PR OO F

Agriculture, Ecosystems and Environment 1803 (2001) 1–21

Carbon, nitrogen and phosphorus allocation in agro-ecosystems of a west African savanna Part III: plant and soil components under continuous cultivation Raphaël J. Manlay a,b,∗ , Jean-Luc Chotte a , Dominique Masse a , Jean-Yves Laurent c , Christian Feller d

a Institute for Research and Development (IRD, ex-ORSTOM), BP1386 Dakar, Senegal Institute of Forestry, Agricultural and Environmental Engineering (ENGREF), BP 5093, 34033 Montpellier Cedex 1, France c Institute for Research and Development (IRD, ex-ORSTOM), BP 5045, 34032 Montpellier, France d Centro de Energia Nuclear na Agricultura (CENA-USP), Institute for Research and Development (IRD, ex-ORSTOM), Caixa Postal 96, 13400-970 Piracicaba, SP, Brazil

ED

b

Received 4 February 2000; received in revised form 20 March 2001; accepted 27 March 2001

Abstract

CO RR E

CT

Carbon (C) and associated nutrient budgets related to land use in agro-ecosystems in west African savannas (WAS) are a matter of both local (sustainability of farming systems) and global (C balance) concern. In a mixed-farming system in southern Senegal, patterns of C, nitrogen (N) and phosphorus (P) allocation in the plant–soil system (down to a 40 cm soil depth) were compared at harvest in 14 plots, six being under semi-permanent cultivation with groundnut (Arachis hypogaea L.), others being under continuous cultivation with millet (Pennisetum glaucum L.), maize (Zea mays L.) or rice (Oryza sativa L.). Carbon stored in the plant–soil system amounted to 25.0, 27.4, 34.9 and 71.9 t per ha, respectively, in groundnut, millet, maize and rice fields. Ninety percent of C and P (total in plant Pt , available P in soil (POD )) and 95% of N of the whole ecosystem were stored in the soil. The high C and nutrient amounts found in rice plots were attributed to the clayey texture of the soil and to seasonal flooding. The lower values for C, N and POD found in soils in the bush ring (groundnut crops) compared to those of the compound ring (millet and maize crops) stemmed from land management. Higher values for C, N and POD in soils in the compound ring were maintained under continuous cultivation thanks to higher organic and nutrient inputs originating from crop residue recycling, manuring and, in the maize plots, spreading of household wastes. In the compound ring, the amount of C stored seemed to depend as much on the amount of C input as on the chemical richness of organic inflow. The effect of land management (bush versus compound ring) on soil properties was generally restricted to the 0–20 cm layer (except for P, cations and pH), and the better soil status in the compound ring was linked to nutrient depletion of the bush ring. From the perspective of global change, the estimated potential of the WAS for C sequestration under continuous cultivation was found to be low. From a methodological point of view, soil carbon status may be considered as a relevant indicator for the fertility of agro-ecosystems in the WAS belt, provided that its biotic components are included, and that both the quality and

∗ Corresponding author. Tel.: +33-467-47-121; fax: +33-467-47-101. E-mail address: [email protected] (R.J. Manlay).

UN

0167-8809/01/$ – see front matter © 2001 Published by Elsevier Science B.V. PII: S 0 1 6 7 - 8 8 0 9 ( 0 1 ) 0 0 2 2 0 - 1

2

R.J. Manlay et al. / Agriculture, Ecosystems and Environment 1803 (2001) 1–21

PR OO F

dynamics of soil organic matter (SOM) (assessment of seasonal variations, and C flows) and soil texture are characterised. © 2001 Published by Elsevier Science B.V. Keywords: Carbon; Continuous cultivation; Manure; Nitrogen; Phosphorus; Plant biomass; Root; Savanna; Senegal; Soil

UN

ED

CO RR E

The study of the carbon (C) cycle and related nutrient dynamics in agro-ecosystems of the west African savanna (WAS) is a matter of both local and global concern, since (1) the sustainability of farming systems in this region relies to a large extent on the integration of crop–livestock and fallowing practices as a way to cycle organic matter (OM) produced on-site (Ker, 1995) and (2) farming systems can act as a source or sink of atmospheric carbon. Both semi-permanent and continuous cultivation coexist in many village territories in the WAS. This is achieved within a compound centred, ring management scheme. However, increasing needs for more cropland call for a shift in traditional practices, which in the current context would lead to annual losses of more than 20–25 kg N and 2.5 kg P per ha in the fields of sub-Saharan Africa (Stoorvogel et al., 1993a). There is much evidence that N and P availability acts as the main chemical factor limiting crop yield in the WAS (Bationo et al., 1998), and past experience shows that at least a part of their supply to the plant should be organically mediated (Pieri, 1989). The conservation of multi-purpose improved savanna fallows is a pillar of the sustainability of farming systems of the WAS. However, intensification is likely to rely on better management of biogeochemical nutrient cycles through crop–livestock integration, mulching and agroforestry in the compound ring (Vierich and Stoop, 1990). Nutrient balances for cropping systems in dry west Africa have been established at local and national scale by Stoorvogel et al. (1993a,b). Nevertheless, these authors admit that further comprehensive local studies are needed to validate their own model and conclusions. Moreover, most studies dealing with the management of soil fertility in continuous and semi-permanent cropping systems in the sub-region were conducted in research centres (Pieri, 1989; Bationo et al., 1998), where environmental conditions (homogeneity of soil features, rational experimen-

tal design) are suitable for the accurate estimation of the parameters responsible for the efficiency of a given practice. Contrastingly, on-farm studies are less well documented, even though soil heterogeneity, pest hazard and a constraining agricultural timetable might offset the effect of a given practice on soil productivity. This work is the third and final part of a study to quantify C, N and P allocation in agro-ecosystems in southern Senegal. Previous papers focused on C, N and P allocation in plant biomass (Manlay et al., 2001a) and in soil (Manlay et al., 2001b) under semi-permanent cultivation. The present paper deals with systems under continuous cultivation. It aims at (1) providing detailed C, N and available P budgets for some plant-soil components; (2) assessing the relationships of soil organic matter (SOM) content with other soil physical and chemical properties, cropping intensity (bush versus compound field), and management of organic inputs (fallow, manure, household waste); (3) appraising the functional complementarity between continuous and semi-permanent cropping systems. The relevance of soil organic status as an indicator of soil quality and of sustainability of WAS cropping systems is discussed.

CT

1. Introduction

2. Methods The study site was described in Manlay et al. (2001a). Settled Peulh people, who are primarily herdsmen, have integrated extensive pastoralism, mostly cattle, in a diversified, partly continuous, partly semi-permanent agricultural system. 2.1. Sampling schemes Sampling was performed at harvest time in 1996 and 1997. Apart from maize that reaches maturity in September, harvest takes place in November and December at the beginning of the dry season. Maximum standing crops are usually recorded at this time, which

R.J. Manlay et al. / Agriculture, Ecosystems and Environment 1803 (2001) 1–21

2.2. Plant and soil analysis

PR OO F

2.3. Estimation of C, N and P inputs in compound fields Some of the annual C and nutrient inputs were estimated for millet and maize plots. For this purpose, some of the data (crop residue and root inputs) were obtained from the present study. Others derived from a more global study carried out at the village territory scale (Manlay, 2000). Briefly, manure inputs were computed by (1) recording livestock grazing trajectories during day straying and corralling places during the night; (2) measuring faecal production and OM intake related to animal weight; (3) analysing total C, N and P contents of dung every month during the dry season. Household wastes consisted of inedible components of cereal panicles, recycled thatch for roof construction, and fuelwood ashes. Figures were computed from enquiries and on-field measurements, waste spreading being assumed to happen on a 30 m-wide band round the compounds.

CO RR E

Cereals sown and fertilising patterns have been constant for more than 15 years in each selected cereal field. The same general sampling design was adopted for GN, MI, MA and RI plots and was described in Manlay et al. (2001a,b): for each plot four 4 m × 4 m subplots were randomly chosen and sampled for plant biomass and soil. Root and soil were sampled at 10 cm increments, down to a depth of 40 cm. Thorough inquiries were made among field owners to check that no mineral fertilizer had been used in the past 10 years, and to obtain an accurate picture of agricultural practices used in each plot. In Sare Yorobana, groundnut, millet, maize, sorghum and rice are usually planted at 0.2 m ×0.4 m, 1.0 m ×1.0 m, 0.5 m ×0.5 m, 1.0 m × 1.0 m and approximately 0.1 m × 0.1 m spacing, respectively.

