Soil Nutrient Composition in Afromontane Forests of Northern Ethiopia

Deforestation in the northern highlands of Ethiopia has left 35,000 forest fragments ranging in size from 3 to 300 ha (Bongers et al 2006). Deforestation produces edges which increase disturbance within the forest such as decreased water availability and increased light. To determine the degree of these edge effects and the nutrient status of these forests, I analyzed the nutrient composition of soils along a gradient from pasture, exterior forest, interior forest and canopy. I examined two forests at different elevations: Bahir Dar (1800 m) and Debre Tabor (2800 m). Deforestation had a strong, negative effect on soil nutrients. The pasture soils had the lowest % nitrogen, % phosphorus and % carbon in comparison to the other habitats. It also had a significantly lower pH. Isotope signatures and bulk density were also significantly impacted. Pasture soils surprisingly had the lowest C:N ratio due to its minimal carbon stocks. Differences in elevation did not significantly impact the effect of deforestation upon the soil. The differences in pasture soils properties and nutrient composition indicate that deforestation has a significant, negative effect on soil fertility and health. Introduction The dry Afromontane forests of Ethiopia can be categorized as “tropical high montane conifer forests.” Aerially they are seen as dots on the northern highlands with average annual temperatures of 14 to 20 oC and annual rainfalls of 700 to 1100mm (Teketay 2005). These forests, which act as sacred groves for the Ethiopian Orthodox Tewahido churches (EOTC) (Wassie 2002), are being rapidly destroyed due to tree harvesting and subsequent conversion to pasture. This deforestation has left 35,000 severely fragmented forests ranging in size from 3 to 300 ha (Bongers et al 2006; Wassie 2002). In 1988, only 2.7% of the original forests remained (Cheng et al 1998). Edge Effects Deforestation generates edges which amplify disturbance within the forest and alter the natural ecosystem. Edge effects can include greater aridity, lower soil moisture levels, dehydration, and damaging winds (Camargo and Kapos (1995), Laurance et al (2000)). Tree mortality can result, allowing solar radiation and winds to penetrate further into the forest, reaching previously unexposed areas (Camargo and Kapos (1995), Laurance et al (1998)). Increased solar radiation and wind can cause ‘soil drought,’ lowering the amount of available water for vegetation (Camargo and Kapos 1995). Soil nutrient health also drastically changes at the forest border due to such edge effects. Didham (1998) reported an increased litter decomposition rate in exterior forest floor soils. Nitrogen net mineralization rates have increased, leading to more available, mobile nitrogen. However, right after disturbances, plants are unable to take up mineralized nitrogen efficiently and thus there is nitrogen loss due to leaching (Vitousek 1982; Aerts and Chapin 2000). This effect was seen in studies where forest edge soils had significantly lower %N and %C than forest interior soils (Johnson and Wedin 1997; Toledo-Aceves and García-Oliva 2007). Closeness to edge also resulted in diminished 1 Baimas-George: Soil Nutrient Composition in Afromontane Forests of Northern Ethi Published by Digital Commons @ Colgate, 2012 ! 161! phosphorus levels, indicating that available phosphorus is inversely related to nearness to edge. Phosphorus loss was due to increased occlusion and subsequent erosion (ToledoAceves and García-Oliva 2007). Thus, I predicted that the forest floor exterior soils would have lower nutrient levels (%N, %P) and a higher C:N ratio than interior forest floor soils Isotope signatures and pH should be similar. Grazing Effects The majority of the deforestation in Ethiopia is due to conversion of forest to pasture land. Tree harvesting and grazing can affect nutrient composition and nutrient processes. Grazing affects the flow of nutrients in ecosystems through stamping, ingestion, urine deposition, redeployment, and export. Tree harvesting leads initially to an increase in the amount of available nutrients which is short-lived due to leaching, erosion, and denitrification (Keenan and Kimmings 1993; Toledo-Aceves and García-Oliva 2007). Overall, the nitrogen cycling rate ends up decreasing (Reiners et al 1994; Vitousek 1984). Phosphorus pools remain high but biologically available forms are not amply accessible due to occlusion and irreversible exportations (Turrión et al (1999), Lavado et al (1995)). The open, exposed pasture will also have a higher rate of decomposition due to higher temperature and wetter climate (Prescott 2002). Due to different environments, soils, and grazing management, the effects of grazing upon soil composition have been found to be contradictory. Neither Lugo and Brown (1993) nor Murty et al (2002) saw significant changes in nutrient soil composition between forest land soils and soils from grazed, unfarmed, pastures. Other studies have seen contrasting effects. Cerri et al (1991) saw a small increase of %C due to low grazing intensity and model farm organization systems. With heavy grazing, soil %C was predicted to decrease as C inputs