Vegetation dynamics, and land use and land cover change in the Bale Mountains, Ethiopia

Download Vegetation dynamics, and land use and land cover change in the Bale Mountains, Ethiopia

Post on 18-Mar-2017

214 views

Category:

Documents

2 download

TRANSCRIPT

<ul><li><p>Vegetation dynamics, and land use and land cover changein the Bale Mountains, Ethiopia</p><p>Yohannes Kidane &amp; Reinhold Stahlmann &amp;Carl Beierkuhnlein</p><p>Received: 11 July 2011 /Accepted: 28 December 2011 /Published online: 26 January 2012# Springer Science+Business Media B.V. 2012</p><p>Abstract Shifts in biological communities are occur-ring at rapid rates as human activities induced globalclimate change increases. Understanding the effects ofthe change on biodiversity is important to reduce lossof biodiversity and mass extinction, and to insure thelong-term persistence of natural resources and naturesservices. Especially in remote landscapes of develop-ing countries, precise knowledge about on-going pro-cesses is scarce. Here we apply satellite imagery toassess spatio-temporal land use and land cover change(LULCC) in the Bale Mountains for a period of fourdecades. This study aims to identify the main driversof change in vegetation patterns and to discuss theimplications of LULCC on spatial arrangements andtrajectories of floral communities. Remote sensingdata acquired from Landsat MSS, Landsat ETM +and SPOT for four time steps (1973, 1987, 2000,and 2008) were analyzed using 11 LULC units definedbased on the dominant plant taxa and cover types ofthe habitat. Change detection matrices revealed thatover the last 40 years, the area has changed from aquite natural to a more cultural landscape. Within a</p><p>representative subset of the study area (7,957.5 km2),agricultural fields have increased from 1.71% to9.34% of the total study area since 1973. Naturalhabitats such as upper montane forest, afroalpinegrasslands, afromontane dwarf shrubs and herbaceousformations, and water bodies also increased.Conversely, afromontane grasslands have decreasedin size by more than half (going from 19.3% to8.77%). Closed Erica forest also shrank from 15.0%to 12.37%, and isolated Erica shrubs have decreasedfrom 6.86% to 5.55%, and afroalpine dwarf shrubsand herbaceous formations reduced from 5.2% to1.56%. Despite fluctuations the afromontane rainfor-est (Harenna forest), located south of the BaleMountains, has remained relatively stable. In conclu-sion this study documents a rapid and ecosystem-specific change of this biodiversity hotspot due tointensified human activities (e.g., deforestation, agri-culture, infrastructure expansion). Specifically, theecotone between the afromontane and the afroalpinearea represent a hotspot of biodiversity loss today.Taking into consideration the projections of regionalclimate warming and modified precipitation regimes,LULCC can be expected to become evenmore intensivein the near future. This is likely to impose unprecedent-ed pressures on the largely endemic biota of the area.</p><p>Keywords Biodiversity loss . Vegetation dynamics .</p><p>Endemism . Elevational gradient . Land use change .</p><p>Remote sensing . Tropical mountains</p><p>Environ Monit Assess (2012) 184:74737489DOI 10.1007/s10661-011-2514-8</p><p>Y. Kidane (*) : R. Stahlmann : C. BeierkuhnleinDepartment of Biogeography, University of Bayreuth,Universitaetsstr. 