ED

• Six bush fields cropped with groundnut (Arachis hypogaea L.) (GN). Four were in biennial rotation with fallow, the two others in irregular rotation with cereal and fallow. These plots were those sampled in Manlay et al. (2001a,b). • Four compound fields permanently cropped with millet (Pennisetum glaucum L.) (MI) with varying intensities of manuring during night corralling and day straying. • Two compound fields under maize (Zea mays L.) (MA), which benefit from household wastes and have the highest manuring rates. • Two seasonally flooded, down-slope rice (Oryza sativa L.) (RI) fields.

moval of organic cements as described in Gavinelli et al., 1995), available P as measured by the Olsen method modified by Dabin (1967) and noted POD ; pH in H2 O and in KCl, Ca, Mg, Na, K, cation exchange capacity or CEC, saturation rate S, mass of five size fractions from mechanical (particle-size) analysis after removal of organic cements (clay, fine and coarse silt denoted FSILT and CSILT, fine and coarse sand denoted FSAND and CSAND), bulk density DENS, mass water content at pF 2.5 and 4.2. Amounts (t per ha) of C, N, POD and Pt stored in soil and vegetation were computed using dry matter (DM) amounts or soil bulk density and C, N, POD and Pt contents shown in Tables 1 and 2. Because of possible bias of soil storage data due to differences in bulk density between plots, statistics on soil elements storage were computed for soil equivalent masses (Ellert and Bettany, 1995). The relative mass of soil fraction 0–50 ␮m obtained by particle-size fractionation was used as an estimate for the abundance of clay + silt size fractions at the subplot scale (see Section 2.4). Even though it is not strictly equivalent to that assessed by mechanical analysis, where organic cements are removed by oxidation, the difference is very small (Feller, 1995).

CT

coincides with the return of livestock kept in the peripheral rangeland during the cropping period. Fourteen fields were chosen for full C, N and P budget assessment. They represent the different types of land management along the toposequence (Fig. 1).

UN

Plant and soil analyses were detailed in Manlay et al. (2001a,b). Variables measured on plant biomass were: dry mass, C, N and total P (denoted Pt ) contents of pod/panicle, stover, bush regrowth and weed, fine and coarse (diameter below and above 2 mm) root, and stump (GN plots only) biomass. Soil variables were: C and N content of non-fractionated soil and of 0–50 and 50–2000 ␮m soil fractions obtained by particle-size fractionation (without re-

3

R.J. Manlay et al. / Agriculture, Ecosystems and Environment 1803 (2001) 1–21

CO RR E

CT

ED

PR OO F

4

UN

Fig. 1. Amounts of carbon, nitrogen and phosphorus in plant and soil of main cash and food crops along a typical toposequence in Sare Yorobana, southern Senegal: (1) used in 3.5. Bu (bush), Com (compound): ring attachment, Fa (fallow rotation), Cor (night corralling), Wa (household waste spreading): practice of fertility management; (2) see Section 2.3 for estimation methods (ND: not determined).

Table 1 Biomass amounts (t DM per ha) in GN, MI, MA and RI plotsa Plot

GN01 GN02 GN03 GN04 GN05 GN06 MI01 MI02 MI03 MI04 MA01 MA02 RI01 RI02

Panicle/ pod

Stover/ haulm

Bush regrowth and weed

Fine roots per soil layer (cm) 0–10

10–20

20–30

30–40

0.66 1.18 0.73 1.24 0.67 1.12 1.41 1.85 2.43 2.44 3.66 5.61 1.30 4.61

1.52 1.86 1.14 1.75 1.18 1.51 4.38 6.28 9.82 9.62 4.71 5.21 2.16 4.45

0.32 0.39 0.75 1.66 1.14 0.82 0.27 0.39 1.32 1.29 1.38 0.97 0.09 0.40

0.19 0.20 0.25 0.33 0.21 0.29 0.32 0.34 0.53 0.70 0.16 0.17 2.91 1.55

0.15 0.17 0.16 0.18 0.09 0.18 0.19 0.23 0.32 0.24 0.05 0.06 0.87 0.36

0.11 0.10 0.12 0.11 0.08 0.10 0.08 0.12 0.16 0.07 0.05 0.05 0.33 0.15

0.07 0.07 0.08 0.07 0.05 0.07 0.11 0.09 0.09 0.07 0.04 0.05 0.26 0.12

5

PR OO F

R.J. Manlay et al. / Agriculture, Ecosystems and Environment 1803 (2001) 1–21

Coarse roots

Stump

Total (stump excluded)

5.18b 3.02b 3.13b 2.99b 2.95b 3.03b 0.24 0.10 0.03 0.07 0.04 0.05 0.26 0.03

NDc 13.8 5.0 3.5 ND ND 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

8.2 7.0 6.4 8.3 6.4 7.1 7.0 9.4 14.7 15.2 10.1 12.2 8.0 11.7

a

ED

Fine roots: diameter ranging 0–2 mm; coarse roots: diameter above 2 mm (stump not included). Computed by using a regression relationship linking coarse root biomass (in t DM per ha) as measured by full excavation (x) to that measured using the coring technique (y): x = 1.73 × y + 2.95; R 2 = 0.6; P {F obs > F th } < 0.05; n = 9 (Manlay et al., 2001a). c ND: not determined.

CT

b

Table 2 Carbon, N and Pt content of plant biomass components in groundnut, millet, maize and rice plots Pod, panicle or ear (mean ± S.E.)

C (g 100 g−1 DM) Groundnut 45.2 Millet 35.3 Maize 35.6 Rice 32.2

± ± ± ±

0.8 0.1 0.4 1.0

N (g 100 g−1 DM) Groundnut 2.87 ± 0.10 Millet 1.23 ± 0.09 Maize 1.21 ± 0.18 Rice 0.63 ± 0.05 Pt (g 100 g−1 DM) Groundnut 0.17 ± 0.01 Millet 0.24 ± 0.01 Maize 0.20 ± 0.05 Rice 0.14 ± 0.04 a

Haulm or stover (mean ± S.E.)

33.6 37.0 35.1 30.7

± ± ± ±

0.2 0.6 0.3 1.2

Bush regrowth and weed (mean ± S.E.) 35.7 35.2 26.4 32.6

± ± ± ±

0.3 0.3 1.5 0.1

CO RR E

Plot

Roots

Fine (mean ± S.E.) 34.1 35.1 34.4 29.8

± ± ± ±

0.5 1.0 1.0 0.2

Coarse (mean ± S.E.) 38.0a 35.1 ± 1.0b 34.4 ± 1.0b 29.8 ± 0.2b

1.70 0.27 0.80 0.41

± ± ± ±

0.06 0.03 0.01 0.01

0.83 1.30 1.58 0.91

± ± ± ±

0.06 0.08 0.11 0.01

1.66 1.04 1.31 0.76

± ± ± ±

0.03 0.04 0.08 0.07

0.35a 1.04 ± 0.04b 1.31 ± 0.08b 0.76 ± 0.07b

0.08 0.07 0.20 0.10

± ± ± ±

0.00 0.00 0.01 0.02

0.06 0.19 0.39 0.21

± ± ± ±

0.00 0.03 0.02 0.02

0.07 0.07 0.08 0.06

± ± ± ±

0.01 0.00 0.00 0.01

0.02a 0.07 ± 0.00b 0.08 ± 0.00b 0.06 ± 0.01b

UN

Estimated as the mean value measured on coarse root biomass of three 1-year-old fallows (Manlay et al., 2001a). Value extrapolated from the content in fine root biomass. Fine roots: diameter ranging 0–2 mm; coarse roots: diameter above 2 mm (stump not included). Groundnut: n = 6; millet: n = 4; maize: n = 2; rice: n = 2. b

R.J. Manlay et al. / Agriculture, Ecosystems and Environment 1803 (2001) 1–21

2.4.1. Multivariate analysis Rapid assessment of relationships between soil variables and cropping practices was done using both correlation and PCA. Data organisation was detailed in Manlay et al. (2001b); dimensions (variables × plot individuals) of data matrices were 23 × 12 and 21 × 12, for soil layers above and below 20 cm, respectively.

3.1. Amounts of carbon, nitrogen and phosphorus in plant biomass of cropped fields Above- and below-ground biomass (AGB and BGB) of cropped fields ranged from 7.2 (in GN plots) to 11.6 (in MI plots) t DM per ha (Table 1). As much biomass was stored in stumps in GN plots (7.5 t DM per ha) as in all remaining plant components (cereal fields being stumpless). Carbon content averaged 34.3 g 100 g per DM, with the lowest values mostly found in rice components (Table 2). The highest N contents were measured in AGB and fine roots of groundnut, highest Pt contents in maize roots. Highest amounts of N and P were found in the biomass of MA fields (stumps in GN plots not included), and more N was stored in the biomass of GN plots than that of MI and RI plots. Vertical partitioning of biomass had the following features: (1) shoot:root ratio was less than one for GN fields, whereas it was over one for cereal plots, equal to, respectively, 2.5, 8.9 and 31.6 for RI, MI, and MA; (2) bush regrowth and weeds accounted for 25 (GN), 7 (MI), 11 (MA) and 3% (RI) of total AGB; (3) panicles made up only 21% of millet standing crop biomass.