from above-ground vegetation would decline. A high persistent grazing intensity has also resulted in a steady reduction of %N in soil (Holdo et al 2007). This effect was confirmed consistent with studies of grazed Pampean grassland that found significant decreases in mineral nitrogen and extractable P content (Lavado et al 1995). Toledo-Aceves and García-Oliva (2007) reported finding similar results: lower concentration of %C and %N and lower microbial C and N in pasture soils versus forest floor soils as well as increases in bulk density and decreases in soil water and porosity (Reiners et al 1994; Milchunas and Lauenroth 1993). Carbon stock changes during conversion of primary forest to pasture have also been investigated as the release of carbon is significant to the global carbon equilibrium and greenhouse gas emissions. Again, results are contradictory depending on soil compaction, clay content, seasonal differences, and pasture supervision. In Brazilian Amazonia, deforestation to pasture has caused a net release of carbon, mostly from biomass (Fearnside 1997b). Twenty-five years after pasture conversion, Veldkamp (1994) found a loss of 2-18% carbon stocks in pasture soils in comparison to primary forest soils of lowland Costa Rica. Similar results were found in Rondonia, with a reduction of 5% in the carbon stock in 3-year-old pasture and a 10% loss in 20-year-old pasture (M. Grzebyk, unpublished data cited by Moraes et al, 1996; p 77). Decreases in carbon stocks were found by Detwiler (1986), Serrao and Falesi (1977) and Houghton et al (1983); these authors found, respectively, an overall decrease of 20% in carbon stock, a decrease of almost 50% in Mato Grosso, and a decline of 25% carbon content with conversion from primary forest to pasture. The decrease in carbon stocks could be due to 2 Colgate Academic Review, Vol. 8, Iss. 1 [2012], Art. 13 http://commons.colgate.edu/car/vol8/iss1/13 ! 162! numerous factors. The loss of cavernous root systems by conversion of forest to pasture could lead to loss of considerable quantities of carbon over short time periods (Nepstad et al (1994)). Minute rises in temperature can considerably amplify the rate of soil respiration, causing a swift drawdown of carbon stocks (Townsend et al 1992). With conversion of forest to pasture, the temperature of the soil rises considerably due to higher solar radiance and thus, it is reasonable to assume a decrease in carbon stocks. Loss of water storage capacity with pasture conversion limits the dispersal of carbon inputs from roots to surface tiers (Cunningham, 1963). Increased decomposition rates will also lead to carbon stock reduction (Post et al 1995). Urine deposition has also had a significant effect on grazing lands as urine spots have increased concentrations of potassium and sodium cations, hydrogen carbonate, and chloride. After deposition, urea hydrolysis and subsequent nitrification begin, escalating the ionic strength of the soil and nitrate to become the dominant ion. Nitrate leaching can then occur at a rapid pace leading to an increase in soil acidity (Haynes and Williams 1992). Lower soil pH can also be a result of an increase of organic matter as organic matter increases cation exchange capacity (CEC) (Haynes and Williams 1993). In regards to these studies, I predicted that pastoral soils would have very low nutrient levels (%N, %P), reduced carbon stocks and a high C:N ratio in comparison to the other soil habitats. It should have a lower pH (Moraes et al 1996), low %"N and high %"C signatures. Canopy and Forest Floor Soils In the canopy, nutrients can be hard to acquire due to limited sources and lack of a large soil pool. Thus epiphytes, through their diverse functional morphologies, must be resourceful and competitive to acquire sufficient nutrients (Nadkarni and Matelson 1991). Epiphytes’ main source of nutrients is found in canopy organic matter (COM) (Hietz et al (2002); Stewart et al (1995)) which is made up of live and dead components: epiphyte and host tree litter, ‘crown humus’ (Jenik 1973), vegetation roots, fungi, invertebrates and microorganisms (Nadkarni et al 2004). The ‘crown humus’ is best described as an ‘arboreal histosol’, made primarily of decomposing epiphytic bryophytes (Nadkarni et al 2002). It also contains captured nutrients from atmospheric sources, mist and dust, litterfall, shedded epiphyte tissues, decaying bark, and the debris of canopy animals (Nadkarni and Longino 1990). Thus, canopy soils is organic and contains a relatively large concentration of carbon and nutrients per unit gram soil. It is very microbially active with microbial populations and mineralization rates similar to the upper horizons of forest floor soils (Vance and Nadkarni (1990), Maffia et al (1993)). Canopy litter dynamics are essential due to the restricted amount of litter as leaching of nutrients or interception of atmospheric nutrients has major effects for epiphytes (Nadkarni and Matelson 1991). Due to the low nutrient availability in the canopy, vegetation with roots in the humus is in high competition with each other, snatching nutrients as soon as they become biologically available. Plants in limited nutrient environments can increase the maximum rate at which they can take up nutrients, allowing for more effective competition (Aerts and Chapin 2000). Conversely, nutrient sources for forest floor vegetation come mainly from the huge nutrient pool of forest floor soils (Cardelús et al 2009). 