European Journal of Soil Biology 42 (2006) S334–S340 http://france.elsevier.com/direct/ejsobi

Original article

Seasonal changes in earthworm diversity and community structure in Central Côte d’Ivoire J.E. Tondoh UFR des sciences de la nature, centre de recherche en écologie, université d’Abobo-Adjamé, BP 23, 4727 Abidjan 23, Ivory Coast Available online 14 September 2006

Abstract This study assessed the impact of seasonal variation in the structure and diversity of earthworm communities of a savanna protected for 27 years in the central region of Côte d’Ivoire. Earthworm species were sampled in 1995 at monthly intervals from January to December on a 95 × 50 m experimental plot, using direct hand-sorting techniques. Each month, 10 monoliths of 1 m2 × 40 cm were randomly selected from a stratified bloc design. Ten earthworm species were collected over the study period. Chuniodrilus zielae from the Eudrilidae family was by far, the most important earthworm species in term of abundance. Although earthworm diversity varied significantly, the effect of seasonal variation was unclear. Sampling efficiency of species richness varied from 80% to 100% regardless of the rainfall variation. On a seasonal time scale the C-score was lower (0.139) than expected (0.154), showing that earthworm communities exhibit a random pattern of organization. There was no evidence of non-random seasonal niche overlap because the Czechanowski index (0.50) was not significantly larger than expected (0.49). © 2006 Elsevier Masson SAS. All rights reserved. Keywords: Tropical earthworms; Diversity; Community structure; Côte d’Ivoire

1. Introduction Compared to temperate countries (Europe, America, etc.) where most species belong to the Lumbricidae family, tropical earthworm communities are highly diverse [18]. Thirteen families have been distinguished with more than hundred species still need to be identified [18]. Population dynamics in tropical earthworms have been documented on both seasonal [6,13,16,17,20,26] and spatial scales [24,14]. However, studies on earthworm diversity and assemblage are not common except for very few investigations [7,13,17,28]. Moreover, little is known about factors affecting the seasonal varia-

E-mail address: [email protected] (J.E. Tondoh).

tion of earthworm diversity which may depend on how diversity is quantified and mainly if a standardized measure of sampling effort has been used [21,22,]. Rather, the crucial role of species diversity in measuring ecosystems functioning [19] emphasizes the need to avoid bias when characterizing diversity at species level. Most studies on earthworm diversity have used diversity indices (Shannon, Simpson indices). However, recent findings in diversity studies revealed the efficiency of accumulation curves to carefully quantify species diversity [2,11,21]. When comparing species richness between communities, species accumulation curves are useful because they allow estimating species richness based on a standardized measure of sampling effort and measurement of inventory efficiency and completeness [22].

1164-5563/$ - see front matter © 2006 Elsevier Masson SAS. All rights reserved. doi:10.1016/j.ejsobi.2006.09.003

J.E. Tondoh / European Journal of Soil Biology 42 (2006) S334–S340

Although it has been a major source of controversy in community ecology for over two decades, the analysis of presence–absence matrices with “null model” randomization tests have been used as an efficient tool for revealing patterns (species co-occurrence, niche overlap, etc.) in natural communities [2,9]. Very few studies on seasonal changes in tropical earthworm diversity have been done using species accumulation curves and null model analysis to detect the signature of species interactions. The purpose of our study was to examine the impacts of seasonal variation on earthworm communities from a savanna protected for 27 years under the assumption that rainfall is one of the major abiotic factors. Species accumulation curves were used to quantify the effects of rainfall variation on earthworm abundance, species richness and community diversity. The analysis of presence–absence matrices based on “null model” randomization tests was also used to evaluate temporal pattern interactions in earthworm communities.

