Macroinvertebrate assemblages associated to break down the process of native and exotic plant leaves in a rain forest stream

Content

Introduction

Study area and methods

Results

Discussion

References

Macroinvertebrate assemblages associated to break down process of native and exotic plant leaves in a rain forest stream

Rodrigo Díaz-Lupanow

Laboratory of soil ecology, Ecology center,

Instituto Venezolano de Investigaciones Científicas

Benthic macroinvertebrate assemblages during leaves breakdown proccess of Ficus maxima and Bambusa vulgaris were investigated. Meshed bags with leaves were placed along a rainforest stream and progressibly recovered dur- ing 24 days, every time chemical variables were measured. Assemblages were compared reducing dimensionality with a NMSD analysis. An ANCOVA was applied to test differences in the remaining mass of the compared leaves, with time as covariate. NMDS resulted in 17% stress after dimensional reduction, allowing to identify assemblages differences throught time and between leaves. Native F. maxima presented faster decomposition and more complex macroin- vertebrate community than the exotic B. vulgaris. Different temporal assem- blages were found depending on the plant, being more evident in native species leaves, while the other was more redundant. This successional patterns alter- ations remark the danger of introduced species to the rain forest ecosystem.

Introduction

Allochthonous organic matter descomposition is an fundamental proccess of water streams en- ergetic income (Benfield 2006). For example, plant leaves play a key role in nutrient circulation, having determinant effect in ecosystem productivity and equilibrium (Marin et al., 2011). In this sense, leaves entering low order streams are subject to physical abrasion, microbial degradation and invertebrate fragmentation (Graca 2001). In the other hand, stream processes are known to affect nutrients cycling at basin scale, making nutrients dynamics of vital importance for ecosystem functioning (McClain et al. 2003). Stream long-term responses to disturbance (Ben- field et al. 2001), acid mine drainage (Niyogi et al., 2001), multiple land use (Sponseller and Benfield 2001) and agricultoral development (Niyogi et al. 2003) has been studied using leaf breakdown experiments.

Exotic species are one of the main reasons of biodiversity loss (Didham et al. 2005). The general impacts of alien plant on community structure and ecosystem processes are well known, but the changes that underlie these impacts are poorly understood (Hladys et al. 2011). This is more notorious in neotropical rain forest stream ecosystems, where the effect of exotic plants on macroinvertebrates has been shortly studied. In the Coastal Range in Venezuela there are many exotic species, and Bambusa vulgaris is one of those. It has some secondary compounds (Bastianoni et al. 2017), that are presumably involved in its defence and can remain active after leaf senescence (Rosenthal and Jensen, 1979).They can interfere with digestion or give a bitter taste to the macroinvertebrates, acting as a deterrent, among those compounds are mainly polyphenolics (tannins, lignins and others) ( Waterman and Mole, 1994).

In same order of ideas, the mencionated toxicity could have an effect on breakdown, both in their community assemblages and in the velocity in whitch labile matter is consumed. In this sense, I first aimed to search for successional patterns occurring in the overall breakdown proc- cess, and then, I will describe the structural and compositional temporal changes in macroinvertebrate community assemblages associated to breakdown of Bambusa vulgaris leaves, and compare it with a native species (Ficus maxima).

Study area and methods

Field data collection:

From november 21 to december 15 2016, a decomposition assay was done in ”Manantial” stream in Altos de Pipe, Miranda State, Venezuela. This creek is immersed in a rain forest of the Coastal range at 1400 m.a.s.l.(1023’18.90”N 6658’4.66”W). It was selected for its small size and easy access, as recomended by Beinfield (2006).

68 meshed bags were placed along the stream at approximately 2 cms under the flow, in contact with botton. Bags were fixed with cords to avoid removal by the current. One half of the bags was filled with near 3 g of Ficus maxima leafs, while the other half using Bambusa vulgaris leafs. 1, 4, 7, 10, 14, 21 and 24 days after experiment started, 8 bags were removed, 4 for each leaf plant species, and chemical variables were measured. Bags were washed and sieved (0.05 mm) in the laboratory to collect macroinvertrebrates, which families were identified using microscopes. The number of individuals of each family was registered. In cases when family recognition was not possible, a more general taxa was used instead. Samples were dried on the stove and weighted.

