Soil carbon stocks and turnovers in various vegetation types and arable lands along an elevation gradient in southern Ethiopia

Geoderma 123 (2004) 177 – 188 www.elsevier.com/locate/geoderma Soil carbon stocks and turnovers in various vegetation types and arable lands along an elevation gradient in southern Ethiopia Mulugeta Lemenih a,*, Fisseha Itanna b b a Wondo Genet College of Forestry, P.O. Box 128, Shashamane, Ethiopia Faculty of Science, Addis Ababa University, P.O. Box 1176, Addis Ababa, Ethiopia Received 22 April 2003; received in revised form 5 January 2004; accepted 3 February 2004 Available online 16 March 2004 Abstract Soil carbon (C) and total N stocks and turnovers were investigated in five vegetation types and following deforestation and conversion of each vegetation types into arable lands along a 37-km elevation transect in southern highlands of Ethiopia. The elevation transect spanned five different eco-climatic zones from semiarid to cool sub-Afroalpine range, each with different vegetation type. Soil C and total N stocks in the upper 0.60 m mineral soil under the natural vegetations varied from 40.3 Mg C ha
1 and 5.3 Mg N ha
1 at the semiarid Acacia woodland (AWL) eco-climatic zone to 234.6 Mg C ha
1 and 20.2 Mg N ha
1 at the humid Podocarpus falcatus forest (PFF) eco-climatic zone, respectively. This trend was directly proportional to the mean annual precipitation and inversely proportional to the mean annual temperature prevailing along the elevation gradient. The soils of the farmlands had significantly lower soil C and total N stocks than the soils under the natural vegetations. Losses of soil C and total N from the upper 0 – 10 cm soil depth following conversion of the natural vegetations to farmlands were highest at the humid PFF eco-climatic zone and lowest at the semiarid AWL eco-climatic zone. The average rates of soil C losses ranged between 2.0% and 3.0% per annum in the sub-humid to humid eco-climatic zones and 0.5 – 1.0% per annum in the semiarid lowland or the cool sub-Afroalpine eco-climatic zones. The results revealed the existence of considerable differences, as large as 191.7 Mg C ha
1, in soil C stocks along the elevation gradient, and wide range of differences in the rate and amount of soil C and total N losses following conversion of natural vegetations into arable lands. D 2004 Elsevier B.V. All rights reserved. Keywords: Afromontane vegetation; Deforestation; Eco-climatic zone; Land use change; Soil C stocks; Soil C loss 1. Introduction Dry forests dominate the forested areas of tropical and sub-tropical landmasses. They account for 42% of * Corresponding author. Present address. Department of Forest Soils, Swedish University of Agricultural Sciences, P.O. Box 7001, 750 07 Uppsala, Sweden. Fax: +46-251-06-20-24-90. E-mail addresses: mulugeta.lemenih@sml.slu.se, wgcf@telecom.net.et (M. Lemenih). 0016-7061/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.geoderma.2004.02.004 the region in contrast to 33% moist forests and 25% wet and rainforests (Murphy and Lugo, 1986). The largest proportion of tropical dry forests is found in Africa where they account for 70 – 80% of the forested area (Murphy and Lugo, 1986; Demel, 1996). The Ethiopian highlands alone contribute to more than 50% of the tropical Afromontane vegetation of Africa (Yalden, 1983; Tamrat, 1993), of which tropical dry Afromontane forests cover the largest part (Demel, 1996). 178 M. Lemenih, F. Itanna / Geoderma 123 (2004) 177–188 Nevertheless, the tropical dry Afromontane forests of Ethiopia are under continuous threats of deforestation. The annual rate of deforestation in these tropical dry Afromontane forest of Ethiopia is estimated to be between 163,000 and 200,000 ha (EFAP, 1994; Reusing, 1998). Land use change that involves deforestation and conversion into farmlands is the principal CO2 emission source in Ethiopia. According to the report of the national inventory of greenhouse gases (GHG), deforestation and land use change account for 81% of the total 16,297 Gg CO2 emissions in Ethiopia (Asress, 1995). However, paucity of empirical data with respect to forest distributions, stocks of soil carbon (C) in the various forest types, rate of deforestation and loss of soil carbon upon conversion and subsequent cultivation of the various vegetation types in the country has limited precise accounting of national CO2 emissions associated with land use changes and forestry. So far, very few quantitative studies have been made on the effects of deforestation and subsequent cultivation on soil carbon in Ethiopia (Solomon et al., 2002). Studies in the tropics have shown significant declines in soil organic matter (SOM) following deforestation and conversion into intensive land uses such as agriculture (Rhoades et al., 2000; Van Noordwijk et al., 1997). However, changes in the amount of SOM following conversion of natural forests to agriculture depend on several factors such as the type of forest ecosystem undergoing change (Rhoades et al., 2000), the post conversion land management (Glaser et al., 2000; Rhoades et al., 2000), the climate (Amelung et al., 1997; Alvarez and Lavado, 1998) and the soil type and texture (Schjønning et al., 1999; Parfitt et al., 1997). Determination of trends in SOM and vegetation cover along moisture and/or temperature gradient is very crucial to the understanding of ecosystem functioning and global climate changes (Ringrose et al., 1998). Elevation gradients are often employed to study the effects of varying climatic variables such as temperature on SOM dynamics (Townsend et al., 1995; Kirschbaum, 1995; Trumbore et al., 1996; Garten et al., 1999). Changes in climatic variables, precipitation and