S335

2.2. Census methods Earthworms sampling was carried out from January to December 1995, on a 95 × 50 m size-plot obtained from an area protected from fire during 27 years. This plot was gridded at 5 m intervals to yield a block system of 10 “column” and 19 “lines”; giving a total of 190 subplots of 25 m2 each. Each month, a total of 10 monoliths of 1 m side and 40 cm depth were randomly selected from the subplots of each column for earthworm sampling. Earthworms were extracted by direct hand-sorting in four successive strata of 10 cm depth [15]. All earthworms were preserved in 4% formaldehyde. Individuals were then separated in the laboratory into species, counted and weighed. 2.3. Data analysis

2. Materials and methods

2.3.1. Earthworm diversity Earthworm diversity analysis was performed in two stages. Firstly, Shannon–Wiener index of diversity [23] was used to measure diversity at sample level.

2.1. Study site

H ¼
∑ pilog2 pi

n

The study site was located in the Natural Reserve of Lamto (6°N, 5°2W) situated in Central Ivory Coast, a transition zone between the semi-deciduous humid forest and the Guinean savanna, which is referred to as the “V Baoulé”. With a surface area of nearly 2500 ha, the Lamto Reserve is characterized by a mosaic of forest and savanna vegetation. The study plot is located in a savanna protected from fire since 1967, similar to a regrowth forest with vegetation composed of edge shrubs and trees. The development of the tree vegetation progressively covers the shrubs and chokes the herbaceous stratum [27]. The succession includes the neotropical invader (Chromolaena odorata) that forms, in some places, dense stands. The bimodal climate is characterized by wet seasons from April to July and from September to October. Mean annual temperature over 10 years (1986–1996) was 28.4 °C while rainfall is in the range of 8.4 mm in January to 189.7 mm in June with an annual total of 1138.1 mm. The soil from the study plot is an oxisol (ferrasol, in FAO classification). It is mainly of sandy texture (75% sand) and slightly acidic [26]. Average organic carbon and total nitrogen concentrations are low (0.86 and 0.076%, respectively); with a C/N ratio of 12.

i¼1

where pi is the frequency of the i species. Secondly, the software EstimateS 6.0b1 [4] was used to estimate the average species accumulation from 500 randomly ordered sequence of monthly monoliths. The estimate number corresponds to species observed in the pooled monoliths [5]. From the same package, the abundance-based coverage estimator (ACE), a nonparametric estimator of true species richness taking into account species present but not observed, was used to generate an estimate of true diversity. 2.3.2. Earthworm species co-occurrence To analyze earthworm species co-occurrence on an annual time scale, we organized the monthly average abundance as a presence absence matrix in which each row was a species and each column, a month. The entries in the matrix indicate the presence (1) or the absence (0) of a species in a particular month. We used the Checkerboard score “C-score” [25] quantitative index of occurrence that measures the extent to which species co-occur less frequently than expected by chance [10]. The larger the C-score, the less coexistence there is between species pairs compared with a random assembled community. A community is structured by competition when the C-score is significantly

S336

J.E. Tondoh / European Journal of Soil Biology 42 (2006) S334–S340

larger than expected. Ecosim software [12] was used to perform a co-occurrence analysis. 2.3.3. Temporal niche overlap To quantify patterns of niche overlap between a given pair of species, the Czechanowski index of niche overlaps between each pair of species [8] was used:

Wiener diversity index. Before performing one-way ANOVA to compare earthworm abundance through months, a log10 transformation was used to normalize the distribution of data. The same test was used to compare the diversity index. Linear regression was used to test the relationships between each response variables and rainfall. These statistical tests were performed using the software Statistica, 1999.

n

O12 ¼ O21 ¼ 1:0
0:5 ∑ jPi1
P12 j i¼1

where:

3. Results 3.1. Abundance

● O12 is the overlap of species 1 on species 2; ● Pi1 is the fraction of the total density for a species 1 that were found in month i. Niche overlap of the entire earthworm community was obtained by calculating the mean and variance of niche overlap among all unique pairs of species in the assemblage, using Ecosim software [12]. The index approaches 0 for species that share no resource state (months) and 1 for species pairs that have identical resource utilization distributions. The statistical significance of the niche overlap patterns was determined by comparing them with an appropriate null model, in which the observed earthworm densities were randomized 5000 times among species. Interspecific competition should cause mean niche overlap to be greater than expected by chance [2]. 2.3.4. Other statistical analysis The response variables in our analyses were earthworm abundance, species richness and Shannon–

Ten earthworm species were collected over the study period: ● ● ● ● ● ● ● ● ● ●

Dichogaster agilis (Megascolecidae); Dichogaster sp. (Megascolecidae); Hyperiodrilus africanus (Eudrilidae); Millsonia lamtoiana (Megascolecidae); Millsonia anomala (Megascolecidae); Chuniodrilus zielae (Eudrilidae); Stuhlmannia porifera (Eudrilidae); Millsonia ghanensis (Megascolecidae); Agastrodrilus sp. (Megascolecidae); Dichogaster terrae-nigrae (Megascolecidae).

Among these species recorded, six (D. agilis, Dichogaster sp., H. africanus, M. anomala, C. zielae, S. porifera) were common because they represented 98.3–99.9% of the overall community of earthworms. C. zielae is by far the most important earthworm species in terms of density. The individuals represented

Fig. 1. Seasonal abundance (ind m−2) patterns of the six most abundant earthworm species.

J.E. Tondoh / European Journal of Soil Biology 42 (2006) S334–S340

39.8–58.2% of the community and the densities were within the range of 181–144.6 ind m−2 (Fig. 1). The lower values of density were recorded in January (44.1 ind m−2), February (32.4 ind m−2), November (57.3 ind m−2) and December (18.1 ind m−2) whereas the higher values were found during the raining season: April (144.6 ind m−2), May (95.6 ind m−2), June (88.5 ind m−2) July (142.6 ind m−2), September (88.4 ind m−2) and October (100.8 ind m−2). H. africanus followed C. zielae in terms of abundance and showed similar seasonal fluctuations. The lowest value (3.5 ind m−2) was recorded in December and the highest (43.7 ind m−2) in July (43.7 ind m−2). The densities of M. anomala, S. porifera, D. agilis and Dichogater sp. were linked and the seasonal fluctuations were similar to those of previous species (Fig. 1). The year-long study showed seasonal fluctuations in the earthworm community abundance (Fig. 2). The lower values were recorded during the drier seasons: January (85.7 ind m−2); February (66.7 ind m−2); November (98.6 ind m−2) and December (40.2 ind m−2) whereas the higher values were found in the more humid months: March (119.9 ind m−2), April (214.0 ind m−2), May (166.7 ind m−2), June (179.5 ind m−2) July (268 ind m−2), August (165.4 ind m−2), September (185.3 ind m−2) and October (173.3 ind m−2). Regardless of seasonal variations, yearly means of the earthworm community abundance was 147 ± 27.7 ind m−2 with C. zielae, H. africanus, M. anomala, S. porifera, D. agilis and Dichogaster sp., representing 53.6%, 14.4%, 10.6%, 7.3%, 6.7% and 6.9% of the community, respectively. The monthly abundance of earthworms was significantly related (R = 0.90, P < 0.0001, N = 12) to rainfall so that the occurrence of 4 months of dry season led to

