Influence of trees on soil organic matter in Mediterranean agroforestry ...
European Journal of Soil Science, June 2007, 58, 728–735
doi: 10.1111/j.1365-2389.2006.00858.x
Influence of trees on soil organic matter in Mediterranean agroforestry systems: an example from the ‘Espinal’ of central Chile C. M UN˜ OZ a , E. Z AGAL b & C. O VALLE c a
Programa de Doctorado en Ciencias de Recursos Naturales, Universidad de La Frontera, Casilla 54-D, Temuco, Departamento de Suelos, Facultad de Agronomı´a, Universidad de Concepcio´n, Casilla 537, Chilla´n, and c Instituto de Investigaciones Agropecuarias CRI Quilamapu, Casilla 426, Chilla´n, Chile b
Summary The ‘Espinal’ agroforestry system of the Mediterranean zone of central Chile, which covers an area of 2000 000 ha, is in various stages of degradation due to human activities. The objective of our study was: (i) to determine the effects of the canopy cover of Acacia caven (‘Espino’) on total soil organic carbon (SOC), soil respiration and the labile components of soil organic matter (microbial biomass, and light fraction); and (ii) to determine the influence of ecosystem degradation on total and labile components of SOC. Soils of the study area are classified as fine, mixed, active, mesic Ultic Palexeralfs, typical of the Mediterranean-type environment. We investigated sites according to the percentage coverage of A. caven canopy: (i) well-preserved Espinal (WPE), 80–51% cover; (ii) good Espinal (GE), 50–26% cover; (iii) degraded Espinal (DE), 25–11% cover; and (iv) very degraded Espinal (VDE), < 10% cover. In addition, a site under native forest (NF) was included to characterize the original state of the zone. Soil samples were taken under and outside the canopy of A. caven at two depths, 0–5 and 5–10 cm. We conclude that the microbial biomass carbon (Cmic), and total and labile components of SOC are influenced by the presence of the A. caven tree, with greater values under than outside its canopy. Under the tree canopy, to a depth of 10 cm, Cmic was less under all the agroforestry systems than in NF (46 and 30% less for WPE and GE, respectively, and 67 and 57% less for DE and VDE). However, there was no clear trend for less Cmic with increased ecosystem degradation, especially outside the canopy. However, the respiration of microbial communities was affected by ecosystem degradation for both soil depths under the tree canopy, e.g. soil respiration in VDE ecosystems was about 50% greater than that found in WPE ecosystems. Increasing the coverage of the A. caven tree in the semiarid ecosystems of central Chile, e.g. changing from VDE to WPE, would result in an eventual, long-term (over several centuries) increase in soil organic C of approximately 50%.
Introduction In arid and semiarid lands, deforestation has devastated large areas of soil, which affects the biology and ecology of the natural ecosystems and puts at risk their sustainability over time. In central Chile, the intensity of soil use by overgrazing and agricultural overexploitation has drastically depleted soil nutrients and depressed primary and secondary productivity as well as eco- and bio-diversity (Ovalle et al., 1999). The Espinal agroforestry system covers more than 2000 000 ha in the unirrigated portions of the Central Depression of Chile. This ecosystem has usually been considered as an artifact due to Correspondence: E. Zagal. E-mail: [email protected] Received 2 September 2005; revised version accepted 27 June 2006
728
several hundred years’ degradation of the indigenous sclerophyllous matorral vegetation, followed by biological invasion of Acacia caven (Fabaceae: Mimosoideae) (Ovalle et al., 1990). The original sclerophyllous vegetation (Maitenus boaria, Quillaja saponaria, Shynus polygamus, Peumus boldus and Cryptocaria alba) is still present but in small patches that are dispersed throughout the area. The resulting anthropogenic formation presents a complex and heterogeneous savanna-like structure, with herbaceous and woody strata, the latter consisting mainly of A. caven, a tree originating in the Chaco region on the eastern side of the Andean cordillera (Aronson, 1992). The herbaceous strata presents a great diversity of annual (mostly exotic, e.g. Avena barbata, Bromus mollis, Aira caryophyllea, Briza minor, Erodium # 2006 The Authors Journal compilation # 2006 British Society of Soil Science
Changes in soil C in a Mediterranean ecosystem 729
cicutarium, Erodium botrys, Hypochoeris glabra, Medicago polymorpha and Lolium multiflorum) and perennial species (mostly native, e.g. Stipa spp., Piptochaetium spp., Nassella spp. and Melica spp.) (Ovalle et al., 1990). The great majority of the exotic species were introduced by the Spaniards during the colonization period and probably were associated with hay, large herbivores and cereal seeds. Presently, there are 300 naturalized species in the Espinal agro-ecosystem (Ovalle et al., 1990). The Espinal area occupies principally the non-irrigated sectors of the central valley and the eastern slope of the Coastal Mountain range (Figure 1). These sectors consist of an arid region with 160–200 mm of annual precipitation with 8–9 dry months, and a subhumid region having up to 1000–1200 mm of precipitation with 4–5 dry months, respectively (Ovalle et al., 1999).
The Espinal area also supports most of the Chilean Mediterranean livestock industry as well as sectors of dryland cereal production (Ovalle et al., 1990; Ovalle et al., 1996) that are normally present in small paddocks (2–20 ha) within a heterogeneous landscape that is dominated by different amounts of A. caven cover. On hillsides where pasture-crop rotations are intense, the Espinal ecosystems are more degraded and there is less coverage of A. caven, greater water erosion and depletion of soil nutrients. By contrast, in flatlands the uninterrupted grazing (and the lack of cropping) has allowed for greater coverage of A. caven and better soil conservation (Ovalle et al., 1996). In the latter areas, most of the A. caven trees of selected paddocks are cut for charcoal or firewood every 40–60 years. Overall, the tree component (A. caven) in Espinal ecosystems appears to favour the formation of patches of improved vegetation under its canopy, where productive herbaceous species are concentrated (Ovalle & Godron, 1989). The labile fraction of SOM (mineralizable C, microbial biomass, light fraction and particulate organic matter) has been suggested as an early indicator of the effects of soil management on the total SOM as it responds to the changes in residue inputs, which affect the microbial biomass population and its activity (Powlson, 1994; Zagal et al., 2001). The lability of SOM is defined as the ease and speed with which it is decomposed by microbes and depends on both chemical recalcitrance (high molecular weight, irregular structure and aromatic structures) and physical protection (stabilization onto clay mineral surfaces or physical protection inside soil aggregates) (Krull et al., 2003). The influence of many land-use systems on soil organic carbon stocks is still poorly understood. Here we report the effects of agroforestry with A. caven based on research in the subhumid area of Mediterranean Chile, where despite a decade of work (e.g. Aronson et al., 2002), there still exists uncertainty about changes in soil C caused by different amounts of A. caven cover. A. caven is a member of a widespread genus in Mediterranean regions in the southern hemisphere (e.g. Chile and Australia) that is often used in agroforestry systems that are based on legume trees. Such systems are common in the Chaco region, which includes northern Argentina, central Paraguay and southern Bolivia. Furthermore, A. caven is one of the most widespread tree species of extra-tropical South America and occurs from about latitude 36° to 18°S, from the Atlantic to the Pacific oceans (Aronson & Ovalle, 1989). Therefore, our objectives were: (i) to determine the effect of the canopy cover of A. caven on total soil organic C, soil respiration and the labile component of the SOM (microbial biomass and light fraction); and (ii) to determine the effect of the degradation of this ecosystem on the content of both total soil organic C and labile C.
Methods Study area and soil sampling Figure 1 Distribution of Acacia caven in the Mediterranean zone of Chile.
