fires kill the Erica lignotubers. They said that the Erica is needed in the system, both as dry season forage, and to fuel the fires, which they assume rejuvenate the pasture. All respondents said that pasture quality and milk production per cow has declined, mainly because of increased livestock density, but also because of year-around grazing. In the past, the heathlands were not grazed in the rainy season because, according to respondents, the cold fog makes cows and children sick, and increases livestock loss to predators. Recently also sheep and goats started grazing the heathlands and they browse the Erica shrubs more intensely, and, according to respondents, this may eventually kill the Erica. Traditionally all grazing land, both above and below the tree line, was communal land, but lately many cattle owners outside the park started building fences to secure longer grass for their own livestock. This was seen as a benefit to the fence owner, but a problem for the community because it causes deforestation. Fencing pasture is illegal but, according to respondents, started following the joint forest management project outside the park. A new phenomenon in the last decade is that it is now possible to rent access to pasture land despite the fact that no grazing land is private.
The traditional ecological knowledge regarding fire ecology and fire management is deep and widespread, both inside and outside the park (Table 2). Respondents could with high precision tell the postfire age of Erica stands in photos. They know that stands younger than five years old (“knee-height”) cannot burn because of lack of fuels, and that 8–14-year-old stands (“waist-height”) are “mature” and easy to burn, and when stands are taller than head-height, they are “overgrown” and need extreme dry weather to burn. They said, “The local community knows more than the agricultural office about fire.” Only one respondent, who worked as a ranger, stated that burning is bad: “As we were trained by the experts, burning Sato is bad and causes mineral depletion ... and kills wildlife.” But later he said that without fire there would be no pasture, and “the heathlands have been burnt since my grandfathers’ time.” The old men complained that children nowadays have little fire knowledge: “children make small dirty burns with messy borders ... full of green areas inside, or with large stumps.” But they said that children on the other hand know other things that the elders don’t know, for example, that the urticating caterpillar will become a moth. Fire management is a male task, nevertheless women also had detailed knowledge about fire ecology, burn practices, and ecological effects on pasture quality and cattle health. One old woman pointed out the fact that it is necessary to have “both young and old stands within daily walking distance of the cattle” to provide a continuous supply of young stands over time. Even though the cattle are not herded in the heathlands, most respondents were well aware of pasture ecology, for example, which herb species were preferred by different animals, and that the proportion grass cover had decreased and unpalatable herbs increased because of overgrazing. Some respondents mentioned the importance of the shrubs for protecting grasses and herbs, and preventing erosion. The respondents were interested in reaching a more constructive dialogue with the authorities on how to stop the dangerous late-season fires, and also to solve the problem of overgrazing.
Our empirical results show that changes in policy and land use has altered fire regimes inside and outside the park in diverging directions. Before the creation of the park, heathland fire sizes were almost equal outside and inside the present park area. Since then, with some delay, fire sizes have decreased outside, but increased inside the park. Our interview results clarify the underlying mechanisms behind these trends: (1) burning is essential for local livelihoods and cannot be excluded from this highly flammable vegetation, (2) the more strongly enforced burn ban inside the national park has reduced ignition rates and hence the number of young stands acting as fuel-breaks, resulting in larger fires, (3) the burn ban has shifted ignitions to later in the dry season, (4) outside the park increased ignition, caused by intensified land use, has decreased fire sizes, and (5) extended droughts have been frequent lately, but large fires occurred mainly inside the park (Fig. 8).
Despite the burn ban, there was no decrease in total burnt/young area in the park, but instead an increase. Three factors control fire sizes in the study landscape. The first is related to the mosaic structure of young stands in the landscape, resulting from preceding decades of ignition (cf. Minnich and Chou 1997). Second, there is often large interannual variation in fire frequency because of variation in dry season weather (cf. Seydack et al. 2007). The Bale heathland climate is normally extremely moist, and fire is normally possible only between December and February. In poor fire years only a few fires occur, and then mainly outside the park. In good fire years, fires are many, both inside and outside the park, but they become larger inside the park. The third factor is the ignition frequency. Anthropogenic ignition is, in this case balanced between local needs to produce food, and the official fire-suppression policy (cf. Pyne 1997, Laris and Wardell 2006, Kull and Laris 2009). Outside the park ignitions were frequent, even under poor burning conditions in bad fire years, as described in interviews and observed in the field. This could explain the few, small fires found outside the park only in poor fire years (Fig. 5).
