Soil CO2 Emissions from Northern Andean Páramo Ecosystems

Jan 26, 2008 - The effects of fallow agriculture on soil organic matter (SOM) dynamics and CO2 emissions were assessed in the tropical Andean páramo ...
0 downloads 0 Views 2MB Size
Research Soil CO2 Emissions from Northern Andean Pa´ramo Ecosystems: Effects of Fallow Agriculture ANA CABANEIRO,* IRENE FERNANDEZ, LUIS PÉREZ-VENTURA, AND TARSY CARBALLAS Departamento de Bioquímica del Suelo, Instituto de Investigaciones Agrobiológicas de Galicia, Consejo Superior de Investigaciones Científicas (CSIC), Apartado 122, E-15780 Santiago de Compostela, Spain

Received June 11, 2007. Revised manuscript received November 27, 2007. Accepted December 10, 2007.

The effects of fallow agriculture on soil organic matter (SOM) dynamics and CO2 emissions were assessed in the tropical Andean páramo ecosystem. Possible changes during the cultivation-fallow cycle were monitored in four areas of the Quebrada Piñuelas valley (Venezuela). Uncultivated soils and plots at different stages of a complete cultivation-fallow cycle were incubated, and SOM mineralization kinetics was determined. Soils exhibited a low SOM mineralization activity, total CO2 evolved never reaching 3% of soil carbon, pointing to a stabilized SOM. Potential soil CO2 effluxes differed significantly according to their plot aspect: northeast (NE)-aspect soils presented higher CO2 effluxes than southwest (SW)-aspect soils. Soil CO2 emissions decreased after ploughing as compared to virgin páramo; low CO2 effluxes were still observed during cropping periods, increasing progressively to reach the highest values after 4–5 y of fallow. In all cases, experimental C mineralization data was fitted to a double exponential kinetic model. High soil labile C pool variability was observed, and two different trends were identified: NE-oriented soils showed more labile C and a wider range of values than SW-facing soils. Labile C positively correlated with CO2 effluxes and negatively with its instantaneous mineralization rate. The instantaneous mineralization rate of the recalcitrant C pool positively correlated with %C evolved as CO2 and negatively with soil C and Al2O3 contents, suggesting the importance of aluminum on SOM stability. The CO2 effluxes from these ecosystems, as well as the proportion of soil C released to the atmosphere, seem to depend not only on the size of the labile C pool but also on the accessibility of the more stabilized SOM. Therefore, fallow agriculture produces moderate changes in SOM quality and temporarily alters the CO2 emission capacity of these soils.

Introduction From a global perspective, terrestrial carbon management may be an essential mitigation component of international climate change strategies. Within terrestrial ecosystems, both vegetation and the pedosphere play important and dynamic roles in the global carbon (C) cycle, contributing greatly to the atmospheric composition. Worldwide estimations of * Corresponding author telephone: +34 981590958; fax: +34 981592504; e-mail: [email protected]. 1408

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 5, 2008

potential contributions of diverse greenhouse gas sources from different types of natural ecosystems and possible changes on soil emissions produced by their transformation in agricultural lands are basic steps needed to obtain a largescale and consistent overall C balance. As a scientific response to the current global warming concern, during the last few years relevant investigations have been implemented to estimate partial contributions of terrestrial ecosystems from different climatic zones to the total atmospheric CO2 concentration (1–4). The páramo is an ecosystem widely found in the tropical Andean mountains developed under drastic daily temperature fluctuations and, in the Northern Andes, this type of ecosystem occupies the altitudinal belt between 2800 and 4800 m above sea level (asl). Páramo soils, generally acidic, stony, and highly organic, were formed approximately 15 000 y ago after the last ice age (5), and several studies regarding their chemical characteristics (6) or SOM turnover (4, 7) were carried out in the past decade. Páramo flora biodiversity are among the majority in the world’s high mountains, associated to its particular geographical distribution. The importance of regulating water availability in the lowlands makes the páramo a high priority area for conservation; however, continuous páramo degradation due to the pressure of an increasing population raises the agricultural frontier altitude by incorporating the virgin páramo into the agricultural cycle (8). The traditional agricultural system in these mountains is known as fallow agriculture with characteristic landscapes in mosaic of cultivated plots, plots in different fallow stages, and never-cultivated areas (8). This land management is widely practiced not only in South America, particularly in Northern and Central Andes (Venezuela, Bolivia, Colombia, and Peru), but also in different areas of Asia and Africa, due to the low inherent capacity of their soils to maintain a continuous production even with fertilizers (8–12). Fallow agriculture, associated to subsistence (13, 14) but considered as ecologically and economically viable under low population density conditions and nonlimited land availability (15), is characterized by a fast soil fertility decline after short cropping periods (1–3 y) and a subsequent slow restoration during long fallow periods (3–20 y). During these fallow periods, a secondary succession takes place, and at early stages exotic herbs (e.g., Rumex acetosella) strongly dominate the new vegetation. At more advanced succession stages, these species are progressively substituted by native herbs (e.g., Lupinus meridanus), later reaching the characteristic páramo life forms (8) mainly dominated by giant caulescent rosettes (Espeletia schultzii) and shrubs (Baccharis prunifolia). The aim of this research was to quantify SOM contents and potential CO2 emissions to the atmosphere from Venezuelan páramo soils as well as to monitor their possible changes as a consequence of the introduction of fallow agricultural practices. The considerable importance of páramo ecosystems, as well as the persistent pressure of land use conversion to fallow agriculture, supports the implementation of a detailed study of soil-atmosphere gas exchanges in these regions to evaluate their role in the global terrestrial C cycle.