3 Baimas-George: Soil Nutrient Composition in Afromontane Forests of Northern Ethi Published by Digital Commons @ Colgate, 2012 ! 163! Canopy nutrient cycling rates and nutrient composition are different from the forest floor due to differences in microclimate, substrates, and nutrient source limitations. Since, temperature and moisture have an effect on soil processes (Prescott 2002), the differences between the microclimates of the canopy and forest floor will result in differences of soil nutrient health. The temperature in the canopy is higher than the forest floor due to increased solar radiation and since, the canopy provides shade and insulation to the forest floor, the ground temperature is ‘buffered’ (Bohlman 1995). The canopy also alters water flow and content as it changes precipitation pathways (Prescott 2002). Due to these differences and others, studies have found disparities in nutrient processes and composition between canopy and forest floor soils. In several studies, the net nitrification rate of the canopy was lower than the rate of the forest floor, thus indicating that nitrogen is more strictly conserved in the canopy as plants take up all available ammonium before nitrifiers can convert it to nitrate (Vance and Nadkarni 1990; Bohlman et al 1995). Nadkarni et al (2002) found canopy soils to have higher %C and %N but comparable %P (mg/g dry weight) to the forest floor in Costa Rican montane forests. These canopy soils were also found to have a lower C:N ratio and a lower pH. It can be hypothesized that canopy soils have higher N levels due to discrete, conserved N cycles and higher net N mineralization rates. Thus, per unit mass, N accessibility is higher in canopy soils than forest floor soils (Cardelús et al 2009). Insignificant differences in available phosphorus between canopy and forest floor soils suggest preservation and strict cycling of phosphorus in both habitats (Cardelús et al 2009; Nadkarni et al 2002). In line with these studies, I predict canopy soils to have a higher %N and %C and similar %P to interior forest floor soils. It should have a lower bulk density and a comparably high pH due to higher levels of organic matter (Haynes and Williams 1993). I predicted canopy soils to have the most depleted %"N and higher %"C values than forest floor soils (Nadkarni 1984; Stewart et al 1995). Elevation Effects I also examined the effect of elevation on soil nutrients by comparing two fragmented forests at different elevations. At higher elevations, colder temperatures decrease rates of decomposition (Aerts and Chapin (2000), Vitousek and Sanford (1986)) and thus there is lower %N and %P available in the soils (Soethe et al (2008) and Tanner et al (1998)), a higher C:N ratio due to limited N, and slow mineralization of plant litter (Soethe et al (2008), Gholz et al (2000)). In line with these studies, I predicted lower %N, %P and higher C:N ratios at the forest site of higher elevation (Debre Tabor; site 2). I also predicted that "N signatures would be lower at higher elevation due to less N losses in a more N-stressed forests; a hypothesis reflected in the studies of Cardelus et al (2010). Material and methods Study Site This study was conducted at two Afromontane forest sites: Bahir Dar and Debre Tabor in the northern highlands of Ethiopia. Site 1, Bahir Dar (37°34'E, 11°48'N), is 8 ha of old-growth Afromontane forest at an elevation of 1950m (Cardelús Fieldwork 2010). It lies on the southern tip of Lake Tana with an average temperature of 18°C during the rainy season (Bahir Dar). Site 2, Debre Tabor (37°59’ E, 11°51’N), is 11.5 ha of old4 Colgate Academic Review, Vol. 8, Iss. 1 [2012], Art. 13 http://commons.colgate.edu/car/vol8/iss1/13 ! 164! growth Afromontane forest at an elevation of 2690m with an average temperature of 17°C during the rainy season (Cardelús Fieldwork 2010; Population of Debre Tabor, Ethiopia). Deforestation and conversion of forest to pastoral lands have been more recent processes in Debre Tabor than Bahir Dar. The dominant wet season of both forests falls between July and September. Annual precipitation is greater than 1,000mm/40 and can increase to 1,500-2,000 mm/60-80 (BBC-Weather Center). The host tree in Bahir Dar was Mimusops Kummel (family Sapotaceae) with a high density of epiphytes per branch. The host tree in Debre Tabor was a Prunus Africana (family Roseceae) with a sparse density of epiphytes per branch (Cardelús Fieldwork 2010). Sample Collection Soil samples were sampled from each of the sites. For forest floor and pastoral sites, the soils were volumetrically sampled with two 30 cm deep cores with a volume of 159.2 cm, 2.5 m from the trunk of the host tree. For canopy soil samples, leaf litter was brushed off the surface of the branch and the soil was volumetrically sampled 80-125 g. In total, there were 28 samples collected (see Table 1). The wet weight, posthomogenization weight, root weight, dry weight, volume and dimensions of each sample were