Samples were collected in the Cauquenes zone, which is in the subhumid Mediterranean portion of the VIIth region of Chile
# 2006 The Authors Journal compilation # 2006 British Society of Soil Science, European Journal of Soil Science, 58, 728–735
730 C. Mun˜oz et al. (latitude 35°58¢S, longitude 72°17¢S, Figure 1). The altitude is 170–180 m above sea level, and the average annual precipitation is 695 mm with 5 months of dry season in the summer. The Cauquenes soil (fine, mixed, active, mesic Ultic Palexeralf) is formed in situ from strongly weathered granitic rocks and has clay textures throughout the profile. The effective depth of the soil varies with the degree of erosion, which ranges from light to very severe. The sample sites were classified according to the percentage of A. caven canopy cover: (i) well-preserved espinal (WPE), 80–51% cover; (ii) good espinal (GE), 50–26% cover; (iii) degraded espinal (DE), 25–11% cover; and (iv) very degraded espinal (VDE), < 10% cover; with 909 154, 604 201, 375 130, 308 51 trees ha–1, respectively. On average, the trees were 30–60 years old and 2.7 m tall. Sampling was done in August and September 2003. In addition, a site under native forest (NF) was included to characterize the original state of the zone. Livestock in the area is mainly ovine, with an average stocking rate of two sheep ha–1 year–1. There also are some areas dedicated to bovine production. Four representative subsites were sampled in each the five ecosystems that were located within a total area of 23 000 ha. At each subsite, replicated soil samples were taken by selecting two trees, and at each tree a soil sample was collected beneath the tree canopy at a distance equal to approximately half of its canopy ratio (mean canopy diameter ¼ 2.2 m), and another collected outside the influence of the tree canopy (at least 1 m to 1.5 m outside). All soil samples from each site were taken at two depths (0–5 and 5–10 cm). Undisturbed soil samples also were collected by the cylinder method to determine bulk density. Bulk density of the undisturbed soil samples was expressed as mass per unit volume. As these soils were free of stones, no correction for rocks fragments was necessary. In the laboratory, the sampled soil was wet sieved ( 2 mm), the roots discarded, and stored at 4°C.
chloroform for 24 hours, and extracted with 100 ml of 0.5 M K2SO4 solution, which was subsequently filtered through a Whatman N°42 filter paper. The extracts were analysed by the N-ninhydrin method (Joergensen & Brookes, 1990) and a factor of 31 used to determine Cmic (Ocio & Brookes, 1990).
Soil respiration Briefly, soil respiration was measured from the CO2 evolved from 25 g soil (dry weight basis) that was incubated at 25°C for 3 days in 1-litre closed jars. The CO2 was trapped in 7.5 ml of 1 M NaOH, which was then titrated with 0.1 M HCl (Alef, 1995).
Soil organic carbon Soil organic carbon was determined with acid dichromate digestion (Sims & Haby, 1971) and presented as t C per 103 t soil ha–1 (Ellert et al., 2002) for the two soil depths. We called this quantity ‘C storage’.
Size fractionation of soil organic matter Size fractionation of soil organic matter followed the procedure of Balesdent et al. (1991), in which the soil is first dispersed mechanically and then separated by wet sieving. Briefly, 50 g air-dried soil was placed in a plastic bottle (250 ml) with 10 small glass beads (6 mm diameter) and 180 ml of distilled water, and shaken (REAX 2, Heidolph Instruments, Schwabach, Germany) at 50 cycles minute–1 for 16 hours. The disrupted soil aggregates were then wet sieved using a stainless steel sieve (212 mm aperture). The light fraction (LF) and retained sand (both 212–2000 mm) were separated by flotation and sedimentation in distilled water.
Statistics Soil preparation We determined soil water content by oven drying at 105°C. Before the start of the incubation experiment for soil respiration or soil microbial biomass determinations all soil samples were conditioned to a gravimetric water content corresponding to 50% water-filled pore space (WFPS) at the measured bulk density (Linn & Doran, 1984).
%WFPS ¼
Soil water content oven-dry bulk density 100 1 ðbulk density=2:65Þ ð1Þ
Soil microbial biomass carbon Soil microbial biomass carbon (Cmic) was estimated by the fumigation-extraction method (Vance et al., 1987). Briefly, 25 g soil (dry weight basis) was fumigated with ethanol-free
Statistical analysis was performed using the bulk density, C storage, Cmic and microbial respiration data, at two depths, for soils under and outside the canopy. One-way ANOVA was made to test for differences between Espinal systems, and Pearson correlation coefficients were determined to explore relationships between Cmic, soil respiration, C storage and the light fraction of soil organic matter.
Results and discussion Bulk density Slightly smaller bulk density of the soils was observed with increased coverage of A. caven, particularly in the 0–5 cm layer under the tree canopy (Table 1). Greater differences were found at the two depths studied when comparing soils from under and outside the canopy, whereby smaller bulk density was found under the trees. This is in agreement with
# 2006 The Authors Journal compilation # 2006 British Society of Soil Science, European Journal of Soil Science, 58, 728–735
Changes in soil C in a Mediterranean ecosystem 731 Table 1 Bulk density (g cm–3) of the ecosystems studied. Mean standard error (n ¼ 8)
Ecosystem NF WPE GE DE VDE
Outside canopy
0–5 cm
5–10 cm
0–5 cm
5–10 cm
1.17 0.09 1.19 0.05 1.27 0.06 1.34 0.05 1.40 0.08
1.48 0.08 1.44 0.05 1.44 0.05 1.50 0.03 1.45 0.05
1.55 0.08 1.34 0.06 1.38 0.05 1.43 0.04 1.43 0.07
1.65 0.05 1.55 0.04 1.55 0.05 1.55 0.03 1.53 0.03
a)
0
20
40
60
80
100
80
100
0-5
Depth /cm
Under canopy
SOC / t per 103 t soil ha–1
NF ¼ native forest; WPE ¼ well preserved Espinal; GE ¼ good Espinal; DE ¼ degraded Espinal; VDE ¼ very degraded Espinal.
5-10
results reported by other authors (Mordelet et al., 1993; Rhoades, 1997).
Organic carbon
SOC / t per 103 t soil ha–1
Microbial biomass carbon The values of microbial biomass carbon (Cmic) were in the range of 366–277 mg C kg–1 outside the canopy and 1105–357 mg
0
b)
20
40
60
0-5
Depth /cm
The total SOC (i.e. ‘C storage’, Figure 2) in the 0–5 cm layer (for both conditions under and outside the canopy) was 50% greater in the WPE soil than the VDE soil. Similar results were obtained when SOC contents were compared between the NF and the WPE and GE ecosystems. Based on earlier investigations (Ovalle et al., 2006), the tree cover would increase by 5% every 10 years (meaning that changing from VDE to WPE would require more than 100 years). In general, SOC values are smaller outside the canopy than under it, with the exception of the two most degraded classes of vegetation (DE and VDE; Figure 2) in the 5–10 cm layer. For example, under natural forest (NF) in the 0–5 cm layer, there was 73% reduction in SOC content outside the canopy compared with under the canopy. Similarly, in the other agroforestry systems at the same depth there was between 28 and 40% less SOC outside the canopy compared with under it. The SOC in all the degraded systems (at both 0–5 and 5–10 cm depths) was significantly less ‘under canopy’ than under canopy in NF, with the smallest values measured in DE and VDE (Figure 2a). For both depths, there was 50% less SOC for WPE and GE and 73% less for DE and VDE compared with NF. Outside the canopy, the SOC in the two most degraded systems (DE and VDE) was significantly less than in natural forest (37–40% reduction of SOC in the 0–5 and 5–10 cm depths, respectively), though this was not the case in the two less degraded systems (WPE and GE). Small differences in SOM (total and fractions) between the different tree densities of the agroforestry system (e.g. between WPE and GE, and DE and VDE) were partly explained by different inputs of plant residues but also by the presence of grazing animals, as they tend to move nutrients through their excreta, thus blurring distinctions between areas.
5-10
Figure 2 Soil organic carbon (t per 103 t soil ha–1) with depth (a) under canopy and (b) outside canopy. NF ¼ native forest; WPE ¼ well preserved Espinal; GE ¼ good Espinal; DE ¼ degraded Espinal; VDE ¼ very degraded Espinal. Error bars indicate one standard error of the mean (n ¼ 8). ¼ NF; ¼ WPE; ¼ GE, ¼ DE; ¼ VDE.
C kg–1 under the canopy, when converted to an area basis (Figure 3). These results were in agreement with those of Zagal et al. (2003), who found values of 350–286 mg C kg–1 for more intensive rotation systems and 932–981 mg C kg–1 for less intensive rotation systems in a silty to silty-loam soil. The Cmic was less (in most cases significantly so) under the tree canopy at both soil depths in all the agroforestry systems compared with NF (46–30% less for WPE and GE, respectively; similarly, 67–57% less for DE and VDE) (Figure 3). However, the differences among the degraded systems were mostly not significantly different, nor was there a clear trend for less Cmic with increasing degree of ecosystem degradation. However, WPE and GE had greater amounts of Cmic than DE and VDE at 0–5 cm depth (the differences were about 45% on
# 2006 The Authors Journal compilation # 2006 British Society of Soil Science, European Journal of Soil Science, 58, 728–735
732 C. Mun˜oz et al.
Figure 3 Microbial biomass carbon (Cmic) (mg C kg–1) in ecosystems studied: (a) under canopy and depth 0–5 cm; (b) under canopy and depth 5–10 cm; (c) outside canopy and depth 0–5 cm; (d) outside canopy and depth 5–10 cm. Error bars indicate one standard error of the mean (n ¼ 8). Ecosystem description as in Figure 2.