Experimental Section Study Site. The study was performed in the Quebrada Piñuelas valley, of glacial origin, located in the Gavidia páramo (8°35′-8°45′ N; 70°52′-70°57′ W), between 3350 and 3700 m asl, within Sierra Nevada (Mérida National Park) in Northern 10.1021/es071392d CCC: $40.75

 2008 American Chemical Society

Published on Web 01/26/2008

TABLE 1. Soil Characteristics (A) and Soil CO2 Emissions Parameters (B) for the 0–15 cm Depth Layers of the Virgin Paramo Ecosystem (VP) from the Four Areas Selected for Studya (A) B-VP R-VP V-VP Y-VP

pH H2Ob

5.7 ( 0.1 80.4 ( 0.1 5.3 ( 0.1 118.0 ( 0.6 5.2 ( 0.1 47.3 ( 0.3 5.6 ( 0.1 61.4 ( 0.1

C/N N (g kg-1)b ratio 4.3 ( 0.2 6.3 ( 0.0 2.5 ( 0.0 3.3 ( 0.1

19 19 19 18

WHC (%)b 56.9 ( 0.5 60.2 ( 0.6 29.8 ( 0.6 40.9 ( 0.4

CEC (cmol kg-1)b Al3+ (cmol kg-1)b Al2O3(g kg-1)b Fe2O3 (g kg-1)b Al2O3/Fe2O3 66.6 ( 1.0 88.8 ( 0.8 38.2 ( 0.2 54.9 ( 0.7

Soil CO2 emissions (mean ( standard deviation)

(B)

B-VP R-VP V-VP Y-VP

C (g kg-1)b

10.3 ( 0.6 9.0 ( 0.0 14.5 ( 0.0 12.5 ( 0.5

22.3 ( 0.6 27.5 ( 0.9 19.1 ( 0.1 21.7 ( 0.4

32.4 ( 0.4 26.1 ( 0.4 19.6 ( 0.3 27.0 ( 0.4

0.69 1.05 0.97 0.80

C mineralization kinetic parametersc (estimated value ( asymptotic standard error)

aspect

potential CO2 effluxes (g C kg-1)

CO2 emission coeff. (% of total C)

C0 (g C kg-1)

K (d-1)

h × 104 (d-1)

R2

NE NE SW SW

2.19 ( 0.21 2.20 ( 0.16 0.86 ( 0.14 1.47 ( 0.21

2.73 ( 0.28 1.87 ( 0.14 1.83 ( 0.29 2.39 ( 0.35

0.52 ( 0.02 0.59 ( 0.01 0.24 ( 0.00 0.20 ( 0.01

0.09 ( 0.01 0.09 ( 0.00 0.13 ( 0.01 0.12 ( 0.01

2.55 ( 0.00 1.64 ( 0.00 1.59 ( 0.00 2.57 ( 0.00

0.999 0.999 0.999 0.999

B, Bárbara; R, Ramón; V, Volcanes; Y, Yaques. b Mean ( standard deviation. c Kinetic parameters estimated according to the first-order model based on the double exponential equation Ct ) C0(1 - e–kt) + (C - C0)(1 - e–ht) proposed by Andrén and Paustian (1987)[C0, labile C pool; k, instantaneous mineralization rate of the labile C pool; h, instantaneous mineralization rate of the recalcitrant C pool; R2, determination coefficient]. a

FIGURE 1. Cumulative kinetics curves of the potential soil CO2 emissions (g Cmineralized/kgdrysoil) (a) and CO2 emission coefficients (100 × Cmineralized/Ctotal) (b) from Venezuelan páramo (VP) soils (B: Bárbara, R: Ramón, V: Volcanes, Y: Yaques), over 12 weeks of incubation. Andes, Venezuela. Páramo soils of Gavidia were developed over pre-Cambrian schists and gneises originated by metamorphism of marine sediments. Their geological and climatic conditions have been reported by Abadín et al (6). Experimental Design. To elucidate the fallow agriculture effects on SOM dynamics and soil CO2 effluxes, two NEoriented sectors (B: Bárbara, R: Ramón) and two SW-oriented sectors (V: Volcanes, Y: Yaques) were selected among 1200 controlled fields described in previous studies (16). Each sector included seven plots of similar topography, parent material, and sun exposure that constituted a crop-fallow chronosequence, with plots at different stages of a complete cultivation-fallow cycle: recently ploughed after long fallow periods (R), 1 y and 2 y potato crop (C1, C2), 1 y fallow (F1), 4 y fallow (F4: Yaques, Volcanes) or 5 y fallow (F5: Bárbara, Ramón), 8 y fallow (F8), and virgin páramo (VP) plots. Soil Analyses. For each plot, 15 soil subsamples were taken randomly from the A horizon (0–15 cm depth), subsequently mixed, and thoroughly homogenized after sieving (4 mm). Afterward, from every whole soil sample, fresh subsamples were used for respirometric experiments, being kept at -18 °C until the start of incubation; the remaining soil was airdried for soil characterization. Organic C was determined by combustion in a Carmhograph12 (Whostoff, Germany). Other main soil characteristics were assessed using methods

