Effect of plant richness on the dynamics of coarse particulate organic matter in a Brazilian Savannah stream

Accepted Manuscript Title: Effect of plant richness on the dynamics of coarse particulate organic matter in a Brazilian Savannah stream Authors: Renan S. Rezende, Mariana A. Sales, Fernanda Hurbath, N´adia Roque, Jos´e F. Gonc¸alves-Junior, Adriana O. Medeiros PII: DOI: Reference:

S0075-9511(16)30157-8 http://dx.doi.org/doi:10.1016/j.limno.2017.02.002 LIMNO 25565

To appear in: Received date: Revised date: Accepted date:

11-10-2016 10-1-2017 6-2-2017

Please cite this article as: Rezende, Renan S., Sales, Mariana A., Hurbath, Fernanda, Roque, N´adia, Gonc¸alves-Junior, Jos´e F., Medeiros, Adriana O., Effect of plant richness on the dynamics of coarse particulate organic matter in a Brazilian Savannah stream.Limnologica http://dx.doi.org/10.1016/j.limno.2017.02.002 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Effect of plant richness on the dynamics of coarse particulate organic matter in a Brazilian Savannah stream.

Renan S. Rezende1,2*; Mariana A. Sales3; Fernanda Hurbath3; Nádia Roque3; José F. GonçalvesJunior2; Adriana O. Medeiros3

1

Program of Postgraduate in Ecology and Conservation, Federal University Rural of Semi-Arid,

CEP: 59.625-900, Rio Grande do Norte, Brazil.; 2

Department of Ecology, Institute of Biology, University of Brasília, 70190-900, Brasília, DF,

Brazil. 3

Department of Botany, Institute of Biology, Bahia Federal University, 40170-115, Salvador,

BA, Brazil;

*Corresponding author. E-mail: [email protected]

Highlights  The plant diversity is important for the quality and quantity (productivity) of CPOM in riparian vegetation.

1

 Vegetation representing Atlantic Forest, Brazilian Savannah and Dry Forest biomes were found in this study.  The dynamics of CPOM were strongly influenced by rainfall and by seasonal characteristics, such as the strong winds at the end of the dry season.  We observed a high productivity riparian zone, but low organic matter accumulation in this riparian zone.  The low number of indicator species was identified for months in association with their phenological responses to seasonal changes.  A high number of indicator species was identified for HI because rainfall water washes organic matter from the soil (bank) of adjacent terrestrial areas.

Abstract: The high plant richness in riparian zones of tropical forest streams and the relationship with an input of organic matter in these streams are not well understood. In this study, we assessed (i) the annual dynamics of inputs of coarse particulate organic matter (CPOM) in a tropical stream; and (ii) the relationship of species richness on riparian vegetation biomass. The fluxes and stock of CPOM inputs (vertical-VI=512, horizontal-HI=1912, and terrestrial-TI=383 g/m2/year) and the benthic stock (BS=67 g/m2/month) were separated into reproductive parts, vegetative parts and unidentified material. Leaves that entered the stream were identified and found to constitute 64 morphospecies. A positive relationship between species richness and litterfall was detected. The dynamics of CPOM were strongly influenced by rainfall and seasonal events, such as strong winds at the end of the dry season. Leaves contributed most to CPOM dynamics; leaf input was more intense at the end of the dry season (hydric stress) and the start of the rainy season (mechanical removal). Our study show an increase of litter input of CPOM by plant diversity throughout the year. Each riparian plant species contributes uniquely to the availability of energy resources, thus highlighting the importance of plant conservation for maintaining tropical streams functioning. Keywords: Organic detritus, Litter, Allochthonous, Brazilian stream.

2

INTRODUCTION Headwater streams in forested areas are characterized by dense riparian vegetation that limits the penetration of light to the streambed, resulting in low stream autotrophic productivity (Vannote et al., 1980). Riparian vegetation also is the main source of energy for food web of these streams by coarse particulate organic matter (CPOM, material >1 mm; Webster and Meyer, 1997). CPOM derived from riparian vegetation includes various parts of plants (e.g. leaves, branches, flowers and fruits; Cummins, 1974). Leaves comprise the largest fraction of CPOM, contributing more than 50% of the total organic matter (Gonçalves and Callisto, 2013; Molinero and Pozo, 2004; Webster and Meyer, 1997). The inputs and forms of CPOM that derive from riparian vegetation vary in quantity and quality and depend on the characteristics of the riparian vegetation and the season (Abelho, 2001; Bambi et al., 2016; Tank et al., 2010). CPOM from riparian vegetation, in the form of litterfall goes into the stream by vertical (VI) and horizontal input (HI). VI was defined as CPOM entering a stream directly, manly by senescent leaves. HI was defined as CPOM entering a stream indirectly by organic matter runoff at the margins to stream water due to wind and animal transport. In addition, terrestrial input (TI), derived from the leaf litter and out of streams flood pulse (part decomposes in the soil), represents the potential stock that can be transported into the stream via HI. The variations in VI, HI and TI that occur at multiple spatial and temporal scales can influence the dynamics of CPOM inputs and stock (Benfield et al., 2001; Gonçalves and Callisto, 2013; Webster and 3

