Agroforestry Systems

, Volume 85, Issue 2, pp 233–245

Can the shrub Chromolaena odorata (Asteraceae) be considered as improving soil biology and plant nutrient availability?

Authors

    • UFR des Sciences de la Nature/Centre de Recherche en EcologieUniversité d’Abobo-Adjamé
  • Ettien F. Edoukou
    • UFR des Sciences de la Nature/Centre de Recherche en EcologieUniversité d’Abobo-Adjamé
  • Jean T. Gonnety
    • UFR des Sciences et Technologie des AlimentsUniversité d’Abobo-Adjamé
  • Aurélie N. A. N’Dri
    • UFR des Sciences et Technologie des AlimentsUniversité d’Abobo-Adjamé
  • Laurenza F. E. Assémien
    • UFR des Sciences et Technologie des AlimentsUniversité d’Abobo-Adjamé
  • Pascal K. T. Angui
    • Laboratoire Géosciences, UFR des Sciences et Gestion de l’EnvironnementUniversité d’Abobo-Adjamé
  • Jérôme E. Tondoh
    • UFR des Sciences de la Nature/Centre de Recherche en EcologieUniversité d’Abobo-Adjamé
    • Laboratoire Sol Eau PlanteCIAT-Mali, IER, CRRA de Sotuba
Article

DOI: 10.1007/s10457-012-9497-5

Cite this article as:
Koné, A.W., Edoukou, E.F., Gonnety, J.T. et al. Agroforest Syst (2012) 85: 233. doi:10.1007/s10457-012-9497-5

Abstract

This paper attempts to provide a new perception of the weed Chromolaena odorata (Asteraceae) which is considered as a plague in agriculture or as a soil fertility indicator. The study was conducted in the forest-savanna transition zone of Côte d’Ivoire and aimed to compare soil biological activity and plant nutrient availability under three well-represented vegetation features, including C. odorata lands (ChrO), shrub savannas (ShrS), and grass savannas (GraS). Each of these vegetation features included five plot replicates (50 m × 50 m size) distributed in the landscape. Soil chemical parameters such as pH, organic matter, soluble phosphorus, exchangeable bases, and biological parameters such as abundance and diversity of earthworms, and soil enzymatic activities were investigated. Composite soil samples were collected and analyzed for chemical and microbial parameters while earthworms were sampled using the Tropical Soil Biology and Fertility 25 × 25 × 30 cm soil monolith method. A principal component analysis showed a clear demarcation between C. odorata plots and the savanna ones. Soluble P and exchangeable bases were significantly higher under ChrO than under both savannas. Earthworm density was twice higher under ChrO (433.3 ± 90.8 ind m−2) than under ShrS (173.9 ± 61.5 ind m−2) and GraS (176.0 ± 40.6 ind m−2) and this was accounted for by the abundance of detritivores and polyhumics. Acid and alkaline phosphatase activities under ChrO (2.9 ± 0.2 and 2.5 ± 0.3 μmol pNP g−1 soil h−1, respectively) were twice higher than under both savannas. Based on the fact that ChrO and ShrS were located on similar soil types and had the same topographic position, we concluded that the establishment of C. odorata in a savanna land and its subsequent high biomass and quality-litter production were mostly the reasons of the improvement in soil biology and plant nutrient availability. Between-savanna comparison showed that ShrS, with higher CEC and exchangeable bases, was somewhat more fertile than GraS, probably because of a better soil physical status. In view of the agronomic potentials of the shrub C. odorata, it may be taken as a basis for improved fallow in Africa.

Keywords

Chromolaena odorataSoluble PExchangeable basesEarthwormsSoil enzymatic activitiesGuinea savannas

Introduction

The forest-savanna transition zone of Côte d’Ivoire, commonly referred to as “V Baoulé” is a mosaic of forest and humid savanna lands where subsistence agriculture is the main activity of population. However, the soils are characterized by low organic matter, phosphorus, and cation contents (Riou 1974; Koné et al. 2008). The finer fraction is also low and consisted of low-activity clays. As a result, agricultural productivity is limited. In this context, the introduction of a plant species or the invasion of lands by a plant species that positively impacts soil organic matter and plant nutrient availability would be beneficial to agriculture.