temperature, along altitudinal gradients, influence the type of vegetation and, consequently, the composition, quantity and turnover of SOM (Garten et al., 1999; Hontoria et al., 1999; Quideau et al., 2001). Garten et al. (1999) emphasized the advantages of elevation gradients for testing the effects of environmental variables on SOC dynamics. The relationship between SOM and altitude has also been investigated and positive correlations were reported (e.g., Sánchez, 1969; Sims and Nielsen, 1986). A typical characteristic of the tropical dry Afromontane vegetation of Ethiopia, like most vegetation distribution in Africa, is a marked tendency towards zonation along an altitudinal gradient (Lundgren, 1971; Friis and Lawesson, 1993; Tamrat, 1993). Given its physical settings and biogeographical position in Africa, understanding the stocks of SOM under the various vegetation types in the dry Afromontane forests of Ethiopia and its dynamics following clearance for agricultural land expansion may contribute significantly as a basis for the future design of SOM management and sustainable agricultural development in the region. In the present work, we studied a climo/ biosequence along a 37-km transect in the southern highlands of Ethiopia with the objectives of (1) assessing soil C and total N stocks variability as affected by climatic gradients and vegetation community along the elevation transect, (2) investigate soil C and total N turnovers in response to deforestation and subsequent cultivation of the natural vegetation along the elevation gradient. 2. Site description 2.1. Location and soil The study was conducted in the Central Ethiopian Rift Valley and its eastern escarpment between 7j22V – 32VN and 38j47V –58VE (Fig. 1). The transect was a 37-km climo/biosequence that spanned a range of 1250 m elevation. It covered sites from semiarid lowland at the Rift Valley plain (1640 m asl) to a cool sub-Afroalpine range on the eastern fringe of the Rift Valley escarpment at 2840 m asl (Table 1). Soils of the elevation transect were closely related to their parent materials and their degrees of weathering (Table 1). The main parent materials are basalt, ignimbrites, lava, gneiss, volcanic ash and pumice. The large volcanoes within the southern Rift Valley belong to the latest period of volcanic activ- M. Lemenih, F. Itanna / Geoderma 123 (2004) 177–188 Fig. 1. Map showing the eastern escarpment of the Ethiopian Central Rift Valley and approximate position of the transect and the sample sites. 179 180 M. Lemenih, F. Itanna / Geoderma 123 (2004) 177–188 Table 1 Major climatic and vegetation characteristics of the elevation transect Eco-climatic zone Altitude (m asl) Natural vegetation type Mean annual rainfall (mm) Mean annual temperature (jC) Soil types Semiarid AWL Dry TWL Sub-humid Afromontane Humid Afromontane Cool sub-Afroalpine 1500 – 1750 1750 – 2100 2100 – 2300 2300 – 2650 >2800 Acacia woodland Dry transitional woodland Podocarpus – Croton mixed forest P. falcatus Forest Montane woodland thickets 716 970 1112 1215 1370 20.4 19.1 16.8 15.2 13.6 Calcic Solanchaks Vitric Andosol Mollic Andosol Mollic Nitosol Eutric Nitosol ities (Mesfin, 1998). The associated lava, ash and pumice are mostly acidic and are also interbedded with lacustrine siltstones and sandstone (Makin et al., 1975). In the Rift Valley plain, the soils are coarse textured, weakly structured, exhibiting high levels of alkalinity in the subsoil developed on lake deposits interbedded with pumice. The highland areas fringing the Rift Valley are characterized by deep moderately weathered dark reddish brown soils of clay loams, which are all associates of the Rift Valley volcanic soils. 2.2. Climate of the elevation transect Variation in climate and the resulting vegetation are the major differences along the elevation transect studied. Climatic variations in the highlands of Ethiopia are generally largely a result of differences in altitude. There are decreases of mean annual temperature and increases of mean annual rainfall with increasing elevation. Mean daily temperature falls with increasing altitude at a rate estimated to be within the range of 0.55– 0.65 jC per 100 m (von Breitenback, 1963; Brown and Cocheme, 1969; Makin et al., 1975). Generally, the climate along the selected elevation transect falls into four eco-climatic zones, which included semiarid, sub-humid, humid and cool and humid zones (Makin et al., 1975). Climatic data for places with meteorological stations along the transect were obtained from the Ethiopian Meteorology Authority (Table 1). Data obtained covered the years from 1987 to 2000. For sites along the transect without meteorological stations, mean annual rainfalls were estimated from existing records of closer places, but along the same altitudinal range and similar ecoclimatic zones, while mean annual temperatures were extrapolated using the above temperature – altitude relationships beginning with known records of the mean annual temperatures of the closest eco-climatic zone. 2.3. Vegetation distribution along the transect In Ethiopia, just like in East Africa as a whole, the natural vegetation shows a marked tendency towards zonation according to altitude and humidity (von Breitenback, 1963; Brown and Cocheme, 1969; Tamrat, 1993). Vegetation changes as a function of climate/altitude along the elevation transect selected (Lundgren, 1971). At the lower altitude end of the transect (approximately 1500– 1750 m asl) is the Rift Valley plain, which is a semiarid lowland (the driest and hottest section of the transect as well). This plain is covered with open Acacia woodlands (AWL), dominated by several leguminous tree species such as Acacia tortilis, Acacia