CO RR E

2.4.2. Univariate analyses As explained in Manlay et al. (2001b), ANOVA were performed on ranks of data and texture was introduced as a covariate in the linear model used for ANOVA. Management × texture interactions were not introduced in the model, since they did not significantly account for the variations in soil properties (data not shown). Two sets of ANOVA were performed. One was aimed at relating soil properties to ‘cultivation intensity’ (bush versus compound fields) and used field plot composites (from subplot replicates) (texture covariate: relative mass of the clay + fine silt fraction, mechanical analysis). The second set of ANOVA was performed on field subplot samples to evaluate the effect of organic management of fertility on SOM status (texture covariate: relative weight of the 0–50 ␮m fraction, particle-size fractionation). Five patterns of soil fertility management that had lasted for more than 10 years were distinguished (see Fig. 1 for correspondence of plot coding, location and amounts of waste and manure inputs).

3. Results

UN

• Bush fields: only fallowed (BuFa), or slightly manured (during night corralling of cattle) (BuCor). • Compound fields: never manured (i.e. never corralled) (Com); manured without household wastes

PR OO F

All statistical analyses were done using SAS software 6.14 (Hatcher and Stepanski, 1994) (procedures GLM and CORR) except principal component analyses (PCA) with ADE-4 software (Thioulouse et al., 1997). Because, Gleysols of the paddy fields were too different from soils of the glacis and plateau, they were not considered for multivariate analyses and analyses of variance (ANOVA).

(ComCor); manured with household wastes (ComCorWa). • Pair-wise T-test (α = 0.05) on least square (LS) means was used to identify management practices that have a similar impact on soil C content and storage.

ED

2.4. Data analysis

CT

6

3.2. Amounts of carbon, nitrogen and phosphorus in the soil of cropped fields

Carbon stored in the soil ranged from 22.3 (GN) to 68.7 (RI) t per ha in the 0–40 cm layer (Fig. 1). Carbon concentration in the 0–20 cm layer was low (GN) to marked (MA). Nitrogen and POD distribution among layers and plots was very similar to that of C. Considering the whole plant–soil system, carbon storage amounted to 25.0, 27.4, 34.9, and 71.7 t per ha in the groundnut, millet, maize and rice fields, respectively. Ninety percent of C and P (Pt + POD ), and 95% of the N of the whole ecosystem were stored in the soil.

R.J. Manlay et al. / Agriculture, Ecosystems and Environment 1803 (2001) 1–21

3.4.1. Chemical and physical properties ANOVA confirmed some of the multivariate analyses findings. Clay + fine silt content was higher in the bush than in the compound fields, especially below 20 cm (Table 3). However, no significant differences were found for other physical properties (except bulk density). All chemical properties (except Na, data not shown) were significantly improved in the 0–10 cm soil layer of the compound fields (Table 4). This was also the case for POD , pH, Mg, K and S in deeper layers. In the 20–40 cm layer, a textural effect was recorded for all chemical properties except POD , Ca, K and S. Consequently, higher C, N contents and CEC values were found in the deepest layers of bush plots. 3.4.2. SOM quality Fine and coarse soil size fractions (Table 5) contributed to the same extent to the increase in total soil organic content (SOC) in compound fields and bush fields. However, relative gains differed between fractions. Carbon content (g kg−1 fraction) improved by 40% between bush and compound fields in the 0–50 ␮m fraction, and by 150% in the coarse fraction. This differential evolution was particularly clear for the maize fields, in which relative gains reached 70 and 300% for the fine and coarse size fractions, respectively. Thus, the increase in total soil C originated mainly from the coarse fraction. C:N varied considerably between fractions, but was not affected by cultivation intensity (Table 5).

CO RR E

3.3.2. The 10–20 cm soil layer Carbon (total and in fractions) and N were not linked to any other variable, while clay was negatively correlated with POD , pH, cations (except Mg) and S. PCA yielded roughly the same results as for the 0–10 cm layer (Fig. 2b1 and 2), but OM variables provided lower loadings to the first PC.

PR OO F

3.3.1. The 0–10 cm soil layer POD content was positively correlated to C (total and in fractions) and N (data not shown). The pH, cations (except Na), CEC and S were related to C, N and POD . Variations in these chemical properties were better linked to the C content of the 0–50 ␮m fraction than to that of the 50–2000 ␮m fraction. Fine elements (clay + fine silt) were correlated with water content only. The first principal component (PC) was clearly defined as ‘SOM-related status’, chemical variables providing the highest loadings (Fig. 2a1). Maize plots were associated with the highest organo-mineral status and GN with the poorest; as in all other layers, MI plots were middle-located along PC1 (Fig. 2a2). Contrastingly, MA, MI and GN clusters did not differ much in their texture, as expressed by their location along the second PC.

3.4. Influence of cropping intensity on soil organic status and other soil properties

ED

Correlation analysis exhibited the following statistically significant links between variables (P{H0 :R s = 0} < 0.05).

POD contents and pH found in MA fields; (3) although nutrient status of MI plots was generally higher than that of GN fields, great variations were found in chemical properties between MA and MI plots.

CT

3.3. Soil organic matter status related to other soil properties

UN

3.3.3. The 20–30 and 30–40 cm soil layers Organic status was positively linked to fine element content, especially in the 30–40 cm layer. Differences in soil properties between plots were mainly fine element content and pH, as testified by loadings of the first PC (Fig. 2c1 and d1). The highest clay, C and N contents were found for GN plots, while MA fields had the best mineral status (Fig. 2c2 and d2). In summary: (1) in the 0–20 cm layer, soil chemical status was strongly controlled by SOM content, and was highest in maize (compound) fields and lowest in GN (bush) fields; (2) in the 20–40 cm layer, texture influenced SOM content, leading to higher C and N content in GN plots; however, other nutrient contents were independent of texture, with highest cation and

7

3.4.3. Soil C, N and P storage Increases in C and N amounts (expressed in soil equivalent masses) between bush and compound fields were significant down to a depth of 30 cm (Table 6). The increase in C was well-balanced between both soil fractions in the 0–10 and 0–20 cm layers (the coarse fraction accounting for half of the gains) (Table 7). Soil POD exhibited a clearer difference, since amounts

R.J. Manlay et al. / Agriculture, Ecosystems and Environment 1803 (2001) 1–21

CO RR E

CT

ED

PR OO F

8

UN

Fig. 2. PC analysis of the soil properties of 12 cropped plots. Correlation circles of the variables (a1, b1, c and d) and projection of the plot replicates (a2, b2, c2 and d2) on plane PC1 × PC2: (a) layer 0–10 cm; (b) layer 10–20 cm; (c) layer 20–30 cm; (d) layer 30–40 cm. RI: relative inertia of the PC. Coding of variables, C: carbon; CA: calcium; CCoarFra: carbon content of the 50–2000 ␮m fraction; CEC: cation exchange capacity; CFinFrac: carbon content of the 0–50 ␮m fraction; CLAY: clay; CLAYFSI: clay + fine silt; CLAYSI: clay + silt; CN: C:N ratio; CSAND: coarse sand; CSILT: coarse silt; DENS: bulk density; FSAND: fine sand; FSILT: fine silt; K: potassium; MG: magnesium; N: nitrogen; NA: sodium; POD : available phosphorus; pH H2O: pH in water; pH KCL: pH in KCl; pF 25 and 42: volumetric water content determined at a suction equivalent to pF 2.5 and 4.2; S: saturation rate; SAND: sand; SI: silt. Coding of plot replicates, GN: groundnut crop; MI: millet crop; MA: maize crop.

9

CO RR E

CT

ED

PR OO F

R.J. Manlay et al. / Agriculture, Ecosystems and Environment 1803 (2001) 1–21

Fig. 2. (Continued).

UN

of POD in the compound fields were about six times higher than in the bush fields, irrespective of the soil layer. Maize plots, which had received the highest rates of manure or household waste, accounted for the majority of these differences: their total amounts of C, N and POD were, respectively, 1.3, 1.4 and 3.3 higher than those of millet crops.