S337

a drastic decrease in density (Fig. 2). There was a marked effect of rainfall variation on total abundance of the earthworm community (Kruskal–Wallis test, P = 0.0000). Densities in all dry months (January, February, March, November, December) were significantly lower (LSD test, P = 0.0000) than in rainy seasons (Fig. 2), the lowest value being recorded in December. 3.2. Diversity 3.2.1. Diversity index The Shannon diversity index (H) was significantly higher in June (2.0), July (2.0), August (2.1) and September (2.1) than in January (1.8), February (1.8), March (1.8), May (1.8); the lowest value of diversity index (1.6) being recorded in April (Table 1). However, there was no significant relationship between earthworm diversity and rainfall (R = 0.14, P > 0.05, N = 12). 3.2.2. Species diversity Cumulative diversity of earthworm during the whole sampling periods varied from seven species in June and November to nine species in September and October. Eight species were recorded in the remaining months (Fig. 3). Surprisingly, the lowest diversity did not correspond to dry months (January, December and February). As a consequence, there was no relationship between species diversity and rainfall (R = 0.02, P = 0.21, N = 12). The non-parametric estimator of potential species number (ACE) was subjected to slight fluctuations (Fig. 3). The highest value (10 species) was recorded in February and December, followed by October and November (9.8). The lowest value (7.5) was recorded

Fig. 2. Monthly variation of the mean earthworm community abundance (ind m−2) and rainfall patterns.

S338

J.E. Tondoh / European Journal of Soil Biology 42 (2006) S334–S340

Table 1 Seasonal variations of log10 abundance and diversity parameters of earthworm communities. Values followed by the same letters are not significantly different (ANOVA, P = 0.00001) Months

Species observed

ACE

January February March April May June July August September October November December

8 8 8 8 8 7 8 8 9 9 7 8

9.1 10.0 9.1 9.1 8.0 7.5 8.0 8.0 9.8 9.8 7.5 10.0

in June and November. The same number (8) of observed and total species richness was recorded in March, July and August (Fig. 3). These results revealed that sampling efficiency of earthworm diversity varied from 80% to 100% regardless of the rainfall variation. In addition, there was no significant relationship between sampling efficiency and monthly variation of rainfall (R = 0.43, P = 0.1634, N = 12). Conversely, monthly pooled individual fluctuations was closely related to rainfall variation (R = 0.91, P < 0.0001, N = 12). 3.3. Community structure 3.3.1. Species co-occurrence On a seasonal time scale the C-score among the sampling month was lower (0.14) than expected by chance (0.15), showing that the earthworm community converged to a random pattern of organization. 3.3.2. Seasonal niche overlap There were no evidence of non-random seasonal niche overlap because the Czechanowski index (0.50) was not significantly (P = 0.93) larger than that expected by chance (0.49). 4. Discussion and conclusions 4.1. Effects of seasonal variation on earthworms abundance The dominance of earthworms from the Eudrilidae family in terms of density was reported by Lavelle [15].

Sampling efficiency (%) 87.8 80.0 87.6 87.7 100.0 93.3 100.0 100.0 92.3 91.6 93.3 80.0

Shannon index

log10 density

1.8 ± 0.05 c 1.8 ± 0.06 c 1.8 ± 0.09 c 1.6 ± 0.08 d 1.8 ± 0.07 c 2.0 ± 0.07 abc 2.0 ± 0.05 abc 2.1 ± 0.04 a 2.1 ± 0.05 a 1.9 ± 0.1 bc 1.9±0.08 bc 2.0 ± 0.1 abc

1.9 ± 0.05 cd 1.8 ± 0.06 d 2.0 ± 0.1 c 2.3 ± 0.05 ab 2.2 ± 0.07 b 2.4 ± 0.04 a 2.2 ± 0.05 b 2.2 ± 0.05 b 2.2 ± 0.05 b 2.2 ± 0.03 b 1.9 ± 0.06 cd 1.5 ± 0.01 e