senegal, Acacia mellifera, Acacia seyal, Acacia persiciflora, Acacia etbaica, Balanites aegyptiaca, Euphorbia candelabrum and other small-leaved deciduous woodland species and shrubs (Getachew, 1999; Mekuria et al., 1999; Demel, 2000). The Acacia woodland gradually gives way to a dry open decidous woodland ecotone of a transitional vegetation type (dry TWL) comprising species such as Acokanthera schimperi, Carissa edulis, Euclea schimperi, Dodonea angustifolia, Oleo africana, Croton macrostachyus, Euphorbia abyssinica, Ficus spp., Balanites aegybtiaca and some Acacia spp. (Eriksson et al., 2003). Located at mid-altitude in the transect, between 2100 and 2600 m asl, are the tropical dry evergreen montane forests. Different plant communities comprise this section often intermingled. At the lower sub-humid portion is a Podocarpus falcatus – Cr. macrostachyus mixed forest (sub-humid PCMF), which gradually grades into the humid zone dominated by P. falcatus forest (humid PFF) (Lundgren, 1971). On reaching approximately 2700 m asl, the M. Lemenih, F. Itanna / Geoderma 123 (2004) 177–188 Rift escarpment begins to flatten and the cool subAfroalpine zone starts. The vegetation of the cool subAfroalpine zone include montane woodland thickets (cool MWLT) often degraded to scrub stage with open grasslands. In the forest section of the sub-Afroalpine region, Arundinaria alpina forest, mixed with Hagenia abyssinica, Hypericum revoltum, Dombiya torrida and few other species predominates. These vegetation communities along the elevation transect are all referred to as ‘‘Montane forests’’ in many classification systems (von Breitenback, 1961, 1963; Brown and Cocheme, 1969; Chapman and White, 1970; White, 1983; Friis, 1992; Tamrat, 1993). Visually, the humid PFF vegetation consists of high above ground biomass (AGB) followed by the sub-humid PCMF and then the dry TWL. The cool MWLT and the semiarid AWL appear to have a comparable but low above ground woody biomass. 2.4. Farming system and cropping The farming systems are dominantly subsistent based on mixed crop – livestock (mostly cattle) production. In the mixed farming system, cattle provide inexpensive and easily accessible inputs required for cultivation such as draught and threshing power, while crop production supplies crop residue as feed supplement for the livestock. However, most of the grazing is carried out in the forest and on communal grazing lands. Manure from the cattle is mainly used for homestead gardens, and those farm fields away from the home garden (which are focus of the present study) often receive no to little manure. Tillage involves ox plowing with simple plows that cultivate the soils to very shallow depths of 15 cm on average (oral communication with the farmers). Major crops grown in the drier parts of the transect are sorghum, maize, haricot-beans and sweet potato, and in the wetter zones of the transect are maize, teff, wheat and barley with only one harvest per year. Fertilize applications are very limited. Whenever applied, the types of fertilizers used in the area are diammonium phosphate (DAP) in combination with urea. Recommended rate of application is 50 kg ha
1 year
1 of each type. However, most farmers use sometimes half of the amount recommended and often none mainly due to economic reasons. The farming systems are also traditional parkland agroforestry system with 181 scattered trees. The trees were preserved from the original forest during clearance. Tree species preserved on farm vary across the eco-climatic zones. However, some of the tree species preserved in these areas have been recognized as soil improvers (Jiregna, 1997). All the farm fields were cleared as slash and burn using simple local tools such as axes and machetes. To determine the ages of the fields, we interviewed farm owners and village elders in the vicinity of each vegetation community. The information from the interviewees was counter checked by interpretation of aerial photographs taken in 1967, 1992 and 1987 over the study area to confirm the ages of the fields. 2.5. Soil sampling and analysis Soil samples were taken from under five vegetation communities and five farmlands of known age that correspond to each of the vegetation communities along the transect. The farmlands were selected within a maximum of 1.25 km distances from the edges of the respective vegetation community. Similarity in topography, parent material, soil types and site orientations were all ensured before soil samples were collected from the farmlands. Pits were dug during site selection and described according to FAO guideline for soil profile description (FAO, 1990) at each site. The results were then used to verify the similarity in soil types between the farmlands and the vegetation sites. These selection criteria, however, did not allow us to find farmlands of similar ages in all of the ecoclimatic zones. Five random points were selected within 30 m of a central point established well inside each site, and pits were dug at each random point to one-m depth. Soil samples were taken from 0– 10, 10 –20, 20 –40 and 40– 60 cm increments from each pit at the vegetation sites, and only from 0 –10 and 10– 20 cm increments at the farmland sites. Soil samples were collected with a core sampler of height 10 cm and diameter 7.2 cm smoothly pushed against the northern faces of the pits. A total of 150 soil samples (five replicates for each sample depth) were collected and analyzed for soil C and total N. Another soil core samples were removed, one from each sample depth, for bulk density determination. The soils samples were air dried and sieved to pass through 2 mm prior to analysis. Organic 182 M. Lemenih, F. Itanna / Geoderma 123 (2004) 177–188 carbon was analyzed according to Walkley & Black method, while total N was analyzed by Kjeldahl method at the national soil research laboratory in Addis Ababa. 2.6. Calculation of soil C and total N stocks and turnovers Soil C (g C m