Our images were dated from (December)/January to February, hence the recorded burnt stands do not include late-season March-April fires in drought years. Late-season fires are instead visible in the distribution of young stands 1–3 years later. For example the largest 2400 ha young patch recorded in 2011, burned in April 2008 (which happens to be the black-smoke fire in Fig. 7b). Unfortunately our empirical data cannot quantify the shift in timing. But respondents stated that soil fires occur only during extended droughts, and these were rare in the past. This corroborates the fact that the heathland climate normally is very moist, and the humus layer rarely dries out (personal observation, between 2005 and 2016). The thick layers of accumulated humus, and the old age of the large E. trimera lignotubers (Johansson et al. 2012) also indicate that late-season smouldering humus fires have been rare in the past.
The patch-mosaic theory proposed by Minnich and Chou (1997) suggests that ignition saturation results in a fine-grained mosaic of young stands acting as fuel breaks in the landscape, exemplified by Mexican Chaparral. This requires ignition saturation, and that young stands really are nonflammable. This is not the case in the Californian Chaparral, where under extreme weather, fires burn through all age-classes, partly because of invasive annual grasses (Keeley and Fotheringham 2001, Brooks et al. 2004). In the Cape Fynbos, Seydack et al. (2007) showed that young stands were fuel-limited and did function as fuel breaks in the low-productivity drier proteoid shrublands, but not in the more productive alpine heathlands, which today are ungrazed. In Patagonia intense livestock grazing altered fuel composition and decreased shrubland flammability (Blackhall et al. 2017). In subalpine heathlands in Victoria, Australia, cattle were reported to have little effect on shrub fuels, because they avoided old stands (Williams et al. 2006). Apart from low site productivity, fuel limitation in young stands can also arise because of intensive range production where livestock consume large quantities of grass and shrub fuel, especially in young stands, as in Bale Mountains. Our results show that stands younger than 5 years do act as fuel breaks, but we do not know whether our 1–3 year old stands could potentially transfer fire if the grass was not grazed short. However, old respondents said that young stands also stopped consecutive fires 40 years ago, when grazing pressure was lower. Young stands are also intensively grazed by large populations of alpine rodents (Vial et al. 2011, Jira et al. 2013). According to our previous experiments (Johansson et al., unpublished data), a two-year-old fenced stand with ~15 cm tall, not fully cured, grass stopped a fire in December 2008.
In the Bale Mountains heathlands, cattle grazing, at moderate stocking density, is an important factor for the fire regime and for biodiversity. Cattle have probably grazed these heathlands since anthropogenic burning started, at least 2000 years ago (Gil-Romera et al. 2019). It is the combination of patch burning and grazing that created this species-rich landscape, and both are necessary for maintaining its resilience.
We did not quantify other landscape metrics than stand size and the proportion of young stands because of methodological constrains (see methods). However, it could have been interesting to compare other spatial characteristics such as mean distance between patches of similar age. Yet, based on qualitative impressions from the imagery and from the field, it is quite clear that small fire sizes also lead to a mosaic pattern with shorter mean distances between patches of similar age.
Young stands provide habitat for many Afroalpine plants (Johansson et al. 2018) and also to grass-eating rodents, important prey for the Simien Wolf (Canis simensis; Vial et al. 2011). Conservationists have not yet embraced the concept of high biological values of cultural landscapes in Africa (Neumann 2014). This has caused conflicts in many national parks regarding fire and grazing (Neumann 1998). We argue that, just like European heathlands, the Bale Mountains heathlands should be understood as an ancient cultural landscape, and managed in the traditional way, but science-based, legally, and controlled.
The reintroduction of patch-burning as a tool for shrubland fuel reduction and/or biodiversity conservation has been controversial (Bradstock et al. 1995, Russell-Smith et al. 2002, Allen 2008). Recently burnt open stands provide habitat to different herbaceous communities than old shady stands (Rundel 1998, Wesche 2006, Allen 2008, Johansson et al. 2018). Hence prescribed burning can increase biodiversity, if young stands are rare in the landscape. But if old stands are rare and harbor threatened species, burning can reduce biodiversity (Bradstock et al. 1995, Syphard et al. 2019). Optimal management should depend on the historic fire regime under which current plant communities assembled. Reintroduction of patch-burning (Fernandes et al. 2013) and livestock grazing (Lovreglio et al. 2014, Marques et al. 2017) has been successful mainly in the Mediterranean basin, maybe because of the short time span since their cessation. Prescribed early-season fires burning under mild fire weather can actually increase the abundance of old patches in the landscape because then they are less flammable (Murphy et al. 2015). Once the fine-grained landscape mosaic has been lost, it can be difficult to restore it in highly flammable vegetation because of the lack of young stands to burn against (Murphy et al. 2015).
In Bale Mountains, while land management diverged in two opposing directions over the last 50 years, climate change should also have influenced the fire regimes. As respondents pointed out, extended droughts have been frequent recently (also cf. Mokria et al. 2017, Mekonnen et al. 2018). This has four implications on the fire regime: a longer dry period to accumulate fires, more late-season fires, reduced primary production increasing fuel build-up time, and the risk of smouldering soil fires killing the Erica lignotubers. Droughts affect both areas equally, but increased fire size has become a problem inside the park only. Outside the park intensified land use pre-emptied heathland fuels and thus potentially partly counteract some effects of climate change (Fig. 8).