average). Outside the canopy there were no significant differences in Cmic between any of the sites, including NF. Analysis of the data (Table 2) showed that Cmic was not consistently correlated with SOC. A close correlation between Cmic and SOC is observed only for the 5–10 cm layer under the canopy and for the 0–5 cm soil outside the canopy, with correlation coefficients of 0.82 (P 0.01). These results were in agreement with Wirth (2001) who found a correlation of 0.766 (P 0.001) between Cmic and SOC. However, a tight correlation between these soil parameters should be expected only for soils where SOC and its constituent fractions are close to equilibrium. Where a change in land use or management has Table 2 Pearson product-moment correlation coefficient between Cmic, SOC, LF and microbial respiration; statistical significance shown: **P 0.01, *P 0.05 Measurement Depth/cm Cmic SOC LF Cmic SOC LF Cmic SOC LF Cmic SOC LF a
0–5 0–5 0–5 5–10 5–10 5–10 0–5 0–5 0–5 5–10 5–10 5–10
Soil organic carbon. Light fraction.
b
Condition Under canopy Under canopy Under canopy Under canopy Under canopy Under canopy Outside canopy Outside canopy Outside canopy Outside canopy Outside canopy Outside canopy
SOCa 0.29
LFb
Respiration
0.15 0.90**
0.53* 0.53** 0.68** 0.22 0.07 0.61* 0.62** 0.63** 0.81** 0.18 0.57** 0.64**
0.82** 0.61* 0.49* 0.35*
0.43 0.81**
0.31
0.85** 0.33
occurred, as in this study, the close relationships would be expected to break down. Expressing Cmic as a percentage of SOC for the 0–5 cm layer of the end-member systems, NF and VDE, gave values of 1.37% and 1.96%, respectively, which suggested that SOC declined proportionally more than Cmic, which was unexpected. It may be that the input of plant residues in the VDE system (or more degraded systems) was mainly from pasture and therefore more readily decomposable than tree residues. Additionally, above-ground tree production in the VDE system is poor, e.g. 1 t ha–1 year–1, which consequently implies a low return of tree residues to the soil (Ovalle et al., 1999). For example, decomposition studies in situ (420 days) of Lolium multiflorum straw and A. caven needles in degraded ecosystems (e.g. DE, litter bags placed outside the canopy) showed that there was greater mass lost for L. multiflorum straw (about 6%) than for A. caven needles (data not shown). On the other hand, the chemical composition of A. caven needles was richer in lignin and N content (26 and 1.83%, respectively) as compared with L. multiflorum straw (7 and 0.41%, respectively).
Soil respiration We found that the respiration rate was significantly greater under the canopy at 0–5 cm in NF with 82 16 (SE) mg C-CO2 g–1 soil, followed by VDE, GE, DE and WPE with 61 8 (SE), 51 10 (SE), 42 7 (SE) and 31 5 (SE) mg C-CO2 g–1 soil, respectively (Figure 4a). In the 5–10 cm layer (Figure 4b), there is no clear difference between ecosystems. Outside the canopy (Figure 4c), there were no significant differences between ecosystems, with a range of 32–45 mg C-CO2 g–1 for 0–5 cm depth,
# 2006 The Authors Journal compilation # 2006 British Society of Soil Science, European Journal of Soil Science, 58, 728–735
Changes in soil C in a Mediterranean ecosystem 733
Figure 4 Microbial respiration (mg C-CO2 g–1 dry soil) in ecosystems studied: (a) under canopy and depth 0–5 cm; (b) under canopy and depth 5–10 cm; (c) outside canopy and depth 0–5 cm; (d) outside canopy and depth 5–10 cm. Error bars indicate one standard error of the mean (n ¼ 8) Ecosystem description as in Figure 2.
while at 5–10 cm depth (Figure 4d) only NF had a significantly greater value with 59 8 (SE) mg C-CO2 g–1, while the mean respiration of the other ecosystems was 16 mg C-CO2 g–1. Correlation analysis (Table 2) indicated that respiration was not consistently related to either SOC or microbial biomass, but was related to the amount of light fraction (LF) in most cases. This could reflect microbial breakdown of a readily decomposable fraction of the plant residues input to soil such as sugars and amino acids. A tight correlation between the amount of LF and respiration rate was also found by Janzen et al. (1992), who reported a correlation coefficient of 0.81 between these soil parameters. Such a relationship could be expected, as the LF of organic matter comprises mainly freshly added organic material from plant debris and manure (Christensen, 1992).
Light fraction The strong correlation between the LF content of soils and soil respiration rates (Table 2), suggests that in these soils LF is an important carbon and energy source for soil microorganisms. This is possible because LF is not protected by clay minerals or other mechanisms, and it is readily accessible to microbial and enzyme activity (Skjemstad et al., 1986; Six et al., 2002). Skjemstad et al. (1986) proposed that the LF is more sensitive than total organic matter content to the effects of cropping practices. In other investigations, LF showed seasonal fluctuations (Boone, 1994; Campbell et al., 1999) in both carbon and nitrogen contents, which were associated with changes in soil moisture, temperature and rainfall. Boone (1994) concluded that the seasonality of new inputs of organic matter to the soil was the main factor affecting the amount of LF extracted from soils under maize. Other investigations have indicated that the LF content of soil organic matter is an important property for determining ecosystem quality and/or stress. For example, Janzen et al. (1992)
found that the LF content increased 2.6-fold with increasing organic matter content, whereas SOC content increased only 1.3-fold. Also, Zagal et al. (2001) found a decrease in LF content as the intensity of soil use increased. We found in this study that with greater intensity of soil use the LF diminishes 3.4-fold with decreases of 1.61-fold of SOC (data not shown) for conditions outside the canopy. This further demonstrates that the LF content is a more sensitive parameter than SOC content to changes in the soil due to land-use intensity. Components of SOM that provide substrate that is physically or chemically more accessible to the microbial biomass are fractions of the total SOM that display proportionally larger changes over the short term and may therefore be useful as indicators. The light fraction has been previously shown to change more rapidly and act as an ‘early warning’ of change in total SOM content, long before these can be measured with statistical certainty (Christensen, 2001). Overall, the microbial biomass, LF content and SOC values were greater under canopy conditions, for both depths. These results might indicate that the soil under the influence of A. caven offers propitious conditions for the development of the microbial populations. However, we found that in Espinal ecosystems with less A. caven cover the microbial activity (e.g. soil respiration) was greater than in ecosystems with more tree cover. For example, values for respiration for the 0–5 cm layer, expressed as CO2-C evolved (mg) per unit of microbial biomass C (g) for NF, GE and VDE, were 120, 76 and 172, respectively. As microbial activity is the result of a combination of the quantity and decomposability of organic input into the soils of the different systems, more trees in GE (604 201 tree ha–1) produced significant inputs that were derived from tree residues (e.g. A. caven needles) and, alternatively, in VDE (308 51 tree ha–1) the inputs were derived of pasture plant species, which were readily decomposed by the microorganisms.
# 2006 The Authors Journal compilation # 2006 British Society of Soil Science, European Journal of Soil Science, 58, 728–735
734 C. Mun˜oz et al. The results regarding the role of A. caven tree cover on the levels of soil carbon in the ‘Espinal’ system of the Mediterranean climate zone of Chile, are in agreement with studies that have been obtained in other agroforestry systems in Mediterranean areas of the world, such as the ‘Dehesas’ of southern Spain and Portugal (Joffre et al., 1999), the savannas of Faidherbia albida of Sudan (Van Noordwijk & Ong, 1999) and the fynbos ecosystems in South Africa (Cocks & Stock, 2001). In these land-use systems, the trees carry out a fundamental function in nutrient and water cycling, and in this regard, these agroforestry systems could be considered similar to the natural ecosystems of these regions (Ewel, 1999; Van Noordwijk & Ong, 1999), thereby conserving an important part of the original function and diversity of the native ecosystem. On the other hand, the results imply that not only is it possible to improve the productivity and sustainability of the Chilean ‘Espinales’, but it could additionally be used as a model for other Mediterranean agroecosystems, such as in Australia where conservation and the management of trees in these systems have not received high priority. In Australia, the mixed original vegetation of woody and herbaceous plants has been replaced by more simplified systems of herbaceous crops and pasture, that maximize the short-term primary productivity, but in the long term have resulted in a loss of sustainability due to soil nutrient depletion and salinity (Ridley & Pannell, 2005).