2) stock for each sample depth were calculated from the following equations: Soil C ðg m
2 Þ ¼ z  pb  c  10 ð1Þ where z = thickness of each sample depth (cm), pb = bulk density (g cm
3) of each sample depth and c = carbon concentration (g C kg
1 soil) of each sample depth. Total N stocks (g N m
2) were also computed with a similar formula. Total stocks to 0.60 m soil depth under each vegetation type were calculated by adding soil C and total N inventories of the 0– 10, 10 –20, 20– 40, 40 –60 cm soil increments. Differences in soil bulk density caused due to shift in land use affect the calculations of soil C stocks by influencing the amount of soils that are sampled from a fixed soil depth (Ellert et al., 2001; Solomon et al., 2002). Such differences in soil bulk density in the present study were accounted for by adjusting the thickness of each sampled layers under the farmlands with respect to equivalent weights of soils under the natural forest. Sampled soil thickness of each depth under the farmlands were adjusted using the equation: zcorrected ¼ ðqforest =qfarmland Þ  z ð2Þ where zcorrected = adjusted thickness of each sampled layer under the farmlands, qforest = bulk density of each sampled soil layer under the natural forest, qfarmland = bulk density of each sampled layer under the farmlands and z = thickness of each sample depth used during field sampling (Solomon et al., 2002). The soil C stocks (kg ha
1) under the natural vegetations were related to mean annual precipitation and mean annual temperature using linear regression functions of the following form: y ¼ ax þ b ð3Þ where y = soil C stock in kg ha
1 to 0.6 m depth mineral soil, a, b = are regression coefficients and x = is the mean annual precipitation (mm) or mean annual temperature (jC). Soil C and total N losses due to shifts in land use were estimated by subtracting the total stocks of the corresponding depths under the natural forest from the stock under the farmland for each eco-climatic zone. Losses/sinks of soil C and total N for each ecoclimatic zone due to land use changes were calculated for each depth by subtracting the soil C and total N stocks under the farmlands from the soil C and total N stocks of the corresponding soil depths of their counterpart natural vegetations, respectively. The calculated losses were finally divided by the ages of the farmlands to obtain average soil C and total N losses per year for each eco-climatic zone. One-way analysis of variance (ANOVA) was used to investigate whether soil C and total N stocks and C/N ratios between the vegetation communities or the farmlands along the transect were significantly different or not. Paired Student’s t tests were also applied for comparing soil C and total N stocks and C/N ratios between the vegetation and the farmland soils along the elevation transect. 3. Results and discussion 3.1. Vegetation community and soil C/total N stocks Soil C and total N stocks in the upper 0.6 m soil differed significantly ( P < 0.05) between the vegetation communities along the elevation gradient. The soil C and total N stocks of the 0.6 m soil depth varied from the minimum of 42.9 Mg C ha
1 and 5.3 Mg N ha
1 under the semiarid AWL to the maximum of 234.6 Mg C ha
1 and 20.1 Mg N ha
1 under the humid PFF, respectively (Table 2). The high soil C and total N contents of the soils in the humid, subhumid and cool sub-Afroalpine ranges of the elevation transect are consistent with a previous study (Lundgren, 1971). These estimates were also within the ranges of estimates for tropical soils of 86 Mg C ha
1 (Brown and Lugo, 1982) and 113 Mg C ha
1 (Post et al., 1982) as well as to the estimates of the global averages for tropical Ultisol (83 Mg C ha
1), Oxisols (97 Mg C ha
1) and Aridisol (42 Mg C ha
1) reported by Kimble et al. (1990), or global averages for Andosols (254 Mg C ha
1) and Nitosols (84 Mg 183 M. Lemenih, F. Itanna / Geoderma 123 (2004) 177–188 Table 2 Mean ( F S.D.) of soil C (%), total N (%) and C/N for each sample layer as well as soil C and N stocks in kg ha
1 of the 0.6 m soil under the different vegetation community along the elevation transect Total N (%) C/N Total soil C (kg C ha
1) Total N (kg N ha
1) 7.00 F 0.99 3.25 F 0.89 2.31 F 0.50 1.57 F 0.94 0.51 F 0.12 0.32 F 0.08 0.21 F 0.04 0.14 F 0.09 13.94 F 2.50 10.07 F 1.20 11.09 F 1.31 12.11 F 2.21 46,700 25,992 45,837 35,168 156,697 4677 3087 4138 3203 15,105 0 – 10 10 – 20 20 – 40 40 – 60 Sum (0 – 60cm) 11.73 F 2.77 7.28 F 1.81 4.67 F 0.89 1.28 F 0.39 1.05 F 0.25 0.63 F 0.17 0.43 F 0.09 0.12 F 0.02 11.17 F 1.45 11.48 F 0.64 10.80 F 1.27 10.30 F 1.42 66,850 50,260 91,591 25,917 234,618 6965 4402 8546 242 20,155 Sub-humid Podocarpus – Croton mixed forest 0 – 10 10 – 20 20 – 40 40 – 60 Sum (0 – 60 cm) 7.09 F 1.29 3.04 F 0.82 1.87 F 0.16 1.07 F 0.46 0.79 F 0.17 0.27 F 0.07 0.18 F 0.03 0.10 F 0.03 9.03 11.24 10.68 10.35 46,820 23,696 37,145 25,894 133,555 5227 2114 3536 2444 13,321 Dry transitional woodland 0 – 10 10 – 20 20 – 40 40 – 60 Sum (0 – 60 cm) 3.90 F 0.63 2.44 F 0.19 1.03 F 0.41 0.66 F 0.09 0.43 F 0.05 0.29 F 0.08 0.11 F 0.03 0.09 F 0.01 8.95 F 1.12 9.17 F 3.96 8.89 F 2.45 7.52 F 0.60 37,507 23,912 22,563 13,566 97,548 4195 2862 2507 1816 11,379 Semiarid Acacia woodland 0 – 10 10 – 20 20 – 40 40 – 60 Sum (0 – 60cm) 1.38 F 0.49 1.08 F 0.15 0.43 F 0.17 0.19 F 0.05 0.20 F 0.07 0.14 F 0.01 0.04 F 0.02 0.02 F 0.01 6.84 F 0.29 7.52 F 0.90 9.34 F 0.94 8.75 F 1.66 14,777 12,718 8439 4362 40,296 2151 1696 931 565 5343 Eco-climatic/ vegetation zones Depth (cm) Cool Montane woodland thicket 0 – 10 10 – 20 20 – 40 40 – 60 Sum (0 – 60 cm) Humid P. falcatus forest Soil C (%) C ha