Our interviews showed that the traditional fire knowledge is still vivid, deep, and widespread among local communities. It is evidence-based, adaptive, and has evolved through long-term interaction with the heathland ecosystem. The contrast between the traditional knowledge and the enforced burn ban has caused confusion and conflict. Even though respondents were aware of the arguments behind the burn ban, they were unable to follow the law in order to feed their livestock. This has caused recurring conflicts, changed ignition patterns, and has resulted in larger fires. This, in combination with frequent droughts, has created a dangerous situation that both communities and authorities wish to end. Respondents emphasized that late-season burning should be stopped, and that this would be facilitated through a dialogue between the park and the communities on legalization and regulation of fire management. Misconceptions among authorities regarding traditional fire management have caused similar problems of fuel build-up and increased fire size and severity in, e.g., Australian savannas (Yibarbuk et al. 2001), Portuguese shrublands (Fernandes et al. 2013), and Brazilian Cerrados (Eloy et al. 2019). In Ethiopia, 10 years of almost total fire exclusion contributed to the exceptionally large heathland fire on Mount Kaka in 2012 (Ararsa Beyene, personal communication). Possibly some of the large heathland fires reported from other East African heathlands could also be partly caused by cessation of (largely undocumented) preceding traditional fire management.
The main aim of this study was to increase our understanding of how changes in policy and land use has altered the fire regimes, and how this has implications for the new REDD+ project, and potentially also for other REDD+ projects in flammable ecosystems. Increasing land scarcity, conflicting goals, and fire exclusion has led to a situation of nonoptimal fire management. The fact that burning is illegal hinders the development of a sound fire management policy. Today the situation in the park is difficult, with increasingly large fires, in combination with recurrent droughts. But the prospective to develop a successful community-based fire management program is better than in many other shrubland systems. The old Erica lignotubers have been resilient to a wide range of fire intervals, the stand structure is not yet completely homogenized, the traditional fire management knowledge is still present, there are no invasive fire-regime altering species, and the fire-weather is not as extreme as in other places. There is a genuine interest among communities to cooperate with authorities to develop an acceptable joint regulation of burning and grazing, as long as they are allowed to remain in the park. The park authority has acknowledged the ecological role of fire in the heathlands. There is a growing awareness of the necessity of science-based prescribed burning to increase long-term carbon storage and biodiversity in savanna systems (Teketay 2000, FAO 2011). Traditional fire management can be successfully integrated in climate mitigation projects. This was first tried in the West Arnhem Land fire abatement project in Australia (Yibarbuk et al. 2001). This concept has now been spread to, e.g., the Brazilian Cerrados (Eloy et al. 2019) and Botswana by The International Savanna Fire Management Initiative, which will continue giving practical workshops in Africa on joint fire management. In the face of climate change, it is necessary to adapt land management strategies in order to reduce wildfire risk (Dearing et al. 2010, Fernandes 2013, van Breugel et al. 2016). Attempts at short-term increase carbon capture, without a thorough understanding of the local social-ecological context of fire and grazing, can lead to large-scale wildfires, carbon loss, and ecosystem degradation in the long term. We suggest that REDD+ projects in flammable ecosystems first investigate the fire history and social context of the systems they aim to restore, and then adopt fire policy based on this, instead of initially enforcing blanket burn bans and grazing restrictions. We suggest that the Bale Mountains REDD+ project abandons the strict fire-exclusion policy, and instead develops a joint fire management program for the heathlands, agreed upon by all stakeholders. This should include prescribed early-season small burns, with varied rotation intervals, and a respected ban on late-season ignition, as well as grazing regulations. In developing this program there is a need to make use of the vivid local traditional fire knowledge, which is globally unique for a flammable shrubland system.
Many thanks to assistant Ayano Abraham who interpreted interviews and collected ground truthing data. Carl Frisk collected ground truthing data. Tadele Kifle transcribed interview audio files to English. Historical satellite images were acquired free of charge on a requisition grant from USGS. Brook Daniel assisted in acquiring and geo-rectifying aerial photos. CIFOR provided office space and support in Ethiopia. Helle Skånes provided GIS advice, Victor Johansson gave advice on figures, and Lowe Börjesson gave advice on qualitative interview interpretation and presentation.
Abera, K., and A. A. Kinahan. 2011. Factors affecting fire extent and frequency in the Bale mountains National Park. Walia 2011:146-157.