30,95440 Bayreuth, Germanye-mail: yohannes-kidane@hotmail.com</p><p>Y. Kidanee-mail: yohannes.kidane@uni-bayreuth.de</p></li><li><p>Introduction</p><p>Tropical mountain ecosystems are biodiversity hot-spots at significant risk in face of land use and landcover change (LULCC) and global warming (Buytaertet al. 2010; Hagedorn et al. 2010). In addition toharboring vital diversity for the biosphere, tropicalmountains are also hot spots of human activities, thusunder increasing pressure from growing populations.LULCC caused by human activities is one of theprimary drivers of terrestrial biodiversity change andloss today (e.g., Sala et al. 2000; Rockstrm et al.2009). It usually results in habitat loss and fragmenta-tion, local extinction, facilitates invasion by alien spe-cies (Lovejoy and Hannah 2005), reduction in soilinfiltration rate (Yimer et al. 2008), and variability insoil carbon stock and the emission of other greenhousegases. In short, LULCC is rapidly transforming manypristine natural habitats beyond their natural range ofvariability (Hannah et al. 2002; Lovejoy and Hannah2005). The recent rate of LULCC is extraordinarilyhigh in tropical countries (Foley et al. 2005), includingEthiopia.</p><p>Ethiopia features extensive high mountain ecosys-tems, which represent ecological islands in a tropicalto subtropical lowland matrix with differing climateand high land use pressure. Mountains cover about43% of the surface of Ethiopia (Woldemariam 1990,cutoff at 1,500 m asl). These mountains harbor aremarkable diversity of endemic fauna and flora. Themountain climate is more favorable for many speciesrelative to the lowlands where arid to semiarid envi-ronments predominate (Hillmann 1990; Messerli et al.1990). Hence, over 85% of the Ethiopian populationand over 75% of livestock are estimated to inhabit theEthiopian mountains today (Amsalu and de Graaff2006).</p><p>As Ethiopia is anciently settled, the question ofpotential pristineness and naturalness of ecologi-cal systems is difficult to answer (see also Chiarucci etal. 2010). Almost all parts of the country are accessedby humans. Areas of low human impact are confinedto the steep escarpments of the Rift Valley, rivergorges, cold afroalpine plateaus, and some sacredareas (Gebregziabher 1991). Yet evidence suggeststhat despite this long history of human impacts,LULCC has accelerated dramatically in recent years,progressively depleting the vegetation (NBSAP 2005).Furthermore, as pointed out in Kreyling et al. (2010),</p><p>this loss is expected to be exacerbated in Ethiopiasmountains with future climate change. Quantificationof LULCC in this unique and highly threatened eco-system is thus extremely urgent.</p><p>The assessment of LULCC has emerged as a fun-damental component of global change research(Turner et al. 2007). It is multidisciplinary in its ap-proach and seeks to understand dynamics in a coupledhumanenvironment system (Turner et al. 2007; Olsonet al. 2008). Here, remote sensing data spanning~40 years have been used to quantify for the first timeLULCC impacts on temporal vegetation dynamics inthe Bale Mountains of Ethiopia. The goal is to identifythe main local drivers of changing vegetation patternsduring this time interval and inform how continuedLULCC will impact the spatial arrangements and tra-jectories of floral communities with future climatechange.</p><p>The Bale Mountains of Ethiopia (Hillman 1986) areof global conservation significance. They are home tothe largest population of the endangered endemicEthiopian wolf (Canis simensis) and of MountainNyala (Tragelaphus buxtoni, Sillero-Zubiri andMacdonalds 1997). The mountains also host a numberof unique plants such as Lobelias (e.g., Lobelia rhyn-chopetalum, Lobelia scebelii, and Lobelia giberroa)and Senecio species (e.g., Senecio nanus, Seneciofresenii, Senecio inornatus, Senecio ochrocarpus,Senecio ragazzi, Senecio schultzii, Senecio subsessilis,and Senecio unionis). Besides, the endemic plant pop-ulations of mountains are important reservoirs of ge-netic diversity (Hillman 1988; Uhlig 1990; NBSAP2005). The rainforests located to the south representthe native habitat of wild coffee. Additionally themountains support numerous ecosystem services forlowland areas including capture, distribution, and reg-ulation of the water supply.