Land Use Change and Mountain Biodiversity - Chapter 2 potx

Miehe and Miehe 1994 presented a detailed study on ericaceous vegetation and on the plant communities within the ericaceous zones of the Bale Mountains.. The present study aims at 1 desc

Trang 1

Part II Effects of Fire on Mountain Biodiversity

3523_book.fm Page 23 Tuesday, November 22, 2005 11:23 AM

Trang 2

Vegetation and Ecology of Treeline Species in the Bale Mountains, Ethiopia, and the Influence of Fire

Masresha Fetene, Yoseph Assefa, Menassie Gashaw, Zerihun Woldu, and Erwin Beck

INTRODUCTION

Uplift and volcanism in the Miocene and Oli-gocene geological periods (between 38 and 7 million BP) resulted in the covering of all the underlying rocks and the formation of the East African mountains that rest like islands on the surrounding hills and plains These Afromon-tane archipelagos are distributed on both sides

of the East African Rift Valley

The Bale Mountains lie in the southeastern part

of the Ethiopian highlands, about 850 km north of the equator The highest peak in Bale, Tulu Dimtu,

is the second highest peak in Ethiopia and the sev-enth in Africa (see Figure 2.1) The East African mountain nearest to the Bale mountains is Mt

Kulal, 550 km south in the Turkana Depression

The vegetation of the Bale Mountains has been the subject of studies by a number of bot-anists and ecologists A full account of the his-tory of botanical exploration of the Bale Moun-tains has been provided by Miehe and Miehe (1994) In a series of publications, Hedberg (1975, 1986) made important analyses of the vegetation and ecology of Afroalpine regions in Ethiopia Weinert (1981), Weinert and Mazurek (1984), and Uhlig (1988) also conducted eco-logical research on the vegetation of the Bale Mountains Miehe and Miehe (1994) presented

a detailed study on ericaceous vegetation and

on the plant communities within the ericaceous

zones of the Bale Mountains The present study attempts to provide a description of plant com-munities in the entire altitudinal range of the Afroalpine and ericaceous zones

The ericaceous belt of the Bale Mountains

is a region most seriously affected by the pro-gressive increase of human activities Cattle and horses put heavy pressure on the vegetation, especially at the lower altitudes The ericaceous bushes are cut for fuel wood and are frequently burned by the local people for various reasons This results in the destruction of the vegetation and in the disappearance of the fauna, and hence leads to a reduction of the region’s biodiversity The present study aims at (1) describing the plant communities of the Afroalpine and erica-ceous zones, (2) documenting the distribution patterns of treeline species and the changes in the structure of ericaceous vegetation with alti-tude, and (3) assessing the incidence and influ-ence of fire on the diversity and composition of vegetation in the ericaceous belt

MATERIAL AND METHODS

D ESCRIPTION OF THE S TUDY A REA

Geology and Climate

The study area is the Harenna Escarpment, located at the southern slopes of the Bale Moun-tains between 6°45 and 7° N and 39°45 and 3523_book.fm Page 25 Tuesday, November 22, 2005 11:23 AM

Trang 3

26 Land Use Change and Mountain Biodiversity

39°40 E The rocks of the volcanic outpourings

are predominantly trachytes but also include

rhyolites, basalts, and associated agglomerates

and tuffs Although adequate information about

glaciations is lacking, the current landforms in

the mountains appear to have resulted from

actions of tectonics and glaciations At least two

glacial periods are documented in the

moun-tains (18,000 BP and 2,000 BP, Bonnefille,

1993)

In contrast to the northern highlands,

south-ern Ethiopia is within the East African climatic

domain, which is highly influenced by

south-easterlies from the Indian Ocean during most

of the year As in most Ethiopian highlands, the intertropical convergence zone (ITCZ) and local altitudinal and topographic influences affect the distribution of the precipitation in the Bale Mountains Annual rainfall in the Bale Mountains ranges between 600 and 1500 mm depending on the relief (see Table 2.1) The diurnal variability in temperature in the Bale Mountains is higher than the seasonal vari-ation A minimum temperature of −15°C was recorded by Hillman (1986) on the Sanetti Pla-teau (3850 m), whereas Miehe and Miehe (1994) recorded a nocturnal minimum temper-ature of –3°C in sparsely vegetated areas of the ericaceous belt Solifluction is common in the

FIGURE 2.1 Map of the study area (From Miehe and Miehe [1994].)