Meyer, 1997). The CPOM inputs and stock are also impact the remobilization of nutrients in the trophic chain of streams (Magana, 2001; Magana and Bretschko, 2003; Rezende et al., 2016). Several studies have focused on elucidating organic matter dynamics in tropical streams (Afonso et al., 2000; Benson and Pearson, 1993; Carvalho and Uieda, 2010; França et al., 2009; Gonçalves et al., 2014; Larned, 2000; Londe et al., 2016; Magana and Bretschko, 2003; Rezende et al., 2016). The dynamics of organic matter in tropical riparian vegetation is typically diver by phenological patterns (França et al., 2009; Gonçalves and Callisto, 2013) and the leaf senescence are often more pronounced during drier periods by hydrical stress (França et al., 2009). However, the relationship between dynamics of organic matter and plant species richness is largely unknown, and most studies consider only the dominant species (Afonso et al., 2000; Carvalho and Uieda, 2010; Magana, 2001). A review by Ives and Carpenter (2007), about stability and diversity in terrestrial systems show a temporal stability of biomass production due to increase of biological diversity, also supported by mathematics models (Valone and Barber, 2008). Therefore, identification of key species for ecosystem functioning and the quantification of plant species richness (and their contribution to ecosystems) in the riparian zone plus the phenological changes that occur throughout the year are important for tropical streams understanding (Rezende et al., 2016; Wantzen et al., 2008). The amount of organic matter input can serve as an effective environmental indicator because of its efficiency and sensitivity in detecting changes in riparian plant species biomass (Gonçalves et al., 2014; Machado et al., 2008). This study may therefore help to resolve important ecological questions, such as (i) how is ecosystem function (e.g., CPOM input) altered by biodiversity change (e.g., plant species contributing to the CPOM input), and (ii) how important are less common species in the functioning of ecosystems? These questions were 4

suggested to be two of the 100 fundamental questions in pure ecology with practical relevance for the conservation of biodiversity and ecosystem function (Sutherland et al., 2013). Therefore, we tested the following hypotheses: (i) the litterfall will be dominated by a small number of species (for more see also Gonçalves et al., 2014; Lisboa et al., 2014; Rezende et al., 2016), and an increase in species richness would not change the CPOM input of riparian vegetation (i.e., by low contribution of infrequent species); (ii) the hydric stress together with first rainfalls have positive effects in the CPOM inputs due the senescence stage of leaves and the removing physical by rain. Our goals were to (i) measure the quantity, (ii) describe species composition, and (iii) identify key species of the inputs and the stock of CPOM over 12 months in a Brazilian Savannah stream.

METHODS Study area This study was conducted in a second-order stretch of the Boiadeiro stream (12°59'43.8" S, 41°19'35.6" W; elevation 900 m) , which is located in Park City Mucugê (PCM) in Chapada Diamantina – Bahia, a transitional area between the Cerrado (Brazilian savannah) and Caatinga (Brazilian dry forest) biomes in northeastern Brazil. The surrounding landscape is composed of rocky fields and outcrops vegetated primarily with species associated with a physiognomy of herbaceous shrubs growing on quartzite soils (rupestrian fields). The climate is mesothermal, with a mean annual temperature of approximately 22°C, low temperatures in the winter, and a rainy season (October to March) and a dry season (April to September). The rivers of this region are channeled through fractures within large rocks and consequently feature a regime of rapid 5

flows during periods of rain (CPRM, 1994). The stretch of the Boiadeiro stream included in this study measured 50 m in length and has a width of approximately 10 m (horizontal distance of the water body between the boundaries of riparian vegetation), with dense bank vegetation and a closed canopy. Rainfall data were obtained from the National Agency of Waters of Brazil meteorological station in the city of Mucugê (station number 1241033; located at 13°1'37.1994" S, 41°13'16.32" W) from the National Agency of Waters of Brazil, available on the website hidroweb (http://hidroweb.ana.gov.br/).

Procedures Litterfall was measured monthly from January to December 2011 (standardized as 30 d with an acceptable deviation of ± 2 d). The organic matter falling directly to the ground (TI) was estimated using 10 nets (1 m2, 1-mm mesh size, located 10-m apart). CPOM entering directly into the river (VI) was measured using 90 buckets (0.53 m2) that were suspended by ropes 2 m above the stream; they were transversely positioned in 5 rows (replicates in the stream), with 3 rows of 6 buckets per site (18 buckets in total) and 10 m between adjacent rows (buckets were the replicates). The bucket bottoms were perforated to allow rainwater to escape. At monthly intervals, the CPOM accumulated in the buckets was retrieved and weighed in situ (wet weight), and the bucket with the highest leaf-litter mass in each row was used for the humidity correction. The horizontal inputs (HI) were collected in nets with an opening of 0.1 m² (0.2 m in height, 0.5 m in width, 1 mm mesh size) that were installed along both banks of the stream (10 on each bank for a total of 20 samples). Two samples of benthic stocks (BS) were collected at each of 5 sampling points (April to December) using a Surber sampler (area of 0.45 m², 250 µm mesh size;