Allen, H. D. 2008. Fire: plant functional types and patch mosaic burning in fire-prone ecosystems. Progress in Physical Geography 32:421-437. https://doi.org/10.1177/0309133308096754
Allen, K. A., P. Denelle, F. M. S. Ruiz, V. M. Santana, and R. H. Marrs. 2016. Prescribed moorland burning meets good practice guidelines: a monitoring case study using aerial photography in the Peak District, UK. Ecological Indicators 62:76-85. https://doi.org/10.1016/j.ecolind.2015.11.030
Anderson, M. K., and J. E. Keeley. 2018. Native peoples’ relationship to the California chaparral. Pages 79-121 in E. C. Underwood, H. D. Safford, N. A. Molinari, and J. E. Keeley, editors. Valuing chaparral: ecological, socio-economic, and management perspectives. Springer International, Cham, Switzerland. https://doi.org/10.1007/978-3-319-68303-4_4
Angassa, A., and G. Oba. 2008. Herder perceptions on impacts of range enclosures, crop farming, fire ban and bush encroachment on the rangelands of Borana, southern Ethiopia. Human Ecology 36:201-215. https://doi.org/10.1007/s10745-007-9156-z
Archibald, S. 2016. Managing the human component of fire regimes: lessons from Africa. Philosophical Transactions of the Royal Society B: Biological Sciences 371:20150346. https://doi.org/10.1098/rstb.2015.0346
Archibald, S., C. E. R. Lehmann, J. L. Gömez-Dans, and R. A. Bradstock. 2013. Defining pyromes and global syndromes of fire regimes. Proceedings of the National Academy of Sciences of the United States of America 110:6442-6447. https://doi.org/10.1073/pnas.1211466110
Archibald, S., A. C. Staver, and S. A. Levin. 2012. Evolution of human-driven fire regimes in Africa. Proceedings of the National Academy of Sciences of the United States of America 109:847-852. https://doi.org/10.1073/pnas.1118648109
Baeza, M. J., V. M. Santana, J. G. Pausas, and V. R. Vallejo. 2011. Successional trends in standing dead biomass in Mediterranean basin species. Journal of Vegetation Science 22:467-474. https://doi.org/10.1111/j.1654-1103.2011.01262.x
Barlow, J., L. Parry, T. A. Gardner, J. Ferreira, L. E. O. C. Aragão, R. Carmenta, E. Berenguer, I. C. G. Vieira, C. Souza, and M. A. Cochrane. 2012. The critical importance of considering fire in REDD+ programs. Biological Conservation 154:1-8. https://doi.org/10.1016/j.biocon.2012.03.034
Bekele, M. 2003. Forest property rights, the role of the state, and intitutional exigency. Dissertation. Swedish University of Agricultural Sciences, Uppsala, Sweden.
Belayneh, A., T. Yohannes, and A. Worku. 2013. Recurrent and extensive forest fire incidence in the Bale Mountains National Park (BMNP), Ethiopia: extent, cause and consequences. International Journal of Environmental Sciences 2:29-39.
Blackhall, M., E. Raffaele, J. Paritsis, F. Tiribelli, J. M. Morales, T. Kitzberger, J. H. Gowda, and T. T. Veblen. 2017. Effects of biological legacies and herbivory on fuels and flammability traits: a long-term experimental study of alternative stable states. Journal of Ecology 105:1309-1322. https://doi.org/10.1111/1365-2745.12796
Bond, W. J., and J. J. Midgley. 2012. Fire and the angiosperm revolutions. International Journal of Plant Sciences 173:569-583. https://doi.org/10.1086/665819
Bowman, D. M. J. S., J. K. Balch, P. Artaxo, W. J. Bond, J. M. Carlson, M. A. Cochrane, C. M. D'Antonio, R. S. DeFries, J. C. Doyle, S. P. Harrison, F. H. Johnston, J. E. Keeley, M. A. Krawchuk, C. A. Kull, J. B. Marston, M. A. Moritz, I. C. Prentice, C. I. Roos, A. C. Scott, T. W. Swetnam, G. R. van der Werf, and S. J. Pyne. 2009. Fire in the Earth system. Science 324:481-484. https://doi.org/10.1126/science.1163886
Bowman, D. M. J. S., J. Balch, P. Artaxo, W. J. Bond, M. A. Cochrane, C. M. D'Antonio, R. DeFries, F. H. Johnston, J. E. Keeley, M. A. Krawchuk, C. A. Kull, M. Mack, M. A. Moritz, S. Pyne, C. I. Roos, A. C. Scott, N. S. Sodhi, and T. W. Swetnam. 2011. The human dimension of fire regimes on Earth. Journal of Biogeography 38:2223-2236. https://doi.org/10.1111/j.1365-2699.2011.02595.x
Bradstock, R. A., A. M. Gill, and R. J. Williams. 2012. Flammable Australia: fire regimes, biodiversity and ecosystems in a changing world. CSIRO Publishing, Melbourne, Australia. https://doi.org/10.1071/9780643104839
Bradstock, R. A., D. A. Keith, and T. D. Auld. 1995. Fire and conservation: imperatives and constraints on managing for diversity. Pages 323-333 in R. A. Bradstock, T. D. Auld, D. A. Keith, R. T. Kingsford, D. Lunney, and D. P. Sivertsen, editors. Conserving biodiversity: threats and solutions. Surrey Beatty & Sons, Australia.