determined and recorded. Each tree’s height and diameter were also measured. Soil Variables All of the soil samples were ground using a coffee grinder (Toastmaster). Total C, N and isotopes ("C and "N) were determined from the soil samples through measurements by the Analytical Elemental Analyzer (Valencia, California) coupled to a Delta Plus Isotope Ratio mass spectrometer (Brenen, Germany). The samples were analyzed for pH using a pH meter (H2O; Thomas 1996). Total P was measured through a dry ash digestion (Jones and Case 1996) and then put through colorimetric analysis using a Powerwave (Astoria Pacific, Clackamas, Oregon). Bulk density was calculated using the dry weights of the soil samples and the volume of the soil core. Statistical Analyses This study was designed to examine differences in nutrient composition of soil habitats on a gradient of canopy to clearing and differences between soil health of sites of different elevations. To determine differences between soil habitats and between sites, we used analysis of variance on normally distributed data. This was followed by a posthoc student’s t test to determine whether the data was significant. Our continuous variable was the nutrient in question, pH, or bulk density and our categorical variable was the soil habitats or site. Results Edge Effects: Interior and Exterior Forest Floor Soils At both forest sites, there were no significant differences in "N, "C, %N, and pH between exterior and interior forest floor soils. However, %C and %P varied between sites. In site 1, the interior forest floor soils had a significantly higher %C then exterior forest floor soils where as in site 2, there was no significant difference in %C between habitats (Fig. 2). This same trend was observed in %P as there was significantly more %P in interior forest than exterior forest 5 Baimas-George: Soil Nutrient Composition in Afromontane Forests of Northern Ethi Published by Digital Commons @ Colgate, 2012 ! 165! The floor soils at site 1 but no significant difference observed between habitats at site 2 (Fig. 4). The exterior forest floor soils had a significantly higher bulk density than interior forest floor soils at both sites (Fig. 5) (Table 1). Pastoral Soils In site 1, the clearing soils had significantly lower %P (Fig. 4), significantly higher bulk density (Fig. 5) and was significantly more acidic with a pH of 4.93 than all other soil habitats. It had significantly lower %C and %N than the interior forest floor and canopy soils although its levels were consistently similar with exterior forest floor soils (Fig. 1, 2). The pasture soils had significantly lower C:N ratios than canopy soils (Fig. 3). The clearing soils had significantly more depleted "N signatures than the interior forest floor soils. Its "N signature was consistent with the exterior forest floor soils and significantly more enriched than the canopy soils. The clearing soils were also significantly more enriched in "C than the canopy soils but consistent with all other habitats (Table 2). In site 2, there were many fewer significant differences in nutrient composition and soil properties between clearing soils and the other soil habitats. There were similar significant differences in "N signatures, pH, and bulk density as site 1. The clearing soils has significantly lower %N than the canopy soils. Its %N was not significantly different from forest floor soils (Fig. 1). The pastoral soils’ %C, however, was significantly lower than all other soil habitats (Fig. 2). The %P was significantly higher than the canopy soils but consistent with other soil samples (Fig. 4). There were no significant differences in "C signatures between clearing soils and other soil habitats (Table 2). Canopy soils vs. Interior Forest Floor Soils There were no significant differences in pH between canopy and forest floor soils. Canopy soils had significantly lower % P values (Fig. 4) and significantly higher %C values than interior forest floor soils at both sites (Fig. 2). Canopy soils in site 1 were significantly higher in %N than interior forest floor soils though there was no significant difference in site 2 (Fig. 1). The canopy soils were significantly more depleted in "N at both sites. There were no significant differences in signatures of "C at both sites. Canopy soils bulk densities were significantly lower than interior forest floor soils at both sites (Fig. 5) (Table 2). Elevation: Between Sites The two sites varied little among nutrient composition and soil properties. Canopy soil at lower elevation (site 1) had a significantly higher bulk density than canopy soil at higher elevation (site 2) (Fig. 5). The interior forest floor soils at higher elevation had a significantly higher %N (Fig. 1), %C (Fig. 2) and more depleted "N signature than interior forest floor soils at lower elevation. The exterior forest soils at higher elevation had significantly higher %P (Fig. 4) and more depleted "N signatures than exterior forest soils at lower elevation. Lastly, the clearing soils at higher elevation had a significantly higher %P than clearing soils at lower elevation (Fig. 4) (Table 2). 6 Colgate Academic Review, Vol. 8, Iss. 1 [2012], Art. 13 http://commons.colgate.edu/car/vol8/iss1/13 ! 