described by Abadín et al (6). All results are expressed per kilogram of oven-dried soil (105 °C). Potential Soil CO2 Emissions. Soils were aerobically incubated following the method of Guckert et al. (17), with subsequent modifications (18), using optimal conditions for microbial activity (28 °C, 75% of water holding capacity) to attain the potential SOM mineralization. For each soil, nine replicates were incubated for 12 weeks. Flask atmospheres were renewed (everyday, every 2 days, or every week depending on flask CO2 concentration) with humidified CO2free air, and the CO2 evolved was trapped as Na2CO3 by bubbling for 2 h in 2 N NaOH. After CO32- precipitation (BaCl2), the remaining NaOH was titrated against HCl (Metrohm736, GP Titrino, Herisau-Switzerland). Potential soil CO2 emissions were quantified by subtraction, using empty flasks incubated as control, and were expressed as grams of CO2-C evolved per kilogram of dry soil (mineralized C) and as percentage of total soil C (CO2 emission coefficient). Kinetic Modeling. To quantify C mineralization kinetic parameters, cumulative data of CO2-C released at different incubation times were fitted to the simple first-order kinetic model of Stanford and Smith (19) and to the double exponential model of Andrén and Paustian (20). In all cases, compared with the simple model that consider a single source of CO2, the data fit better to the double exponential model VOL. 42, NO. 5, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1409

FIGURE 2. Cumulative kinetics curves of the potential soil CO2 emissions (g Cmineralized/kgdrysoil) from all soils of the four sectors at different stages of the complete fallow-cultivation cycle. that considers two C pools of different lability (C0: labile C; C-C0: recalcitrant C) with different instantaneous mineralization rates (k for labile C and h for recalcitrant C). This second model [Ct ) C0(1 - e–kt) + (C - C0)(1 - e–ht)], where Ct is the cumulative C released after time t, was fitted to the data using nonlinear parameter estimation procedures in SPSS 14.0 (2005), and to avoid parameter estimation errors, Updegraff et al. (21) convergence criteria were applied. Statistical Analyses. Data were analyzed by one-way ANOVA, and least significant differences were established using Tukey’s tests. Principal components analysis (PCA) with varimax rotation was applied to determine the main factors controlling soil CO2 emissions. All statistical analyses were performed using the computer software SPSS 14.0 (2005).

Results and Discussion Soil CO2 Emissions from Páramo Soils. Virgin páramo soils were moderately acidic (Table 1a), with high exchangeable H+ and Al3+ contents. Similar C/N ratios, with values of about 19 in all cases, suggest that their SOM could be considered a moder humus type, according to Duchaufour’s humus classification (22). All soils exhibited considerable amounts of free aluminum oxides that surpass 20 g Al2O3 kg-1 in most cases, with values similar to or even higher than values of free iron oxides. High water holding capacities (WHC), low nutrient availability (CEC), and soil bulk densities lower than 1 g cm-3 were also observed. These soil characteristics agreed with findings of other authors for soils of the same region (6, 7). Remarkable total soil C differences were found among the four páramo soils, with SOM content being nearly double for northeasterly aspect sectors (Bárbara, Ramón) as compared with southwesterly aspect soils (Volcanes, Yaques) where greater afternoon irradiation boosts soil warming. This 1410

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 5, 2008

effect of topographic aspect on soil properties is widely recognized, and aspect-dependent SOM content differences were reported (23). Total CO2 released from páramo soils were relatively low, with their potential CO2 effluxes ranging from 0.86 to 2.20 g C kg-1. We found an expected relationship between potential soil CO2 effluxes from virgin páramo soils and soil C contents (Table 1), because NE-aspect sectors with higher SOM contents also exhibited higher quantities of CO2-C evolved during incubation in comparison with SWaspect sectors (ANOVA, P < 0.01, n ) 3); the sector with the lowest C content (Volcanes) also showed the lowest soil CO2 effluxes (ANOVA, P < 0.05, n ) 3). Emissions of similar magnitude ∼1.5 g C kg-1drysoil) have also been reported for páramo soils from the same area (24). The C mineralization kinetics over the incubation period were quite similar for all virgin páramo soils (curve slants having always a progressive decline), despite the previously indicated differences in the overall CO2 effluxes (Figure 1a). Cumulative experimental data of CO2 emissions from all soils were excellently fitted to a double exponential kinetic model that considers two soil C pools of different lability. Table 1b shows the total soil CO2 emissions after 12 weeks of incubation as well as the kinetic parameters estimated with this model. Important differences on the labile C pool (C0) were obtained, the two virgin páramo soils with the highest C contents exhibiting the largest labile C pools, with about twice the values obtained for the less organic soils; these higher C0 values are responsible for the highest mineralization curve slant during the first incubation days (Figure 1a), despite the lower instantaneous mineralization rate of their labile C pool (k). Kinetic parameters of similar magnitude have also been reported by other authors for páramo soils in the same area (24). The CO2 emission coefficient, which expresses CO2-C