Conclusions 1 Microbial biomass, respiration rate, light fraction and SOC are influenced by the presence of the A. caven tree, with values being greater under its canopy than outside its canopy. 2 Under the tree canopy, to a depth of 10 cm, microbial biomass carbon (Cmic) was less under all the agroforestry systems than in NF. However, there was no clear trend for less Cmic with increased ecosystem degradation, especially in the soil from outside the canopy. 3 The respiration of microbial communities was affected by the stage of degradation of the ecosystems, and was related to the relative proportions and decomposability of plant inputs (pasture versus tree residues) in the different systems. 4 Microbial respiration is correlated with LF content, indicating that this provides a substrate that is physically or chemically more accessible to the microbial biomass. 5 Increasing the cover of the A. caven tree in the current semiarid ecosystems of central-south Mediterranean Chile would result in an eventual long-term (over several centuries) increase in soil organic C of approximately 50%.
Acknowledgements This research was made possible thanks to the financing granted by FONDECYT (Project No. 1030883) and the Andes Foundation (No. C-13755 28). We also thank the personnel of the
Department of Soils of the Universidad de Concepcio´n (Chile) for assistance with soil analysis.
References Alef, K. 1995. Estimation of soil respiration. In: Methods in Applied Soil Microbiology and Biochemistry (eds K. Alef & P. Nannipieri), pp. 464–467. Academic Press, London. Aronson, J. 1992. Evolutionary biology of Acacia caven (Leguminosae, Mimosoideae): infraspecific variation in fruit and seed characters. Annals of the Missouri Botanical Garden, 79, 958–968. Aronson, J. & Ovalle, C. 1989. Study of the natural variability, biogeography and potential for genetic improvement of Acacia caven. Bulletin of the International Group for the Study of Mimosideae, 17, 111–121. Aronson, J., Ovalle, C., Avendan˜o, J., Longeri, L. & Del Pozo, A. 2002. Agroforestry tree selection in central Chile: biological nitrogen fixation and early plant growth in six dryland species. Agroforestry Systems, 56, 155–166. Balesdent, J., Pe´traud, J. & Feller, C. 1991. Effets des ultrasons sur la distribution granulome´trique des matie`res organiques des sols. Science du Sol, 29, 95–106. Boone, R.D. 1994. Light-fraction soil organic matter: origin and contribution to net nitrogen mineralization. Soil Biology and Biochemistry, 26, 1459–1468. Campbell, C.A., Biederbeck, V.O., Wen, G., Zentner, R.P., Schoenau, J. & Hahn, D. 1999. Seasonal trend in selected soil biochemical attributes: effects of crop rotation in the semiarid prairie. Canadian Journal of Soil Science, 79, 73–84. Christensen, B.T. 1992. Physical fractionation of soil and organic matter in primary particle size and density separates. Advances in Soil Science, 20, 1–90. Christensen, B.T. 2001. Physical fractionation of soil and structural and functional complexity in organic matter turnover. European Journal of Soil Science, 52, 345–353. Cocks, M.P. & Stock, W.D. 2001. Field patterns of nodulation in fifteen Aspalathus species and their ecological role in fynbos vegetation of southern Africa. Basic and Applied Ecology, 2, 115–125. Ellert, B.H., Janzen, H.H. & Entz, T. 2002. Assessment of a method to measure temporal change in soil carbon storage. Soil Science Society of America Journal, 66, 1687–1695. Ewel, J. 1999. Natural systems as model for the design of sustainable systems of land use. Agroforestry Systems, 45, 1–21. Janzen, H.H., Campbell, C.A., Brandt, S.A., Lafond, G.P. & TownleySmith, L. 1992. Light fraction organic matter in soils from long term crop rotations. Soil Science Society of America Journal, 56, 1799–1806. Joergensen, R.G. & Brookes, P.C. 1990. Ninhydrin-reactive nitrogen measurements of microbial biomass in 0.5 M K2SO4 soil extracts. Soil Biology and Biochemistry, 22, 1023–1027. Joffre, R., Rambal, S. & Ratte, J. 1999. The Dehesa system of southern Spain and Portugal as a natural ecosystem mimic. Agroforestry Systems, 45, 57–79. Krull, E.S., Baldock, J.A. & Skjemstad, J.O. 2003. Importance of mechanism and processes of the stabilization of soil organic matter for modelling carbon turnover. Functional Plant Biology, 30, 207–222. Linn, D.M. & Doran, J.W. 1984. Effect of water-filled pore space on carbon dioxide and nitrous oxide production in tilled and nontilled soils. Soil Science Society of America Journal, 48, 1267–1272.
# 2006 The Authors Journal compilation # 2006 British Society of Soil Science, European Journal of Soil Science, 58, 728–735
Changes in soil C in a Mediterranean ecosystem 735
Mordelet, P., Abbadie, L. & Menaut, J.C. 1993. Effects of tree clumps on soil characteristics in a humid savanna of West Africa. Plant and Soil, 153, 103–111. Ocio, J.A. & Brookes, P.C. 1990. An evaluation of methods for measuring the microbial biomass in soils following recent additions of wheat straw and the characterization of the biomass that develops. Soil Biology and Biochemistry, 22, 685–694. Ovalle, C., Aronson, J., Del Pozo, A. & Avendan˜o, J. 1990. The espinal: agroforestry systems of the Mediterranean-type climate region of Chile. Agroforestry Systems, 10, 213–239. Ovalle, C., Aronson, J., Del Pozo, A. & Avendan˜o, J. 1999. Restoration and rehabilitation of mixed espinales of central Chile. Arid Soil Research and Rehabilitation, 13, 369–381. Ovalle, C., Del Pozo, A., Casado, M.A., Acosta, B. & de Miguel, J.M. 2006. Consequences of landscape heterogeneity on grassland diversity and productivity in the Espinal agroforestry system of central Chile. Landscape Ecology, 21, 585–594. Ovalle, C. & Godron, M. 1989. Influencia del a´rbol sobre la vegetacio´n herba´cea en matorrales de Acacia caven. Rol del animal (ovinos) en las interacciones a´rbol-pasto. Stvdia Œcologica, 6, 225–243. Ovalle, C., Longeri, L., Aronson, J., Herrera, A. & Avendan˜o, J. 1996. N2-fixation, nodule efficiency and biomass accumulation after two years in three Chilean legume trees and Tagasaste (Chamaecytisus proliferus subsp. palmensis). Plant and Soil, 179, 131–140. Powlson, D.S. 1994. The soil microbial biomass: before, beyond and back. In: Beyond the Biomass. Compositional and Functional Analysis of Soil Microbial Communities (eds K. Ritz, J. Dighton & K.E. Giller), pp. 3–20. John Wiley and Sons, Chichester. Rhoades, C.C. 1997. Single-tree influences on soil properties in agroforestry: lessons from natural forest and savanna ecosystems. Agroforestry Systems, 35, 71–94.
Ridley, A.M. & Pannell, D.J. 2005. The role of plants and plantbased research and development in managing dryland salinity in Australia. Australian Journal of Experimental Agriculture, 45, 1341– 1355. Sims, J.R. & Haby, V.A. 1971. Simplified colorimetric determination of soil organic matter. Soil Science, 112, 137–141. Six, J., Conant, R.T., Paul, E.A. & Paustian, K. 2002. Stabilization mechanisms of soil organic matter: implications for C-saturation of soils. Plant and Soil, 241, 155–176. Skjemstad, J.O., Dalal, R.C. & Barron, P.F. 1986. Spectroscopic investigations of cultivation effects on organic matter of Vertisols. Soil Science Society of America Journal, 50, 354–359. Van Noordwijk, M. & Ong, C.K. 1999. Can the ecosystem mimic hypotheses be applied to farms in African savannahs? Agroforestry Systems, 45, 131–158. Vance, E.D., Brookes, P.C. & Jenkinson, D.S. 1987. An extraction method for measuring soil microbial biomass C. Soil Biology and Biochemistry, 19, 703–707. Wirth, S.J. 2001. Regional-scale analysis of soil microbial biomass and soil basal CO2-respiration in northeastern Germany. In: Sustaining the Global Farm – 10th International Soil Conservation Organization Meeting (eds D.E. Stott, R.H. Mohtar & G.C. Steinhardt), pp. 486–493. West Lafayette, IN, USA. ISCO-USDA-Purdue University. Zagal, E., Rodrı´ guez, N., Vidal, I. & Flores, B. 2001. La fraccio´n liviana de la materia orga´nica de un suelo volca´nico bajo distinto manejo agrono´mico como ı´ ndice de cambios de la materia orga´nica. Agricultura Te´cnica (Chile), 62, 284–295. Zagal, E., Rodrı´ guez, N., Vidal, I. & Maureira, M. 2003. Biomasa microbiana por los me´todos reactivo nitro´geno ninhydrina y absorbancia ultravioleta (UV280) en distintos sistemas de cultivo. Revista de la Ciencia Del Suelo Y Nutricio´n Vegetal (Chile), 3, 17–28.