166! Discussion High deforestation rates have a significant, negative impact on soil nutrient composition (Laurance et al (2000); Laurance et al (1998); Camargo and Kapos (1995)). Typically, as forests shrink, edge effects increase, leading to considerable losses of soil nutrients (Johnson and Wedin (1997); Toledo-Aceves and Garcia-Oliva (2007)). In order to evaluate the extent of deforestation on soil nutrient composition in Ethiopia Afromontane forests, I examined whether soil nutrient composition significantly differed along a gradient from canopy, interior, exterior to pasture soils. We found that nutrient status across all habitats was low although pasture soils were significantly lower in nutrient composition than the other habitats. Edge effects impact the entire forests due to their small sizes. Edge Effects: Interior and Exterior Forest Floor Soils The exterior forest floor soils had significantly lower % phosphorus than the interior forest floor soils at Bahir Dar: site 1 (Figure 2). Edge soils have lower phosphorus levels as a result of decreased mineralization and decomposition due to drier soil from high evapotranspiration (Camargo and Kapos 1995). There was no significant difference between phosphorus compositions at Debre Tabor: site 2. Site 2 was more recently deforested and even though the edge soils have lower % phosphorus, it has not been exposed to edge effects long enough for there to be a significant difference. Over time, I predict that a significant difference will develop. The interior soils had a higher % carbon than exterior soils at site 2 (Figure 4). The exterior soils receive more sun exposure due to a decreased canopy cover buffer. Since small rises in temperature are significant enough to increase the rate of soil respiration, resulting in quick drawdown of carbon stocks (Townsend et al 1992), exterior soils predictably have less % carbon than interior forest soils. There was no significant difference in pH between habitats. Thus, it can be inferred that there were not significant differences in amount of rainfall or in nitrate leaching. No significant differences in % nitrogen also suggest that nitrate leaching did not have a dominant effect in edge soils. There were similar signatures of "N because both interior and exterior forest soils obtain nitrogen from substantially fractionated nitrogen forest floor soils (Hietz et al (2002); Stewart et al (1995); Cardelus et al (2009)) (Figure 6). Similar signatures of "C again imply that there were no significant differences in amount of rainfall as both habitats were equally water stressed. Forest exterior soils had a significantly higher bulk density as it is exposed to increased solar radiation and wind desiccation (Camargo and Kapos 1995) (Figure 6). There were not significant differences among every nutrient or soil property studied between the interior and exterior soils. Along with the fact that there was also a low nutrient composition throughout the entire forest, it can be presumed that edge effects are permeating the entire forest. Since both forests sampled were of small size (8 and 11 ha) and edge effects can impact up to 100 meters into a forest (Laurance et al 1998), edge effects are impacting the forest interior and canopy. Therefore, the forest edge and interior are affected similarly as reflected in some of the results. 7 Baimas-George: Soil Nutrient Composition in Afromontane Forests of Northern Ethi Published by Digital Commons @ Colgate, 2012 ! 167! Pastoral Soils The pastoral soils had a significantly lower nutrient composition than the other habitats. The pasture had the lowest % phosphorus in site 1 due to occlusion and subsequent erosion. Phosphorus gets tightly bound into compounds with iron or aluminum hydrous oxides, fixing phosphorus into an insoluble solution that will move with soil particulates in runoff (Jobbagy and Jackson 2001). Almost 90% of phosphorus can be lost through this process (Tiessen 1995). However, in site 2, the pasture had higher % phosphorus than the canopy soils. Site 2 was more recently deforested and therefore, the soils have not been exposed to the disrupted ecosystem conditions long enough for significant losses in phosphorus. The pasture soils were the most acidic with a pH of 4.93 at site 1 and a pH of 5.37 at site 2. Due to animal urine deposition, nitrogen leaching, and increased exposure to rainfall, the pH of the pasture is significantly lower than the other habitats. When rain exceeds evapotranspiration (as it does in pastoral lands), the rain leaches the soil of basic calcium and magnesium cations, replacing them with acidic aluminum and iron cations (Finzi et al 1998). Futhermore, nitrate leaching is partnered with loss of Ca2+ and Mg2+, again increasing the soil’s acidity (Steele et al 1984). Site 2 was not as acidic as site 1 because deforestation happened more recently at site 2. The pasture had the lowest carbon stocks: 3.3% in site 1 (Figure 5) and 2.9% in site 2 compared to primary forest, which is consistent with findings in other studies (Detwiler (1986); Veldkamp (1994); Reiners et al (1994); Van Dam et al (1997)). Low % carbon in the pasture is a result of high grazing intensity, increased temperatures