Native to tropical America, Chromolaena odorata (L.) King and Robinson (Asteraceae) was introduced to Africa, precisely in Nigeria in 1937 and thereafter spread throughout Central and West Africa (Obatolu and Agboola 1993). It is a perennial shrub forming dense tangled bushes 1.5–2.0 m in height (Roder et al. 1995). The species rapidly colonize perturbed spaces due to its high production of wind-dispersed seeds (Gautier 1992). In Côte d’Ivoire, the invasion started in the southeast in the early 1950s and the weed was observed for the first time in the forest-savanna transition zone in the early 1970s (Vuattoux 1976). Because C. odorata may be an invasive weed in agricultural systems, most of the researchers consider it as a plague. It is also considered as a threat to conservation since it reduces the biodiversity of ecosystems (Goodall and Erasmus 1996). As a result, many research studies have been undertaken to investigate this weed species with the goal to eradicate it (Taillez 1992; Huguenin et al. 1992; Gunasekera and Rajapakse 1994; Setter and Campbell 2002; Muniappan et al. 2005; Kluge and Zachariades 2006). On the contrary, a positive perception among farmers was reported about the species in West and Central Africa. Some of them consider C. odorata as a soil fertility indicator (Dove 1986; Kouassi 2010) while others found it to improve soil fertility and crop yields (Slaats 1995; Koutika and Rainy 2010). For instance, in central Côte d’Ivoire, up to 80% of farmers have their farms in C. odorata fallows (Bouadi 2009). Moreover, some studies have reported on the influence of the shrub on improving soil organic matter, highlighting its agronomic potentials (Obatolu and Agboola 1993; Quansah et al. 2001; Koutika et al. 2004). However, the mechanism of this improvement remains poorly understood. In order to improve our understanding, investigations involving soil chemical, biological, and microbial properties and their interactions with vegetation attributes are required.

Kizito et al. (2007) and Dossa et al. (2010) found shrub bushes in savanna lands to represent distinct ecosystem units with better soil quality. According to the authors, they improve microclimate and water availability, leading to enhanced biological activity and biogeochemical cycling. The abundance and adaptability of invasive species to various environments coupled with their rapid growth rate and high vegetative matter turn over makes them good candidate for soil fertility improvement (Obatolu and Agboola 1993). It’s known that soil organisms such as earthworms provide essential services for the sustainable functioning of terrestrial ecosystems (Scheu 2003; Lavelle et al. 2006). These ecological services include control of nutrient cycling (Araujo et al. 2004; Koné et al. 2012a), carbon sequestration in soils, regulation of soil physical structure, and water retention capacity (Blanchart et al. 1997). Also, enzymatic activities are directly involved in nutrient cycling (Acosta-Martinez et al. 2007).

This paper compares soil properties in three well-represented vegetation features in the forest-savanna transition zone of Côte d’Ivoire. More specifically, it focused on the beneficial impacts of short-term C. odorata fallows on soil fertility and examines the underlying mechanisms. The main hypothesis of this study was that C. odorata bushes improve soil organic matter, soil biological activity, and plant nutrient availability as a result of their high biomass production, the favorable microhabitat they create and the quality-litter they provide to soil. To address these questions, comparisons were made between savanna lands and surrounding C. odorata fallows, through soil chemical (pH, total C, total N, soluble P, cations), macrofaunal (abundance and diversity of earthworms), and microbiological (enzymatic activities) properties.

Materials and methods

The study area

The study site was located in the “V Baoulé”, a humid savannas zone in central Côte d’Ivoire and laid between 6′10–6′15 N and 4′55–5′00 W. The vegetation structure of the region is characterized by a mosaic of secondary forests, savannas, C. odorata fallows and various agroecosystems. Savannas are the most represented vegetation feature. They are characterized by the presence of grass species such as Hyparrhenia diplandra, Imperata cylindrica, Loudetia simplex, and Andropogon schirensis as well as shrub species, such as Bridelia ferruginea, Cussonia barteri, Crossopteryx febrifuga, Terminalia glaucescens, and Annona senegalensis. An important trait of the landscape is the presence of the Borassus aethiopum palm trees (Menaut 1971). As observed in many tropical regions, farmers from the “V baoulé” used to burn savanna when the grass cover become high and dry, primarily because of its obstructing character. Savannas are also burnt for hunting. This practice occurs annually between December and March, although it may occur occasionally. As for the C. odorata bushes, they may stop wildfire, especially when they contain sufficient water content (Gautier 1992).

The climate is of a subequatorial type with four seasons: a long dry season from December to February; a long wet season from March to July; a short dry season in August and a short wet season from September to November. The area has a nearly constant temperature throughout the year, averaging 27°C and an annual rainfall of about 1,200 mm.

Soils are moderately leached Ferralsols, with granite being the main bedrock. Upper layers are generally of sandy texture (60–80% of the elements have sizes higher than 500 μm). Clays consist of illites and slightly crystallized kaolinites, with a low adsorption capacity (Riou 1974).

Experimental layout

The study was conducted in farmer plots distributed over the study site of the project AJAMSA (AIRES-Sud 7030), around the village of Ahérémou-2. It consisted of 13 plots of 1,302 m2 (42 m × 31 m) size each, established broadly in three main vegetation features: the natural shrub savanna, the natural grass savanna, and C. odorata fallows, represented 3, 5, and 5 times, respectively. As a common characteristic, the plots were demarcated on sites with a low slope (>5%). In addition, neither machinery nor chemical fertilizers were used on these plots. Specific descriptions are as follows:
  • C. odorata fallows (referred to as ChrO): as indicated by their name, they were dominated by the weed species C. odorata that develops into a dense thicket in almost pure stands. The plots were 4- to 5-year old and were previously cultivated with crop plants such as yam (Dioscorea spp) or cassava (Manihot esculenta). In general, they were located either on the upper or the middle slopes in the landscape.