TABLE 2. Soil CO2 Emissions and C Mineralization Kinetic Parameters Estimated According to the First-order Model Based on the Double Exponential Equation Ct = C0(1 - e–kt) + (C - C0)(1 - e–ht) Proposed by Andren and Paustian (1987) for Agricultural Plots from Northeasterly Aspect Sectors (B: Barbara, R: Ramon) at Different Stages of a Crop-Fallow Chronosequencea C mineralization kinetic parameters (estimated value ( asymptotic standard error)

soil CO2 emissions (mean ( standard deviation)

B-R B-C1 B-C2 B-F1 B-F5 B-F8 R-R R-C1 R-C2 R-F1 R-F5 R-F8

C (g C kg-1)

C/N ratio

potential CO2 effluxes (g C kg-1)

CO2 emission coeff. (% of total C)

C0 (g C kg-1)

K (d-1)

h × 104 (d-1)

R2

76.1 ( 0.1 76.9 ( 0.0 47.4 ( 0.0 79.6 ( 0.0 99.2 ( 0.0 71.0 ( 0.4 100.3 ( 0.0 84.9 ( 1.2 101.5 ( 0.2 89.2 ( 0.1 97.8 ( 0.0 113.9 ( 0.2

18 17 15 18 16 17 18 20 17 21 17 21

1.29 ( 0.16 1.06 ( 0.21 1.18 ( 0.24 1.36 ( 0.29 2.02 ( 0.41 1.43 ( 0.30 1.52 ( 0.05 1.42 ( 0.23 1.70 ( 0.18 1.45 ( 0.19 2.08 ( 0.27 1.76 ( 0.15

1.71 ( 0.21 1.38 ( 0.26 2.49 ( 0.51 1.70 ( 0.37 2.04 ( 0.42 2.43 ( 0.43 1.51 ( 0.05 1.67 ( 0.27 1.67 ( 0.16 1.63 ( 0.21 2.13 ( 0.28 1.54 ( 0.13

0.25 ( 0.01 0.25 ( 0.03 0.41 ( 0.07 0.28 ( 0.03 0.41 ( 0.02 0.38 ( 0.03 0.31 ( 0.01 0.24 ( 0.00 0.41 ( 0.01 0.44 ( 0.01 0.43 ( 0.03 0.57 ( 0.02

0.14 ( 0.01 0.07 ( 0.01 0.04 ( 0.01 0.10 ( 0.02 0.15 ( 0.02 0.09 ( 0.01 0.11 ( 0.02 0.15 ( 0.01 0.09 ( 0.01 0.08 ( 0.00 0.09 ( 0.00 0.08 ( 0.01

1.68 ( 0.00 1.28 ( 0.00 1.98 ( 0.00 2.85 ( 0.00 1.99 ( 0.00 2.30 ( 0.00 1.44 ( 0.00 1.66 ( 0.00 1.52 ( 0.00 1.38 ( 0.00 2.03 ( 0.00 1.25 ( 0.00

0.999 0.999 0.999 0.997 0.998 0.999 0.998 0.998 0.999 0.999 0.999 0.999

a R: recently ploughed soil after a long fallow period; C1 and C2: 1 and 2 y potato crop plots; F1, F5, and F8: 1, 5, and 8 y fallow plots, respectively.

TABLE 3. Soil CO2 Emissions and C Mineralization Kinetic Parameters Estimated According to the First-order Model Based on the Double Exponential Equation Ct = C0(1 - e–kt) + (C - C0)(1 - e–ht) Proposed by Andren and Paustian (1987) for Agricultural Plots from Southwesterly Aspect Sectors (V: Volcanes, Y: Yaques) at Different Stages of a Crop-Fallow Chronosequencea C mineralization kinetic parameters (estimated value ( asymptotic standard error)

soil CO2 emissions (mean ( standard deviation)

V-R V-C1 V-C2 V-F1 V-F4 V-F8 Y-R Y-C1 Y-C2 Y-F1 Y-F4 Y-F8

C (g C kg-1)

C/N ratio

potential CO2 effluxes (g C kg-1)

CO2 emission coeff. (% of total C)

C0 (g C kg-1)

K (d-1)

h × 104 (d-1)

R2

62.3 ( 0.0 46.1 ( 0.1 76.5 ( 0.1 58.1 ( 0.1 45.6 ( 0.1 46.9 ( 0.0 82.9 ( 0.3 80.1 ( 0.3 79.3 ( 0.1 58.9 ( 0.1 69.5 ( 0.0 51.9 ( 0.0