Trang 4

Diversity of Afroalpine Vegetation and the Influence of Fire 27

Afroalpine area and in the upper parts of the

ericaceous vegetation

Recently, the ericaceous and the Afroalpine

areas have been subjected to increasing grazing

pressure The number of livestock varies in the

wet and dry seasons (the maximum is 46/km2

in the plateau and minimum is less than 2/km2)

(Hilman, 1986; Gottelli and Sillerio-Zubiri,

1992) Poaching of mountain nyalas and small

antelopes is also common in the area These

activities are accompanied by deliberate setting

of bush fires for hunting, and clearing and

improvement of pastures (Miehe and Miehe,

1994)

There is evidence of early settlements in

some valleys and plains in the area Recently,

with the construction of an all-weather road

tra-versing the plateau, there is an increase in barley

cultivation in the ericaceous and Afroalpine

veg-etation However, the highlands of the Bale

Mountains are still less densely populated than

the Semien Mountains of northwestern Ethiopia

(see Table 2.2) For instance, barley is cultivated

in Bale at 600 to 800 m lower than in Semien

This is due to the transhumant mode of living

in the Bale Mountains

V EGETATION S AMPLING

The current study considers vegetation in the

ericaceous belt of the Bale Mountains along an

altitudinal gradient ranging from 3000 to 4200

m Transects were laid out based on

homoge-neity of the vegetation (Mueller-Dombois and

Ellenberg, 1974) Relevés of 15 m × 15 m size

were established at 50-m vertical distance

(alti-tude) Within each altitudinal level, replicate

relevés were put with minimum lateral distance

of 20 m Within each relevé, a subplot of 2 m

× 2 m was made for the herbaceous vegetation All vascular plants in each relevé were recorded We estimated abundance for single species using the 9-level ordinal cover abun-dance scale following Braun Blanquet as mod-ified by Van der Maarel (Van der Maarel, 1979) The height of trees and shrub species, diameter

at breast height (DBH) for trees, and the diam-eter at stump height (DSH) for shrubs were also recorded in all relevés

S OIL AND E NVIRONMENTAL D ATA

The rainfall measurements were compiled for the time of fieldwork and for the previous 11 months Climate data of the area from previous studies were also considered (Miehe and Miehe, 1994) For each plot, information on altitude, slope, inclination, soil surface, and vegetation cover, etc., were collected Soil sam-ples were collected from the topsoil and at a depth of 30 cm from the surface of each relevé Soil moisture, texture, pH, and total nitrogen were determined for each sample

I NCIDENCE OF F IRE

Records on incidence of recent fires were gath-ered In addition to the information obtained from the local people, the incidence of fire was assessed from the presence or absence of Bryum argenteum (a moss that grows after fire), char-coal, and remnants of charred twigs and ligno-tubers The presences of each of these indicators were summed for each relevés, yielding a com-bined index of fire incidence

TABLE 2.1

Annual rainfall for northern (n) and southern (s) slopes of Bale Mountains and on Sanetti Plateau (P)

Chorchora (n) 3500 1086 1985–1991 Goba (n) 2720 925 1968–1980 Koromi (P) 3850 1051 1985–1991 Mena (s) 1250 387 1983–1988 Rira (s) 3000 848 1987–1990 Tulu Konteh (P) 4050 852 1985–1991

Source: From Hilman (1986); Miehe and Miehe (1994)

Trang 5

D ATA A NALYSIS

Vegetation data were analyzed with hierarchical

syntax clustering using agglomerative method

with optimization (Podani, 2000) A

resem-blance matrix was calculated with the similarity

ratio:

Sij=1-∑i xij xik/ (∑i xij2 +∑i xik2 - ∑i xij xik) where S(i, j) in row i and column j is the distance

between observations i and j Species-wise cover

abundance values were used to classify

vegeta-tion communities In classifying the

communi-ties, the subject group averages were used to

eval-uate the degree of dissimilarities among the

relevés Both the vegetation data and the

environ-mental variables were analyzed with canonical

correspondence analysis (CCA) using CANOCO

(ter Braak and Smilaur, 1998) to explore the

cor-relation between vegetation and environmental

variables Species richness and relative

abun-dance were analyzed using the Shannon–Weaver

index of diversity (Krebs, 1989)

RESULTS AND DISCUSSION

P LANT C OMMUNITIES

The southern slope of the Harenna Escarpment

with its montane forest between 1500 and 2800

m is more gentle than the ericaceous vegetation

above this altitude The Bale Mountains have high floral and faunal diversity as well as ende-micity The floristic composition of the area has been reported by Friis (1986); Hedberg (1986); Negatu and Tadesse (1986); Woldu et al (1989); Gashaw and Fetene (1996); and Bussman (1997)

A total of 60 relevés were sampled at the northwestern side of the Bale Mountains The hierarchical classification gave six major plant communities The first of these is the Knipho-fia– Euphorbia–Alchemilla community (3400

to 3500 m) In this community, Kniphofia foli-osa, Euphorbia dumalis, and Alchemilla abys-sinica were the characteristic species At the next altitudinal level, we find the Alchemilla haumannii community (3700 to 4000 m) This community is dominated by A haumannii, which sometimes forms pure stands

On the southeastern side of the Bale Mountains, a total of 110 relevés were sam-pled, in which 84 species of vascular plants were encountered Eight of these were trees and shrubs, and the rest were herbaceous plants The ericaceous vegetation was grouped into three altitudinal subzones fol-lowing previous works: lower subzone (3000

to 3400 masl), central subzone (3400 to 3600 masl), and the upper subzone (3600 to 4000 masl) (see also Hedberg, 1951; Miehe and Miehe, 1994) Thirteen community types were identified from the cluster analysis The communities were named based on the

spe-TABLE 2.2

Population density in Semien and Bale Mountains (persons/km 2 ) based on the census data taken for each zone (district) and the woredas (subdistricts) circumscribed by the mountains

Zone/Woreda

Year

Semien Mountains North Gondar zone 49.8 51.2 52.6

Debark 91.7 94.4 97 Bale Mountains Bale zone 22.1 22.7 23.4

Kokosa 160 164.5 169 Dodola 91.2 94 96.9 Adaba 52.2 53.7 55.3 Sinana Dinsho 90.3 93.2 96.2 Goba 44 45.8 47.6 Menana Harenna Bulqi 14.1 14.5 14.9

Source: Central Statistical Authority (2001)

Trang 6

cies with the highest cover abundance The

distribution of the communities varied in the

lower (3000 to 3400 m), central (3400 to 3600

m), and upper (3600 to 4200 m) subzones of

the ericaceous belt Some of the community

types occurred in the entire altitudinal range

(3000 to 4200 m), whereas others were

restricted to certain ranges Plant diversity

showed an inverse bell-shaped pattern The

upper and the lower subzones had higher

diversities than the central one The complete

list of the communities and their respective

distribution, diversity, and evenness in the

three subzones are given in Table 2.3

The Schefflera volkensii–Erica

trim-era–Discopodium penninervium community

(altitude range, 3100 to 3300 m) is found at the

lowermost part of the ericaceous subzone The

emergent tree in this community is Schefflera

volkensii. Higher up in the lower ericaceous

subzones, the Erica trimera–Hagenia

abyssin-ica–Hypericum revolutum community occurs

The characteristic species for this community

are Erica trimera, Trifolium acaule, Hypericum

revolutum, Hagenia abyssinica, and

Discopo-dium penninervium At lower altitudes

(between 3000 and 3200 m), this community

forms a subcommunity that is characterized by

the dominance of Hagenia abyssinica and

Hypericum revolutum Another community also

common at the lower subzone of the ericaceous

belt is the Erica

trimera–Polystichum–Hyperi-cum revolutum community (Plate 2.1a) The

characteristic species of this community

include Erica trimera, the codominant tree

Hypericum revolutum, the most common fern

Polystichum sp., Discopodium penninervium,

and Cynoglossum amplifolium.