# 2006 The Authors Journal compilation # 2006 British Society of Soil Science, European Journal of Soil Science, 58, 728–735
doi: 10.1111/j.1365-2389.2006.00858.x
Influence of trees on soil organic matter in Mediterranean agroforestry systems: an example from the ‘Espinal’ of central Chile C. M UN˜ OZ a , E. Z AGAL b & C. O VALLE c a
Programa de Doctorado en Ciencias de Recursos Naturales, Universidad de La Frontera, Casilla 54-D, Temuco, Departamento de Suelos, Facultad de Agronomı´a, Universidad de Concepcio´n, Casilla 537, Chilla´n, and c Instituto de Investigaciones Agropecuarias CRI Quilamapu, Casilla 426, Chilla´n, Chile b
Summary The ‘Espinal’ agroforestry system of the Mediterranean zone of central Chile, which covers an area of 2000 000 ha, is in various stages of degradation due to human activities. The objective of our study was: (i) to determine the effects of the canopy cover of Acacia caven (‘Espino’) on total soil organic carbon (SOC), soil respiration and the labile components of soil organic matter (microbial biomass, and light fraction); and (ii) to determine the influence of ecosystem degradation on total and labile components of SOC. Soils of the study area are classified as fine, mixed, active, mesic Ultic Palexeralfs, typical of the Mediterranean-type environment. We investigated sites according to the percentage coverage of A. caven canopy: (i) well-preserved Espinal (WPE), 80–51% cover; (ii) good Espinal (GE), 50–26% cover; (iii) degraded Espinal (DE), 25–11% cover; and (iv) very degraded Espinal (VDE), < 10% cover. In addition, a site under native forest (NF) was included to characterize the original state of the zone. Soil samples were taken under and outside the canopy of A. caven at two depths, 0–5 and 5–10 cm. We conclude that the microbial biomass carbon (Cmic), and total and labile components of SOC are influenced by the presence of the A. caven tree, with greater values under than outside its canopy. Under the tree canopy, to a depth of 10 cm, Cmic was less under all the agroforestry systems than in NF (46 and 30% less for WPE and GE, respectively, and 67 and 57% less for DE and VDE). However, there was no clear trend for less Cmic with increased ecosystem degradation, especially outside the canopy. However, the respiration of microbial communities was affected by ecosystem degradation for both soil depths under the tree canopy, e.g. soil respiration in VDE ecosystems was about 50% greater than that found in WPE ecosystems. Increasing the coverage of the A. caven tree in the semiarid ecosystems of central Chile, e.g. changing from VDE to WPE, would result in an eventual, long-term (over several centuries) increase in soil organic C of approximately 50%.
Introduction In arid and semiarid lands, deforestation has devastated large areas of soil, which affects the biology and ecology of the natural ecosystems and puts at risk their sustainability over time. In central Chile, the intensity of soil use by overgrazing and agricultural overexploitation has drastically depleted soil nutrients and depressed primary and secondary productivity as well as eco- and bio-diversity (Ovalle et al., 1999). The Espinal agroforestry system covers more than 2000 000 ha in the unirrigated portions of the Central Depression of Chile. This ecosystem has usually been considered as an artifact due to Correspondence: E. Zagal. E-mail: [email protected] Received 2 September 2005; revised version accepted 27 June 2006
728
several hundred years’ degradation of the indigenous sclerophyllous matorral vegetation, followed by biological invasion of Acacia caven (Fabaceae: Mimosoideae) (Ovalle et al., 1990). The original sclerophyllous vegetation (Maitenus boaria, Quillaja saponaria, Shynus polygamus, Peumus boldus and Cryptocaria alba) is still present but in small patches that are dispersed throughout the area. The resulting anthropogenic formation presents a complex and heterogeneous savanna-like structure, with herbaceous and woody strata, the latter consisting mainly of A. caven, a tree originating in the Chaco region on the eastern side of the Andean cordillera (Aronson, 1992). The herbaceous strata presents a great diversity of annual (mostly exotic, e.g. Avena barbata, Bromus mollis, Aira caryophyllea, Briza minor, Erodium # 2006 The Authors Journal compilation # 2006 British Society of Soil Science
Changes in soil C in a Mediterranean ecosystem 729
cicutarium, Erodium botrys, Hypochoeris glabra, Medicago polymorpha and Lolium multiflorum) and perennial species (mostly native, e.g. Stipa spp., Piptochaetium spp., Nassella spp. and Melica spp.) (Ovalle et al., 1990). The great majority of the exotic species were introduced by the Spaniards during the colonization period and probably were associated with hay, large herbivores and cereal seeds. Presently, there are 300 naturalized species in the Espinal agro-ecosystem (Ovalle et al., 1990). The Espinal area occupies principally the non-irrigated sectors of the central valley and the eastern slope of the Coastal Mountain range (Figure 1). These sectors consist of an arid region with 160–200 mm of annual precipitation with 8–9 dry months, and a subhumid region having up to 1000–1200 mm of precipitation with 4–5 dry months, respectively (Ovalle et al., 1999).
The Espinal area also supports most of the Chilean Mediterranean livestock industry as well as sectors of dryland cereal production (Ovalle et al., 1990; Ovalle et al., 1996) that are normally present in small paddocks (2–20 ha) within a heterogeneous landscape that is dominated by different amounts of A. caven cover. On hillsides where pasture-crop rotations are intense, the Espinal ecosystems are more degraded and there is less coverage of A. caven, greater water erosion and depletion of soil nutrients. By contrast, in flatlands the uninterrupted grazing (and the lack of cropping) has allowed for greater coverage of A. caven and better soil conservation (Ovalle et al., 1996). In the latter areas, most of the A. caven trees of selected paddocks are cut for charcoal or firewood every 40–60 years. Overall, the tree component (A. caven) in Espinal ecosystems appears to favour the formation of patches of improved vegetation under its canopy, where productive herbaceous species are concentrated (Ovalle & Godron, 1989). The labile fraction of SOM (mineralizable C, microbial biomass, light fraction and particulate organic matter) has been suggested as an early indicator of the effects of soil management on the total SOM as it responds to the changes in residue inputs, which affect the microbial biomass population and its activity (Powlson, 1994; Zagal et al., 2001). The lability of SOM is defined as the ease and speed with which it is decomposed by microbes and depends on both chemical recalcitrance (high molecular weight, irregular structure and aromatic structures) and physical protection (stabilization onto clay mineral surfaces or physical protection inside soil aggregates) (Krull et al., 2003). The influence of many land-use systems on soil organic carbon stocks is still poorly understood. Here we report the effects of agroforestry with A. caven based on research in the subhumid area of Mediterranean Chile, where despite a decade of work (e.g. Aronson et al., 2002), there still exists uncertainty about changes in soil C caused by different amounts of A. caven cover. A. caven is a member of a widespread genus in Mediterranean regions in the southern hemisphere (e.g. Chile and Australia) that is often used in agroforestry systems that are based on legume trees. Such systems are common in the Chaco region, which includes northern Argentina, central Paraguay and southern Bolivia. Furthermore, A. caven is one of the most widespread tree species of extra-tropical South America and occurs from about latitude 36° to 18°S, from the Atlantic to the Pacific oceans (Aronson & Ovalle, 1989). Therefore, our objectives were: (i) to determine the effect of the canopy cover of A. caven on total soil organic C, soil respiration and the labile component of the SOM (microbial biomass and light fraction); and (ii) to determine the effect of the degradation of this ecosystem on the content of both total soil organic C and labile C.
Methods Study area and soil sampling Figure 1 Distribution of Acacia caven in the Mediterranean zone of Chile.