16 18 18 18 17 18 19 18 16 15 16 17

0.84 ( 0.08 1.10 ( 0.17 1.17 ( 0.11 1.21 ( 0.13 1.33 ( 0.13 1.03 ( 0.16 1.20 ( 0.14 1.45 ( 0.29 1.43 ( 0.38 1.19 ( 0.24 1.63 ( 0.25 1.17 ( 0.28

1.35 ( 0.13 2.38 ( 0.37 1.53 ( 0.15 2.08 ( 0.23 2.92 ( 0.29 2.20 ( 0.33 1.46 ( 0.17 1.82 ( 0.29 1.80 ( 0.48 2.01 ( 0.41 2.35 ( 0.36 2.26 ( 0.53

0.16 ( 0.01 0.29 ( 0.01 0.26 ( 0.01 0.29 ( 0.02 0.26 ( 0.00 0.19 ( 0.03 0.10 ( 0.01 0.22 ( 0.02 0.24 ( 0.01 0.20 ( 0.00 0.11 ( 0.04 0.22 ( 0.03

0.10 ( 0.00 0.10 ( 0.01 0.12 ( 0.01 0.11 ( 0.01 0.11 ( 0.00 0.13 ( 0.00 0.14 ( 0.02 0.05 ( 0.00 0.09 ( 0.00 0.09 ( 0.01 0.21 ( 0.01 0.06 ( 0.02

1.31 ( 0.00 2.11 ( 0.00 1.43 ( 0.00 1.92 ( 0.00 2.84 ( 0.00 2.16 ( 0.00 1.64 ( 0.00 1.90 ( 0.00 1.85 ( 0.00 2.07 ( 0.00 2.73 ( 0.00 2.28 ( 0.00

0.999 0.999 0.999 0.999 0.999 0.998 0.998 0.998 0.998 0.997 0.998 0.998

a R: recently ploughed soil after a long fallow period; C1 and C2: 1 and 2 y potato crop plots; F1, F4 and F8: 1, 4, and 8 y fallow plots, respectively.

evolved as % of total soil C, allows the comparison of soils with different C contents, because this index mainly depends on SOM quality. Virgin páramo soils from all sectors exhibited low CO2 emission coefficients, never reaching 3% (Figure 1b). These values (Table 1b) seem to be determined by h, the instantaneous mineralization rate of the recalcitrant C pool (as suggested by their nearly significant Pearson correlation, r ) 0.942). Low CO2 emission coefficients pointed to a highly stabilized SOM, probably due to the important Al content of these soils. Stabilizing effects of Al on SOM have been suggested earlier by several authors either in laboratory experiments (25) or for natural soils with high Al contents (26, 27). Boudot et al. (28) also stated that Al complexes predominate in moderately acidic soils and that both SOM accumulation and limitation of translocation processes should be due to Al rather than Fe. Thus, in our case, the páramo soil from Bárbara, which exhibited the lowest Al2O3/ Fe2O3 ratio (Table 1a), showed a significantly higher CO2 emission coefficient (ANOVA, P < 0.05, n ) 3) in comparison with other sectors with higher Al2O3 proportion, such as Ramón and Volcanes (Figure 1b).

Effects of Fallow Agriculture on Soil CO2 Emissions. The impact of converting high altitude páramo ecosystems into agriculture on the potential for soil CO2 emissions can be estimated by comparing virgin páramo soils with plots at different stages of a complete cultivation-fallow cycle. Aspect-related differences in soil CO2 effluxes among the sectors studied were found. Figure 2 shows cumulative curves of potential soil CO2 emissions for all sectors, including cultivated and fallow plots as well as virgin páramo and recently ploughed plots. Similarly to the virgin páramo emissions, soils from Bárbara and Ramón presented higher CO2 effluxes than plots from the other two sectors at the same stage of the cycle, with soils from Volcanes showing the smallest potential CO2 emissions. Although cultivation caused soil C decreases in northeasterly-aspect sectors, followed by some recovery during fallow periods and opposite trends in southwesterly-aspects (Tables 1, 2, and 3), similar tendencies for potential CO2 emissions in all sectors appeared (Figure 3a). After the greater (NE-aspect) or lesser (SW-aspect) decline in ploughing stage, potential soil CO2 effluxes increased from C2 to F4-F5 due to the recovery of SOM VOL. 42, NO. 5, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1411

FIGURE 3. Total carbon mineralized at the end of the incubation (a) and labile carbon contents (b) of soils from Bárbara (B), Ramón (R), Volcanes (V) and Yaques (Y) at different stages of the complete fallow-cultivation cycle. VP: virgin páramo; R: recently ploughed; C1 and C2: 1 and 2 y culture soils; F1, F4, F5, and F8: 1, 4, 5, and 8 y fallow soils, respectively (NE: North East orientation, SW: South West orientation). Lines are third-order polynomial regressions. mineralization activity, followed by a slight decrease after 8 y of fallow. SOM and C mineralization decreases have also been reported for Venezuelan páramo soils after 3 y of potato crop (24). When multiple correlation analyses were applied to all plots from the four sectors, some relationships between soil CO2 emissions and several soil characteristics were found. Potential soil CO2 effluxes correlated negatively with Al3+ (P < 0.05) and positively with soil C (P < 0.001), N (P < 0.001), WHC (P < 0.001), and CEC (P < 0.001). Other significant relationships were not consistent for all sectors when analyzed together but appeared in some cases when sectors were separately considered. For example, although Ramón and Bárbara soil CO2 effluxes were positively correlated with soil pH (P < 0.01), this correlation was not significant for Volcanes and Yaques. Experimental curves of soil CO2 emissions always exhibited a progressive slant decline during incubation and fitted well to the double exponential model, with R2 values higher than 0.997, although some differences in the SOM mineralization kinetics were observed among sectors (Figure 2). Values of the kinetic parameters estimated according to this 1412