  • Shrub savanna (referred to as ShrS): the plots were dominated by the grass species H. diplandra, and to a lesser extent, I. cylindrica. Shrub species such as B. ferruginea, C. febrifuga, T. glaucescens, and A. senegalensis were also represented. These plots derived from savanna sites which were burnt (wildfire) 1 year before the start of the study. They were located either on the upper or the middle slopes and were never used before for crop production.

  • Grass savanna (referred to as GraS): the plots were dominated by the grass species Loudetia simplex and I. cylindrica. Like the previous savanna sites, they were burnt (wildfire) 1 year before the start of the study and were never used before for crop production. However, they were located either on the middle or the lower slopes in the landscape.

C. odorata lands are sought after by farmers, because they consider their soils to be more fertile than those of savannas. However, when farmers are short of these lands, they established farms in the shrub savanna or, as a last resort, in the grass savanna. At the end of the cropping period, shrub savanna soils are commonly invaded by C. odorata while grass savanna soils are invaded by I. cylindrica (Menaut 1971; Gautier 1992). Once established in the savanna, C. odorata develops into a dense thicket and spread gradually.

Since ChrO is the only treatment that has ever been cultivated, it is noteworthy to mention that the observed treatment effects may not be due solely to C. odorata, but also to cultivation and land management activities. Also, it is of importance to note that the influence of C. odorata on soil parameters should be best perceived when comparing soil characteristics from ChrO and ShrS because of similarity in type and topography. Indeed, a difference in these parameters (especially clay content) can affect soil chemical and biological properties (Hassink 1997).

Soil characterization

Samplings were conducted in July 2009, at the end of the long rainy season. Top (0–10 cm) and sub (10–20) soils were sampled for chemical and enzymatic activities analyses. The samples were taken at five distinct points distributed over each plot, using an auger. For a given layer, the five samples were pooled and thoroughly mixed as a single composite sample, which was further subdivided into two parts. The first part was immediately stored in an ice chest and taken to the laboratory where it was kept in a refrigerator at a temperature of 4°C for enzyme analyses. The second part was air-dried for 1 week at room temperature, passed through 2 mm sieve and then kept in plastic containers for later chemical analyses. For each soil layer, the bulk density was determined on core samples obtained using the cylinder method. These cores were weighed before and after oven-drying for 48 h at 105°C to obtain moisture content and bulk density. Soil morphological characterization was conducted on each plot by digging a pit of 1 m large, 1.5 m long, and 1.3 m depth.

Total soil N was extracted according to Nelson and Sommers (1980) and determined using Technicon autoanalyzer (Technicon Industrial Systems 1977). Total C was determined using a modified Anne method (Nelson and Sommers 1982). Available phosphorus was extracted according to the Bray-1 procedure (Olsen and Sommers 1982) and determined using a Technicon AutoAnalyzer (Technicon Industrial Systems 1977). Exchangeable bases were extracted using the standard ammonium acetate (pH 7) buffer and measured by atomic absorption spectrometry (Thomas 1982). Cation exchange capacity (CEC) was obtained using standard methods (Anderson and Ingram 1993). Soil pH was determined using a glass electrode in 1:2.5 soil:water ratio.

Litter characterization

Litter collection was carried out concurrently to the soil sampling for chemical analysis. The chemical composition of leaf-litter from each plot was determined on composite samples obtained by mixing litter materials collected at different points at the soil surface. For the savanna plots, chemical analyses were done on senesced leaves cut from the grass tufts since these leaves remained linked to grass tufts together with the green leaves.

Carbon and nitrogen contents were determined by dry combustion using a CHN autoanalyzer (EA1112 Thermo Finnigan Series, France) for dried 4 mg and ground samples (<0.1 mm). Phosphorus was determined following chloride acid (HCl) digestion and analysing extracts for orthophosphate by the molybdenum blue colorimetric method (Murphy and Riley 1962). Major cations were extracted with ammonium acetate buffer (1N; pH 7; litter:extractant ratio of 1:20, g:ml), and determined by atomic absorption spectrophotometry techniques (VARIAN SPECTRAA 220 SF model).

Enzymatic activity assessment

Four enzymatic activities were assessed because of the key role they play in the biogeochemical cycles of carbon, phosphorus and nitrogen. They include acid and alkaline phosphatases, β-glucosidase and N-acetyl-β-d-glucosaminidase activities. Acid and alkaline phosphatase activities were determined in 0.5 g of air-dried soil (<2 mm) using acetate (pH 5.6) and Tris–HCl (pH 8) buffers with p-Nitrophenyl phosphate (10 mM) and incubated at 37°C for 1 h. The reaction was stopped by adding 1 ml 0.5 M CaCl2 and 2 ml 0.5 M NaOH to the mixture, which was then centrifuged according to Tabatabai and Bremner (1969). The released p-Nitrophenol was determined using a spectrophotometer at 410 nm. The enzymatic activity was expressed in micromoles of p-nitrophenol/g soil/h. The β-glucosidase and N-acetyl-β-d-glucosaminidase activities were determined following the methods of Eivazi and Tabatabai (1988) and Parham and Deng (2000), respectively.