3.5. Influence of the management of organic inputs on soil organic status Significant differences in soil physical properties related to the five organic practices were evidenced at the subplot level, especially in soil bulk density (Table 8). In the 0–10 cm layer, treatments ranked as follows with regard to carbon content: ComCorWa >

10

R.J. Manlay et al. / Agriculture, Ecosystems and Environment 1803 (2001) 1–21

Layer (cm)

Bush fields mean (±S.E.) (n = 6)

Clay + fine silt content (%) 0–10 11.3 ± 10–20 17.1 ± 20–30 22.5 ± 30–40 29.4 ±

1.2 1.3 2.5 4.3

Compound fields mean (±S.E.) (n = 6)

10.6 14.5 18.0 21.6

± ± ± ±

F

Cultivation intensity

1.0 0.9 1.2 2.3

0.1 2.3 1.3 1.8

1.02 1.13 1.42 2.03

6.03 5.82 6.28 7.58

± ± ± ±

0.44 0.59 0.78 0.99

0.0 0.1 0.1 0.7

pF 4.2 (g H2 O 100 g−1 soil) 0–10 2.70 ± 10–20 4.03 ± 20–30 5.35 ± 30–40 7.30 ±

0.34 0.46 0.75 1.22

2.82 3.03 3.78 4.82

± ± ± ±

0.37 0.23 0.28 0.66

0.5 0.1 2.1 0.9

Bulk density (kg DM−3 ) 0–10 1.52 10–20 1.47 20–30 1.54 30–40 1.45

0.01 0.03 0.03 0.03

1.47 1.54 1.54 1.56

± ± ± ±

0.02 0.01 0.01 0.01

P{H0 :F obs > F th = 0} < 0.05. P{H0 :F obs > F th = 0} < 0.01. ∗∗∗ P{H :F 0 obs > F th = 0} < 0.001. ∗∗

UN

Overall

7.3∗ 15.2∗∗ 45.2∗∗∗ 134.0∗∗∗

3.7 8.9∗∗ 26.1∗∗∗ 83.6∗∗∗

33.1∗∗∗ 24.9∗∗∗ 36.0∗∗∗ 184.2∗∗∗

16.6∗∗ 16.4∗∗ 24.9∗∗∗ 114.6∗∗∗

2.6 0.3 0.1 0.4

3.6 2.7 0.4 13.3∗∗

the soil. Storage of SOC was further improved when nutrient-rich household waste was added.

CO RR E

Com = ComCor > BuFa = BuCor. In the 10–20 cm layer, ComCorWa had the highest organic content, while other treatments exhibited similar values. Trends became less obvious below this depth, due to textural differences between treatments. As a matter of fact, the contribution of management to SOC variability was significant for layers 0–10 and 20–30 cm only. Trends in carbon storage were similar to those in C content. Carbon, N and P fluxes linked to organic inputs were computed for the Com, ComCor and ComCorWa plots (Fig. 3). Big differences (ratios of 1–4 and 6) in C input were recorded between the three treatments. Due to different N and P contents in plant biomass, manure and household wastes, values for nutrient inputs varied even more widely, resulting in a decrease in C:N (from 49 to 18) and C:P (from 425 to 134) ratios from Com to ComCorWa plots. Manuring thus increased SOC content only in the soil surface layer, as well as C and nutrient inputs to

5.1 5.2∗ 0.7 23.9∗∗∗

Texture

CT

∗

ED

pF 2.5 (g H2 O 100 g−1 soil) 0–10 6.95 ± 10–20 8.02 ± 20–30 8.80 ± 30–40 11.35 ±

± ± ± ±

PR OO F

Table 3 Effect of cultivation intensity and texture (clay + fine silt content) on soil physical properties

4. Discussion 4.1. Agroecological implications of plant biomass distribution in cropped fields In cropped fields in Sare Yorobana, much of the plant biomass was stored in vegetative components. This was particularly noticeable for roots and non-crop AGB of groundnut bush fields as a result of bush regrowth and weed competition, and to a lesser extent, for millet and rice fields. Harvest indexes were particularly low in cereals (0.13 in millet, assuming that grains accounted for 61% of panicle biomass, Manlay, 2000), compared to indexes usually found for cereals in northern agricultural systems (0.45, Smil, 1999).

11

PR OO F

R.J. Manlay et al. / Agriculture, Ecosystems and Environment 1803 (2001) 1–21

Table 4 Effect of cultivation intensity and texture (clay + fine silt content) on soil chemical propertiesa Layer (cm)

Bush fields mean (±S.E.) (n = 6)

Compound fields mean (±S.E.) (n = 6)

F

Cultivation intensity

Texture

Overall

0.9 0.8 1.4 134.1∗∗∗

15.6∗∗ 0.5 2.8 103.2∗∗∗

6.2∗ 1.0 9.5∗ 56.2∗∗∗

18.9∗∗∗ 0.5 7.2∗ 48.7∗∗∗

Carbon (g kg−1 ) 0–10 10–20 20–30 30–40

4.49 3.66 3.32 3.40

± ± ± ±

0.19 0.16 0.17 0.24

7.54 4.17 2.75 2.62

± ± ± ±

1.00 0.50 0.19 0.15

31.0∗∗∗ 0.7 2.3 8.3∗

Nitrogen (g kg−1 ) 0–10 10–20 20–30 30–40

0.36 0.31 0.30 0.32

± ± ± ±

0.02 0.02 0.02 0.03

0.65 0.34 0.24 0.24

± ± ± ±

0.11 0.04 0.02 0.03

33.6∗∗∗ 0.2 1.1 7.3∗

PhosphorusOD (10−3 g kg−1 ) 0–10 2.52 ± 10–20 1.80 ± 20–30 1.73 ± 30–40 1.52 ±

0.21 0.20 0.21 0.15

16.32 11.08 8.62 8.30

± ± ± ±

6.43 4.20 2.64 2.47

22.9∗∗ 20.4∗∗ 19.2∗∗ 19.4∗∗

pH (H2 O) 0–10 10–20 20–30 30–40

5.92 5.44 5.23 5.20

± ± ± ±

0.08 0.13 0.17 0.17

6.52 6.45 6.25 6.07

± ± ± ±

0.06 0.09 0.16 0.20

pH (KCl) 0–10 10–20 20–30 30–40

5.15 4.63 4.44 4.34

± ± ± ±

0.12 0.17 0.17 0.17

5.93 5.69 5.42 5.21

± ± ± ±

0.12 0.14 0.20 0.24

Ca (meq 100 g−1 soil) 0–10 1.32 ± 0.13 10–20 1.06 ± 0.17 20–30 0.99 ± 0.13 30–40 1.03 ± 0.14

2.75 1.84 1.35 1.20

± ± ± ±

0.39 0.27 0.16 0.15

1.03 0.69 0.63 0.66

± ± ± ±

0.25 0.21 0.22 0.24

± ± ± ±

± ± ± ±

0.03 0.04 0.04 0.05

UN

K (meq 100 g−1 soil) 0–10 0.04 ± 0.00 10–20 0.04 ± 0.00 20–30 0.03 ± 0.00 30–40 0.03 ± 0.00

55.5∗∗∗ 26.4∗∗∗ 58.2∗∗∗ 7.5∗

10.2∗ 6.7∗ 21.5∗∗ 4.3

34.8∗∗∗ 27.5∗∗∗ 58.5∗∗∗ 9.5∗∗

18.0∗∗ 25.2∗∗∗ 10.6∗ 5.1

4.0 5.7∗ 10.0∗ 2.7

11.7∗∗ 25.4∗∗∗ 15.5∗∗ 6.3∗

13.2∗∗ 4.6 3.2 0.2

0.0 1.3 0.7 0.3

6.7∗ 5.0∗ 2.8 0.2

0.13 0.09 0.06 0.05

28.1∗∗∗ 26.8∗∗∗ 47.0∗∗∗ 80.0∗∗∗

0.1 0.5 4.7 12.8∗∗

14.1∗∗ 14.9∗∗ 23.5∗∗∗ 40.0∗∗∗

0.05 0.03 0.04 0.04

28.5∗∗∗ 21.6∗∗ 25.5∗∗∗ 28.2∗∗∗

0.1 0.1 0.1 0.6

14.2∗∗ 14.0∗∗ 14.0∗∗ 15.1∗∗

CT

ED

11.4∗∗ 19.2∗∗∗ 12.2∗∗ 15.3∗∗

CO RR E

Mg (meq 100 g−1 soil) 0–10 0.36 10–20 0.32 20–30 0.35 30–40 0.40

0.2 3.4 0.5 1.5

R.J. Manlay et al. / Agriculture, Ecosystems and Environment 1803 (2001) 1–21

PR OO F

12

Table 4 (Continued) Layer (cm)