Brooks, M. L., C. M. D'Antonio, D. M. Richardson, J. B. Grace, J. E. Keeley, J. M. DiTomaso, R. J. Hobbs, M. Pellant, and D. Pyke. 2004. Effects of invasive alien plants on fire regimes. BioScience 54:677-688. https://doi.org/10.1641/0006-3568(2004)054[0677:EOIAPO]2.0.CO;2
Butz, R. J. 2009. Traditional fire management: historical fire regimes and land use change in pastoral East Africa. International Journal of Wildland Fire 18:442-450. https://doi.org/10.1071/WF07067
Dass, P., B. Z. Houlton, Y. P. Wang, and D. Warlind. 2018. Grasslands may be more reliable carbon sinks than forests in California. Environmental Research Letters 13:7. https://doi.org/10.1088/1748-9326/aacb39
Dearing, J. A., A. K. Braimoh, A. Reenberg, B. L. Turner, and S. van der Leeuw. 2010. Complex land systems: the need for long time perspectives to assess their future. Ecology and Society 15(4):21. https://doi.org/10.5751/ES-03645-150421
Eloy, L., B. A. Bilbao, J. Mistry, and I. B. Schmidt. 2019. From fire suppression to fire management: advances and resistances to changes in fire policy in the savannas of Brazil and Venezuela. Geographical Journal 185:10-22. https://doi.org/10.1111/geoj.12245
Environmental Systems Research Institute (ESRI). 2011. ArcGIS Desktop. ESRI, Redlands, California, USA.
Evangelista, P., P. Swartzinski, and R. Waltermire. 2007. A profile of the mountain nyala (Tragelaphus buxtoni). African Indaba 5(2).
Fernandes, P. M. 2013. Fire-smart management of forest landscapes in the Mediterranean basin under global change. Landscape and Urban Planning 110:175-182. https://doi.org/10.1016/j.landurbplan.2012.10.014
Fernandes, P. M., G. M. Davies, D. Ascoli, C. Fernández, F. Moreira, E. Rigolot, C. R. Stoof, J. A. Vega, and D. Molina. 2013. Prescribed burning in southern Europe: developing fire management in a dynamic landscape. Frontiers in Ecology and the Environment 11:E4-E14. https://doi.org/10.1890/120298
Fetene, M., Y. Assefa, M. Gashaw, Z. Woldu, and E. Beck. 2006. Diversity of Afroalpine vegetation and ecology of treeline species in the Bale Mountains, Ethiopia, and the influence of fire. Pages 25-38 in E. M. Spehn, M. Liberman, and C. Korner, editors. Land use change and mountain biodiversity. CRC Press-Taylor & Francis Group, Boca Raton, Florida, USA. https://doi.org/10.1201/9781420002874-2
Food and Agriculture Organization (FAO). 2011. Community-based fire management: a review. FAO, Rome, Italy.
Gil-Romera, G., C. Adolf, B. M. Benito, L. Bittner, M. U. Johansson, D. A. Grady, H. F. Lamb, B. Lemma, M. Fekadu, B. Glaser, B. Mekonnen, M. Sevilla-Callejo, M. Zech, W. Zech, and G. Miehe. 2019. Long-term fire resilience of the Ericaceous Belt, Bale Mountains, Ethiopia. Biology Letters 15:20190357. https://doi.org/10.1098/rsbl.2019.0357
Gill, A. M. 1975. Fire and the Australian flora: a review. Australian Forestry 38:4-25. https://doi.org/10.1080/00049158.1975.10675618
Gustafsson, J. 2009. Habitat and plant selection of livestock in a fire-managed Afro-alpine heathland in Ethiopia. Thesis. Swedish University of Agricultural Science, Uppsala, Sweden.
Hedberg, O. 1964. Features of Afroalpine plant ecology. Swedish Phytogeographical Society, Uppsala, Sweden.