Earthworm sampling and identification

Earthworm sampling was conducted in the same period as the soil sampling, following the TSBF method (Anderson and Ingram 1993). In each plot, five distinct soil monoliths of 25 × 25 × 30 cm size each were extracted and earthworm specimens were hand-sorted. They were immediately stored in a 4% formaldehyde solution until they were identified. Specimens were identified to species level or, when this proved difficult, to numbered morpho-species. They were counted and grouped following their feeding habits (at species level) since this has implications in nutrient cycling. There were mainly the detritivores (or litter feeders) which feed at or near the soil surface on plant litter and the geophagous which feed deeper in the soil and derive their nutrition from soil organic matter and dead roots ingested with mineral soil (Lee 1985). The geophagous were divided into three groups, on the basis of their dependence upon soil organic matter (Lavelle 1981): the polyhumics which feed on decaying residues mixed with little mineral soil, the mesohumics which feed on soil fairly rich in organic matter and the oligohumics which feed on organic matter-poor soil.

Data analyses

Earthworm diversity was estimated through the species richness (average and cumulative number of species) and the Shannon–Weaver index of diversity (H′) (Pielou 1966): \( H^{\prime } = - \sum {Pi.\log_{2} Pi} \), with Pi = \( \tfrac{ni}{N} \). ni is the number of individuals of a species i and N, the total number of worms in a soil monolith.

The data were subjected to the U-test of Mann–Whitney and the Kruskal–Wallis ANOVA (non parametric tests), because of the uneven number of replicates between the three systems. These statistical analyses were processed using the STATISTICA ver. 6.0 software program (Statistica, Tulsa, OK). In addition, a Principal Component Analysis (PCA) was run, using the ADE 4 program package (Thioulouse et al. 1997) to determine soil parameters that best explained discrimination between the vegetation features.

Results

Quality of leaf-litters

Leaf-litter of C. odorata had significantly higher quality, as indicated by the higher N, P, and cations contents as well as the lower C:N ratio (Table 1). Significant differences were also observed for C and P between the two savannas, the shrub savanna (which is primarily composed of H. diplandra and I. cylindrica) showing the highest contents.
Table 1

Quality parameters (mean ± standard error) of leaf-litters collected in the three vegetation features

Litter parameters

ChrO

ShrS

GraS

p (K–Wallis)

C (%)

43.1 ± 0.6

50.0 ± 0.8

40.6 ± 0.3

0.04

N (%)

1.8 ± 0.1

0.3 ± 0.0

0.2 ± 0.0

0.004

P (g kg−1)

1.4 ± 0.2

0.5 ± 0.1

0.2 ± 0.0

0.003

Ca (g kg−1)

22.1 ± 0.6

6.5 ± 0.3

5.6 ± 0.2

0.004

Mg (g kg−1)

7.7 ± 0.6

2.0 ± 0.3

1.8 ± 0.0

0.004

K (g kg−1)

4.2 ± 0.7

1.0 ± 0.1

0.9 ± 0.1

0.004

C/N

23.9 ± 0.9

147.1 ± 11.3

191.1 ± 9.1

0.01

ChrOC. odorata, ShrS Shrub savanna, GraS Grass savanna. K–Wallis Kruskal–Wallis Anova

Physico-chemical characteristics of soils

The observation of soil profiles on plots revealed two main soil types based on either color or texture: (1) yellow soils, with two textures (from upper to deeper horizons): sandy with low clay to sandy clay (10–15% clay), or sandy clay to clayey sand (15–25% clay), observed under both C. odorata (ChrO) and shrub savanna (ShrS) and (2) light-colored soils, at foot-slopes and very low in clay content (approximately 5–8% clay), observed under Grass savanna (GraS).

Soil bulk density was found to be significantly lower under C. odorata than under both savannas, irrespective of the considered soil layers. The top soil moisture was higher under C. odorata than under the two savannas, which showed similar values (Table 2).
Table 2

Physico-chemical properties (mean ± standard error) of soils from the three vegetation features

Soil parameters

Soil layers

ChrO

ShrS

GraS

p (K–Wallis)

pH

L1

7.2 ± 0.1

6.7 ± 0.1

6.5 ± 0.1

0.03

L2

7.0 ± 0.1

6.4 ± 0.1

6.2 ± 0.1

0.005

Organic C (g kg−1)

L1

16.8 ± 2.5

7.7 ± 1.3

7.4 ± 1.4

0.007

L2

11.0 ± 2.6

5.3 ± 0.9

4.9 ± 0.3

NS

Total N (g kg−1)

L1

1.0 ± 0.2

0.4 ± 0.1

0.3 ± 0.0

0.007

L2

0.6 ± 0.1

0.3 ± 0.0

0.4 ± 0.1

NS

C:N

L1

17.8 ± 3.0

18.5 ± 1.1

25.1 ± 6.7

NS

L2

26.4 ± 5.2

16.3 ± 4.8

20.5 ± 9.2

NS

Soluble P (mg kg−1)