CEC (meq 0–10 10–20 20–30 30–40

Bush fields mean (±S.E.) (n = 6) 100 g−1

soil) 2.31 2.49 2.70 2.95

± ± ± ±

Compound fields mean (±S.E.) (n = 6)

0.10 0.08 0.12 0.36

3.69 2.78 2.48 2.65

Saturation rate (%) 0–10 74.5 ± 5.2 10–20 57.3 ± 7.4 20–30 51.5 ± 4.9 30–40 54.8 ± 9.9

109.5 98.2 89.3 81.3

± ± ± ±

± ± ± ±

F

0.48 0.29 0.14 0.26

Cultivation intensity

Texture

17.2∗∗ 1.0 0.0 5.3∗

2.1 1.4 9.3∗ 107.7∗∗∗

9.3∗∗ 0.9 5.2∗ 55.6∗∗∗

2.4 3.1 4.3 3.3

18.8∗∗∗ 10.6∗∗ 22.9∗∗∗ 5.6∗

33.6∗∗∗ 9.5∗ 28.8∗∗∗ 3.7

3.8 5.0 6.8 5.7

a

Overall

Figures for Na not shown, since no significant change was found between bush and compound fields. P{H0 :F obs > F th = 0} < 0.05. ∗∗ P{H :F 0 obs > F th = 0} < 0.01. ∗∗∗ P{H :F 0 obs > F th = 0} < 0.001. ∗

Carbon content 0–10 10–20

Fraction (␮m)

Bush fields mean (±S.E.)

Compound fields mean (±S.E.)

F

Cultivation intensity

Texture

Overall

fraction) 0–50 50–2000

15.23 ± 0.77 1.11 ± 0.10

24.52 ± 2.55 3.59 ± 1.06

13.6∗∗ 20.7∗∗

0.3 4.2

7.1∗ 11.8∗∗

0–50 50–2000

10.13 ± 0.41 0.66 ± 0.19

11.16 ± 0.96 0.91 ± 0.22

0.1 0.7

0.1 0.5

0.2 0.4

3.14 ± 0.15 0.87 ± 0.07

5.12 ± 0.58 2.82 ± 0.80

1.9 2.9

12.2∗∗ 7.6∗

2.70 ± 0.10 0.48 ± 0.14

3.02 ± 0.39 0.65 ± 0.15

0.2 0.9

0.0 0.7

0.1 0.6

(g kg−1

10–20

0–50 50–2000

C/N on fractions 0–10 0–50 50–2000 10–20 ∗

0–50 50–2000

CO RR E

Carbon content (g kg−1 soil) 0–10 0–50 50–2000

CT

Layer (cm)

ED

Table 5 Effect of cultivation intensity and texture (clay + fine silt content) on SOM quality as assessed by C content in fine- and coarse-size soil fractions

23.3∗∗∗ 13.1∗∗

12.7 ± 0.3 29.4 ± 2.3

13.1 ± 0.4 24.5 ± 1.8

0.2 0.8

0.0 0.1

0.1 0.5

11.9 ± 0.5 35.2 ± 4.6

13.0 ± 0.7 28.6 ± 4.0

1.5 0.1

0.1 0.7

1.1 0.7

P{H0 :F obs > F th = 0} < 0.05. P{H0 :F obs > F th = 0} < 0.01. ∗∗∗ P{H :F 0 obs > F th = 0} < 0.001. ∗∗

UN

Heavy allocation to vegetative biomass may be seen as a limit to any improvement in grain yield. In practice, it provides farmers with various agro-ecological benefits, which may be classified as immediate and direct.

1. Immediate benefits (yielding commodities). The above-ground biomass of cropped fields is a versatile resource (fence making, roofing), but its main use is as forage for animals. Its high feed value is testified by livestock foraging behaviour in the

R.J. Manlay et al. / Agriculture, Ecosystems and Environment 1803 (2001) 1–21

13

Bush fields mean (±S.E.) (n = 6)

Carbon storage (t per ha) 0–10 6.9 0–20 12.5 0–30 17.5 0–40 22.6

± ± ± ±

Compound fields mean (±S.E.) (n = 6)

0.3 0.5 0.6 0.7

11.5 17.7 21.8 25.7

Nitrogen storage (kg per ha) 0–10 546 ± 26 0–20 1026 ± 52 0–30 1471 ± 78 0–40 1955 ± 117

983 1486 1847 2209

PhosphorusOD storage (kg per ha) 0–10 3.9 ± 0.3 0–20 6.6 ± 0.5 0–30 9.2 ± 0.7 0–40 11.5 ± 0.9

± ± ± ±

± ± ± ±

24.9 41.5 54.3 66.8

F

1.5 2.2 2.4 2.3

Texture

Overall

31.0∗∗∗ 18.7∗∗ 5.7∗ 3.1

0.9 0.3 0.1 0.9

15.6∗∗ 9.7∗∗ 3.0 1.6

4.8 3.4 1.8 2.8

16.6∗∗ 6.9∗ 3.0 1.9

0.2 0.2 3.0 2.4

11.4∗∗ 11.5∗∗ 19.5∗∗∗ 18.4∗∗∗

30.0∗∗∗ 13.3∗∗ 5.8∗ 2.8

164 216 225 221

± ± ± ±

Cultivation intensity

22.9∗∗ 19.1∗∗ 23.7∗∗∗ 21.2∗∗

9.8 16.0 19.8 22.9

a

Reference plot for equivalent soil mass calculation was set as GN03. P{H0 :F obs > F th = 0} < 0.05. ∗∗ P{H :F 0 obs > F th = 0} < 0.01. ∗∗∗ P{H :F 0 obs > F th = 0} < 0.001. ∗

ED

Layer (cm)

PR OO F

Table 6 Effect of cultivation intensity and texture (clay + fine silt content) on soil C, N and POD storage (computed in equivalent soil masses) for non-fractionated soila

Layer (cm)

Fraction (␮m)

Bush fields mean (±S.E.)

CT

Table 7 Effect of cultivation intensity and texture (clay + fine silt content) on soil C storage (t C per ha, computed in equivalent soil masses) for two soil-size fractionsa Compound fields mean (±S.E.)

F

Cultivation intensity

Texture

Overall

0–50 50–2000

5.40 ± 0.28 1.48 ± 0.10

7.51 ± 0.63 3.97 ± 1.01

10.0∗

0.2 0.0

6.9∗ 5.0∗

0–20

0–50 50–2000

10.19 ± 0.44 2.29 ± 0.28

12.73 ± 1.15 5.00 ± 1.19

7.0∗ 5.8∗

1.8 0.1

3.7 3.0

a

CO RR E

0–10

13.8∗∗

Reference plot for equivalent soil mass calculation was set as GN03. P{H0 :F obs > F th = 0} < 0.05. ∗∗ P{H :F 0 obs > F th = 0} < 0.01. ∗

UN

compound ring of the village within the first days of common grazing (Ickowicz et al., 1997). Feed value is not only related to DM amounts, but also to crop species and weed content. The nitrogen-rich tissues of the groundnut are harvested as haulms and used as a feed supplement for animals in the farmyard. The higher N and Pt contents in weeds than in stover (except in groundnut fields), and Lamers et al. (1996) indicate a much higher feed value for weeds than for millet mixed with these

weeds. The use of weeds as feed supplement for animals should thus be worked out, especially when combined with less valuable forage such as cereal residues. Common grazing results in incomplete uptake of crop residues, due to trampling of weed and stover, and to tainting by urine. This means that only 20–60% of the biomass left on site is browsed by livestock in the subregion (Quilfen and Milleville, 1983; Manlay, 2000).

CO RR E

CT

ED

PR OO F

R.J. Manlay et al. / Agriculture, Ecosystems and Environment 1803 (2001) 1–21

UN

14

15

CO RR E

CT

ED

PR OO F

R.J. Manlay et al. / Agriculture, Ecosystems and Environment 1803 (2001) 1–21

Fig. 3. Carbon, nitrogen and phosphorus organic and fuelwood ash inputs in compound fields under three different patterns of organic management of fertility. Root exudation of C and biological fixation of N not included.

UN

2. Indirect benefits (maintaining farming system viability). Common grazing saves labour, not forage, but unbrowsed biomass is not lost for the ecosystem. As in faecal production, the return of associated carbon and nutrients to the soil sustains the fer-

tility of the compound ring. In the bush ring, high allocation of carbon to rooting systems and bush regrowth is certainly the keystone of fertility maintenance in the low-input, semi-permanent cropping systems of the bush fields (Pieri, 1989; Brown et al.,

R.J. Manlay et al. / Agriculture, Ecosystems and Environment 1803 (2001) 1–21

This study suggests that land use has little effect on soil physical properties. Differences in texture of the deepest layers result from the distinct positions of bush and compound fields along the toposequence, these in turn being related to the location of the village and not to farmers’ perception of different soil properties. Although textural differences are not significant (Table 3), they hinder any interpretation of the effects of management on SOM status below a depth of 20 cm.