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 5, 2008

model (Tables 1, 2, and 3) show a wide range of variation for the labile C pool (C0) that goes from 0.10 (Y-R) to 0.59 g kg-1drysoil (R-VP). This variability of the labile organic content (Figure 3b) is related with slope orientation; NE-aspect sectors showed high C0 values in virgin páramo and longest fallow plots, whereas SW-aspect sectors showed similar labile C contents in all crop and fallow plots. Topographic aspect induces microclimatic differences, which along with chemical and physical soil characteristics, regulate SOM decomposition (23). Behavior resemblances of both soil labile C and potential soil CO2-C effluxes (Figure 3) was confirmed by the strong positive correlation found between them (Figure 4a). Total soil C, N, WHC, and CEC, variables previously found to be correlated with soil CO2 emissions, were consequently positively correlated with the kinetic parameter C0 (P < 0.01). Instantaneous mineralization rate of the labile pool (k), which showed a considerable range of variation (from 0.04 to 0.21 d-1 for B-C2 and Y-F4, respectively), positively correlated with soil pH (P < 0.05) and negatively with C0 (P < 0.05), indicating that the labile organic pool not only varied in size but also in quality. Instantaneous mineralization rate of the

FIGURE 4. Relationships found between the labile carbon contents and the potential soil CO2 emissions (a) and between the instantaneous mineralization rate of the recalcitrant carbon pool h and the soil CO2 emission coefficient (b) for all plots from Bárbara, Ramón, Volcanes and Yaques sectors. recalcitrant pool (h) varied between 1.25 × 10-4 (R-F8) to 2.85 × 10-4 d-1 (B-F1), and although its variability was much lower than for k, its importance is not negligible because recalcitrant C always represents a large proportion of the total SOM and exhibited significant negative correlations with soil C (P < 0.05) and Al2O3 contents (P < 0.01). Soil CO2 emission coefficients were also low for agricultural plots, because even in the most active sector (Volcanes) this coefficient did not surpass 3% of the total soil C (Tables 2 and 3). This coefficient correlated negatively with soil C (P < 0.01), N (P < 0.05), and Al2O3 (P < 0.001) and positively with soil pH (P < 0.001). The CO2 emission coefficient mainly depends on the accessibility of the more recalcitrant SOM since a very strong positive correlation was found between this index and the instantaneous mineralization rate of the recalcitrant pool (Figure 4b) kinetic parameter previously mentioned to be negatively correlated with Al2O3. This supports the important role of Al2O3 content on SOM stability, possibly by decreasing the recalcitrant pool biodegradability (25, 26). To establish the significance of the differences among the CO2 emission coefficients of plots from the same sector, an ANOVA was applied. In Bárbara sector B-R, B-C1 and B-F1 exhibited values significantly lower than those of B-VP, B-F8, and B-C2 plots (P < 0.05), whereas B-F5 showed intermediate values. The unexpected high values of the CO2 emission coefficient exhibited by B-C2 are undoubtedly the consequence of its notably lower SOM content in comparison with the rest of the plots from this sector. The Ramón sector, with a narrower range of values, also exhibited significant differences among plots in their CO2 emissions, with the R-F5 plot showing a CO2 emission coefficient significantly higher than the rest of plots from this sector (P < 0.05), with