</p><p>Until the early 1960s, the Bale Mountains area wassparsely populated (Stephens et al. 2001). Hence, pop-ulation pressure in the area was low. In 1971 the BaleMountain National Park (BMNP) was established,covering a large portion of the massif. It comprisesan area 2,200 km2 of afroalpine habitat and includespart of the afromontane rainforests. The establishmentof this park was targeted at protecting the afroalpinehabitat and its endemic mammalian fauna (Hillman1988). Today, BMNP is listed as one of theUNESCO 200 worldwide Bio-Regions (Umer et al.2007) and represents a potential world heritage site.</p><p>7474 Environ Monit Assess (2012) 184:74737489</p></li><li><p>Furthermore, the afroalpine ecosystems of the massifare one of 34 recognized Conservation InternationalBiodiversity Hotspots and are listed as an ImportantBird Area by Birdlife International (NBSAP 2005).</p><p>Ecological studies of the Bale Mountains focusedon documenting the biodiversity, fauna, and flora ofthe area (e.g., Bonnefille 1983; Gebregziabher 1988,1991; Woldu et al. 1989; Messerli et al. 1990; Uhlig1990; Uhlig and Uhlig 1991; Friis 1992; Miehe andMiehe 1994; Mohamed-Saleem and Woldu 2002;Darbyshire et al. 2003; Umer et al. 2007). Some stud-ies focused on issues of exploitation, degradation, andmanagement (e.g., Getahun 1984; Messerli et al. 1990;Grepperud 1996; Taddese 2001; Amente 2005).Wesche et al. (2000) and Wesche (2003) have investi-gated the role of fire and drought. Until now, there has</p><p>been no integrated quantitative assessment of howbiodiversity and LULCC developed over time.</p><p>Materials and methods</p><p>Study area</p><p>Orography</p><p>The Bale Mountains are located in southeast Ethiopiain the Ormia regional state (Fig. 1), between 0629 N,3903 E and 0710 N, 4000 E. The mountainscomprise one of most extensive high altitude plateaus(above 3,000 m in elevation), the largest contiguousmountain massif of over 2,600 km2 in Africa, and</p><p>Fig. 1 The study area with the borders of the Bale Mountains National Park. The extents of major altitudinal gradients in meters abovesea level and important locations are indicated. The whole above 4,000 m is the Sanetti plateau which is the main afroalpine zone</p><p>Environ Monit Assess (2012) 184:74737489 7475</p></li><li><p>one of the last remaining pristine afroalpine biodiver-sity hotspots on the continent (Hillman 1988;Laurenson et al. 1998). The massif exhibits a steepgradient of ecological zones ranging from lowlandsemideserts, savannas and grasslands, and tropicalrainforests to afroalpine vegetation. At the plateauthere are several high peak summits. The highest pointmount Tullu Deemtu reaches 4,385 m asl of themassif.</p><p>Climate</p><p>The Bale Mountains are located at the convergence ofthe wet east African and dry northeast African moun-tains of southeast Ethiopia. The climate of the moun-tains varies from north to south mainly due thedifferences in elevation, aspect, and the influences oflowland hot air masses (Uhlig 1990). Historically, thearea has experienced a high degree of climate variabil-ity and change (Umer et al. 2007). These past climaticchanges have played a crucial role in shaping thecontemporary vegetation. The current climate is char-acterized by a short dry season (November toFebruary) and a long period of rainfall and high mois-ture (March to October). The wet season rainfall pat-tern is slightly bimodal, with a peak from April to Mayfollowed by a second peak from September to October(Woldu et al. 1989). Recently, the area has also wit-nessed an increase in the frequency and severity ofexceptional droughts (e.g., year 2000 see Wesche2003).</p><p>Rain comes to the Bale Mountains from two differ-ent sources: the equatorial westerly and the IndianOcean monsoon (Uhlig 1990; Miehe and Miehe1994). Along altitude, precipitation increases up toan elevation of 3,850 m asl, but decreases then againtowards the summits (Hillman 1986). The northernpart of the mountain range exhibits 8001,000 mmof annual rainfall and a wet season from June toSeptember. The southern part is more humid, with asubtropical climate and 1,0001,500 mm of annualrainfall (Woldu et al. 1989).