Samples were collected in the Cauquenes zone, which is in the subhumid Mediterranean portion of the VIIth region of Chile
# 2006 The Authors Journal compilation # 2006 British Society of Soil Science, European Journal of Soil Science, 58, 728–735
730 C. Mun˜oz et al. (latitude 35°58¢S, longitude 72°17¢S, Figure 1). The altitude is 170–180 m above sea level, and the average annual precipitation is 695 mm with 5 months of dry season in the summer. The Cauquenes soil (fine, mixed, active, mesic Ultic Palexeralf) is formed in situ from strongly weathered granitic rocks and has clay textures throughout the profile. The effective depth of the soil varies with the degree of erosion, which ranges from light to very severe. The sample sites were classified according to the percentage of A. caven canopy cover: (i) well-preserved espinal (WPE), 80–51% cover; (ii) good espinal (GE), 50–26% cover; (iii) degraded espinal (DE), 25–11% cover; and (iv) very degraded espinal (VDE), < 10% cover; with 909 154, 604 201, 375 130, 308 51 trees ha–1, respectively. On average, the trees were 30–60 years old and 2.7 m tall. Sampling was done in August and September 2003. In addition, a site under native forest (NF) was included to characterize the original state of the zone. Livestock in the area is mainly ovine, with an average stocking rate of two sheep ha–1 year–1. There also are some areas dedicated to bovine production. Four representative subsites were sampled in each the five ecosystems that were located within a total area of 23 000 ha. At each subsite, replicated soil samples were taken by selecting two trees, and at each tree a soil sample was collected beneath the tree canopy at a distance equal to approximately half of its canopy ratio (mean canopy diameter ¼ 2.2 m), and another collected outside the influence of the tree canopy (at least 1 m to 1.5 m outside). All soil samples from each site were taken at two depths (0–5 and 5–10 cm). Undisturbed soil samples also were collected by the cylinder method to determine bulk density. Bulk density of the undisturbed soil samples was expressed as mass per unit volume. As these soils were free of stones, no correction for rocks fragments was necessary. In the laboratory, the sampled soil was wet sieved ( 2 mm), the roots discarded, and stored at 4°C.
chloroform for 24 hours, and extracted with 100 ml of 0.5 M K2SO4 solution, which was subsequently filtered through a Whatman N°42 filter paper. The extracts were analysed by the N-ninhydrin method (Joergensen & Brookes, 1990) and a factor of 31 used to determine Cmic (Ocio & Brookes, 1990).
Soil respiration Briefly, soil respiration was measured from the CO2 evolved from 25 g soil (dry weight basis) that was incubated at 25°C for 3 days in 1-litre closed jars. The CO2 was trapped in 7.5 ml of 1 M NaOH, which was then titrated with 0.1 M HCl (Alef, 1995).
Soil organic carbon Soil organic carbon was determined with acid dichromate digestion (Sims & Haby, 1971) and presented as t C per 103 t soil ha–1 (Ellert et al., 2002) for the two soil depths. We called this quantity ‘C storage’.
Size fractionation of soil organic matter Size fractionation of soil organic matter followed the procedure of Balesdent et al. (1991), in which the soil is first dispersed mechanically and then separated by wet sieving. Briefly, 50 g air-dried soil was placed in a plastic bottle (250 ml) with 10 small glass beads (6 mm diameter) and 180 ml of distilled water, and shaken (REAX 2, Heidolph Instruments, Schwabach, Germany) at 50 cycles minute–1 for 16 hours. The disrupted soil aggregates were then wet sieved using a stainless steel sieve (212 mm aperture). The light fraction (LF) and retained sand (both 212–2000 mm) were separated by flotation and sedimentation in distilled water.
Statistics Soil preparation We determined soil water content by oven drying at 105°C. Before the start of the incubation experiment for soil respiration or soil microbial biomass determinations all soil samples were conditioned to a gravimetric water content corresponding to 50% water-filled pore space (WFPS) at the measured bulk density (Linn & Doran, 1984).
%WFPS ¼
Soil water content oven-dry bulk density 100 1 ðbulk density=2:65Þ ð1Þ
Soil microbial biomass carbon Soil microbial biomass carbon (Cmic) was estimated by the fumigation-extraction method (Vance et al., 1987). Briefly, 25 g soil (dry weight basis) was fumigated with ethanol-free
Statistical analysis was performed using the bulk density, C storage, Cmic and microbial respiration data, at two depths, for soils under and outside the canopy. One-way ANOVA was made to test for differences between Espinal systems, and Pearson correlation coefficients were determined to explore relationships between Cmic, soil respiration, C storage and the light fraction of soil organic matter.
Results and discussion Bulk density Slightly smaller bulk density of the soils was observed with increased coverage of A. caven, particularly in the 0–5 cm layer under the tree canopy (Table 1). Greater differences were found at the two depths studied when comparing soils from under and outside the canopy, whereby smaller bulk density was found under the trees. This is in agreement with
# 2006 The Authors Journal compilation # 2006 British Society of Soil Science, European Journal of Soil Science, 58, 728–735
Changes in soil C in a Mediterranean ecosystem 731 Table 1 Bulk density (g cm–3) of the ecosystems studied. Mean standard error (n ¼ 8)
Ecosystem NF WPE GE DE VDE
Outside canopy
0–5 cm
5–10 cm
0–5 cm
5–10 cm
1.17 0.09 1.19 0.05 1.27 0.06 1.34 0.05 1.40 0.08
1.48 0.08 1.44 0.05 1.44 0.05 1.50 0.03 1.45 0.05
1.55 0.08 1.34 0.06 1.38 0.05 1.43 0.04 1.43 0.07
1.65 0.05 1.55 0.04 1.55 0.05 1.55 0.03 1.53 0.03
a)
0
20
40
60
80
100
80
100
0-5
Depth /cm
Under canopy
SOC / t per 103 t soil ha–1
NF ¼ native forest; WPE ¼ well preserved Espinal; GE ¼ good Espinal; DE ¼ degraded Espinal; VDE ¼ very degraded Espinal.
5-10
results reported by other authors (Mordelet et al., 1993; Rhoades, 1997).
Organic carbon
SOC / t per 103 t soil ha–1
Microbial biomass carbon The values of microbial biomass carbon (Cmic) were in the range of 366–277 mg C kg–1 outside the canopy and 1105–357 mg
0
b)
20
40
60
0-5
Depth /cm
The total SOC (i.e. ‘C storage’, Figure 2) in the 0–5 cm layer (for both conditions under and outside the canopy) was 50% greater in the WPE soil than the VDE soil. Similar results were obtained when SOC contents were compared between the NF and the WPE and GE ecosystems. Based on earlier investigations (Ovalle et al., 2006), the tree cover would increase by 5% every 10 years (meaning that changing from VDE to WPE would require more than 100 years). In general, SOC values are smaller outside the canopy than under it, with the exception of the two most degraded classes of vegetation (DE and VDE; Figure 2) in the 5–10 cm layer. For example, under natural forest (NF) in the 0–5 cm layer, there was 73% reduction in SOC content outside the canopy compared with under the canopy. Similarly, in the other agroforestry systems at the same depth there was between 28 and 40% less SOC outside the canopy compared with under it. The SOC in all the degraded systems (at both 0–5 and 5–10 cm depths) was significantly less ‘under canopy’ than under canopy in NF, with the smallest values measured in DE and VDE (Figure 2a). For both depths, there was 50% less SOC for WPE and GE and 73% less for DE and VDE compared with NF. Outside the canopy, the SOC in the two most degraded systems (DE and VDE) was significantly less than in natural forest (37–40% reduction of SOC in the 0–5 and 5–10 cm depths, respectively), though this was not the case in the two less degraded systems (WPE and GE). Small differences in SOM (total and fractions) between the different tree densities of the agroforestry system (e.g. between WPE and GE, and DE and VDE) were partly explained by different inputs of plant residues but also by the presence of grazing animals, as they tend to move nutrients through their excreta, thus blurring distinctions between areas.
5-10
Figure 2 Soil organic carbon (t per 103 t soil ha–1) with depth (a) under canopy and (b) outside canopy. NF ¼ native forest; WPE ¼ well preserved Espinal; GE ¼ good Espinal; DE ¼ degraded Espinal; VDE ¼ very degraded Espinal. Error bars indicate one standard error of the mean (n ¼ 8). ¼ NF; ¼ WPE; ¼ GE, ¼ DE; ¼ VDE.
C kg–1 under the canopy, when converted to an area basis (Figure 3). These results were in agreement with those of Zagal et al. (2003), who found values of 350–286 mg C kg–1 for more intensive rotation systems and 932–981 mg C kg–1 for less intensive rotation systems in a silty to silty-loam soil. The Cmic was less (in most cases significantly so) under the tree canopy at both soil depths in all the agroforestry systems compared with NF (46–30% less for WPE and GE, respectively; similarly, 67–57% less for DE and VDE) (Figure 3). However, the differences among the degraded systems were mostly not significantly different, nor was there a clear trend for less Cmic with increasing degree of ecosystem degradation. However, WPE and GE had greater amounts of Cmic than DE and VDE at 0–5 cm depth (the differences were about 45% on
# 2006 The Authors Journal compilation # 2006 British Society of Soil Science, European Journal of Soil Science, 58, 728–735
732 C. Mun˜oz et al.
Figure 3 Microbial biomass carbon (Cmic) (mg C kg–1) in ecosystems studied: (a) under canopy and depth 0–5 cm; (b) under canopy and depth 5–10 cm; (c) outside canopy and depth 0–5 cm; (d) outside canopy and depth 5–10 cm. Error bars indicate one standard error of the mean (n ¼ 8). Ecosystem description as in Figure 2.