Hemp, A. 2005. Climate change-driven forest fires marginalize the impact of ice cap wasting on Kilimanjaro. Global Change Biology 11:1013-1023. https://doi.org/10.1111/j.1365-2486.2005.00968.x
Hillman, J. C. 1988. The Bale Mountains National Park Area, southeast Ethiopia, and its management. Mountain Research and Development 8:253-258. https://doi.org/10.2307/3673456
Hobbs, R. J., and C. H. Gimingham. 1987. Vegetation, fire and herbivore interactions in heathland. Advances in Ecological Research 16:87-173. https://doi.org/10.1016/S0065-2504(08)60088-4
Jacob, M., S. Annys, A. Frankl, M. De Ridder, H. Beeckman, E. Guyassa, and J. Nyssen. 2015. Tree line dynamics in the tropical African highlands - identifying drivers and dynamics. Journal of Vegetation Science 26:9-20. https://doi.org/10.1111/jvs.12215
Jira, G., A. Bekele, G. Hemson, and B. Mundanthra. 2013. Rodents in fire affected heather shrublands in Bale Mountains National Park, Ethiopia. Journal of King Saud University - Science 25:289-295. https://doi.org/10.1016/j.jksus.2013.05.006
Johansson, M., C. A. Frisk, S. Nemomissa, and K. Hylander. 2018. Disturbance from traditional fire management in subalpine heathlands increases Afro-alpine plant resilience to climate change. Global Change Biology 24:2952-2964. https://doi.org/10.1111/gcb.14121
Johansson, M. U., M. Fetene, A. Malmer, and A. Granström. 2012. Tending for cattle: traditional fire management in Ethiopian montane heathlands. Ecology and Society 17(3):19. https://doi.org/10.5751/ES-04881-170319
Johansson, M. U., and A. Granström. 2014. Fuel, fire and cattle in African highlands: traditional management maintains a mosaic heathland landscape. Journal of Applied Ecology 51:1396-1405. https://doi.org/10.1111/1365-2664.12291
Johnson, C. N., L. D. Prior, S. Archibald, H. M. Poulos, A. M. Barton, G. J. Williamson, and D. M. J. S. Bowman. 2018. Can trophic rewilding reduce the impact of fire in a more flammable world? Philosophical Transactions of the Royal Society B: Biological Sciences 373(1761). https://doi.org/10.1098/rstb.2017.0443
Jury, M. R. 2016. Determinants of southeast Ethiopia seasonal rainfall. Dynamics of Atmospheres and Oceans 76:63-71. https://doi.org/10.1016/j.dynatmoce.2016.08.004
Keeley, J. E. 2002. Native American impacts on fire regimes of the California coastal ranges. Journal of Biogeography 29:303-320. https://doi.org/10.1046/j.1365-2699.2002.00676.x
Keeley, J. E., W. J. Bond, R. A. Bradstock, J. G. Pausas, and P. W. Rundel. 2012. Fire in Mediterranean ecosystems: ecology, evolution and management. Cambridge University Press, Cambridge, UK. https://doi.org/10.1017/CBO9781139033091
Keeley, J. E., and C. J. Fotheringham. 2001. History and management of crown-fire ecosystems: a summary and response. Conservation Biology 15:1561-1567. https://doi.org/10.1046/j.1523-1739.2001.t01-1-00186.x
Keith, D. A., D. Lindenmayer, A. Lowe, J. Russel-Smith, S. Barrett, N. Enright, B. Fox, G. Guerin, D. Panton, M. Tozer, and C. Yates. 2014. Heathlands. Pages 283-334 in D. Lindenmayer, E. Burns, N. Thurgate, and A. Lowe, editors. Biodiversity and environmental change: monitoring, challenges and direction. CSIRO Publishing. https://doi.org/10.1071/9780643108578
Khatun, K., E. Corbera, and S. Ball. 2016. Fire is REDD+: offsetting carbon through early burning activities in south-eastern Tanzania. Oryx 51:43-52. https://doi.org/10.1017/S0030605316000090
Kull, C. A., and P. Laris. 2009. Fire ecology and fire politics in Mali and Madagascar. Pages 171-226 in M. A. Cochrane, editor. Tropical fire ecology: climate change, land use, and ecosystem dynamics. Springer-Verlag, Berlin, Germany. https://doi.org/10.1007/978-3-540-77381-8_7
Laris, P., and D. A. Wardell. 2006. Good, bad or ‘necessary evil’? Reinterpreting the colonial burning experiments in the savanna landscapes of West Africa. Geographical Journal 172:271-290. https://doi.org/10.1111/j.1475-4959.2006.00215.x
Lovreglio, R., O. Meddour-Sahar, and V. Leone. 2014. Goat grazing as a wildfire prevention tool: a basic review. iForest - Biogeosciences and Forestry 7:260-268. https://doi.org/10.3832/ifor1112-007
Marques, D., M. Fachada, and H. Viana. 2017. Synergies between goat grazing and shrub biomass in mountain areas. Pages 155-175 in J. Simões and C. Gutiérrez, editors. Sustainable goat production in adverse environments: Volume I: welfare, health and breeding. Springer International, Cham, Switzerland. https://doi.org/10.1007/978-3-319-71855-2_10
Mekonnen, Z., H. Kassa, T. Woldeamanuel, and Z. Asfaw. 2018. Analysis of observed and perceived climate change and variability in Arsi Negele District, Ethiopia. Environment Development and Sustainability 20:1191-1212. https://doi.org/10.1007/s10668-017-9934-8
Miehe, G., and S. Miehe. 1994. Ericaceous forests and heathlands in the Bale mountains of South Ethiopia: ecology and man’s impact. Stiftung Walderhaltung in Afrika, Hamburg, Germany.