Statistics:

NMDS analysis:

Nonmetric Multidimensional Scaling (NMDS), with stable solution from random starts, axis scaling and obtation of species scores, was done with vegan package in R. Jaccard’s in- dices were computed as 2B/(1+B), where B is BrayCurtis dissimilarity. This is considered a robust measure of ecological distance (Faith et al. 1987). In addition, Kruskal’s Non-metric Multidimensional Scaling was applied to the community dissimilarity matrix using MASS R package. The procceadure results in scaled values and an associated stress (S) for the transformation. This is used for feature extraction, considerated to be one of the most important steps in machine learning and topological data analysis (Galili 2016), mainly used as ordination method (Minchin 1987). The Stress (S) is calculated as:

Abbildung in dieser Leseprobe nicht enthalten

S statistic serves as goodness of fit (Oksanen 2015). It consists in a equation of non-linear monotone transformation of the observed dissimilarities (Theta d) and the ordination distances d. The main goal of this process is to plot multivariate data points in two dimensions, thus revealing the structure of the dataset by visualizing the relative distance of the observations.

Function stressplot was used to draw a Shepard plot, where ordination distances are faced against community dissimilarities. The fit is shown as a monotone step line and allows to obtain two correlation like statistics of goodness of fit, which are based on stress (R² = 1S² ). The fit-based R² is the correlation between the fitted values (d) and ordination distances d, or between the step line and the points. The two correlations are calculated on the residuals, but there is a difference in their null models. In linear fit, the null model is that all ordination distances are equal, and the fit is a flat horizontal line. No-linear models do not have that assumption. Finally, rows with only zeros and species observed less than 2 times were removed from the NMDS community matrix.

Leaves remaining mass was corrected with the manipulation error and temperature effect. An analysis of covariance (ANCOVA) was performed to search for differences between remain- ing mass of the two species, with time as covariate. Linear model assumptions were checked before applying the test. In this sense, normality, homocedasticity and independence were ac- complished.

Table 1: Chemical features of ”Manantial” stream during experimental time Mean sd

Abbildung in dieser Leseprobe nicht enthalten

Hierarchical cluster analysis on the set of species Jaccard’s dissimilarities was carried out. The algorithm proceeds iteratively, at each stage joining the two most similar clusters, contin- uing until there is just a single cluster. At each stage distances between clusters are recom- puted by the LanceWilliams dissimilarity update formula according to the particular clustering method being used. Ward’s minimum variance method was used to find compact, spherical clusters (Murtagh et al. 2014).

Water chemistry: Every sample day, temperature, dissolved oxygen (DO), conductivity, pH, and total solids dissoved (TSD) were measured using a submersible sensor. Also, at day 14, sporulating hyphomycetes were identifed and measured in both plant leaves.

Results

R² equal to 0.97 for non-metric fit was found, while R² was 0.86 for linear fit. In addition, a stress of 0.17 was found after data dimensionallity reduction with NMDS.

Constrained and unconstrained PCA, CCA and RDA were tried out, but not enought vari- ance explanation was obtained (results not shown). Only when the variance-covariance matrix was used, high percent of the variance was explained, but there was not good ordination since big magnitude differences occur between species abundance. In such a cases, it is necessary to use the correlation matrix instead, but not fair enough variance was explained. NMDS is shown in figure 3, and it reflects the obtained ordination of the macroinvertebrate communities through

Abbildung in dieser Leseprobe nicht enthalten

Figure 1: NMDS ordination plot with ellipses for litter source . It shows macroinvertebrate assemblages differences in both litters througt time stages.

time.

Macroinvertebrate families richness was higher during native F. maxima breakdown than in exotic B. vulgaris (47 to 34 resp.). Chironomidae presented the highest relative total abundance (296 ind.), proportionally distributed in both kind of leaves. Same was found in Hydropsy- chidae, but sum of individuals was near 50 for each kind of leaf. Simmilar proportions of macroinvertebrates family individals occured in Planaria group, Coenagrionidae, Stratiomydae and Ceratopogonidae. The second highest abundance within families was registered for Calo- moceratidae, 111 individuals associated to F. maxima while 36 to B. vulgaris. This trend of preference for F. maxima rather than B. vulgaris was also observed in annelids, acari, dipters, Collembola, Empipidae, Elmidae and Hemiptera. By the other hand, highest abundances of

Abbildung in dieser Leseprobe nicht enthalten

Figure 2: NMDS ordination plot with ellipses for litter source . It shows macroinvertebrate assemblages differences depending on plant species litter.

Ptilodactylidae, Perlidae and Dixidae. Fig. 2 compresses the bynary Jaccard’s distances between the samples and shows polygons depending on leaves source.