1) reported by Batjes (1996). However, these cited estimates refer to the upper 100 cm soil depth. The values obtained for total N stocks are also within the ranges of global averages for a range of mineral soils reported by Batjes (1996) and also to those given for the Amazon basin (Moraes et al., 1995). The difference in soil C and total N stocks of the soils between the vegetation communities did not always follow the same pattern as with the visually observable AGB along the elevation transect. For instance, despite the large biomass differences between the sub-humid PCMF and cool MWLT, the former being larger than the latter, the soil of the cool MWLT had 19.6 Mg ha
1 soil C and 1.8 Mg ha
1 total N larger than the soil of the sub-humid PCMF. Similarly, the soil of the cool MWLT had nearly three fold greater soil C and total N stocks than the soil of F 0.44 F 1.31 F 1.41 F 1.68 the semiarid AWL, despite the comparable biomass between the two vegetation communities. Soil C and total N stocks increased from the semiarid AWL to the humid PFF, with the exception of the cool sub-Afromontane eco-climatic zone, which was in a direct proportion to the amount of precipitation prevailing and inversely proportional to the mean annual temperature. Likewise, the higher soil C and total N stocks at the cool sub-Afroalpine soil compared to the soil C and total N stocks at the semiarid AWL, dry TWL and sub-humid PCMF eco-climatic zones that may reflect the role of temperature in controlling soil C storage. However, the soil C and total N stocks of the soil at the cool sub-Afroalpine eco-climatic zone are low compared to the soil at the humid PFF eco-climatic zone, although the former received the largest precipitation and mildest temper- 184 M. Lemenih, F. Itanna / Geoderma 123 (2004) 177–188 overall soil C and total N stocks and dynamics along such an elevation gradient. Regression analysis yielded a significant ( P < 0.05) positive correlation (r = 0.69) between the soil C stocks along the elevation gradient and the mean annual precipitation, and significant ( P < 0.05) negative correlation (r = 0.71) between the soil C stocks and the mean annual temperature (Fig. 2). This is more or less consistent with other studies such as Post et al. (1982), where soil C stock is shown to increase with increasing precipitation and decrease with increasing temperature. The soil C distribution with soil depth along the elevation transect varied with precipitation gradient (Fig. 3). Soils in the drier section of the transect, from the semiarid to the sub-humid eco-climatic zones, had more than 50% of their soil C held in the upper 20 cm soil depth, while the soils in the more precipitation zones, humid and cool eco-climatic zones, had even distribution of soil C along the whole sampled soil depths. This could be probably due to leaching of dissolved soil C through the soil profile due to the high precipitation prevailing at these eco-climatic zones. Fig. 2. Relationships between total soil C and mean annual rainfall (a) and mean annual temperature (b). ature. This is understood to be due to the joint interactions between precipitation, temperature and amount of net primary production in controlling the 3.2. Carbon/nitrogen (C/N) ratios Soil C/N ratios differed significantly between the vegetation communities ( P < 0.001) and also between the farmlands ( P < 0.01) but not with soil depth within each vegetation community. Paired t test did not show Fig. 3. Distribution of soil C along the soil depths as % of the total soil C of the upper 0.6 m for each vegetation community along the elevation transect. 185 M. Lemenih, F. Itanna / Geoderma 123 (2004) 177–188 significant C/N ratio differences between the vegetation and the farmland soils. On the other hand, the soil C/N ratios of both the natural vegetation and the farmland soils decreased with drop in elevation. The soil at the cool sub-Afroalpine range had the highest C/N ratio of 13.9 in the surface 0 –10 cm layer and an average of 11.8 for the entire depth, while the soil at the semiarid AWL had the lowest C/N ratio of 6.8 in the surface layer and an average of 8.1 for the entire depth, a trend consistent with a previous work (Lundgren, 1971). The reason for this trend could be due to the N-fixing behavior of the acacia trees that predominate the semiarid AWL eco-climatic zone. Most Acacia spp. in the lowland vegetation are Nfixing. Some of these species still coexisted with other species in the vegetation at the dry TWL eco-climatic zone as well. This may also verify the possible differences between the vegetation communities in terms of their organic matter quality and total soil C and N stock variations. 3.3. Losses of soil C and total N due to land use changes Soil C and total N stocks in the farmland soils were significantly lower ( P < 0.05) than the soil C and total N stocks of the soils under the natural vegetation (Table 3). This shows that conversions of the natural vegetations into farmlands induce reductions of soil C and total N stocks. The reduction was most pronounced in the surface 0 –10 cm soil depth (Table 4). Soil C and total N also declined in the 10– 20 cm soil depth, but not at all of the eco-climatic zones. Generally, losses were minimum for the semiarid AWL and sub-Afroalpine vegetation ranges and highest for the humid and sub-humid vegetation zones. The rates of soil C losses in the top 0– 10 cm were about 2.0 – 3.0% per annum in the dry to humid forest ranges and between 0.5% and 1.0% per annum in the semiarid lowland and cool sub-Afroalpine eco-climatic zones. The highest losses were 53% (1773 kg ha