from high solar radiation, and soil damage during pasture clearing (Nepstad et al (1994); Elmore and Asner (2006)). Increased decomposition (Schlesinger (1984); Post et al (1995)), amplified oxidation of organic matter, elimination of topsoil by clearing (Detwiler 1986), and loss of water storage capacity (Cunningham, 1963) also contribute to a decline in carbon stocks. The pasture soils also had significantly lower % nitrogen than the interior and canopy at site 1 (Figure 3). Low nitrogen levels are a result of nitrate leaching and erosion (Lavado et al (1995); Holdo et al (2007)). Lower rates of nitrogen mineralization (and sometimes nitrification) are also commonly seen in conversion of forest to pasture. Ammonium pools are higher in pastures than forests due to slower rates of plant uptake of ammonium. Therefore, the overall nitrogen cycling rate is less in pastures (Reiners et al 1994). At site 2, the pasture soils were only significantly lower in %nitrogen to the canopy. The forest floor soils were not significantly different in %N to the pasture because site 2 was deforested more recently than site 1. Over time, edge effects will have a greater impact further into the forest and will mirror the trends of site 1. Interestingly, the pasture soils had the lowest C:N ratio; an unexpected result (Figure 1). Usually a low C:N ratio reflects a high percentage of nitrogen. However, the pasture soils have the lowest % nitrogen. The low ratio can be explained by the low % carbon found in the pasture. The canopy soils had the highest C:N ratio; another unpredicted result. This can again be explained by the fact that although the canopy does have the highest % nitrogen, it also has the highest % carbon, resulting in a elevated C:N ratio . The pasture soils had low "N signatures due to leaching (Figure 7). The signature was higher than in the canopy because canopy soils will have the most depleted "N signatures due to fractionation and higher dependence on atmospheric and canopy only derived nitrogen sources. The pasture soils had higher "C signatures than the canopy at 8 Colgate Academic Review, Vol. 8, Iss. 1 [2012], Art. 13 http://commons.colgate.edu/car/vol8/iss1/13 ! 168! site 1 because of severe water stress from increased radiation and wind desiccation (Neill et al (1996); Milchunas and Lauenroth (1993)). At site 2, there were no significant differences in "C signatures again because of site 2’s more recent deforestation. The pasture had a significantly higher bulk density than all other habitats at both sites (Figure 6). Its high bulk density is due to animal stomping and eroded soil particles (Reiners et al 1994). Typically pastures will have a higher bulk density than terrestrial soils by 13% (Murty et al 2002). The Ethiopian pasture soils had bulk densities that were 47% (site 1) and 50% (site 2) greater than interior terrestrial soils. This could be a result of higher grazing intensity or older sites as bulk density increases with age (Van Dam et al 1997). Canopy versus Interior Forest Floor Soils Canopy soils had significantly lower % phosphorus than interior forest floor soils at both sites. Since the canopy has very limited nutrient sources, phosphorus is strictly cycled and conserved. The interior forest floor has a much larger, nutrient rich soil pool and thus has higher % phosphorus in soil. Canopy soils also had a significantly higher % carbon. Canopy soils are highly organic, containing a large pool of carbon because of shed epiphyte tissue, decaying bark, and canopy animals’ debris (Nadkarni and Longino (1990); Vance and Nadkarni (1990)). This is consistent with findings in other studies (Nadkarni et al (2002); Cardelus et al (2009)). The canopy soils had significantly higher % nitrogen in site 1 (Figure 3). Canopy soils has the highest % nitrogen per unit gram of soil due to extremely strict nitrogen cycling as reflected in low to negative nitrification rate which is consistent with other studies (Bohlman et al (1995); Cardelus et al (2009)). There was no significant difference in % nitrogen between canopy and interior soils in site 2. This could reflect the amount of time since deforestation, as site 2 was deforested more recently. The canopy soils had a more depleted "N signature due to fractionation and higher dependence on atmospheric and canopy only derived nitrogen sources (Figure 7). The interior forest floor soils had more enriched signatures because forest floor soils obtain nitrogen after substantial fractionation (Hietz et al 2002). There was no difference in "C signatures between the canopy and interior floor soils although it was predicted the canopy soils would have higher values due to restricted water storage capacity as a result of the small, limited amount of canopy soils (Nadkarni (1984); Stewart et al (1995)). Due to the small size of the forest, the edge effect of evapotranspiration has a strong effect on the forest interior, leading to equal water stress levels between the canopy and interior soils. Canopy soils had a significantly lower bulk density at both sites as it is very organic, loose, and porous (Figure 6). Forest interior soils are mineral soil and thus less organic, resulting in a slightly higher bulk density. There was no difference in pH between the canopy and interior soils. I had predicted that the canopy soils would be more acidic due to high levels of organic matter which lead to an increase in CEC (Hayes and Williams 1993). However, the canopy soils had very low % carbon in comparison to a healthy forest and this possibly had a positive effect on the pH, making it more basic and closer to forest floor pH. 9 Baimas-George: Soil Nutrient Composition in Afromontane Forests of Northern Ethi Published by Digital Commons @ Colgate, 2012 ! 