UN

CO RR E

4.2.1. Organic status In the current socio-economic context of Sare Yorobana, improvement in the OM status of crops under continuous cultivation can only occur in fields very close to the compounds. This general result is in agreement with the few surveys held in the subregion at the village territory scale (Prudencio, 1993; Peters and Schulte, 1994). Outside research stations, the effect of manuring alone on soil carbon status in the WAS is variable, but is often reported to be of moderate impact, and restricted to surface soil (Powell, 1986; Feller, 1995), or even nil or very small (Schleich, 1986; Diouf, 1990). But, that lack of precise information on manuring intensity makes comparisons with this study difficult. Soil organic carbon content is controlled by local climate and soil conditions, fire and grazing intensity, as well as the amount and chemical quality of the OM added to the soil. In the present study, no simple relation can be drawn between amounts of carbon stored belowground and those added to the soil every year (Figs. 1 and 3; Table 8). This may result from different C:N and C:P ratios of organic and ash amendments, since these ratios (among other factors) drive the kinetics of decomposition of organic residues (Pieri, 1989). Biochemical properties of structural carbon (“neutral detergent fibre:cellular content” ratio as recommended by Feller, 1995) or enzyme content (Mathur, 1982) might also account for the different/varying effects of stover, roots, household wastes and manure on SOC content.

PR OO F

4.2. Organic practices of fertility management and soil properties

4.2.2. Soil nutrient status of cropped fields in compound and bush rings Unlike C and N content, soil P availability and cationic content are markedly improved in the compound ring at least down to a depth of 40 cm (suggesting that deeper sampling should be made to fully assess nutrient storage variations between the two rings). This improvement stems mainly from exogenous inputs to the compound fields. Organic transfers improve soil chemical properties in three different ways: they (1) are a net source of carbon and nutrients; (2) contribute to a large extent to the gain in CEC; (3) stimulate mechanisms of biological fixation (Feller, 1995; Asadu et al., 1997). For instance, increased availability of phosphorus results either from an increase in the amount of total P, or from the release of unavailable P, previously adsorbed on clay and released by substitution with organic compounds or a modification in soil pH (Jones and Wild, 1975; Feller, 1995). The way SOM improves soil properties depends on what form it is in (Feller et al., 2000). In tropical soils with 1:1-type clay, stable SOM complexed with clay and silt in the 0–50 ␮m fraction influences soil capacity for storage and exchange of nutrients. Carbon stored in the fraction 50–2000 ␮m has a higher turnover rate and vehicles biological functions (supply of energy and nutrients to soil microflora and macrofauna). Correlation analyses of C content from coarse and fine size fractions with other chemical variables, as well as the C:N ratios of both fractions, are in agreement with this interpretation. The present study thus provides evidence that both the storing–exchanging and biological functions of OM are affected to the same extent by cropping intensity. However, some of the improvement in the mineral status and consequently fertility of compound fields can not be related to SOM accumulation, especially below a depth of 20 cm.

ED

1994; Manlay et al., 2001b). Contrastingly, maize, a crop that demands high soil chemical status, exhibits the lowest investment in its rooting system.

CT

16

4.3. An insight at the village scale: agro-ecological complementarity of semi-permanent and continuous cultivation In Sare Yorobana organic practices of fertility management under continuous cultivation have a greater effect than fallowing on improving soil chemical status (Manlay et al., 2001b). Marked differences concern POD , Ca, K, CEC, S and pH. However, it

R.J. Manlay et al. / Agriculture, Ecosystems and Environment 1803 (2001) 1–21

4.4. Methodological considerations: the relevance of SOM as an indicator of soil quality and agricultural sustainability in west African savannas

PR OO F

ED

1. Particular attention must be paid to SOM biological functions. In this connection, this study confirmed that the coarse size fraction (20–2000 or 50–2000 ␮m) is the most relevant fraction (Feller et al., 2000). 2. Living soil components (fauna, microflora and roots) should be precisely characterised, and the root component should be included in SOC computation. 3. A ‘fertile’ savanna soil should be viewed as a living, dynamic system. From a bio-thermodynamic point of view, living systems are open, self-organised systems maintained far from thermodynamic equilibrium (death). This implies the establishment of a continuous flow of energy (and nutrients) through the soil, necessarily coupled with matter (carbon) dissipation (Toussaint and Schneider, 1998; Straskraba et al., 1999). In this perspective, withdrawal of carbon from the soil system (and subsequent low carbon sequestration potential) through faunal and microbial catabolism, and termite-induced spatial redistribution, is the price to pay to maintain soil organisation and adequate functioning (Perry et al., 1989). As a result, emphasis should be put on the assessment of carbon flows (carbon gaseous emissions, fine root and litter production), and not only on stocks.

CO RR E

Patterns of carbon allocation and dynamics under continuous and semi-permanent cultivation at Sare Yorobana (this study; Manlay et al., 2001b) call into question the link between SOM status as defined by the soil carbon content, and fertility of savanna agro-ecosystems. In the tropics, this relation has been seriously questioned, since its quantification is a difficult task (Pieri, 1989; Greenland et al., 1992). For the sandy soils of the WAS, specific methodological precautions must be taken before any agro-ecological interpretations are made of SOC content for two main reasons.

soil texture resulting in low nutrient availability and unstable soil structure (Jones and Wild, 1975; Pieri, 1989). Biota response to these constraints has been (1) control of soil stability and porosity by perennial rooting systems and fauna and microflora; (2) conservative management of nutrients, protected either in root biomass or in stable organic compounds (Menaut et al., 1985; Chotte et al., 1995). Thus, in savanna soils biological mechanisms play a crucial role in processes leading to plant nutrition. This has three implications as follows.

CT

is difficult to compare the efficiency of fallowing and manuring/waste spreading for the improvement of agro-ecosystem fertility since: (1) the increase in SOM content in the compound fields is mainly restricted to the 0–10 cm layer; (2) in Sare Yorobana, fallowing is still an efficient means to rapidly improve soil C content (5.2 g kg−1 in fallows aged 1–10 years, 5.9 g kg−1 in fields in the compound ring), N, POD , Mg contents and CEC (Manlay et al., 2001b); (3) the maintenance of good soil properties in the compound ring relies mostly on the conservation of the bush and savanna rings, as these zones feed cattle during the wet season thus avoiding any damage to cropped fields. In addition, during the dry season, they provide farmers with manure, since 60% of the forage needs of cattle are satisfied by the vegetation in this area (Manlay, 2000). The potential of livestock-mediated organic transfers to balance nutrient exportation by crops has been well demonstrated for areas drier than the Kolda region (de Leeuw et al., 1994). In Sare Yorobana, it apparently accounts for 80–85% of anthropogenic returns to the soil resulting from farming activity (Manlay, 2000). Inputs of nutrients to the compound ring are also mediated by crop harvest, as well as by the collection of fuelwood, which represents a great sink of phosphorus for fallow ecosystems (Manlay et al., 2001a).

UN

4.4.1. SOM content only provides limited information about the agro-ecological role of carbon The productivity of most tropical ecosystems, including WAS, is limited by erratic rainfall patterns and

17

4.4.2. Influence of factors other than management practices on SOC content 4.4.2.1. Texture. Textural heterogeneity can bias agro-ecological diagnostic based on soil carbon content analysis alone. To overcome this drawback in soils with low-activity clay, Feller (1995) proposed to linearly relate SOC to the clay + fine silt content. Applying this relation to the 23 field and fallow plots

R.J. Manlay et al. / Agriculture, Ecosystems and Environment 1803 (2001) 1–21

ED

PR OO F

18

Fig. 4. Assessment of soil quality of 23 crop and fallow plots (this study and Manlay et al., 2001b as predicted by the criterion of Feller (1995) based on carbon content and fine texture. Young fallow: aged 0–9 years; old fallow: 10-year-old or more.