1 year
1) soil C and 68% (235.3 kg ha
1 year
1) total N in the humid PFF eco-climatic zone followed by 51.25% (1748 kg ha
1 year
1) soil C and 47% (179.2 kg ha
1 year
1) total N in the dry TWL eco-climatic zone (Table 4). There were also large soil C and total N losses in the sub-humid PCMF zone. Losses were minimum for the semiarid AWL and sub-Afroalpine vegetation ranges and highest for the humid and sub-humid vegetation zones (Table 4). The reason for the low soil C and total N losses in the sub-Afroalpine eco-climatic zone could be attributed Table 3 Mean F S.D. of soil C (%) and total N (%) and C/N for each sampled soil layer as well as soil C and total N in (kg ha
1) of the 0.2 m soil depth under the farmlands along the elevation transect Eco-climatic/vegetation zones Farmland age Depth (cm) Soil C (%) N (%) C/N Soil C (kg C ha
1) N (kg N ha
1) Cool Montane woodland thicket 22 0 – 10 10 – 20 Sum (0 – 20 cm) 5.79 F 2.36 3.69 F 1.44 0.48 F 0.05 0.39 F 0.11 11.81 F 3.83 9.53 F 2.63 52,518 31,205 83,723 4373 3273 7646 Humid P. falcatus forest 20 0 – 10 10 – 20 Sum (0 – 20 cm) 5.49 F 1.84 4.35 F 1.03 0.39 F 0.07 0.42 F 0.07 11.42 F 3.03 12.78 F 3.56 31,323 32,210 63,533 2259 2506 4765 Sub-humid Podocarpus – Croton mixed forest 20 0 – 10 10 – 20 Sum (0 – 20 cm) 4.21 F 0.71 2.14 F 0.79 0.42 F 0.02 0.21 F 0.04 9.95 F 1.34 9.90 F 2.04 26,953 16,570 43,523 2700 1673 4373 Dry transitional woodland 11 0 – 10 10 – 20 Sum (0 – 20 cm) 1.91 F 0.61 2.00 F 0.11 0.23 F 0.03 0.24 F 0.03 8.09 F 1.55 8.36 F 0.89 18,284 19,594 37,878 2224 2361 4585 Semiarid Acacia woodland 15 0 – 10 10 – 20 Sum (0 – 20 cm) 1.59 F 0.25 1.42 F 0.12 0.22 F 0.02 0.21 F 0.02 6.96 F 0.44 6.77 F 0.18 16,984 16,737 33,721 2027 2210 4237 186 M. Lemenih, F. Itanna / Geoderma 123 (2004) 177–188 Table 4 Losses or gains of soil C and total N from 0 – 10 and 10 – 20 cm soil depths following conversion of forestlands to arable lands along the elevation transect (+ signs denote soil C or total N gains and those without indicate soil C or total N losses) Eco-climatic/ vegetation zone Depth (cm) Losses/sinks Soil C (%) Cool Montane woodland thicket Humid P. falcatus forest Sub-humid Podocarpus – Croton mixed forest Dry transitional woodland Semiarid Acacia woodland 0 – 10 8.97 10 – 20 +11.94 Total N (kg ha
1) (%) (kg ha
1) 4457.94 +3103.76 6.51 304.70 +6.01 +185.67 0 – 10 10 – 20 53.00 35,452.50 48.97 24,612.00 68.00 4706.00 43.08 1896.40 0 – 10 10 – 20 42.00 19,867.81 30.00 7126.38 48.00 2526.82 21.00 441.31 0 – 10 10 – 20 0 – 10 10 – 20 51.25 19,222.88 47.00 1970.78 13.16 2969.67 17.50 500.52 10.00 1419.67 17.00 356.17 +7.67 +975.98 +30.00 +513.94 to the low mean annual temperature prevailing at such high altitude. These results show that deforestation and subsequent cultivation in the humid eco-climatic zone releases more CO2 to the atmosphere than zones that are either at the dry and low altitudes or cool and moist high altitudes along the elevation gradient studied. Generally, these estimated losses are within the ranges of soil C losses due to conversion from natural to agricultural ecosystems reported from various geographical regions, and also to the world average 1% year
1 (Lal, 2001). The large soil C loss at the humid PFF eco-climatic zone could be due to the high initial soil C stock under the natural vegetation (Mann, 1986; Lal, 2001) coupling the favorable climatic conditions, combination of sufficient moisture and moderately high mean annual temperature, which probably favor rapid organic matter mineralization at this eco-climatic zone. On the other hand, the low soil C and total N losses in the sub-Afroalpine eco-climatic zone could be attributed to the low mean annual temperature prevailing at such high altitude. However, the relatively low loss at the semiarid AWL eco-climatic zone was probably due to soil moisture limitation that hindered rapid organic matter decomposition (Garten et al., 1999) or due to the calcareous nature of the soil (Spain et al., 1983; Mendoza-Vega, 2002). Precipitation is so low at the semiarid Rift Valley plain that crop production often suffers failure. Therefore, soil moisture is rarely sufficient to sustain soil microbial activity for extended period each year at this part of the elevation gradient to enhance large soil C losses. N-fixing trees have been shown to store more soil C than non-fixing species (Resh et al., 2002). The high N availability, as a result of N fixation by the acacia trees, might have also reduced organic matter decomposition rate and lower loss of soil C at the AWL eco-climatic zone. Equally possible is the calcareous nature of the soil at this end of the transect. The presence of CaCO3 in soil is also known to protect soil organic matter from decomposition (Duchaufour, 1982; Spain et al., 1983; Sanchez et al., 1989; Mendoza-Vega, 2002). There were soil C or total N gains in the sub-soils of the farmlands at the two ends of the elevation transect. The total N gains in the sub-soil for the farmland at the semiarid eco-climatic zone, which is characterized by N-fixing vegetation species, may be due to the several factors that retard the decomposition of the SOM. The gain of soil C in the sub-soils of the farmland at the sub-Afroalpine zone is not evident, probably due to soil C leaching owing to the excessive precipitation and the flatter topography of the fringe. In general, the present subsistent mode of agricultural production systems across the elevation transect generally resulted in net losses of large amounts of soil C and total N. Although, few studies have been made so far in Ethiopia on organic matter dynamics, a recent investigation on soil C change following deforestation and 30 years of subsequent cultivation at the subhumid forest range of the present elevation transect indicated a 63% soil C (40.2 Mg ha