During experimental time, Chironomidae abundance grew up exponentially, from 28 in the early stage to 205 at the end. Simmilar kind of poblational time pattern was registered for Hydropsychidae, whitch had 16 individuals at the fourth day and ended with 47. Dixidae had high 17 individuals at final stage, but only 3 and 1 at early and intermediate periods respec- tively. In contrast, strong decay at the last stage was found in Calomoceratidade, Coenagrion- idae and Elmidae. Several families presented highest numbers of individuals at intermediate phase, strongly represented in Collembola, Tipulidae and Stratiomydae, and more softly in Ptilodactylidae, anelids, hemiptera and Ceratopogonidae. Diptera was not present during inter-

Abbildung in dieser Leseprobe nicht enthalten

Figure 3: NMDS ordination plot with polygons successional assemblages in B. vulgaris litter.

mediate stage while acarus were only present in the first days.

Fig. 3 shows assemblages of macroinvertebrates in B. vulgaris leaves during three temporal stages. An interception zone between the stages relates to temporal generalist species, present during the entire experiment. This is the case for the late stage assemblage, conformed by species present at the early and intermediate species in B. vulgaris. By the other hand, early stage assemblage polygon has zones with no superposition with intermediate polygon.

In fig. 4 temporal assemblages using F. maxima were separed by polygons. At contrary that found in B. vulgaris, late stage shows little superposition with early stage, meaning a clear differentation between temporal assemblages. Intermediate stage is conformed by a mix of early and late stage, but with most of its area separated from the others. The overlapped area of the three polygons is relativelly small.

Abbildung in dieser Leseprobe nicht enthalten

Figure 4: NMDS ordination plot with polygons successional assemblages in F. maxima litter.

Chironomidae had its highest abundance during B. vulgaris late break down process. Same but more intense occurred with Dixidae individuals, whitch presented 80% of during late stage in B. vulgaris. While Calomoceratidae had more than 50 individuals in B. vulgaris early and intermediate stages, that number did not exceeded 15 in any period in F. maxima leaves . An- tagonistically, Collembola and anelids were more abundant in F. maxima intermediate phase. More over, Ptilodactylide was more numerosous during the last days of expermimental time in textit B vulgaris litter, while in F. maxima occurred at intermediate time. Such time traspotition was also observed in planaria, being found in early and intermediate stages in F. maxima, but predominantly late in B vulgaris.

In fig. 5 it is possible to appreciate assemblages both temporal and between litter differences. There is a big overlapping zone near Chironomidae and Ceratopogonidae. Early assemblage in F. maxima can be differenciated from intermediate and late stages in B. vulgaris, as well as late stage in same F. maxima. Both leaves presented overlapping late areas but the polygon is greater for F. maxima.

The cluster shown in fig. 6 confirms what had been told about the assemblages pattern differences, both in temporal and between species. Early and late stages in B. vulgaris are prone to be grouped. There is a big group mainly conformed by intermediate stages of both species. There is also a very big cluster having mostly F. maxima early stage grouped, with no late stage branches within it.

Fig. 7 represents leaves remaining mass after correction during 24 days on field. Half of the mass was lost after 7 days in F. maxima litter, while in B. vulgaris, experimental time was not enought to superate such magnitude.

In table 2, significant differences (alpha < 0.01) were found between species litter and between sampling days. As well, interaction specie:time was significant. 6 observations were deleted due to missingness. Also, mean squared of the residuals is quite low.

Abbildung in dieser Leseprobe nicht enthalten

Figure 5: NMDS ordination plot with ellipses of temporal stages for both species. It shows signs of succession proccess.

Table 2: Summary of ANCOVA results

Abbildung in dieser Leseprobe nicht enthalten

Figure 6: Community cluster analysis based on Jaccard distance

Abbildung in dieser Leseprobe nicht enthalten

Figure 7: Percentage of remaining mass throught time in leaves of a exotic and a native species. Red and blue lines are the prediction of linear regression. Standard deviation is shown around the trend. Sl. stands for slope while R2 for adjusted R squared.

Abbildung in dieser Leseprobe nicht enthalten

Table 3: Litter sporulating hyphomycetes. Data facilitated by M. Mendoza.

Abbildung in dieser Leseprobe nicht enthalten

Already leaves are colonised by microorganisms when fall from trees (Paul and 1983), but, associated biomass of microorganisms greatly increases after leaves enter the streams ( Suberkropp et al.1983). This increase is attributed to the colonisation by aquatic hyphomycetes (Barlocher, 1992). In the table 3 is shown the different fungi species and their sporulation rate found in the compared plant species. It is possible to appreciatte higher sporulation rate in F. maxima than in B. vulgaris. Bacterial colonization were not studied because leaves has been shown to be relatively unimportant in terms of biomass ( Baldy et al. 1995). While Lunulospara curvula had the lowest sporolation rate in B. vulgaris, it was the most prolific in the native species leaves. In constrast, Tetracladium marchalianum mantained high sporulation rates in both litter.