</p><p>The afroalpine habitats are characterized by strongdiurnal temperature fluctuations and night frost. Onthe Sanetti Plateau, Hillmann (1990) recordedextremes of a diurnal temperature range of 40C(15C to +26C) during the dry season. The diurnalamplitude in temperature varies between the wet anddry season. The wet season is warmer at night and</p><p>cooler by day compared to the dry season (Admasu etal. 2004). Here, the coldest time of the year is also thedriest, the main wet season being associated withconvergence of northeast and southwest airstreams(Bonnefille 1983). Consequently, the biota experienceextreme temperature variation within each 24-h periodduring the whole year, experiencing summer everyday and winter every night (Hedberg 1964).</p><p>Geology and geomorphology</p><p>The Bale Mountains are fragmented due to numerousvolcanic plugs, peaks, alpine lakes, and rushing moun-tain streams that descend into deep rocky gorges ontheir way to the lowlands. The uppermost part, theSanetti plateau, is an isolated area covering 211 km2</p><p>at an altitude of about 4,000 m bordered by abruptescarpments to the south. The north and northeast aredeeply dissected valleys descending to the northernslope, while to the west lava flows form spectacularbluffs (Osmaston et al. 2005).</p><p>Bale Mountain geology is characterized by a high-altitude volcanic plateau over much older volcanicmaterial formed during the spreading of the EastAfrican Rift Valley system. The petrography is domi-nated by alkali basalt and tuffs, with occasional rhyo-lites (Uhlig and Uhlig 1991). The mountains werelocally glaciated, which shaped their recent geomor-phology (Osmaston et al. 2005).</p><p>Soils in the area tend to be shallow, gravelly, andrecently derived from volcanic rock exposed sinceglacial retreat (Sillero-Zubiri and Macdonald 1997).Soils consist of a relatively silty loam of reddishbrown to black color (Woldu et al. 1989). Soils locatedon top of stratigraphically youngest units derive main-ly from the Miocene basalt and trachyte lavas that layover Mesozoic sediments (Umer et al. 2007).</p><p>Data collection</p><p>We focus on a subset area of the Bale Mountains thatranges from 3939 E/739 N and 4009 E/647 N.Data consist of preprocessed Landsat and SPOTimages of four time steps (1973, 1987, 2000, and</p><p>Fig. 2 Time series of raw satellite images used for the analysiswith their respective year of acquisition. a The 1973 MSS imageof path 180 and row 055; b the 1987 MSS image of paths 167and 168, and row 055; c the 2000 ETM + image of paths 167and 168, and row 055; and d the 2008 SPOT_5 image with path138 to 142 and rows 135 to 137</p><p>7476 Environ Monit Assess (2012) 184:74737489</p></li><li><p>Environ Monit Assess (2012) 184:74737489 7477</p></li><li><p>2008). The Landsat Multispectral Scanner (MSS) andEnhanced Thematic Mapper Plus (ETM+) data wereacquired from United States Geological Survey(USGS) EarthExplorer homepage (http://edcsns17.cr.usgs.gov/NewEarthExplorer). SPOT images fromFebruary 2008 were acquired from SPOT Planet ac-tion (http://www.planet-action.org; Fig. 2). We chosehigh-resolution (2.5 m) SPOT data instead of morerecent Landsat images for the 2008 time slice due tothe known failure of the Scan Line Protector on boardLandsat 7 on May 21, 2003 (USGS 2009).</p><p>All images selected for analysis were from the endof the dry season (end of January and beginning ofFebruary, Table 1). Dry season images are preferablebecause they are more likely to be cloud free and theirspectral properties are less affected by moisture.Seasonal and sensor coherence allows mapping simi-lar vegetation phenology and similar atmospheric con-ditions to reduce sun angle...</p></li></ul>

Recommended

View more >