As in other studies [3,16,18,26], variations in earthworm abundance were under the influence of seasonal fluctuations. Maximum values were observed during the raining season. The increase in density during the humid period follows reproductive peaks, which, in turn results in growth of individuals. The latter depends on the soil water content. Thus, in the humid period, the earthworms have accomplished most of their life cycle. In our study, rainfall determines temporal occurrence patterns of earthworms. These results are consistent with the findings of Jiménez et al. [13]. Abbott [1] described a similar pattern of earthworm dynamic across the continental Australia. He found out that few earthworms occurred in unfavorable climatic contexts with annual rainfall less than 400 mm. 4.2. Effects of seasonal variation on earthworm diversity The number of earthworm species recorded (10) from the savanna protected from fire was higher than the value of eight found during the 1971–1972 period by Lavelle [15]. The increase in species richness following 27 years of protection effort was likely to be due to the presence of H. africanus, a peregrine species and Dichogaster sp., another native species. H. africanus is thought to have been introduced through soil from ornamental plants in a garden located near the fire-protected savanna which was later colonized. The apparition of Dichogaster sp., a species that feeds on litter and detritus can be explained through the increase of soil litter quantity. Although diversity fluctuated significantly with seasonal rainfall variation, there was not a clear pattern of

J.E. Tondoh / European Journal of Soil Biology 42 (2006) S334–S340

S339

Fig. 3. Earthworm species accumulation curves during the sampling period. ○ = observed species; ▪ = ACE.

earthworm diversity variation linked with rainfall. Only individuals obtained by the accumulation curves were significantly related to rainfall. This result shows that the number of individuals is an appropriate measure of sampling effort to compare species accumulation curves among months as has been reported by [29,22].

Species richness sampling efficiency seems not to be dependent on seasonal change, since there was no significant relationship between “ACE” and rainfall. However, the maximum (100%) of sampling efficiency was reached in May, July and August, months characterized by a high rainfall. This finding pointed out the precious

S340

J.E. Tondoh / European Journal of Soil Biology 42 (2006) S334–S340

role of the rain in inventorying efficiently tropical earthworm populations.

[13]

4.3. Effects of seasonal change on the community structure [14]

On a monthly time scale, there was no evidence of seasonal niche partitioning. Niche overlap and pairwise species associations were random. Using Pianka index, [17] identified size differences, time and space and food partitioning to be responsible for the restriction of niche-overlap in Lamto earthworm communities. Seasonal species co-occurrence patterns were random, showing that species were less subject to competitive interaction [9,10]. This result should help to address specific research topics aiming at studying efficiently the interactive relations within earthworm communities.

[15]

[16]

[17]

[18]

[19]

References [1]

I. Abbott, Distribution of the native earthworm fauna of Australia: a continent-wide perspective, Aust. J. Soil Res. 32 (1994) 117–126. [2] M. Albrecht, N.J. Gotelli, Spatial and temporal niche partitioning in grassland ants, Oecologia 126 (2001) 134–141. [3] S.A. Bennour, G.A. Nair, Density, biomass and vertical distribution of Aporrectodea calignosa (Savigny, 1929) (Oligochaeta, Lumbricidae) in Benghazi, Lybia. Biol. Fertil. Soils 24 (1997) 102–105. [4] R.K. Colwell, Statistical Estimate of Species Richness and Shared from Samples. Version 6.0b1. (2000) http://viceroy. eeb.ucon.edu/estimates. [5] R.K. Colwell, J.A. Coddington, Estimating terrestrial biodiversity through extrapolation, Philos. T. Roy. Soc. B 345 (1994) 101–118. [6] C.M. Dash, K.B. Senapati, Cocoon morphology, hatching and emergence pattern in tropical earthworms, Pedobiologia (Jena) 20 (1980) 316–324. [7] T. Decaëns, J.J. Jiménez, Earthworm communities under an agricultural intensification gradient in Colombia, Pl. Soil 240 (2002) 133–143. [8] P. Feinseinger, E.E. Spears, R.W. Poole, A simple measure of niche breadth, Ecology 62 (1981) 27–32. [9] N.J. Gotelli, Null model analysis of species co-occurrence patterns, Ecology 81 (2000) 2606–2621. [10] N.J. Gotelli, A.E. Arnett, Biogeographic effects of red fire ant invasion, Ecol. Lett. 3 (2000) 257–261. [11] N.J. Gotelli, R.K. Colwell, Quantifying biodiversity: procedures and pitfalls in the measurement and comparison of species richness, Ecol. Lett. 4 (2001) 379–391. [12] N.J. Gotelli, G.L. Entsminger, Ecosim: Null Models Software for Ecology, Version 5.0. Acquired Intelligence Inc and