1) and 60% total N (2.8 Mg ha
1) declines in the 0 – 10 cm soil layer (Solomon et al., 2002). In another report, soil C and total N declined by 51.2% and 59.2%, respectively, within 53 years after conversion to arable lands along the same sub-humid forest chain (Mulugeta et al., unpublished). 4. Conclusion The empirical data obtained from this study showed that although deforestation and subsequent cultivation results in significant reduction of soil C M. Lemenih, F. Itanna / Geoderma 123 (2004) 177–188 stocks, the rates and magnitudes of losses varied considerably between eco-climatic zones. This implies that national estimates of CO2 emissions on the basis of few data sets will be highly unrealistic, particularly given the wide topographic, vegetation, climatic and edaphic variability in Ethiopia. Therefore, we suggest that attempts should be made to acquire reliable information on the types of forest available, annual area cleared for each type and the amount of carbon losses associate with conversion of each forest type for reliable accounting of national scale CO2 emissions as related to land use change and forestry in Ethiopia. Acknowledgements We would like to extend our thanks to Girmay Z., the staff of Munessa-Shashamane Forest Industry Enterprise and all staff members of the Abijata-Shalla sanctuary for their cooperation during the fieldwork. Thanks are due to the farmers who allowed sample collection from their farmlands. We are very grateful to the Department of Forest Soils of the Swedish University of Agricultural Sciences (SLU) for providing facilities during the preparation of writing of the manuscript. References Alvarez, R., Lavado, R.S., 1998. Climate, organic matter and clay content relationships in the Pampa and Chaco soils, Argentina. Geoderma 83, 127 – 141. Amelung, W., Flach, K.W., Zech, W., 1997. Climate effects on soil organic matter composition in the Great Plains. Soil Sci. Soc. Am. J. 61, 115 – 123. Asress, W., 1995. Ethiopia: greenhouse gas emissions and sources. In: Ramos-Mañé, C., Benioff, R. (Eds.), Interim Report on Climate Change Country Studies. http://www.gcrio.org/CSP/pdf/ Ethiopia_Inven_Vuladap_mitop.pdf. Batjes, N.H., 1996. Total carbon and nitrogen in the soils of the world. Eur. J. Soil Sci. 47, 151 – 163. Brown, L.H., Cocheme, J. 1969. A study of the agroclimatology of the highlands of Eastern Africa. FAO/UNESCO/WMO. Brown, S., Lugo, A.E., 1982. The storage and production of organic matter in tropical forests and their role in the global carbon cycle. Biotropica 14, 161 – 187. Chapman, J.D., White, F., 1970. The Evergreen Forests of Malawi. Commonwealth Forestry Institute, Oxford University, UK. Demel, T. 1996. Seed ecology and regeneration in dry Afromontane 187 forests of Ethiopia. Doctoral thesis. Swedish University of Agricultural Sciences, Umeå, Sweden. Demel, T., 2000. Vegetation types and forest fire management in Ethiopia. Proceedings: Round table conference on integrated forest fire management in Ethiopia. 19 – 20 Sept. 2000. Ministry of Agriculture With GTZ and GFMC, Addis Ababa, Ethiopia, pp. 1 – 35. Duchaufour, P., 1982. Pedology. George Allen & Unwin, London, pp. 28 – 66. EFAP (Ethiopian Forestry Action Program) 1994. Ethiopian forestry action program. The Challenge for Development, vol. 2. Ministry of Natural Resources Development and Environmental Protection, Addis Ababa. Ellert, B.H., Janzen, H.H., McConkey, B., 2001. Measuring and comparing soil carbon storage. In: Lal, R., Kimble, J.M., Stewart, B.A. (Eds.), Assessment Methods for Soil Carbon. Advances in Soil Science. CRC Press, Boca Raton, pp. 131 – 146. Eriksson, I., Demel, T., Granström, A., 2003. Response of plant communities to fire in an acacia woodland and a dry Afromontane forest, southern Ethiopia. For. Ecol. Manag. 177, 39 – 50. FAO, A., 1990. Guidelines for soil profile description. Food and Agricultural Organization of the United Nations, Rome, Italy. Friis, I., 1992. Forests and forest trees of Northeast tropical Africa. Kew Bull., Add. Ser. 15, 1 – 396. Friis, I., Lawesson, J., 1993. Altitudinal zontation in the forest tree flora of Northeast tropical Africa. Opera Bot. 121, 125 – 127. Garten, C.T., Post, W.M., Hanson, P.J., Cooper, L.W., 1999. Forest soil carbon inventories and dynamics along an elevation gradient in the southern Appalachian Mountains. Biogeochemistry 45, 115 – 145. Getachew, E. 1999. Assessment of fuelwood Resources in Acacia Woodlands in the Rift Valley of Ethiopia. Towards the development of planning tools for sustainable management. Doctoral thesis. Swedish University of Agricultural Sciences, Umeå. Glaser, B., Turrion, M.B., Solomon, D., Ni, A., Zech, W., 2000. Soil organic matter quantity and quality in mountain soils of the Alay Range, Kyrgyzia, affected by land use change. Biol. Fertil. Soils 31, 407 – 413. Hontoria, C., RodrÍguez-Murillo, C.J., Saa, A., 1999. Relationships between soil organic carbon and site characteristics in peninsular Spain. Soil Sci. Soc. Am. J. 63, 614 – 621. Jiregna, G. 1997. Decomposition and nutrient release from leaves of Croton macrostchyus and Millettia ferruginea for soil improvement in agroforestry systems. MSc thesis. Report No. 1997:24. Swedish University of Agricultural Sciences, Skinnskatteberg. Kimble, J., Cook, T., Eswaran, K., 1990. Organic matter in soils of the tropics. Proceedings of the Symposium on the Characterization and Role of Organic Matter in Different Soils. Int. Congr. Soil Sci. 14th, Kyoto, Japan. 