UN

CO RR E

4.4.2.2. Seasonal dynamics. Another often-neglected factor that influences SOC is the date of sampling. Fast SOM turnover rates under tropical climates (Greenland et al., 1992) could lead to significant seasonal variations in total SOC content. These variations may affect agro-ecosystems differently, depending on the contribution of the coarse-size fraction to total SOC content. These seasonal variations, occurring in the WAS as “flushes of mineralisation”, have been well studied for their impact on nutrient availability (Jones and Wild, 1975). But apart from Fearnside and Barbosa (1998), few authors have warned of their methodological implications on SOC measurement. And yet significant seasonal variations in SOC content have already been reported for some tropical successional ecosystems (Saxena and Ramakrishnan, 1986; Zarin et al., 1998). The lack of information concerning the sampling date in most SOC-related

field studies illustrates this problem. Significant dependence of SOC content on sampling date in sandy soils in the WAS is all the more probable because: (1) root biomass, which fuels much of the SOM reservoir and contributes up to 30% of total underground OM, undergoes drastic seasonal variations (Manlay et al., 2001a,b); (2) carbon amounts in the soil, and annual flows through the soil are of the same order of magnitude (this study; Manlay et al., 2001b); (3) carbon inputs in fallow and savanna cropped ecosystems occur in flushes over time (Menaut et al., 1985). As a result, the low potential of soils in the various agro-ecosystems in Sare Yorobana to store carbon may reflect a levelling off of SOC pools after the flush of mineralisation at the onset of the rainy season. This hypothesis is supported by results of SOM fractionation (Section 3.4.2; Manlay et al., 2001b; Feller and Beare, 1997). Moreover, for the nearby station of Sefa, mid-Casamance, Blondel (1971) reported N mineralisation up to 160 kg per ha; since mineralisation mostly concerns OM in fresh plant residues and in the soil coarse size fraction (C:N > 30), this would correspond to 6 t C per ha, that is, at least a third of C amounts stored in the 0–20 cm soil layer of cropped fields in Sare Yorobana.

CT

described in this study and in Manlay et al. (2001b) would attribute good soil quality to 2/3 of the plots (Fig. 4). But, while proving satisfactory for the compound fields, this criterion does not allow for separation between bush fields and old fallow plots, although the latter are more fertile in the farmers’ opinion.

ED

From a methodological and conceptual point of view, a shift still has to be made from a static/structural perception to a dynamic/energetic perception of the role of OM as a support for nutrient-limited tropical agro-ecosystems. Clearly, in Sare Yorobana, much of the beneficial effect of organic management of soil fertility due to fallowing and manuring relies on mineralisation rather than humification processes, which means that SOM content alone is a questionable indicator of the fertility of savanna sandy soils, farming system sustainability, or even environmental quality for the purpose of global change. The present attempt to quantify carbon and nutrient allocation in the plant–soil component related to patterns of organic management of fertility suggests that there is some room for manoeuvre by farmers in the effort to control and improve the carbon and nutrient balance of plots under fallowing or manuring. Obviously, these opportunities are closely related to land and livestock availability. Given the collective organisation of land tenure and social structures in west Africa, further operational appraisal of carbon and nutrient availability and management options should be carried out at the holding and village levels (Defoer et al., 1998; Woomer et al., 1998). As population increases, new intensification pathways will have to be defined. Some of them can be easily implemented by farmers themselves provided achievement is not hindered by lack of available labour, but others such as tree planting or nutrient recapitalisation are beyond the peasant’s control, for they depend on tenurial and subsidy policies. Consequently, any win–win strategy aimed at replenishing carbon and nutrient stocks in African smallholder farming for both fertility enhancement and global change will have to be endorsed by society as a whole and not just by farmers (Izac, 1997).

UN

CO RR E

Limited perspectives for soil carbon sequestration enhancement in traditional cropping systems of the WAS belt are suggested here. Apart from intrinsic soil textural features, the reasons for this are: (1) the limited stability of organic carbon gains in soils following intensive manure and residue management; (2) the high level of carbon inputs required, which are greater than the supply of OM (Williams et al., 1995); (3) the decreasing use of mineral fertilisers as a result of the abandonment of subsidy policies, despite the fact that fertilisers are needed to stabilise SOM and decrease the need for cropped land. Because, sustainable and manageable sequestration of carbon occurs mainly in live woody biomass, tree–crop integration and the use of local mineral fertiliser such as rock phosphate should be promoted. Such low-cost improvements could double the amount of carbon stored in the whole agricultural system in the east African highlands (Woomer et al., 1998). However, the contribution of savanna ecosystems to global mitigation of greenhouse gas release does not necessarily imply local sequestration of carbon (Brown and Lugo, 1990). Settling people in savannas and in the Sahel is one way to slow down further deforestation and subsequent carbon release in nearby, wetter ecozones. This can be achieved by sustainable intensification of cropping. Favouring carbon sequestration is relevant in this context. Maximising carbon fluxes as a way to prevent the ecological breakdown of the system should also be considered. For this purpose, more efficient ways should be sought to manage organic resources with the aim of limiting carbon losses during short bio-geochemical cycles (fire, residue sale). The success of such practices will very much rely on better integration of livestock in cropping strategies, and the adoption by the farmers of agricultural practices such as hay-making, cover crops, slash-and-mulch, compost or no-till practices. The biological activity of soil ‘engineers’, such as termites and earthworms is another promising field of study. In fact, the better savanna cropping systems mimic natural ecosystems in the way they store and recycle carbon, nutrients and energy, the more sustainable they will be (Ewel, 1999).

5. Conclusion

CT

4.5. Savanna cropping systems and the global carbon cycle

19

PR OO F

R.J. Manlay et al. / Agriculture, Ecosystems and Environment 1803 (2001) 1–21

Acknowledgements The authors acknowledge critical and helpful comments by two anonymous reviewers, J. Aronson, C. Floret, A.-M. Izac, C. Millier, J.-C. Remy and D. Richard as well as G. Ciornei and J. Fardoux for pro-

R.J. Manlay et al. / Agriculture, Ecosystems and Environment 1803 (2001) 1–21

UN

CO RR E

Asadu, C.L.A., Diels, J., Vanlauwe, B., 1997. A comparison of the contributions of clay, silt, and organic matter to the effective CEC of soils of sub-Saharan Africa. Soil Sci. 162, 785–794. Bationo, A., Lompo, F., Koala, S., 1998. Research on nutrient flows and balances in west Africa: state-of-the-art. Agric. Ecosyst. Environ. 71, 19–35. Blondel, D., 1971. Contribution à la connaissance de la dynamique de l’azote en sol ferrugineux tropical à Séfa (Sénégal). L’Agronomie Tropicale 26, 1334–1353. Brown, S., Lugo, A.E., 1990. Tropical secondary forests. J. Trop. Ecol. 6, 1–32. Brown, S., Anderson, J.M., Woomer, P.L., Swift, M.J., Barrios, E., 1994. Soil biological processes in tropical ecosystems. In: Woomer, P.L., Swift, M.J. (Eds.), The Biological Management of Tropical Soil Fertility. Wiley, Chichester, pp. 15–46. Chotte, J.-L., Blanchart, E., Lavelle, P., 1995. Gestion durable des terres en milieu tropical. Régulation biologique des processus de décomposition de la matière organique. In: Ganry, F., Campbell, B. (Eds.), Sustainable Land Management in African Semi-Arid and Subhumid Regions, Proceedings of the SCOPE Workshop, Dakar, Sénégal, 15–19/11/1993, CIRAD, pp. 89–97. Dabin, B., 1967. Application des dosages automatiques à l’analyse des sols. 3Éme Partie. Cah. ORSTOM, Sér. Pédol. 5, 257–286. de Leeuw, P.N., Reynolds, L., Rey, B., 1994. Nutrient transfers from livestock in west African agricultural systems. In: Powell, J.M., Fernandez-Rivera, S., Williams, T.O., Renard, C. (Eds.), Livestock and Sustainable Nutrient Cycling in Mixed Farming Systems of Sub-Saharan Africa, Proceedings of an International Conference, Addis Ababa, Ethiopia, 22–26/11/1993, International Livestock Centre for Africa (ILCA), pp. 371–391. Defoer, T., de Groote, H., Hilhorst, T., Kante, S., Budelman, A., 1998. Participatory action research and quantitative analysis for nutrient management in southern Mali. A fruitful marriage? Agric. Ecosyst. Environ. 71, 215–228. Diouf, M., 1990. Diagnostic agronomique en parcelles paysannes. Une méthode d’amélioration des systèmes de culture. In: Savanes d’Afrique, Terres Fertiles? Montpellier, France, 10–14/12/1990, Ministère de la Coopération — CIRAD, pp. 123–143.