Table 1 remarks that initial Carbon and Nitrogen are little higher in B. vulgaris than F. maxima, but ratio between this elements is almost equal. By the other side, polyphoids are quite higher in B. vulgaris. In table 4 I show the results of Alessia et al. (2017)

Table 4: Some chemical features of F. maxima and B. vulgaris leaves. Mean (+-sd, n=3). Extracted from Bastianoni et al. (2017)

Abbildung in dieser Leseprobe nicht enthalten

Discussion

Both NMDS and cluster analysis showed that, although it is the same community of macroin- vertebrates associated to the breakingdown process for both species, there appeared temporal differences depending on leaf type. Moreover, this temporal assemblages are evidence of suc- cessional differences. While F. maxima had different early and late community assemblages, in the exotic plant the late stage was predominantly colonized by temporal generalist species.

Native F. maxima presented faster decomposition and more complex macroinvertebrate community than the exotic B. vulgaris. This greater diversity in the macroinvertebrate com- munity was associated with faster breakdown process. In the early stage, shredders like Calo- moceratidae had preferences for the native species. Calomoceratidae shredders are known to discriminate among the variety of leaves normally found in the stream. This may be related to differences in leaf toughness, plant nutrient content of leaves and the presence of secondary compounds (Graca 2001). Concentrations of lignin, tannin, cellulose, hemicellulose, nitrogen, and carbon change during leaf litter breakdown, and this has been associated with subsequent colonization and activity of decomposer flora and fauna (Hunter et al. 2003). In this sense, this study is a field confirmation of what has been found under laboratory conditions by Schulze and Walker (1997), that shredders are very selective on their feeding. More over, Ribeiro and Vieira (2013) found that Chironomidae and Oligochaeta were the most common macroinvertebrates involved on organic matter decomposition in streams in Brazil. Like that results, the agrupation done in this study showed that Chiromidae is the one of the main species of the process.

Some experimental issues can be considered. For example, erosion of litter mass caused by water flow could not be measuared. Also, hymenoptera and centipede were not part of the assemblages, they conformed particulate organic matter in the system and should be removed from the analysis. In addition, if the identification of the macroinvertebrates had have been up to the species taxon level, the temporal changes in community structure and composition may be more contrastating than observed in the experiment. With such information, functional successional assemblages during the process could be more precisely described. The big abundance of Chironomidae was also the cause of the ordination problems that PCA had compared with NMDS, because the last does a previous distance scaling.

Cummins et al. (2005) proposed a classification for functional groups. In that proposal, Chironomidae and Stratiomyiidae appears inside the gathering collectors group and Hydropsy- chidae as filtering collectors. Predators like acarus, whitch used the experimental leaves as hunting area at the beggining, while Hemiptera mostly at intermediate, showed temporal use of the alocthonous organic matter. Simmilarly, when shredders cannot obtain more matter from the leaves, they dissociate from community in order to colonize new feeding areas. This tem- poral differences in the community assemblages were more evident in the native species, while the exotic leaves were, in a temporal sense, colonized mainly by temporal generalist families.

The polyphenoids present in B. vulgaris may be one of the main reasons why assemblages and final decomposition rate differences that had been disscussed. Also, typical parallel form of nerves in monocotyledons can also be a factor that helps explaining the results, in the sense that shredders may be prone to choose textures more paladable for their fisionomy. In addition, the conditionation of leaves by bacteria and fungi can be selective, and this is key step in leaf litter decomposition (Pascoal and Cassio, 2004). This selectivity can affect upwards trophic trame. This can produce a cascade effect that determines in big part the final community assemblages and their temporal dynamics. Finally, if exotic B. vulgaris, or simmilar species, becomes more abundant in the neotropical rainforest ecosystem, the alterations of succesional dynamics that had been shown in this report could affect stream functionality and productivity.