L1

16.1 ± 6.1

6.4 ± 2.2

5.9 ± 0.7

0.04

L2

7.3 ± 3.5

5.0 ± 2.1

4.1 ± 0.7

NS

CEC (cmolc kg−1)

L1

8.4 ± 0.9

4.1 ± 0.4

2.6 ± 0.1

0.005

L2

6.7 ± 0.8

3.9 ± 0.8

2.9 ± 0.5

0.005

K+ (cmolc kg−1)

L1

0.2 ± 0.0

0.1 ± 0.0

0.1 ± 0.02

NS

L2

0.1 ± 0.0

0.1 ± 0.0

0.1 ± 0.0

NS

Mg2+ (cmolc kg−1)

L1

5.2 ± 0.7

2.1 ± 0.2

1.1 ± 0.1

0.005

L2

4.1 ± 0.6

2.2 ± 0.6

1.5 ± 0.3

0.005

Ca2+ (cmolc kg−1)

L1

2.5 ± 0.3

1.0 ± 0.1

0.5 ± 0.1

0.005

L2

1.9 ± 0.3

1.0 ± 0.3

0.7 ± 0.1

NS

Bulk density (g cm−3)

L1

1.0 ± 0.0

1.2 ± 0.0

1.2 ± 0.0

0.03

L2

1.1 ± 0.0

1.2 ± 0.0

1.2 ± 0.0

0.03

Moisture (%)

L1

11.8 ± 2.0

9.8 ± 1.6

4.9 ± 0.9

0.03

L2

6.9 ± 1.6

7.9 ± 1.2

6.4 ± 1.0

NS

ChrOC. odorata, ShrS shrub savanna, GraS grass savanna

NS no significant difference, KWallis Kruskal–Wallis Anova

Soils layers: L1: 0–10 cm, L2: 10–20 cm soil layers

Soil pH differed significantly among vegetations irrespective of the considered soil layers; being slightly acidic under both savannas and neutral under C. odorata (ChrO). Total soil C and N were highly variable between the treatments in the two soil layers, ChrO showing the highest values. However, ShrS and GraS were not significantly different. The C:N ratios did not show any significant difference between treatments, although the top soil in ChrO had a lower C:N ratio than the savannas. Soluble P was significantly higher for ChrO than the savannas in the top soil. The CEC was significantly higher for ChrO than the savannas. In addition, soil CEC was higher for ShrS than GraS. Except for K+, exchangeable bases varied significantly between the three treatments, showing higher values under ChrO than ShrS and GraS.

Earthworm abundance and diversity

Overall, 19 earthworm species and morpho-species were collected on the studied plots (Table 3). Sixteen species and morpho-species were collected under both ChrO and GraS while only 11 species were collected under ShrS. The average species number varied significantly between treatments with the highest in C. odorata (Table 4). This was not the case for the cumulative species number and the Shannon–Weaver index of diversity even if values tended to be higher under ChrO and GraS. As for the total density of earthworms, it was 2.5 times higher under ChrO than GraS and ShrS (Table 4).
Table 3

List of species and morpho-species of earthworms collected in the three vegetation features

Family

Species

Feeding group

Density-based presence

ChrO

ShrS

GraS

Acanthodrilidea

Dichogaster leroyi

Detritivore

**

 

*

Dichogaster Baeri

Detritivore

*

*

*

Dichogaster saliens

Detritivore

****

*

*

Dichogaster papillosa

Detritivore

**

*

*

Dichogaster notabilis

Detritivore

*

 

*

Dichogaster sp

Detritivore

**

 

**

Dichogaster terrae-nigrae

Oligohumic

*

*

*

Millsonia lamtoiana

Detritivore

*

  

Millsonia omodeoi

Mesohumic

**

**

**

Millsonia schlegeli

Mesohumic

*

 

**

Millsonia sp

Mesohumic

  

*

Millsonia ghanensis

Oligohumic

  

*

Eudrilidea

Eudrilus eugeniae

Detritivore

*

  

Agastrodrilus multivesiculatus

Oligohumic

*

*

*

Agastrodrilus opisthogynus

Oligohumic

 

*

 

Hyperiodrilus africanus

Polyhumic

*

*

*

Stuhlmannia zielae

Polyhumic

*****

*****

****

Stuhlmannia sp

Polyhumic

*

**

**

Stuhlmannia palustris

Polyhumic

***

*

**

ChrOC. odorata, ShrS Shrub savanna, GraS Grass savanna

Density: * 1–10, ** 11–30, *** 31–50, **** 51–100, ***** >100 individuals m−2

Table 4

Total density and diversity parameters (mean ± standard error) of earthworm communities in the three vegetation features

Vegetation features

Density (ind. m−2)

Average species number (species m−2)

Cumulative species number (species plot−1)