PR OO F

References

Ellert, B.H., Bettany, J.R., 1995. Calculation of organic matter and nutrients stored in soils under contrasting management regimes. Can. J. Soil Sci. 75, 529–538. Ewel, J.J., 1999. Natural systems as models for the design of sustainable systems of land use. Agroforest. Syst. 45, 1–21. Fearnside, P.M., Barbosa, R.I., 1998. Soil carbon changes from conversion of forest to pasture in Brazilian Amazonia. Forest Ecol. Manage. 108, 147–166. Feller, C., 1995. La Matière Organique Dans Les Sols Tropicaux à Argile 1:1. Recherche de Compartiments Fonctionnels. Une Approche Granulométrique. Collection TDM, Vol. 144. ORSTOM, Paris. Feller, C., Beare, M.H., 1997. Physical control of soil organic matter dynamics in the tropics. Geoderma 79, 69–116. Feller, C., Balesdent, J., Nicolardot, B., Cerri, C., 2000. Approaching functional soil organic matter pools through particle-size fractionation. Examples for tropical soils. In: Lal, R., Kimble, J.M., Follett, R.F., Stewart, B.A. (Eds.), Assessment Methods for Soil Carbon, pp. 53–67. Floret, C. (Ed.), 1998. Raccourcissement du Temps de Jachère, Biodiversité et Développement Durable en Afrique Centrale (Cameroun) et en Afrique de l’Ouest (Mali, Sénégal): Final Report. Comm. des Communautés Européennes. Contrat TS3-CT93-0220 (DG 12 HSMU), and IRD Paris. Gavinelli, E., Feller, C., Larre-Larrouy, M.C., Bacyé, B., Djegui, N., Nzila, J.D.D., 1995. A routine method to study soil organic matter by particle-size fractionation: example for tropical soils. Commun. Soil Sci. Plant Anal. 26, 1749–1760. Greenland, D.J., Wild, A., Adams, D., 1992. Organic matter dynamics in soils of the tropics — from myth to complex reality. In: Lal, R., Sanchez, P.A. (Eds.), Myths and Science of Soils of the Tropics. Soil Science Society of America, Madison, pp. 17–33. Hatcher, L., Stepanski, E.J., 1994. A step-by-step approach to using the SAS system for univariate and multivariate statistics. SAS Institute Inc., Cary, NC. Ickowicz, A., Usengumuremyi, J., Bastien, D., de Choudens, N., 1997. Spatial analysis of land use by cattle herds in a village of the Sudanese zone in Senegal. Application for grazing system improvement. In: Proceedings of the XVIIIth International Grassland Congress, Winnipeg-Saskatoon, Canada, 8–19/6/97, in press. Izac, A.-M.N., 1997. Ecological economics of investing in natural resource capital in Africa. In: Replenishing Soil Fertility in Africa. American Society of Agronomy and Soil Science Society of America, Madison, USA, pp. 237–251. Jones, M.J., Wild, A., 1975. Soils of the West African Savanna. Technical Communications, Vol. 55. Commonwealth Agricultural Bureaux, Farnham Royal. Ker, A., 1995. Farming Systems in the African Savanna. A Continent in Crisis. International Development Research Centre (IDRC), Ottawa. Lamers, J., Buerkert, A., Makkar, H.P.S., von Oppen, M., Becker, K., 1996. Biomass production, and feed and economic value of fodder weeds as by-products of millet cropping in a Sahelian farming system. Exp. Agric. 32, 317–326. Manlay, R.J., 2000. Organic matter dynamics in mixed-farming systems of the west African savanna: a village case

ED

viding laboratory analyses. This work received financial support from the EEC Project “Reduction of the Fallow Length, Biodiversity and Sustainable Development in Central and West Africa” (TS3-CT93-0220, DG12 HSMU) (Floret, 1998), the Inter-Institutional Inciting Action ORSTOM-CNRS-CIRAD-INRA “bio-functioning of tropical soils and sustainable land management”, the French Institute for Research and Development (IRD, ex-ORSTOM) and the French Institute of Forestry, Agricultural and Environmental Engineering (ENGREF).

CT

20

R.J. Manlay et al. / Agriculture, Ecosystems and Environment 1803 (2001) 1–21

CT

ED

PR OO F

Saxena, K.G., Ramakrishnan, P.S., 1986. Nitrification during slash-and-burn agriculture (Jhum) in north-eastern India. Acta Oecol. 7, 307–319. Schleich, K., 1986. Le fumier peut-il remplacer la jachère? Possibilité d’utilisation du fumier: exemple de la savane d’Afrique Occidentale. Rev. Élev. Méd. Vét. Pays Trop. 39, 97–102. Smil, V., 1999. Crop residues: agriculture’s largest harvest. Bioscience 49, 299–308. Stoorvogel, J.J., Smaling, E.M.A., Janssen, B.H., 1993a. Calculating soil nutrient balances in Africa at different scales. I. Supra-national scale. Fert. Res. 35. Stoorvogel, J.J., Smaling, E.M.A., Janssen, B.H., 1993b. Calculating soil nutrient balances in Africa at different scales. II. District scale. Fert. Res. 35, 237–250. Straskraba, M., Jorgensen, S.E., Patten, B.C., 1999. Ecosystems emerging. 2. Dissipation. Ecol. Model. 117, 3–39. Thioulouse, J., Chessel, D., Dolédec, S., Olivier, J.M., 1997. ADE-4: a multivariate analysis and graphical display software. Stat. Comput. 7, 75–83. Toussaint, O., Schneider, E.D., 1998. The thermodynamics and evolution of complexity in biological systems. Comp. Biochem. Physiol. A 120, 3–11. Vierich, H.I.D., Stoop, W.A., 1990. Changes in west African Savanna agriculture in response to growing population and continuing low rainfall. Agric. Ecosyst. Environ. 31, 115– 132. Williams, T.O., Powell, J.S., Fernandez-Rivera, S., 1995. Soil fertility maintenance and food crop production in semi-arid west Africa: is reliance on manure a sustainable strategy? Outlook Agric. 24, 43–47. Woomer, P.L., Palm, C.A., Qureshi, J.N., Kotto-Same, J., 1998. Carbon sequestration and organic resource management in African smallholder agriculture. In: Lal, R., Kimble, J.M., Follett, R.F., Stewart, B.A. (Eds.), Management of Carbon Sequestration in Soil. CRC Press, Boca Raton, pp. 153– 173. Zarin, D.J., Duchesne, A.L., Hiraoka, M., 1998. Shifting cultivation on the tidal floodplains of Amazonia: impacts on soil nutrient status. Agroforest. Syst. 41, 307–311.

UN

CO RR E

study from south Senegal. Ph.D. thesis, ENGREF, Paris (http://www.engref.fr/ANGLAIS/Accueil.htm). Manlay, R.J., Ka¨ıré, M., Masse, D., Chotte, J.-L., Ciornei, G., Floret, C., 2001a. Carbon, nitrogen and phosphorus allocation in agro-ecosystems of a west African savanna. I. The plant component under semi-permanent cultivation. Agric. Ecosyst. Environ., in press. Manlay, R.J., Masse, D., Chotte, J.-L., Feller, C., Ka¨ıré, M., Fardoux, J., Pontanier, R., 2001b. Carbon, nitrogen and phosphorus allocation in agro-ecosystems of a west African savanna. II. The soil component under semi-permanent cultivation. Agric. Ecosyst. Environ., in press. Mathur, S.P., 1982. The role of soil enzymes in the degradation of organic matter in the tropics, subtropics and temperate zones. In: Non-Symbiotic Nitrogen Fixation and Organic Matter in the Tropics, Proceedings of the Transactions of the 12th International Congress of Soil Science, New Delhi, India, 8–16/2/1982, pp. 125–136. Menaut, J.-C., Barbault, R., Lavelle, P., Lepage, M., 1985. African savannas: biological systems of humification and mineralization. In: Tothill, J.C., Mott, J.J. (Eds.), Ecology and Management of the World’s Savannas. Australian Academic Science, Canberra, pp. 14–33. Perry, D.A., Amaranthus, M.P., Borchers, J.G., Borchers, S.L., Brainerd, R.E., 1989. Bootstrapping in ecosystems. Bioscience 39, 230–237. Peters, J.B., Schulte, E.E., 1994. Soil test survey of the Gambia — an overview. Commun. Soil Sci. Plant Anal. 25, 1713–1733. Pieri, C., 1989. Fertilité des Terres de Savanes. Ministère de la Coopération — CIRAD, Paris (English edition: Pieri, C., 1992. Fertility of Soils: A Future for Farming in the West African Savannah. Springer Series in Physical Environment. Springer, Berlin). Powell, J.M., 1986. Manure for cropping: a case study from central Nigeria. Exp. Agric. 22, 15–24. Prudencio, C.Y., 1993. Ring management of soils and crops in the west African semi-arid tropics — the case of the Mossi farming system in Burkina Faso. Agric. Ecosyst. Environ. 47, 237–264. Quilfen, J.P., Milleville, P., 1983. Résidus de culture et fumure animale. Un aspect des relations agriculture-élevage dans le Nord de la Haute-Volta. L’Agronomie Tropicale 38, 206–212.

21