Minnich, R. A., and Y. H. Chou. 1997. Wildland fire patch dynamics in the chaparral of southern California and northern Baja California. International Journal of Wildland Fire 7:221-248. https://doi.org/10.1071/WF9970221
Mokria, M., A. Gebrekirstos, A. Abiyu, M. Van Noordwijk, and A. Bräuning. 2017. Multi-century tree-ring precipitation record reveals increasing frequency of extreme dry events in the upper Blue Nile River catchment. Global Change Biology 23:5436-5454. https://doi.org/10.1111/gcb.13809
Murphy, B. P., M. A. Cochrane, and J. Russell-Smith. 2015. Prescribed burning protects endangered tropical heathlands of the Arnhem Plateau, northern Australia. Journal of Applied Ecology 52:980-991. https://doi.org/10.1111/1365-2664.12455
Neumann, F. H., L. Scott, and M. K. Bamford. 2011. Climate change and human disturbance of fynbos vegetation during the late Holocene at Princess Vlei, Western Cape, South Africa. Holocene 21:1137-1149. https://doi.org/10.1177/0959683611400461
Neumann, R. P. 1998. Imposing wilderness: struggles over livelihood and nature preservation in Africa. University of California Press, Berkeley, California, USA.
Neumann, R. P. 2014. Stories of nature’s hybridity in Europe: implications for forest conservation in the Global South. Pages 31-44 in S. B. Hecht, K. D. Morrison, and C. Padoch, editors. The social lives of forests: past, present, and future of woodland resurgence. The University of Chicago Press, Chicago, Illinois, USA. https://doi.org/10.7208/chicago/9780226024134.003.0004
Oromia Forest and Wildlife Enterprise. 2016. Bale Mountains Eco-region REDD+ project monitoring and implementation report. Oromia Forest and Wildlife Enterprise, Farm Africa, SOS Sahel, Addis Ababa, Ethiopia.
Parr, C. L., C. E. R. Lehmann, W. J. Bond, W. A. Hoffmann, and A. N. Andersen. 2014. Tropical grassy biomes: misunderstood, neglected, and under threat. Trends in Ecology & Evolution 29:205-213. https://doi.org/10.1016/j.tree.2014.02.004
Pausas, J. G., and S. Fernández-Muñoz. 2012. Fire regime changes in the Western Mediterranean Basin: from fuel-limited to drought-driven fire regime. Climatic Change 110:215-226. https://doi.org/10.1007/s10584-011-0060-6
Pausas, J. G., and J. E. Keeley. 2009. A burning story: the role of fire in the history of life. BioScience 59:593-601. https://doi.org/10.1525/bio.2009.59.7.10
Pausas, J. G., and J. E. Keeley. 2014. Abrupt climate-independent fire regime changes. Ecosystems 17:1109-1120. https://doi.org/10.1007/s10021-014-9773-5
Pyne, S. J. 1997. Vestal fire: an environmental history, told through fire, of Europe and Europe’s encounter with the world. University of Washington Press, Seattle, Washington, USA.
Pyne, S. J. 2016. Fire in the mind: changing understandings of fire in Western civilization. Philosophical Transactions of the Royal Society B: Biological Sciences 371:8. https://doi.org/10.1098/rstb.2015.0166
R Core Team. 2014. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.