12 – 18 Aug. 1990. ISSS, Wageningen the Netherlands, pp. 250 – 258. Kirschbaum, M.U.F., 1995. The temperature dependence of soil organic matter decomposition, and the effect of global warming on soil organic matter storage. Soil Biol. Biochem. 27, 753 – 760. Lal, R., 2001. World cropland soils as a source or sink for atmospheric carbon. Adv. Agron. 71, 145 – 191. Lundgren, B., 1971. Soil Studies in a Montane Forest in Ethiopia. 188 M. Lemenih, F. Itanna / Geoderma 123 (2004) 177–188 Research Notes Nr11. Department of Forest ecology and Forest Soils, Royal College of Forestry, Stockholm. Makin, M.J., Kingham, J.J., Waddams, A.E., Brichall, C.J., Tamene, T. 1975. Land Resources Study 21. Development Prospects in the Southern Rift Valley, Ethiopia. Land Resources Division, Great Britain Ministry of Overseas Development No. 21. pp. 270 + maps. Mann, L.K., 1986. Changes in soil carbon storage after cultivation. Soil Sci. 142, 279 – 288. Mekuria, A., Demel, T., Olsson, M., 1999. Soil seed flora, germination and regeneration pattern of woody species in an Acacia woodland of the Rift Valley in Ethiopia. J. Arid Environ. 43, 411 – 435. Mendoza-Vega, J. 2002. Influences of land use/land cover and soil types on amounts of soil organic carbon and soil characteristics. Doctoral thesis. Swedish University of Agricultural Sciences, Uppsala, Sweden. Mesfin, A., 1998. Nature & Management of Ethiopian Soils. Alemaya University of Agriculture, Ethiopia. Moraes, J.L., Cerri, C.C., Mellilo, J.M., Kichlighter, D., Neill, C., Skole, D.l., Steudler, P.A., 1995. Soil carbon stocks of the Brazilian Amazon Basin. Soil Sci. Soc. Am. J. 59, 244 – 247. Mulugeta Lemenih, Erik, K., Mats, O. (unpublished). Assessing soil chemical and physical property responses to deforestation and subsequent cultivation in smallholders farming system in Ethiopia. Swedish University of Agricultural Sciences, Department of Forest soils, Uppsala Sweden. Murphy, P.G., Lugo, A.E., 1986. Ecology of tropical dry forest. Ann. Rev. Ecolog. Syst. 17, 67 – 88. Parfitt, R.L., Theng, B.K.G., Whiton, T.S., Shepherd, T.G., 1997. Effects of clay minerals and land use on organic matter pools. Geoderma 75, 1 – 12. Post, W.M., Emaanuel, W.R., Zinke, P.J., Stangenberger, A.G., 1982. Soil carbon pools and world life zones. Nature (London) 298, 156 – 159. Quideau, S.A., Chadwick, Q.A., Benesi, A., Graham, R.C., Anderson, M.A., 2001. A direct link between forest vegetation type and soil organic matter composition. Geoderma 104, 41 – 60. Resh, S.G., Binkley, D., Parrotta, A.J., 2002. Greater soil carbon sequestration under nitrogen-fixing trees compared with Eucalyptus species. Ecosystems 5, 217 – 231. Reusing, M. 1998. Monitoring of Forest Resources in Ethiopia. Government of the Federal Democratic Republic of Ethiopia, MOA, Natural Resources Management & Regulatory Department, in cooperation with GTZ, Addis Ababa, Ethiopia. Rhoades, C.C., Eckert, G.E., Coleman, D.C., 2000. Soil carbon differences among forest, agriculture and secondary vegetation in lower montane Ecuador. Ecol. Appl. 10, 497 – 505. Ringrose, S., Matheson, W., Vanderpost, C., 1998. Analysis of soil organic carbon and vegetation cover trends along the Botswana Kalahari Transect. J. Arid Environ. 38, 379 – 396. Sánchez, B., 1969. Relaciones cuantitatives entre la altitude, la acidez y la material orgánica en suelos cultivados. An. Edafol. Agrobiol. 28, 355 – 365. Sanchez, P.A., Palom, C.A., Szott, L.T., Cuevas, E., Lal, R., 1989. Organic input management in tropical agroecosystems. In: Coleman, D.C., Oades, J.M., Uehara, G.D. (Eds.), Dynamics of Soil Organic Matter in Tropical Ecosystems. Univ. of Hawaii Press, Honolulu, HI, pp. 125 – 152. Schjønning, R., Thomsen, I.K., Møberg, J.P., de Jonge, H., Kristensen, K., Christensen, B.T., 1999. Turnover of organic matter in differently textured soils: I. Physical characteristics of structurally disturbed and intact soils. Geoderma 89, 177 – 198. Sims, Z.R., Nielsen, G.A., 1986. Organic carbon in Montana soils as related to clay content and climate. Soil Sci. Soc. Am. J. 50, 1269 – 1271. Solomon, D., Fritzsche, F., Lehmann, J., Tekalign, M., Zech, W., 2002. Soil organic matter dynamics in the subhumid agroecosystems of the Ethiopian highlands: evidence from natural 13C abundance and particle-size fractionation. Soil Sci. Soc. Am. J. 66, 969 – 978. Spain, A.V., Isbell, R.F., Probert, M.E., 1983. Soil organic matter. Soils, An Australian Viewpoint. Academic Press, London. Tamrat, B., 1993. Vegetation ecology of remnant Afromontane forests on the Central Plateau of Shewa, EthiopiaActa Phytogeographica Suecica, vol. 79. Opulus Press, Uppsala, Sweden, pp. 1 – 59. Townsend, A.R., Vitousek, P.M., Trumbore, S.E., 1995. Soil organic matter dynamics along gradients in temperature and land use on the island of Hawaii. Ecology 76, 721 – 733. Trumbore, S.E., Vitousek, P.M., Amundson, R.R., 1996. Rapid exchange between soil carbon and atmospheric carbon dioxide driven by temperature change. Science 272, 393 – 396. von Breitenback, F., 1961. Forests and woodlands of Ethiopia, a geobotanical contribution to the knowledge of the principal plant communities of Ethiopia, with special regard to forestry. Ethiop. For. Rev. 1, 5 – 16. von Breitenback, F., 1963. The indigenous trees of Ethiopia. Ethiopian Forestry Association, Addis Ababa. Van Noordwijk, M., Cerri, C., Woomer, P.L., Nugroho, K., Bernoux, M., 1997. Soil carbon dynamics in the humid tropical forest zone. Geoderma 79, 187 – 225. White, F., 1983. The vegetation of Africa. A descriptive memoir to accompany the UNESCO/AETFAT/UNSO vegetation map of Africa. Nat. Resour. Res. (Paris) 20, 1 – 356. Yalden, D.W., 1983. The extent of high ground in Ethiopia compared to the rest of Africa. Sinet, Ethiop. J. Sci. 6, 35 – 39.