H

ChrO

433.3 ± 90.8

4.7 ± 0.7

8.6 ± 1.2

1.6 ± 0.2

ShrS

173.9 ± 61.5

2.8 ± 0.6

5.7 ± 0.9

1.0 ± 0.2

GraS

176.0 ± 40.6

3.8 ± 0.6

8.0 ± 1.3

1.5 ± 0.2

p

0.03

0.04

NS

NS

ChrOC. odorata, ShrS Shrub savanna, GraS Grass savanna

H′ Shannon–Weaver index of diversity, NS no significant difference

Figure 1 shows the abundance of the earthworms based on their feeding groups. Significant differences were observed between vegetations, particularly concerning detritivores (or litter feeders) and polyhumics. The two groups, with 120.3 ± 41.1 ind m−2 and 277.1 ± 66.4 ind m−2, respectively, were significantly more abundant under ChrO than savannas. Although differences were not significant, the detritivores were present in lower density under ShrS (4.3 ± 1.1 ind m−2) than under GraS (31.1 ± 15.9 ind m−2). Conversely, the polyhumic worms were present in higher density under ShrS (133.3 ± 60.8 ind m−2) than under GraS (95.4 ± 28.3 ind m−2). Among polyhumics, Stuhlmannia zielae was by far, the most represented with 227.8 ± 63.8 ind m−2 under ChrO, 113.1 ± 50.14 ind m−2 under ShrS and 55.0 ± 9.4 ind m−2 under GraS. Detritivores were dominated by Dichogaster saliens, particularly in ChrO (51.84 ± 38.5 ind m−2). However, mesohumics tended to be more abundant in GraS than ShrS and ChrO. The density of Millsonia omodeoi remained almost constant (approx. 26 ind m−2) under the three vegetation covers while the density of M. schlegeli increased under GraS (17.3 ± 11.3 ind m−2). The oligohumics were the less represented, with no significant change between treatments.
https://static-content.springer.com/image/art%3A10.1007%2Fs10457-012-9497-5/MediaObjects/10457_2012_9497_Fig1_HTML.gif
Fig. 1

Abundance of earthworm feeding groups in the three vegetation features. ChrOC. odorata , ShrS Shrub savanna, GraS Grass savanna. Vertical bars represent standard errors

Enzymatic activities

Enzymatic activities assessment in soil samples brought out acid and alkaline phosphatases, β-glucosidase and N-acethyl-β-d-Glucosaminidase both in soil under C. odorata and natural savannas. Phosphatases activities were more important than those of β-glucosidase and N-acethyl-β-d-Glucosaminidase. Between-treatment comparisons revealed significant variations in all the enzymatic activity levels (p < 0.05), except for N-acethyl-β-d-Glucosaminidase. Values were significantly higher under ChrO than under both the savannas (Fig. 2). Both acid and alkaline phosphatase activities in ChrO (2.9 ± 0.2 and 2.5 ± 0.3 μmol pNP g−1 soil h−1, respectively) were more than two times higher than in ShrS and GraS. As for N-acethyl-B-d-Glucosaminidase and ß-Glucosidase activities, they were three times higher under ChrO (0.3 ± 0.1 and 0.6 ± 0.1 μmol pNP g−1 soil h−1, respectively) than ShrS and GraS. No difference was observed between savannas.
https://static-content.springer.com/image/art%3A10.1007%2Fs10457-012-9497-5/MediaObjects/10457_2012_9497_Fig2_HTML.gif
Fig. 2

Enzymatic activities in soils from the three vegetation features. ChrOC. odorata , ShrS Shrub savanna, GraS Grass savanna. Acid/Alkaline Phos Acid/Alkaline phosphatase, β-Glu β-glucosidase, N-acethyl-β-d-GluN-acethyl-β-d-Glucosaminidase. Vertical bars represent standard errors

Soil parameters explaining discrimination between plots

A normalized Principal Components Analysis (PCA) was carried out to determine soil parameters that best accounted for the discrimination between the three vegetation features. Axes 1 and 2 explained 51.6 and 14.6% of the total variability between the vegetations, respectively, thereby explaining 66.2% of variability. With the exception of the C:N ratio, mesohumic and oligohumic worm densities and soluble P, all the considered soil parameters were significantly linked to axis 1, thus explaining plots discrimination along this axis (Fig. 3a). There were mainly C and N (r < −0.85), the enzymatic activities (r < −0.8), the CEC (r = 0.9), the total earthworm density (r = 0.75) and the bulk density (r = 0.9). Discrimination along Axis 2 was rather explained by the mesohumic density (r = 0.7) and soluble P (r = 0.7). The projection of data in the factorial plane 1–2 (Fig. 3b) revealed a clear demarcation (Monte-Carlo test, p < 0.001) of plots, primarily along axis 1, into three broad groups corresponding to the three ecosystems. However, the two groups of savanna plots could easily be merged into one group, due to their close positions in the plane. Soils under C. odorata were characterized by higher SOM, exchangeable bases and enzymatic activities, and were more conducive to earthworm growth, particularly detritivores and polyhumic worms.
https://static-content.springer.com/image/art%3A10.1007%2Fs10457-012-9497-5/MediaObjects/10457_2012_9497_Fig3_HTML.gif
Fig. 3