169! Elevation Between Sites There were not many significant differences between the two sites of different elevations. It is hard to accurately evaluate the few significant differences because the sites were deforested at different points in time and thus this confounding variable will skew analysis of nutrient differences in regards to elevation. Furthermore, differences in grazing intensity and management will alter nutrient composition and soil properties. However, it is apparent that regardless of elevation, deforestation has a negative impact on soil nutrient composition. Conclusion These data exemplify the impacts of deforestation on soils from a gradient of canopy to pastoral land. My data demonstrates that deforestation has a significant, negative effect on soil nutrient composition and this impact increases with time since deforestation. The pasture soils are most heavily affected as evidenced by their significantly lowest nutrient levels. The low nutrient levels in the canopy and interior soils indicated that edge effects are permeating the entire forests. Due to the small sizes of the sampled forests, edge effects can impact the interior. Therefore, the remaining, fragmented forests are severely influenced when trees are harvested. Continual tree harvesting will destroy these forests and make regeneration highly improbable. Future research should examine larger forests, several hundred ha in diameter. This research could better illuminate the impact of edge effects and see how far they impact into the forest. These studies could also establish whether forest regeneration is a possibility in larger forests. Acknowledgments I would like to thank Dr. Catherine Cardelús for her mentoring, laboratory help, and sample collection. I would also like to thank Dr. William Peck for his help and patience with the elemental analyzer and mass spectrometer. I would like to thank Colgate University for laboratory equipment and funds. 10 Colgate Academic Review, Vol. 8, Iss. 1 [2012], Art. 13 http://commons.colgate.edu/car/vol8/iss1/13 !170!ReferencesAerts, R. and F.S. Chapin III. The Mineral Nutrition of Wild Plants Revisited: AReevaluation of Processes and Patterns. Advances in Ecological Research. 30 (1-67):2000."Bahir Dar, Ethiopia." GAISMA. Web. 16 Sept. 2010.."BBC Weather Centre World Weather – Country Guides Ethiopia." BBC -Homepage.Web.16 Sept. 2010..Bohlman, Stephanie A., Teri J. Matelson, Nalini M. Nadkarni. Moisture and temperaturepatterns of canopy humus and forest floor soil of a montane cloud forest, Costa Rica.Biotropica 21,1(13-19):1995.Bongers, Frans. Alemayehu Wassie, Frank J. Sterck, Tesfaye Bekele, and DemelTeketay. Ecological Restoration and Church Forests in northern Ethiopia. Journal ofDrylands 1. 1(2006): 35-44.Camargo, J. L. C. and V. Kapos. Complex Edge Effects on Soil Moisture andMicroclimate in Central Amazonian Forest. Journal of Tropical Ecology 11.2(1995): 205-221.Cardelús, Catherine L. and Michelle Mack. The Nutrient Status of Epiphytes and theirHost Trees along an Elevational Gradient in Costa Rica. Plant Ecology. 207(25-37):2010.Cardelús, Catherine L., Michelle C. Mack, Carrie Woods, Jennie DeMarco and KathleenK.Treseder. The Influence of Tree Species on Canopy Soil Nutrient Status in aTropical Lowland Wet Forest in Costa Rica. 2009.Cerri CC, Volkoff B, Andreaux F Nature and behavior of organic matter in soils undernatural forest, and after deforestation burning and cultivation, near Manaus. ForestEcology and Management, 38(247-257):1991.Cheng, Sheauchi, Yukio Hiwatashi, Hideki Imai, Mitsuru Naito, and Tatsuka Numata.Deforestation and degradation of natural resources in Ethiopia: Forest managementimplications from a case study in the Belete-Gera Forest. Journal of ForestResearch 3,4(199-204):1998.Cunningham, R.H. The effect of clearing a tropical forest soil. J. Soil Science14(344):1963.Detwiler, R.P. Land use change and the global carbon cycle: the role of the tropical soils.Biogeochemistry 2(67-93):1986.Didham, Raphael K. Altered leaf-litter decomposition rates in tropical forest fragments.Oecologia 116(397-406):1998.Elmore, Andrew and Gregory Asner. Effects of grazing intensity on soil carbon stocksfollowing deforestation of a Hawaiian dry tropical forest. Global Change Biology12:9(1761-1771):2006.Fearnside, P.M. Greenhouse gases from deforestation in Brazilian Amazonia: netcommitted emissions. Climate Change 35, 3(321-360):1997.Finzi, Adrien C., Charles D. Canham, and Nico van Breemen. Canopy Tree-SoilInteractions within Temperate Forests: Species Effects on pH and Cations.Ecological Applications, 8(447-454):1998. 