the exception of R-VP, which did not differ significantly from any of them. The Volcanes plots showed the highest differences in CO2 emission coefficients; again, the recently ploughed plot (V-R) was one of the less-active treatments, together with V-C2, whereas V-F4 was the highest one (P < 0.05), with the rest of plots exhibiting intermediate values. Finally, in the Yaques sector, Y-R showed significantly lower values than the other plots, whereas Y-F4, Y-F8, and Y-VP presented values significantly higher than the rest of plots (P < 0.05), with Y-F1 and both Y-C1 and Y-C2 plots showing in-between emission coefficients. To determine the main factors influencing soil CO2 emissions, a PCA was applied using variables related to soil organic and mineralogical composition. The three PCA first components explained more than 70% of the total variance, with the two first components being associated to SOM characteristics and the third one to soil mineralogical composition. Figure 5a shows the variables in the space outlined by the two first components, which explained more than 45% of total variance, the first one being defined by potential soil CO2 emissions associated with SOM contents (C and N) and WHC as well as labile C on its positive extreme and opposite, on the negative arm, by exchangeable Al3+. Component II shows a strong polarization; the CO2 emission coefficient, along with instantaneous mineralization rate of the recalcitrant C pool and soil pH, are in the negative side, and the exchangeable H+ and Al2O3 are in the positive one. These results confirm the importance of SOM and its biochemical characteristics in the explanation of the total variance, and they corroborate that the soil aluminum contents have a strong influence on SOM stabilization, thus having a direct effect on the potential soil-atmosphere CO2 exchanges. Figure 5b shows the plots distribution with regard to the first two components, and the location on this plane allows plot differentiation according to their high or low potential CO2 emissions. It was observed that the grouping of plots depends on both the sector and the stage in the cultivation-fallow cycle. Thus, Bárbara and Ramón plots (NE) were generally located in the positive arm of factor I due to their high SOM content and relatively high CO2 emissions, whereas Volcanes and Yaques plots (SW) appeared mainly in the negative one. Besides, most virgin páramo plots and F4-F5 plots appeared in the spatial region determined by positive values of factor I and negative values of factor II (Figure 5b, below the dashed line) as a consequence of their higher SOM mineralization activity. The results confirm earlier findings (7) that soil CO2 emissions from Northern Andean páramo ecosystems to the atmosphere mainly depend on SOM biological accessibility. Our study elucidates that such accessibility was largely determined by two key factors: (i) the topographic aspect that strongly controls the SOM content along with the weights of labile and recalcitrant C pools and (ii) soil aluminum that acts as a chemical SOM stabilizing agent. Moreover, cultivation-fallow cycle monitoring makes clear that this type of agriculture merely produces moderate changes in SOM quality and, consequently, the CO2 emission capacity of these soils remains only temporary altered, mainly due to both the previously quoted SOM stability and the low-impact characteristics of agricultural systems with long fallow periods that permit soil recovery. Therefore, potential soil CO2 emission quantification in these highly sensitive mountain ecosystems provides support information for better management practices and contributes to the evaluation of the role of páramo ecosystems in the global terrestrial C cycle, not only in the contemporary climatic conditions but also in a global warming scenario. VOL. 42, NO. 5, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1413

FIGURE 5. Planes defined by factors I and II of the principal components analysis. (a) Variable weightings: C, soil carbon content; N, soil nitrogen content; C/N, C/N ratio; pH, pH in H2O; Fe2O3, Fe oxides; Al2O3, Al oxides; WHC, water holding capacity; H+, exchangeable hydrogen; Al3+, exchangeable aluminum; CO2, potential soil CO2 emissions; CO2 coef., soil CO2 emission coefficient; Co, soil labile carbon pool; k, instantaneous mineralization rate for the labile organic fraction; h, instantaneous mineralization rate for the recalcitrant organic fraction. (b) Plots’ distribution: Bárbara (B), Ramón (R), Volcanes (V), and Yaques (Y) at different stages of the complete fallow-cultivation cycle (VP, virgin páramo; R, recently ploughed; C1 and C2, 1 and 2 y culture soils; F1, F4, F5, and F8: 1, 4, 5, and 8 y fallow soils, respectively).

Acknowledgments This research was framed within European INCO-DC programme (No. ERBIC18CT98-0263) and was supported by CYTED (project XII.4). Field database with known crop-fallow status pertained to J. Smith; we thank her for allowing us to use it. We thank J. Salmonte, R. Tovar, and N. Leite for technical assistance and Dr. González-Prieto and J. Abadín for soil sampling and characterization.

Literature Cited (1) Davidson, E. A.; Ishida, F. Y.; Neptad, D. C. Effects of an experimental drought on soil emissions of carbon dioxide, methane, nitrous oxide and nitric oxide in a moist tropical forest. Global Change Biol. 2004, 10, 718–730. (2) Li, Y.; Xu, M.; Zou, X.; Zhang, Y. Comparing of soil organic carbon dynamics in plantations and secondary forests in wet tropics in Puerto Rico. Global Change Biol. 2005, 11, 239–248. (3) Epron, D.; Nouvellon, Y.; Deleporte, Ph.; Ifo, S.; Kazotti, G.; ThongoM’Bou, A.; Mouvondy, W.; SaintAndre, L.; Roupsard, O.; Jourdan, Ch.; Hamel, O. Soil carbon balance in a clonal Eucalyptus plantation in Congo: effects of logging on carbon inputs and soil CO2 efflux. Global Change Biol. 2006, 12, 1021– 1031. 1414