Principal component analysis (PCA) on physical, chemical, and biological characteristics of soils from the three vegetation features: correlation circle (a) and projection of plots in the factorial planes 1 and 2 (b) Ew-T Dens earthworm total density, Pac acid phosphatase, Pal alkaline phosphatase, β-Glu β-glucosidase, NAGN-acethyl-β-d-Glucosaminidase. ChrOC. odorata; ShrS Shrub savanna, GraS Grass savanna

Discussion

In this study, the investigated soil parameters showed higher values under C. odorata bushes (ChrO) than both the shrub (ShrB) and grass (GraS) savannas. The mechanisms involved in this trend included chemical, biological, and physical factors. However, it is of importance to note that the impact of C. odorata will be best perceived when comparing ChrO and ShrS, because these treatments were located at the same position in the landscape (upper slope and middle slope) and laid on the same soil type. Indeed, a difference in topography or clay content may affect soil chemical and biological properties (Hassink 1997).

Detailed studies of the role of C. odorata in the vegetation dynamic in the forest-savanna transition zone in Côte d’Ivoire showed that C. odorata can develop on shrub savanna soils (Menaut 1971; Gautier 1992), provided that the grass-cover is reduced or suppressed. This is confirmed by the fact that these soils are commonly invaded by C. odorata when they are left to fallow. In this region, land (shrub savanna either C. odorata fallow) conversion into food crops results in a significant drop in soil fertility within a 2 or 3-year period as a result of the sandy status of soils. However, farmers do go back to the same land after approximately a 4- or 5-year fallow period, particularly when the dominant plant species is C. odorata, suggesting some restoration of soil fertility by the shrub. The suppression of the grass cover during cropping period is probably the main modification that benefit to the shrub since such condition is conducive to the germination of its seeds. Afterward, the weed gradually spreads into the savanna land (Gautier 1992), forming dense thickets in almost pure stands and processes leading to soil transformation take place. Certainly, there is some possibility that the level of soil fertility observed under C. odorata was due to soil properties that were pre-existing to its establishment, but it is unlikely that this overshadows the impact of the shrub on soil fertility because of the reasons mentioned above.

The C contents in savanna soils were lower than 11 mg kg−1, defined as the threshold of fertility level for tropical soils by Lal (1997), reflecting their low nutrient-level nature. C. odorata bushes continuously produced abundant and quality-litter (high N and P contents, low C:N ratio) that may explain the higher levels of plant nutrients. The species may yield up to 13 Mg dry matter ha−1 while grass biomasses recorded in savannas were around 6 Mg ha−1 (results not shown) which mostly returns to soil at the end of the vegetation cycle of annual species or after wildfire as ashes. Based on the classification made by Jamaludheen and Kumar (1999), C. odorata leaf-litter, with 1.8% N, can be considered as N-rich whereas leaf-litter on savanna plots (0.2–0.4% N) were N-poor. The N content in C. odorata leaf-litter was comparable to that of Tithonia diversifolia (1.76%) which is another invasive weed species from the Asteraceae family, and which was proposed to be used as green manure (Cong and Merckx 2005; Olabode et al. 2007). The N content (1.8%) in leaf-litter from young C. odorata leaf-litter is comparable, even higher than those of some herbaceous legume species such as Mucuna pruriens (1.6%) and Pueraria phaseoloides (1.1%) (Koné et al. 2008). In older fallows (10–12 years), leaf-litter may contain up to 2.5% N (Koné et al. 2012b) which is higher than N content in leaf-litter from the herbaceous legume Lablab purpureus (2.1%) (Koné et al. 2012a). As a result, C. odorata leaf-litter presumably experienced a faster turnover than those in the natural savannas as reported by Koné (2009). In a decomposition study, the author found that C. odorata leaf-litter completely decomposed within 3 months, whereas about 20% of the initial mass of H. diplandra leaf-litter remained undecomposed after 6 months. In addition to the higher P content, the faster decomposition of C. odorata leaf-litter is a probable explanation of the soluble P level recorded under C. odorata. Indeed, the P derived from plant residues was reported to constitute a major part of soil soluble P when plant residues input is high (Ha et al. 2007). Regarding the savannas, the difference in soil texture was a probable reason, though partly, why the CEC and the exchangeable bases were higher under ShrS than under GraS. Obatolu and Agboola (1993) and Koutika and Rainy (2010) also reported a beneficial effect of C. odorata on soil soluble P (in Nigeria) and exchangeable K (in Cameroon), respectively. The increase in pH under C. odorata could be linked to contents in Ca2+ (Dossa et al. 2010).