Rundel, P. W. 1998. Landscape disturbance in Mediterranean-type ecosystems: an overview. Pages 3-22 in P. W. Rundel, G. Montenegro, and F. M. Jaksic, editors. Landscape disturbance and biodiversity in Mediterranean-type ecosystems. Springer, Berlin, Germany. https://doi.org/10.1007/978-3-662-03543-6_1
Russell-Smith, J., P. G. Ryan, and D. C. Cheal. 2002. Fire regimes and the conservation of sandstone heath in monsoonal northern Australia: frequency, interval, patchiness. Biological Conservation 104:91-106. https://doi.org/10.1016/S0006-3207(01)00157-4
Sass-Klaassen, U., C. Couralet, Y. Sahle, and F. J. Sterck. 2008. Juniper from Ethiopia contains a large-scale precipitation signal. International Journal of Plant Sciences 169:1057-1065. https://doi.org/10.1086/590473
Schüler, L., A. Hemp, W. Zech, and H. Behling. 2012. Vegetation, climate and fire-dynamics in East Africa inferred from the Maundi crater pollen record from Mt Kilimanjaro during the last glacial-interglacial cycle. Quaternary Science Reviews 39:1-13. https://doi.org/10.1016/j.quascirev.2012.02.003
Seydack, A. H. W., S. J. Bekker, and A. H. Marshall. 2007. Shrubland fire regime scenarios in the Swartberg Mountain Range, South Africa: implications for fire management. International Journal of Wildland Fire 16:81-95. https://doi.org/10.1071/WF06015
Strömberg, C. A. E. 2011. Evolution of grasses and grassland ecosystems. Annual Review of Earth and Planetary Sciences 39:517-544. https://doi.org/10.1146/annurev-earth-040809-152402
Syphard, A. D., T. J. Brennan, and J. E. Keeley. 2019. Drivers of chaparral type conversion to herbaceous vegetation in coastal Southern California. Diversity and Distributions 25:90-101. https://doi.org/10.1111/ddi.12827
Teketay, D. 2000. Vegetation types and forest fire management in Ethiopia. Pages 1-35 in Proceedings: Round Table Conference on Integrated Forest Fire Management in Ethiopia, Addis Ababa, Ethiopia.
Tesfaye, Y., A. Roos, B. J. Campbell, and F. Bohlin. 2012. Factors associated with the performance of user groups in a participatory forest management around Dodola Forest in the Bale Mountains, southern Ethiopia. Journal of Development Studies 48:1665-1682. https://doi.org/10.1080/00220388.2012.714123
Umer, M., H. F. Lamb, R. Bonnefille, A.-M. Lézine, J.-J. Tiercelin, E. Gibert, J.-P. Cazet, and J. Watrin. 2007. Late Pleistocene and Holocene vegetation history of the Bale Mountains, Ethiopia. Quaternary Science Reviews 26:2229-2246. https://doi.org/10.1016/j.quascirev.2007.05.004
van Breugel, P., I. Friis, S. Demissew, J.-P. B. Lillesø, and R. Kindt. 2016. Current and future fire regimes and their influence on natural vegetation in Ethiopia. Ecosystems 19:369-386. https://doi.org/10.1007/s10021-015-9938-x
van Wilgen, B. W. 2013. Fire management in species-rich Cape fynbos shrublands. Frontiers in Ecology and the Environment 11:e35-e44. https://doi.org/10.1890/120137
Vial, F., D. W. Macdonald, and D. T. Haydon. 2011. Limits to exploitation: dynamic food web models predict the impact of livestock grazing on Ethiopian wolves Canis simensis and their prey. Journal of Applied Ecology 48:340-347. https://doi.org/10.1111/j.1365-2664.2010.01943.x
Watson, C., S. Mourato, and E. J. Milner-Gulland. 2013. Uncertain emission reductions from forest conservation: REDD in the Bale Mountains, Ethiopia. Ecology and Society 18(3):6. https://doi.org/10.5751/ES-05670-180306
Wesche, K. 2006. Is Afroalpine plant biodiversity negatively affected by high-altitude fires? Pages 39-50 in E. M. Spehn, M. Liberman, and C. Korner, editors. Land use change and mountain biodiversity. CRC Press, Taylor & Francis Group, Boca Raton, Florida, USA. https://doi.org/10.1201/9781420002874-3
Wesche, K., G. Miehe, and M. Kaeppeli. 2000. The significance of fire for Afroalpine Ericaceous vegetation. Mountain Research and Development 20:340-347. https://doi.org/10.1659/0276-4741(2000)020[0340:TSOFFA]2.0.CO;2
Williams, J. 2013. Exploring the onset of high-impact mega-fires through a forest land management prism. Forest Ecology and Management 294:4-10. https://doi.org/10.1016/j.foreco.2012.06.030
Williams, R. J., C. H. Wahren, R. A. Bradstock, and W. J. Muller. 2006. Does alpine grazing reduce blazing? A landscape test of a widely-held hypothesis. Austral Ecology 31:925-936. https://doi.org/10.1111/j.1442-9993.2006.01655.x
Yibarbuk, D., P. J. Whitehead, J. Russell-Smith, D. Jackson, C. Godjuwa, A. Fisher, P. Cooke, D. Choquenot, and D. M. J. S. Bowman. 2001. Fire ecology and Aboriginal land management in central Arnhem Land, northern Australia: a tradition of ecosystem management. Journal of Biogeography 28:325-343. https://doi.org/10.1046/j.1365-2699.2001.00555.x
Zeleke, T. T., F. Giorgi, G. T. Diro, and B. F. Zaitchik. 2017. Trend and periodicity of drought over Ethiopia. International Journal of Climatology 37:4733-4748. https://doi.org/10.1002/joc.5122