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 5, 2008

(4) Pansu, M.; Sarmiento, L.; Metselaar, K.; Hervé, D.; Bottner, P. Modelling the transformations and sequestration of soil organic matter in two contrasting ecosystems of the Andes. Eur. J. Soil Sci. 2007, 58, 775–785. (5) Sarmiento, L; Monasterio, M.; Montilla, M. Succession, regeneration and stability in high Andean ecosystems and agroecosystems: the rest-fallow strategy in the Páramo de Gavidia, Mérida, Venezuela. Geogr. Bernensia 1991, 151–157. African studies Series A8. (6) Abadin, J.; González-Prieto, S. J.; Sarmiento, L.; Villar, M. C.; Carballas, T. Successional dynamics of soil characteristics in a long fallow agricultural system of the high tropical Andes. Soil Biol. Biochem. 2002, 34, 1739–1748. (7) Coûteaux, M. M.; Sarmiento, L.; Bottner, P.; Acevedo, D.; Thiéry, J. M. Decomposition of standard plant material along an altitudinal transect (65–3968 m) in the tropical Andes. Soil Biol. Biochem. 2002, 34, 69–78. (8) Sarmiento, L.; Llambi, L. D.; Escalona, A.; Marquez, N. Vegetation patterns, regeneration rates and divergence in an old-field succession of the high tropical Andes. Plant Ecol. 2003, 166, 63–74. (9) Ramakrishnan, P. S. Shifting Agriculture and Sustainable Development: An Interdisciplinary Study from North-eastern India. United Nations Educational, Scientific and Cultural Organization (UNESCO): París, 1992.

(10) Floret, C.; Pontanier, R.; Serpantie, G. La Jachère en Afrique Tropicale (Dossier MAB 16). United Nations Educational, Scientific and Cultural Organization (UNESCO): Paris, 1993. (11) Sarmiento, L.; Monasterio, M.; Montilla, M. Ecological bases, sustainability and current trends in traditional agriculture in the Venezuelan high Andes. Mt. Res. Dev. 1993, 13, 167–176. (12) Pestolazzi, H. Sectoral fallow systems and the management of soil fertility: the rationality of indigenous knowledge in the high Andes of Bolivia. Mt. Res. Dev. 2000, 20, 64–71. (13) Conklin, H. Hanunoo Agriculture; Food and Agricultural Organization of the United Nations: Roma, 1957. (14) Raintree, J. B.; Warner, K. Agroforestry pathways for the intensification of shifting cultivation. Agrofor. Syst. 1986, 4, 39– 54. (15) Ingram, J.; Swift, M. Tropical Soil Biology and Fertility (TSBF) Programme: Report of the Fourth TSBF Interregional Workshop, Harare, Zimbabwe 1988. In Biology International,Special Issue 20; International Union of Biological Sciences: Paris, 1989. (16) Smith, J. K. Die Auswirkungen der Intensivierung des Ackerbaus im Paramo de Gavidia-Landnutzungswandel an der oberen Anbaugrenze in den venezolanischen Andean, M. Sc. Thesis, Geographisches Inst., Universität Bonn, Bonn, 1995. (17) Guckert, A.; Roger, P.; Jacquin, F. Contribution a l’utilisation de techniques radioisotopiques pour l‘étude de la matière organique du sol. Bull. l’ENSAIA Nancy 1968, 10, 69–100. (18) Fernández, I.; Cabaneiro, A.; Carballas, T. Organic matter changes immediately after a wildfire in an atlantic forest soil and comparison with laboratory soil heating. Soil Biol. Biochem. 1997, 29, 1–11. (19) Standford, G; Smith, S. J. Nitrogen mineralization potentials of soils. Soil Sci. Soc. Am. Proc. 1972, 36, 465–472.

(20) Andren, O.; Paustian, K. Barley straw decomposition in the field: a comparison of models. Ecology 1987, 68, 1190–1200. (21) Updegraff, K.; Pastor, J.; Bridgham, S. D.; Johnston, A. Environmental and substrate controls over carbon and nitrogen mineralization in northern wetlands. Ecol. Appl. 1995, 5, 151– 163. (22) Duchaufour, P. Pedologie. 1. Pédogenèse et classification; Masson: Paris, 1977. (23) Yimer, F.; Ledin, S.; Abdekadir, A. Soil organic carbon and total nitrogen stocks as affected by topographic aspect and vegetation in the Bale Mountains, Ethiopia. Geoderma 2006, 135, 335–344. (24) Sarmiento, L.; Bottner, P. Carbon and nitrogen dynamics in two soils with different fallow times in the high tropical Andes: indications for fertility restoration. Appl. Soil Ecol. 2002, 19, 79–89. (25) Fernández, I.; Cabaneiro, A.; González-Prieto, S. J. Partitioning CO2 Effluxes from an Atlantic Pine Forest Soil between Endogenous Soil Organic Matter and Recently Incorporated13CEnriched Plant Material. Environ. Sci. Technol. 2006, 40, 2552– 2558. (26) Boudot, J. P.; Bel Hadj, B. A.; Choné, T. Carbon mineralization in Andosols and aluminium-rich soils. Soil Biol. Biochem. 1986, 18, 457–461. (27) Fernández, I.; Cabaneiro, A.; Carballas, T. Carbon mineralization dynamics in soils after wildfires in two Galician forests. Soil Biol. Biochem. 1999, 31, 1853–1865. (28) Boudot, J. P.; Bel Hadj, B. A.; Steiman, R.; Seigle-Murandi, F. Biodegradation of synthetic organo-metallic complexes of iron and aluminium with selected metal to carbon ratios. Soil Biol. Biochem. 1989, 21, 961–966.

ES071392D

VOL. 42, NO. 5, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1415