At the central subzone of the ericaceous

belt, we find the Erica trimera–Hypericum

revolutum–Alchemilla abyssinca community

In this community, the dominance of Erica

trimera is conspicuous in the upper layer of

the canopy Another community of the central

subzone is the Erica trimera–Cynoglossum

amplifolium–Discopodium penninervium

community

Among communities of the upper subzone,

we find the Haplocarpha rueppellii–Alchemilla

microbetula–Alchemilla pedata community

(3300 to 3900 m) and the Satureja

pdoxa–Asplenium aethiopicium-Geranium

ara-bicum community In the former, Haplocarpha

rueppellii, Alchemilla microbetula, Alchemilla

pedata, Myosotis abyssinica, and Discopodium penninervium are the characteristic species, whereas the characteristic species in the latter community are Satureja paradoxa, Asplenium aethiopicium, Geranium arabicum, Crepis rueppellii, and Stachys aculeolata. The upper part of the ericaceous belt had a patchy appear-ance with more openings Depending on the microsite factors, the diversity was comparable with the lower part (2.35 ± 0.048 for the lower and 2.10 ± 0.05 for the upper) and was greater than in the central subzones

D ENSITY AND F REQUENCY OF T REELINE

S PECIES

A total of eight tree and shrub species were recorded, out of which Erica trimera was found in almost all relevés, whereas one spe-cies (Pittosporum viridiflorum) was recorded

in one relevé only and is not shown in Figure 2.2 Erica trimera and Hypericum revolutum

showed similar trends in frequency in the

lower and central subzone (Figure 2.2) H.

revolutum was absent in the upper part of the

ericaceous subzone At the lower ericaceous

subzone, the frequency of E trimera was

lower because of the competitive strength of the other montane woodland species (Miehe and Miehe, 1994) However, it is an important component of all three subzones of the erica-ceous belt and no other species, including

Erica arborea, showed such a wide distribu-tion Erica arborea was not found below 3200

m Rapanea melanophloeos, H revolutum, and D penninervium were constituents of both

the lower and central subzone but not of the

upper subzone Schefflera volkensii is

restricted to the lower part of the ericaceous

belt, and Hagenia abyssinica attained its

high-est frequency in the lower subzone The den-sity of the treeline species showed a similar trend as the frequency

The height of treeline species decreased with increasing altitude (Table 2.4) The most

notable change was observed for E trimera.

The regression analysis (Figure 2.3) showed a strong inverse relation between altitude and height (R2 = 0.60) This could be attributed to the decrease in temperature with increasing alti-tude

Trang 7

TABLE 2.3

The distribution of the diversity (H), evenness of the 13 community types, and average value for incidence of fire in lower, central, and

upper subzones of ericaceous vegetation

Community Types

Lower (3000–3400)

Central (3400–3600)

Upper (3600–4200)

Species Number

Shannon

Index:

revolutum

abyssinica

amplifolium–Discopodium penninervium

penninervium

abyssinica–Cynoglossum amplifolium

arabicum

microbetula–Alchemilla pedata

abyssinica

haumanni

aethiopicum–Alchemilla abyssinica

aethiopicum–Geranium arabicum

Trang 8

FIGURE 2.2 Frequency of treeline species in three ericaceous subzones: Schefflera volkensii (SV); Rapanea

melanophloeos (RM); Hypericum revolutum (HR); Discopodium penninervium (DP); Hagenia abyssinica (HA); Erica trimera (ET); Erica arborea (EA) The three ericaceous subzones are the lower (3000 to 3400 masl); middle

(3400 to 3600 masl); and upper (3600 to 4000 masl) zones.