average). Outside the canopy there were no significant differences in Cmic between any of the sites, including NF. Analysis of the data (Table 2) showed that Cmic was not consistently correlated with SOC. A close correlation between Cmic and SOC is observed only for the 5–10 cm layer under the canopy and for the 0–5 cm soil outside the canopy, with correlation coefficients of 0.82 (P 0.01). These results were in agreement with Wirth (2001) who found a correlation of 0.766 (P 0.001) between Cmic and SOC. However, a tight correlation between these soil parameters should be expected only for soils where SOC and its constituent fractions are close to equilibrium. Where a change in land use or management has Table 2 Pearson product-moment correlation coefficient between Cmic, SOC, LF and microbial respiration; statistical significance shown: **P 0.01, *P 0.05 Measurement Depth/cm Cmic SOC LF Cmic SOC LF Cmic SOC LF Cmic SOC LF a
0–5 0–5 0–5 5–10 5–10 5–10 0–5 0–5 0–5 5–10 5–10 5–10
Soil organic carbon. Light fraction.
b
Condition Under canopy Under canopy Under canopy Under canopy Under canopy Under canopy Outside canopy Outside canopy Outside canopy Outside canopy Outside canopy Outside canopy
SOCa 0.29
LFb
Respiration
0.15 0.90**
0.53* 0.53** 0.68** 0.22 0.07 0.61* 0.62** 0.63** 0.81** 0.18 0.57** 0.64**
0.82** 0.61* 0.49* 0.35*
0.43 0.81**
0.31
0.85** 0.33
occurred, as in this study, the close relationships would be expected to break down. Expressing Cmic as a percentage of SOC for the 0–5 cm layer of the end-member systems, NF and VDE, gave values of 1.37% and 1.96%, respectively, which suggested that SOC declined proportionally more than Cmic, which was unexpected. It may be that the input of plant residues in the VDE system (or more degraded systems) was mainly from pasture and therefore more readily decomposable than tree residues. Additionally, above-ground tree production in the VDE system is poor, e.g. 1 t ha–1 year–1, which consequently implies a low return of tree residues to the soil (Ovalle et al., 1999). For example, decomposition studies in situ (420 days) of Lolium multiflorum straw and A. caven needles in degraded ecosystems (e.g. DE, litter bags placed outside the canopy) showed that there was greater mass lost for L. multiflorum straw (about 6%) than for A. caven needles (data not shown). On the other hand, the chemical composition of A. caven needles was richer in lignin and N content (26 and 1.83%, respectively) as compared with L. multiflorum straw (7 and 0.41%, respectively).
Soil respiration We found that the respiration rate was significantly greater under the canopy at 0–5 cm in NF with 82 16 (SE) mg C-CO2 g–1 soil, followed by VDE, GE, DE and WPE with 61 8 (SE), 51 10 (SE), 42 7 (SE) and 31 5 (SE) mg C-CO2 g–1 soil, respectively (Figure 4a). In the 5–10 cm layer (Figure 4b), there is no clear difference between ecosystems. Outside the canopy (Figure 4c), there were no significant differences between ecosystems, with a range of 32–45 mg C-CO2 g–1 for 0–5 cm depth,
# 2006 The Authors Journal compilation # 2006 British Society of Soil Science, European Journal of Soil Science, 58, 728–735
Changes in soil C in a Mediterranean ecosystem 733
Figure 4 Microbial respiration (mg C-CO2 g–1 dry soil) in ecosystems studied: (a) under canopy and depth 0–5 cm; (b) under canopy and depth 5–10 cm; (c) outside canopy and depth 0–5 cm; (d) outside canopy and depth 5–10 cm. Error bars indicate one standard error of the mean (n ¼ 8) Ecosystem description as in Figure 2.
while at 5–10 cm depth (Figure 4d) only NF had a significantly greater value with 59 8 (SE) mg C-CO2 g–1, while the mean respiration of the other ecosystems was 16 mg C-CO2 g–1. Correlation analysis (Table 2) indicated that respiration was not consistently related to either SOC or microbial biomass, but was related to the amount of light fraction (LF) in most cases. This could reflect microbial breakdown of a readily decomposable fraction of the plant residues input to soil such as sugars and amino acids. A tight correlation between the amount of LF and respiration rate was also found by Janzen et al. (1992), who reported a correlation coefficient of 0.81 between these soil parameters. Such a relationship could be expected, as the LF of organic matter comprises mainly freshly added organic material from plant debris and manure (Christensen, 1992).
Light fraction The strong correlation between the LF content of soils and soil respiration rates (Table 2), suggests that in these soils LF is an important carbon and energy source for soil microorganisms. This is possible because LF is not protected by clay minerals or other mechanisms, and it is readily accessible to microbial and enzyme activity (Skjemstad et al., 1986; Six et al., 2002). Skjemstad et al. (1986) proposed that the LF is more sensitive than total organic matter content to the effects of cropping practices. In other investigations, LF showed seasonal fluctuations (Boone, 1994; Campbell et al., 1999) in both carbon and nitrogen contents, which were associated with changes in soil moisture, temperature and rainfall. Boone (1994) concluded that the seasonality of new inputs of organic matter to the soil was the main factor affecting the amount of LF extracted from soils under maize. Other investigations have indicated that the LF content of soil organic matter is an important property for determining ecosystem quality and/or stress. For example, Janzen et al. (1992)
found that the LF content increased 2.6-fold with increasing organic matter content, whereas SOC content increased only 1.3-fold. Also, Zagal et al. (2001) found a decrease in LF content as the intensity of soil use increased. We found in this study that with greater intensity of soil use the LF diminishes 3.4-fold with decreases of 1.61-fold of SOC (data not shown) for conditions outside the canopy. This further demonstrates that the LF content is a more sensitive parameter than SOC content to changes in the soil due to land-use intensity. Components of SOM that provide substrate that is physically or chemically more accessible to the microbial biomass are fractions of the total SOM that display proportionally larger changes over the short term and may therefore be useful as indicators. The light fraction has been previously shown to change more rapidly and act as an ‘early warning’ of change in total SOM content, long before these can be measured with statistical certainty (Christensen, 2001). Overall, the microbial biomass, LF content and SOC values were greater under canopy conditions, for both depths. These results might indicate that the soil under the influence of A. caven offers propitious conditions for the development of the microbial populations. However, we found that in Espinal ecosystems with less A. caven cover the microbial activity (e.g. soil respiration) was greater than in ecosystems with more tree cover. For example, values for respiration for the 0–5 cm layer, expressed as CO2-C evolved (mg) per unit of microbial biomass C (g) for NF, GE and VDE, were 120, 76 and 172, respectively. As microbial activity is the result of a combination of the quantity and decomposability of organic input into the soils of the different systems, more trees in GE (604 201 tree ha–1) produced significant inputs that were derived from tree residues (e.g. A. caven needles) and, alternatively, in VDE (308 51 tree ha–1) the inputs were derived of pasture plant species, which were readily decomposed by the microorganisms.
# 2006 The Authors Journal compilation # 2006 British Society of Soil Science, European Journal of Soil Science, 58, 728–735
734 C. Mun˜oz et al. The results regarding the role of A. caven tree cover on the levels of soil carbon in the ‘Espinal’ system of the Mediterranean climate zone of Chile, are in agreement with studies that have been obtained in other agroforestry systems in Mediterranean areas of the world, such as the ‘Dehesas’ of southern Spain and Portugal (Joffre et al., 1999), the savannas of Faidherbia albida of Sudan (Van Noordwijk & Ong, 1999) and the fynbos ecosystems in South Africa (Cocks & Stock, 2001). In these land-use systems, the trees carry out a fundamental function in nutrient and water cycling, and in this regard, these agroforestry systems could be considered similar to the natural ecosystems of these regions (Ewel, 1999; Van Noordwijk & Ong, 1999), thereby conserving an important part of the original function and diversity of the native ecosystem. On the other hand, the results imply that not only is it possible to improve the productivity and sustainability of the Chilean ‘Espinales’, but it could additionally be used as a model for other Mediterranean agroecosystems, such as in Australia where conservation and the management of trees in these systems have not received high priority. In Australia, the mixed original vegetation of woody and herbaceous plants has been replaced by more simplified systems of herbaceous crops and pasture, that maximize the short-term primary productivity, but in the long term have resulted in a loss of sustainability due to soil nutrient depletion and salinity (Ridley & Pannell, 2005).