References

1. Baldy, V., M. Gessner and E. Chauvet. 1995. Bacteria, fungi and the break- down of leaf litter in a large river. Oikos 74: 93102
2. Bastianoni A. J. Montoya, C. Mendez and J. Paulini. Evaluation of leaf litter leaching through commonly-used mathematicacl models. In revision.
3. Barlocher, F. (ed.). 1992. The ecology of aquatic hyphomycetes. Springer- Verlag, Berlin, 225 pp.
4. Benfield, E. F., J. R. Webster, J. L. Tank, and J. J. Hutchens. 2001. Long- term patterns in leaf breakdown in streams in response to watershed logging. International Revue of Hydrobiologie 86:467474.
5. Didham, R., J., Tylianakis, M., Hutchinson, R., Ewers, N., Gemmell. 2005. Are invasive species the drivers of ecological change?. DOI: dx.doi.org/10.1016/j.tree.2005.07.006. 6. Faith, D. P, Minchin, P. R. and Belbin, L. (1987). Compositional dissimilarity as a robust measure of ecological distance. Vegetatio 69, 5768.
7. Galili, T. 2016. Multidimensional Scaling with R (from Mastering Data Anal- ysis with R). Rbloggers. https://www.r-bloggers.com/multidimensional-scaling- with-r-from-mastering-data-analysis-with-r/
8. Graa, M. 2001. The Role of Invertebrates on Leaf Litter Decomposition in Streams a Review. International Review of Hydrobiology, 86: 383393. doi:10.1002/1522-2632(200107)86:4/5¡383::AID-IROH383¿3.0.CO;2-D
9. Hladys, S. , Abjornsson, K., Chauvet, E., Dobson, M., Elosei, A., Ferreira, V., Fleituch, T., Gessner, M. O., Giller, P.S, Gulis, V., Hutton, S.A., Lacoursiere, J.O., 2011. Stream ecosystem functioning in an agricultureal landscape? the importance of terrestrial-aquatic linkages. Adv. Ecol. Res. 44, 211-273.
10. Hunter M., Sina A., Catherine M. Pringle, David C. Coleman, Relative ef- fects of macroinvertebrates and habitat on the chemistry of litter during de- composition, Pedobiologia, Volume 47, Issue 2, 2003, Pages 101-115, ISSN 0031-4056, http://dx.doi.org/10.1078/0031-4056-00174.
11. Cummins K., R. Merrit and P. Andrade. 2005. The use of invertebrate func- tional groups to characterize ecosystem attributes in selected streams and rivers in south Brazil. Studies on Neotropical Fauna and Environment. 40(I):69-
12. McClain, M., Boyer, E., Dent, C., Gergel E., Grimm N., Groffman P., Hart S., Harvey J., Johnston C., Mayorga E., McDowell H., and Pinay G. 2003. Biogeochemical Hot Spots and Hot Moments at the Interface of Terrestrial and Aquatic Ecosystems. Ecosystems 6: 301. doi:10.1007/s10021-003-0161- 9.
13. Minchin, P.R. (1987) An evaluation of relative robustness of techniques for ecological ordinations. Vegetatio 69, 89107.
14. Murtagh, Fionn and Legendre, Pierre (2014). Ward’s hierarchical agglomera- tive clustering method: which algorithms implement Ward’s criterion? Journal of Classification 31 (forthcoming).
15. Niyogi, D. K., W. M. Lewis, Jr., and D. M. McKnight. 2001. Litter breakdown in mountain streams affected by mine drainage: biotic mediation of abiotic 18 controls. Ecological Applications 11:506516.
16. Niyogi, D. K., K. S. Simon, and C. R. Townsend. 2003. Breakdown of tussock grass in streams along a gradient of agricultural development in New Zealand. Freshwater Biology 48:16981708.
17. Pascoal C. and F. Cassio. 2004. Contribution of Fungi and Bacteria to leaf litter decomposition in a polluted river. Appl Environ Microbiol. 2004 Sep; 70(9): 52665273. doi: 10.1128/AEM.70.9.5266-5273.2004
18. Paul, M. and J., Meyer. 1996. Fungal biomass of 3 leaf litter species during decay in an Appalachian stream. J. N. Am. Benthol. Soc. 15: 421432.
19. Ribeiro A. and N. Vieira. 2013. Spacial and temporal variation of degradation of organic matter by benthic macroinvertebrates. Journal of environmental protection. DOI:10.4236/jep.2013.411A001
20. Rosenthal, G. A. and D. Jensen, 1979. Herbivores: their interaction with secondary plant metabolites, Academic press, N. Y
21. Schulze, D. and K. Walker. 1997. Riparian eucalypts and willows and their significance for aquatic invertebrates in the River Murray, South Australia. Regul. Rivers Res. Manage. 13: 557577.
22. Suberkropp K., T. L. Arsufii and J. P. Anderson. 1983. Comparison of degradative ability, enzymatic activity, and palability of aquatic hyphomycetes grown on leaf litter. Appl. Envir. Microbiol. 46: 237244.
23. Waterman, P. and S. Mole. 1994. Analysis of phenolic plant metabolites. Blackwell Scientific Publications, Oxford, 238pp