Carbon sequestration in C. odorata biomass followed by its accumulation into soil through ready decomposition of leaf-litter can explain the improvement in soil C, but that for N, P and the exchangeable bases might be carried out by additional factors. Plant root geometry and morphology are known to be essential in maximizing nutrient uptake, because root systems that have higher ratios of surface area to volume will more effectively explore a larger volume of soil (Lynch 1995). In this respect, C. odorata contributed more to nutrient cycling from deeper to surface soil layers than did the savanna grass species. The luxuriant vegetative growth of this weed coupled with the spreading root systems, extracts larger quantities of nutrient elements from the soil, and may act as a nutrient pump (Obatolu and Agboola 1993). In addition, C. odorata could have developed greater symbiotic associations with mycorrhizal fungi that allowed subsequent greater P acquisition and accumulation in the soil surface layer through litter decomposition. Indeed, in a research work conducted in Cameroon, Onguene (2007) observed that roots of C. odorata were more colonized (62%) by arbuscular mycorrhizal fungi than those of the other plant species (14–49%). Also, one could hypothesize that the new soil conditions yielded by the invasion by C. odorata boosted the growth of free-living N-fixing bacteria (Azotobacter, Azospirillum) and this may explain N content recorded in C. odorata leaf-litter. However, further studies are needed to confirm these hypotheses. Despite the dense thicket formed by C. odorata, the new soil conditions allowed the growth of other plant species among which some legume as observed on the plots. For instance, Mucuna pruriens and Crotalaria pubescens were observed in places, though in low density, and may have contributed to some extent to soil N increase.

Plant residues are incorporated into soil in the form of organic matter after having been broken down and decomposed by soils organisms such as earthworms and microbes. However, it is well known that the decomposing activity of these organisms is driven by litter chemistry (Loranger et al. 2002; Goma-Tchimbakala and Bernhard-Reversat 2006; Koné et al. 2012a). The density of earthworms under C. odorata was increased, primarily based on the quality of its leaf-litter. The difference between C. odorata plots and the savanna ones were most striking when comparing the densities of detritivores and polyhumics. This was because these worm species feed on litter or on soil rich in organic matter, respectively. Earthworms are known to participate in litter decomposition and accumulation in soil, create and maintain soil structure by digging burrows and modifying aggregation (Lavelle et al. 2006). A consequence of these activities is the improvement of plant nutrient availability in soil (Bhadauria and Saxena 2010). The neutral pH and the maintenance of a higher soil moisture under the thick C. odorata cover were additional factors that could support the growth and the activity of the earthworms (Tondoh 2006; Mainoo et al. 2008). Indeed, the densities of earthworms were found to be lower in more acidic soils (Mboukou-Kimbasta et al. 2007). In the present study, the density of mesohumics and oligohumics were not impacted by C. odorata. At the species level, Millsonia schlegeli (mesohumic) and Agastrodrilus multivesiculatus (oligohumic) appeared to be more abundant in grass and shrub savanna, respectively. These worms feed on mineral soil and as such, they were not much influenced by litter production as the two first groups. There is some possibility that the lower (but not significantly) diversity of earthworm observed in the shrub savanna was due to the lower number of plot investigated. This assumption is possible since plots were distant from each other and consequently, some other worm species could have been encountered, had we investigated additional plots (Magurran 2004). Unfortunately, among the plots farmers lent us for field activities, only three were located in shrub savanna.

The higher quality of litter products from C. odorata probably made them more easily accessible to microorganisms. This could be an explanation to the fact that the enzymatic activities were more important under C. odorata than the savanna plots, regardless of the continuous contribution of litter in the soil. The levels of soil N and soluble P were consistent with these activities. The enhancement in enzymatic activities observed in C. odorata could also be linked to earthworms feeding activities. Indeed, these microbial activities have been reported to increase in biogenic structures produced by these soil organisms (Mora et al. 2003). Therefore, the impact of earthworms on the parameter probably increased as their density increased. Acid and alkaline phosphatases can be considered as the most active enzymes in the soil investigated; this is in line with the observations of Acosta-Martinez et al. (2007) when evaluating the impact of soil use on the activity of certain hydrolase enzymes. However on savanna plots, litter quality was very low as shown by the higher C:N ratio (147.2–191.1) and litter-fall was rather insignificant over the year, occurring only at the end of the cycle of annual grass species or when the savanna was burnt (wildfire). The unsteady organic supply to the soil may explain the low enzymatic activity in savanna soils.

Conclusion

This study showed that invasion of soils by the shrub C. odorata in savanna environment leads to significant improvement in biology and plant nutrient availability. This was attributed to the thickness of the bushes it forms, the quantity and quality of its residue products and probably nutrient uptake by its roots. These conditions were conducive to earthworm and microbial growth and activities. Thus, the study gives evidence of the important role C. odorata can play in low-input agriculture, and should thereby provide new perception among researchers. Considering the agronomic potentials of the shrub C. odorata, it may be taken as a basis for fallow improvement for the well-being of farmers in Africa.

Acknowledgments

This study was funded by the French Ministry of Foreign and European Affairs, through the research program AIRES-Sud, implemented by the Institut de Recherche pour le Développement (IRD-DSF). We are grateful to R. Orendo-Smith for his contribution in editing the text, and the two anonymous referees for their valuable inputs. Special thanks to the farmers from the village Ahérémou-2 for their active participation in the project AJAMSA.

Copyright information

© Springer Science+Business Media B.V. 2012