Conclusions 1 Microbial biomass, respiration rate, light fraction and SOC are influenced by the presence of the A. caven tree, with values being greater under its canopy than outside its canopy. 2 Under the tree canopy, to a depth of 10 cm, microbial biomass carbon (Cmic) was less under all the agroforestry systems than in NF. However, there was no clear trend for less Cmic with increased ecosystem degradation, especially in the soil from outside the canopy. 3 The respiration of microbial communities was affected by the stage of degradation of the ecosystems, and was related to the relative proportions and decomposability of plant inputs (pasture versus tree residues) in the different systems. 4 Microbial respiration is correlated with LF content, indicating that this provides a substrate that is physically or chemically more accessible to the microbial biomass. 5 Increasing the cover of the A. caven tree in the current semiarid ecosystems of central-south Mediterranean Chile would result in an eventual long-term (over several centuries) increase in soil organic C of approximately 50%.
Acknowledgements This research was made possible thanks to the financing granted by FONDECYT (Project No. 1030883) and the Andes Foundation (No. C-13755 28). We also thank the personnel of the
Department of Soils of the Universidad de Concepcio´n (Chile) for assistance with soil analysis.
References Alef, K. 1995. Estimation of soil respiration. In: Methods in Applied Soil Microbiology and Biochemistry (eds K. Alef & P. Nannipieri), pp. 464–467. Academic Press, London. Aronson, J. 1992. Evolutionary biology of Acacia caven (Leguminosae, Mimosoideae): infraspecific variation in fruit and seed characters. Annals of the Missouri Botanical Garden, 79, 958–968. Aronson, J. & Ovalle, C. 1989. Study of the natural variability, biogeography and potential for genetic improvement of Acacia caven. Bulletin of the International Group for the Study of Mimosideae, 17, 111–121. Aronson, J., Ovalle, C., Avendan˜o, J., Longeri, L. & Del Pozo, A. 2002. Agroforestry tree selection in central Chile: biological nitrogen fixation and early plant growth in six dryland species. Agroforestry Systems, 56, 155–166. Balesdent, J., Pe´traud, J. & Feller, C. 1991. Effets des ultrasons sur la distribution granulome´trique des matie`res organiques des sols. Science du Sol, 29, 95–106. Boone, R.D. 1994. Light-fraction soil organic matter: origin and contribution to net nitrogen mineralization. Soil Biology and Biochemistry, 26, 1459–1468. Campbell, C.A., Biederbeck, V.O., Wen, G., Zentner, R.P., Schoenau, J. & Hahn, D. 1999. Seasonal trend in selected soil biochemical attributes: effects of crop rotation in the semiarid prairie. Canadian Journal of Soil Science, 79, 73–84. Christensen, B.T. 1992. Physical fractionation of soil and organic matter in primary particle size and density separates. Advances in Soil Science, 20, 1–90. Christensen, B.T. 2001. Physical fractionation of soil and structural and functional complexity in organic matter turnover. European Journal of Soil Science, 52, 345–353. Cocks, M.P. & Stock, W.D. 2001. Field patterns of nodulation in fifteen Aspalathus species and their ecological role in fynbos vegetation of southern Africa. Basic and Applied Ecology, 2, 115–125. Ellert, B.H., Janzen, H.H. & Entz, T. 2002. Assessment of a method to measure temporal change in soil carbon storage. Soil Science Society of America Journal, 66, 1687–1695. Ewel, J. 1999. Natural systems as model for the design of sustainable systems of land use. Agroforestry Systems, 45, 1–21. Janzen, H.H., Campbell, C.A., Brandt, S.A., Lafond, G.P. & TownleySmith, L. 1992. Light fraction organic matter in soils from long term crop rotations. Soil Science Society of America Journal, 56, 1799–1806. Joergensen, R.G. & Brookes, P.C. 1990. Ninhydrin-reactive nitrogen measurements of microbial biomass in 0.5 M K2SO4 soil extracts. Soil Biology and Biochemistry, 22, 1023–1027. Joffre, R., Rambal, S. & Ratte, J. 1999. The Dehesa system of southern Spain and Portugal as a natural ecosystem mimic. Agroforestry Systems, 45, 57–79. Krull, E.S., Baldock, J.A. & Skjemstad, J.O. 2003. Importance of mechanism and processes of the stabilization of soil organic matter for modelling carbon turnover. Functional Plant Biology, 30, 207–222. Linn, D.M. & Doran, J.W. 1984. Effect of water-filled pore space on carbon dioxide and nitrous oxide production in tilled and nontilled soils. Soil Science Society of America Journal, 48, 1267–1272.
# 2006 The Authors Journal compilation # 2006 British Society of Soil Science, European Journal of Soil Science, 58, 728–735
Changes in soil C in a Mediterranean ecosystem 735
Mordelet, P., Abbadie, L. & Menaut, J.C. 1993. Effects of tree clumps on soil characteristics in a humid savanna of West Africa. Plant and Soil, 153, 103–111. Ocio, J.A. & Brookes, P.C. 1990. An evaluation of methods for measuring the microbial biomass in soils following recent additions of wheat straw and the characterization of the biomass that develops. Soil Biology and Biochemistry, 22, 685–694. Ovalle, C., Aronson, J., Del Pozo, A. & Avendan˜o, J. 1990. The espinal: agroforestry systems of the Mediterranean-type climate region of Chile. Agroforestry Systems, 10, 213–239. Ovalle, C., Aronson, J., Del Pozo, A. & Avendan˜o, J. 1999. Restoration and rehabilitation of mixed espinales of central Chile. Arid Soil Research and Rehabilitation, 13, 369–381. Ovalle, C., Del Pozo, A., Casado, M.A., Acosta, B. & de Miguel, J.M. 2006. Consequences of landscape heterogeneity on grassland diversity and productivity in the Espinal agroforestry system of central Chile. Landscape Ecology, 21, 585–594. Ovalle, C. & Godron, M. 1989. Influencia del a´rbol sobre la vegetacio´n herba´cea en matorrales de Acacia caven. Rol del animal (ovinos) en las interacciones a´rbol-pasto. Stvdia Œcologica, 6, 225–243. Ovalle, C., Longeri, L., Aronson, J., Herrera, A. & Avendan˜o, J. 1996. N2-fixation, nodule efficiency and biomass accumulation after two years in three Chilean legume trees and Tagasaste (Chamaecytisus proliferus subsp. palmensis). Plant and Soil, 179, 131–140. Powlson, D.S. 1994. The soil microbial biomass: before, beyond and back. In: Beyond the Biomass. Compositional and Functional Analysis of Soil Microbial Communities (eds K. Ritz, J. Dighton & K.E. Giller), pp. 3–20. John Wiley and Sons, Chichester. Rhoades, C.C. 1997. Single-tree influences on soil properties in agroforestry: lessons from natural forest and savanna ecosystems. Agroforestry Systems, 35, 71–94.
Ridley, A.M. & Pannell, D.J. 2005. The role of plants and plantbased research and development in managing dryland salinity in Australia. Australian Journal of Experimental Agriculture, 45, 1341– 1355. Sims, J.R. & Haby, V.A. 1971. Simplified colorimetric determination of soil organic matter. Soil Science, 112, 137–141. Six, J., Conant, R.T., Paul, E.A. & Paustian, K. 2002. Stabilization mechanisms of soil organic matter: implications for C-saturation of soils. Plant and Soil, 241, 155–176. Skjemstad, J.O., Dalal, R.C. & Barron, P.F. 1986. Spectroscopic investigations of cultivation effects on organic matter of Vertisols. Soil Science Society of America Journal, 50, 354–359. Van Noordwijk, M. & Ong, C.K. 1999. Can the ecosystem mimic hypotheses be applied to farms in African savannahs? Agroforestry Systems, 45, 131–158. Vance, E.D., Brookes, P.C. & Jenkinson, D.S. 1987. An extraction method for measuring soil microbial biomass C. Soil Biology and Biochemistry, 19, 703–707. Wirth, S.J. 2001. Regional-scale analysis of soil microbial biomass and soil basal CO2-respiration in northeastern Germany. In: Sustaining the Global Farm – 10th International Soil Conservation Organization Meeting (eds D.E. Stott, R.H. Mohtar & G.C. Steinhardt), pp. 486–493. West Lafayette, IN, USA. ISCO-USDA-Purdue University. Zagal, E., Rodrı´ guez, N., Vidal, I. & Flores, B. 2001. La fraccio´n liviana de la materia orga´nica de un suelo volca´nico bajo distinto manejo agrono´mico como ı´ ndice de cambios de la materia orga´nica. Agricultura Te´cnica (Chile), 62, 284–295. Zagal, E., Rodrı´ guez, N., Vidal, I. & Maureira, M. 2003. Biomasa microbiana por los me´todos reactivo nitro´geno ninhydrina y absorbancia ultravioleta (UV280) en distintos sistemas de cultivo. Revista de la Ciencia Del Suelo Y Nutricio´n Vegetal (Chile), 3, 17–28.
# 2006 The Authors Journal compilation # 2006 British Society of Soil Science, European Journal of Soil Science, 58, 728–735