[20]

[21]

[22]

[23] [24]

[25] [26]

[27]

[28]

[29]

Kesey-Bear, Burlington, VT (1999) http://homepages.together. net/~gentsmin/ecosim.htm. J.J. Jiménez, A.G. Moreno, T. Decaëns, P. Lavelle, M.J. Fisher, R.J. Thomas, Earthworm communities in native savannas and man-made pastures of Eastern Plains of Colombia, Biol. Fertil. Soils 28 (1998) 101–110. J.J. Jiménez, J.-P. Rossi, P. Lavelle, Spatial distribution of earthworms in acid-soil savannas of eastern plains of Colombia, Applied Soil Ecology 17 (2001) 267–278. P. Lavelle, Les vers de terre de la savane de Lamto (Côted’Ivoire) : Peuplements, populations et fonction dans l’écosystème, Thèse de Doctorat, université Paris-VI. Publi Labo Zool ENS 12, 301 p., 1978. P. Lavelle, Relations entre types écologiques et profiles démographiques chez les vers de terre de la savane de Lamto, Rev. Ecol. Biol. Sol 16 (1979) 85–101. P. Lavelle, The soil fauna of tropical savannas. II. The earthworms, in: F. Bourlière (Ed.), Tropical Savanna, Elsevier Scientific Publishing, Amsterdam, 1983, pp. 485–504. P. Lavelle, I. Barois, Potential use of earthworms in tropical soils, in: C.A. Edward, A. Neuhauser (Eds.), Earthworms in Waste and Environmental Management, 1988, pp. 73–279. M. Loreau, S. Naeem, P. Inchausti, J. Bengtsson, J.P. Grime, A. Hector, D.U. Hooper, M.A. Huston, D. Raffaelli, B. Schmid, D. Tilman, D.A. Wardle, Biodiversity and ecosystem functioning: current knowledge and future challenges, Science 294 (2001) 804–808. P.S. Madge, Field and laboratory studies on the activities of two species of tropical earthworms, Pedobiologia (Jena) 9 (1969) 188–214. D.J. McCabe, N.J. Gotelli, Effects of disturbance frequency, intensity, and area on assemblages of stream macroinvertebrates, Oecologia 124 (2000) 270–279. C.E. Moreno, G. Halffter, Assessing the completeness of bat biodiversity inventories using species accumulation curves, J. Appl. Ecol. 37 (2000) 149–158. E.C. Pielou, The measurement of diversity in different types of biological collections, J. Theor. Biol. 13 (1966) 213–226. J.-P. Rossi, Rôle fonctionnel de la distribution spatiale des vers de terre dans une savane humide de Côte-d’Ivoire. Thèse de Doctorat, université Paris-VI, 240 p., 1998. L. Stone, A. Roberts, The checherboard score and species distributions, Oecologia 85 (1990) 74–79. J.E. Tondoh, Démographie et fonctionnement des populations du vers de terre Hyperiodrilus africanus (Eudrilidae) dans une savane protégée du feu en moyenne Côte-d’Ivoire. Thèse de Doctorat, université Paris-VI, 209 p., 1998. R. Vuattoux, Observation sur l’évolution des strates arborées et arbustives dans la savane de Lamto (Côte-d’Ivoire), Ann. Univ. Abidjan E 1 (1970) 285–315. A. Wills, I. Abbott, Landscape-scale species richness of earthworms in the Porongurup range, Western Australia: influence of aspect, soil fertility, and vegetation type, Biol. Fertil. Soils 39 (2003) 94–102. S.J. Willott, Species accumulation curves and the measure of sampling effort, J. Appl. Ecol. 38 (2001) 485–487.