TABLE 2.4

Height and DBH of five treeline species at (1) lower, 3000–3400 m, (2) central, 3400–3600

m, and (3) upper, 3600–4200 m, subzones of the ericaceous belt in Harenna Escarpment, Bale Mountains

2 1.75 ± 1.28

2 4.04 ± 0 1.12 ± 0.80

3 — 0.95 ± 0

2 12.32 ± 12.38 5.30 ± 4.53

3 10.00 ± 8.20 2.08 ± 1.98

2 16.74 ± 10.11 7.37 ± 5.86

2 14.34 ± 7.81 8.96 ± 6.60

0 20 40 60 80 100

SV RM HR DP HA ET EA

3600-4200 m 3400-3600 m 3000-3400 m

Trang 9

R ELATIONS BETWEEN D ISTRIBUTION AND

E COLOGICAL C HARACTERS OF T REELINE

S PECIES AND T HEIR E NVIRONMENTAL

F ACTORS

The Pearson correlation analysis revealed a

strong positive correlation between altitude and

slope (0.8) and an even stronger negative cor-relation between altitude and pH Percent silt

and clay showed negative correlations at r = − 0.6 and −0.4, respectively The correlation coef-ficients of the environmental parameters are given in Table 2.5

An ordination biplot was made for all envi-ronmental variables The biplot diagram of the

FIGURE 2.3 Biplot diagram showing the correlations of environmental parameters in the canonical ordination

space.

TABLE 2.5

Pearson’s correlation coefficient matrix for the nine environmental variables

Slope 0.725

Aspect −0.001 −0.216

Moisture −0.33 −0.299 0.668

pH −0.785 0.629 −0.206 −0.375

N 0.028 0.129 −0.066 0.013 −0.196

Sand 0.638 0.871 −0.349 −0.457 0.637 0.226 0.150

Clay −0.201 −0.469 0.554 0.202 −0.480 −0.338 0.186 −0.418

Silt −0.605 −0.733 0.126 0.41 −0.475 −0.089 −0.248 −0.902 −0.014

Note: The magnitude indicates the degree of correlation Positive signs indicate positive correlation and negative signs

indicate inverse relation Numbers in bold indicate significant correlation at p < 0.05.

Altitude (m)

2800 3000 3200 3400 3600 3800 4000 4200

0 5 10 15 20

Trang 10

environmental variables reflects approximately

the Pearson’s correlation coefficients (Figure

2.4)

R ECENT I NCIDENCE OF F IRE

Recent incidence of fire showed an increasing

tendency with increasing altitude (Figure 2.5)

Fire incidence was not common in rocky areas

with big boulders The incidence was lower in

areas with high cover of epiphytes, due,

per-haps, to the convective cloud from Harenna that

leads to the formation of thick epiphytic cover,

playing a crucial role in insulation Highly

dis-turbed sites were avoided intentionally in this

study However, even in the relatively less

dis-turbed vegetation, there was some evidence for

recent occurrence of fire, especially at the upper

subzone of the ericaceous vegetation Incidence

of fire was more common at the upper part of the ericaceous vegetation This is an indication that fire had little influence on the physiognomy

of the lower part of ericaceous vegetation This

is in agreement with other investigations (Wesche, 2002) The highest incidence of fire

was recorded in the Satureja paradoxa–Asple-nium aethiopicum–Geraparadoxa–Asple-nium arabicum

com-munity at the upper subzone of the ericaceous vegetation The absence of indicators for fire

incidence in the Senecio fresenii–Alchemilla abyssinica–Cynoglos-sum amplifolium

com-munity does not necessarily show the complete absence of fire in those localities Alternatively,

it may indicate the disappearance of the indi-cators of fire, which might be due to more severe disturbance

FIGURE 2.4 Regression analysis of the correlation of average height of E trimera with altitude.

Định dạng
Số trang	15
Dung lượng	2,27 MB