11Baimas-George: Soil Nutrient Composition in Afromontane Forests of Northern Ethi Published by Digital Commons @ Colgate, 2012 !171!Gholz, Henry L., David A. Wedin, Stephen M. Smitherman, Mark E. Harmon, andWilliam J.Long-term dynamics of pine and hardwood litter in contrastingenvironments: toward a global model of decomposition. Global Change Biology6(751-765):2000.Haynes, R. J. and P. H. Williams. Changes in Soil Solution Composition and pH inUrine Affected Areas of Pasture. Journal of Soil Sciences. 43 (323-334): 1992.Haynes, R.J. and P.H. Williams. Nutrient cycling and soil fertility in the grazed pastureecosystem. Advances in Agronomy 49(119-199):1993.Hietz, Peter, Wolfgang Wanek, Rita Wania, and Nalini M. Nadkarni. Nitrogen-15natural abundance in a montane cloud forest canopy as an indicator of nitrogencycling and epiphyte nutrition. Oecologia. 131 (350-355): 2002.Holdo, Ricardo M., Robert D. Holt, Michael B. Coughenour, and Mark E. Ritchie. PlantProductivity and Soil Nitrogen as a Function of Grazing, Migration and Fire in anAfrican savanna. Journal of Ecology. 95 (115-128): 2007.Houghton, R.A., J.E. Hobbie, J.M. Melillo, B. Moore, B.J. Peterson, G.R. Shaver, andG.M. Woodwell. Changes in the carbon content of terrestrial biota and soils between1860 and 1980: a net release of CO2 to the atmosphere. Ecological Monographs53(235-262):1983.Jenik, J. Root systems of tropical trees Preslia 45(250-264):1973.Jobbagy, Esteban G. and Robert B. Jackson. The Distribution of Soil Nutrients withDepth: Global Patterns and the Imprint of Plants. Biogeochemistry, 53, 1(51-77):2001.Johnson, N.C. and D.A. Wedin. Soil carbon,nutrients and mycorrhizae during conversionof dry tropical forest to grassland. Ecol Appl 7)171-182):1997.Keenan, R.J. and J.P. Kimmins. The ecological effects of clear-cutting. Environ. Rev.1(121-144):1993.Schlesinger, W.H. Soil organic matter: a source of atmospheric CO2. The Role ofTerrestrial Vegetation in the Global Carbon Cycle (Chapter 4): 1984.Serrao, E.A.S. and I.C. Falesi. Pastures of the Tropical Brazilian Forests. EmpresaBrasileira de Pesquisa Agropecuaria, Belem, Para, Brazil, 63 pp.Soethe, Nathalie, Johnannes Lehmann and Christof Engels. Nutrient availability atdifferent altitudes in a tropical montane forest in Ecuador. Journal of TropicalEcology 24(397-406):2008.Steele, K.W., M.J. Judd, and P.W. Shannon. Leaching of nitrate and other nutrients froma grazed pasture. New Zealand Journal of Agricultural Research 27,1(5-11):1984.Stewart, G.R., S. Schmidt, L.L. Handley, M.H. Turnbull, P.D. Erskine, and C.A. Joly.15N natural abundance of vascular rainforest epiphytes: implications for nitrogensource and acquisition. Plant, Cell, and Environment 18,1(85-90):1995.Tanner, E.V.J., P.M. Vitousek, and E. Cuevas. Experimental investigation of nutrientlimitation of forest growth on wet tropical mountains. Ecology 79(10-22):1998.Teketay, Demel. Seed and Regeneration Ecology in dry Afromontane Forests ofEthiopia: I. Seed Production—Population Structures. Tropical Ecology 46. 1 (29-44):2005.Thomas GW. Soil pH and soil acidity. In: Sparks DL, Page CN, Helmke PA, LoeppartRH, Soltanpour PN, Tabatabai MA, Johnson CT, Sumner ME (eds) Methods of soil 12Colgate Academic Review, Vol. 8, Iss. 1 [2012], Art. 13 http://commons.colgate.edu/car/vol8/iss1/13 !172!analysis. Part 3. Chemical Methods-SSSA Book Series no. 5 Soil Science Society ofAmerica and American Society of Agronomy, Madison. (475-490):1996. Tiessen, H. Phosphorus in the global environment: transfers, cycles and management.SCOPE 54(462-467):1995.Toledo-Aceves, Tarin and Felipe García-Oliva. Effects of forest-pasture edge on C, Nand P associated with Caesalpinia eriostachys, a dominant tree species in a tropicaldeciduous forest in Mexico. Ecol Res 23(271-280):2007.Townsend, A.R., P.M. Vitousek, E.A. Holland. Tropical soils could dominate the short-term carbon cycle feedbacks to increase global temperatures. Climate Change22(293-303):1992.Turrión, M.B., B. Glaser, D. Solomon, A. Ni, W. Zech. Effects of deforestation onphosphorus pools in mountain soils of the Alay Range, Khyrgyzia. Biol Fertil Soils31(134-142):2000.Van Dam, Douwe, Edzo Veldkamp, Nico Van Breeman. Soil organic carbon dynamics:variability with depth in forested and deforested soils under pasture in Costa Rica.Biogeochemistry 39(343-375):1997. Vance, E.D. and N.M. Nadkarni. Microbial biomass and activity in canopy organicmatter and the forest floor of a tropical cloud forest. Soil Biology and Biochemistry22(677-684):1990.Veldkamp, E. and A.M. Weitz. Uncertainty analysis of "C method in soil organicmatter studies. Studies Biology Biochemistry 26,6(153-160):1994.Vitousek, P.M. Litterfall, nutrient cycling, and nutrient limitation in tropical forests.Ecology 65(285-298):1984.Vitousek, Peter M., James R. Gosz, Charles C. Grier, Jerry M. Melillo, and William A.Reiners. A Comparative Analysis of Potential Nitrification and Nitrate Mobility inForest Ecosystems. Ecological Monographs. 52, 2 (155-177):1982.Vitousek, P.M. and R.L. Sanford, Jr. Nutrient cycling in moist tropical forest. AnnualReview of Ecology and Systematics 17(137-167):1986. 13Baimas-George: Soil Nutrient Composition in Afromontane Forests of Northern Ethi Published by Digital Commons @ Colgate, 2012 !173!Appendix Site CanopySoilsInterior Forest Floor SoilsExterior