6

total = 10 samples). The sample design was similar to that described by Sales et al. (2015) and Rezende et al. (2016). The material collected from each compartment (VI, HI, TI and BS) were taken to the laboratory and dried in an oven at 60°C for 72 h and weighed (precision = 0.005 g); then, the material was sorted into reproductive parts (flowers, fruits), vegetative parts (branches, leaves) and unidentified material (others). Leaves were identified to species according to the Angiosperm Phylogeny Group II system (APGII, 2003), and CPOM that could not be identified to species were sorted into morphospecies. Each leaf species was weighed to determine each species contribution to composition of the CPOM. We consider the most representative plant species in terms of the biomass of leaf material in the inputs and the stock as those species that contribute in excess of 5% of the total annual contribution. In the Table 1, the literature search included personal literature databases and journal indices (in Web of Science -n=15- and Scopus -n=12-). The search terms used in online databases were “litterfall” and “tropical streams” and “riparian vegetation”. Studies were included if they satisfied the following criteria: (i) data of litterfall; (ii) performed in natural riparian zones of tropical and subtropical; (iii) data of number of plant species and/or leaf percentage in litterfall.

Data analysis The contribution of high species richness of riparian vegetation to ecosystem functioning was tested by linear regression of the natural logarithm of the litterfall input along months (Crawley, 2007). A factorial two-way repeated measures ANOVA (RM-ANOVA) was used to test for differences in CPOM mass (dependent variable) among the 3 inputs (VI, HI and TI) over 7

time (months/categorical variables) for each plant part (leaves, reproductive parts, branches and miscellaneous). The collecting instruments (buckets in VI; nets in HI and TI) were used as the repeated measurements in the RM-ANOVA (Crawley, 2007). The RM-ANOVA is commonly used in experiments with different plot sizes and error variances (see also chapter 11 of Crawley, 2007); in our study the factor of repeated measures was applied to the samples, which were collected monthly (intrinsic correction of fixed position of the collection device). A contrast analysis was used to assess the differences among the categorical variables; associations between some variables were evaluated using linear correlations (Crawley, 2007). Data normality was assessed with a Kolmogorov-Smirnov test, and the homogeneity of variance was assessed using Levene’s test; when necessary, the values were Ln transformed if needed to meet these assumptions. The variation in species composition of CPOM (dependent variable) was tested among inputs (VI, HI and TI) and months (categorical variable) by a multivariate analysis of variance (PerMANOVA). For the PerMANOVA, we used a Bray-Curtis resemblance matrix and 10,000 permutations to obtain the pseudo-F value; p < 0.05 was considered significant (Adonis function, vegan package for R; Oksanen et al., 2008). To determine the key plant species in terms of leaf contribution to VI, HI, TI and BS, data for all plant species identified were subjected to an indicator species analysis (Dufrêne and Legendre, 1997), which provides an indicator value (IV) ranging between 0 (for a non-indicator) to 1 (for a perfect indicator); in the OM input of the study stream. It is important to emphasize that this analysis was based on the frequency and abundance of the plant species as determined by their leaf contributions to CPOM to the stream and selection of the species with the highest contribution to litterfall inputs. The significance of the IV was determined by performing a Monte Carlo test with 10,000 permutations (p < 0.05).

8

RESULTS Plant community contribution We identified 64 plant morphospecies that contributed to the dynamics of CPOM in the riparian zone of the stretch of the Boiadeiro stream included in this study. We detected a positive relationship between increasing species richness and litterfall biomass over the months of sampling (linear regression; p < 0.01, r2 = 0.97, n = 12; Figure 1). Of the 64 morphospecies, 11 could not be identified at the species level; the other 53 (30 tree, 17 shrub, 3 herb, 2 liana and 1 vine), which were distributed among 30 families and 45 genera (Table MS 1), are common species of the Amazonian Forest, Atlantic Rain Forest, Cerrado and Caatinga. Bonnetia stricta, Doliocarpus elegans, Laplacea fructicosa, Richeria grandis, and Vochysia acuminata contributed 66% of the total leaf material entering the stream. According to the VI, B. stricta, D. elegans, Clusia criuva, L. fructicosa, R. grandis, Tibouchina barnebyana and V. acuminata contributed 95% of leaf input. Seven species – D. elegans, Ilex theezans, L. fructicosa, R. grandis, Tapirira obtusa, T. barnebyana and V. acuminata – accounted for 77% of total TI leaf input, whereas HI was dominated by B. stricta, L. fructicosa, R. grandis, and V. acuminata (52.6%), and 68% of the BS was composed of material from B. stricta, L. fructicosa, R. grandis, T. obtusa and V. acuminata.

Dynamics of coarse particulate organic matter

9

Leaves were more common than other plant parts in all inputs and stocks, and they accounted for 74%, 44%, 70% and 57% of VI, HI, TI and BS, respectively. The second most common part were branches (8%, 15%, 14% and 33%, respectively), followed by miscellaneous plant material (7%, 35%, 6% and 7%, respectively) and flowers/fruits (11%, 6%, 10% and 3%, respectively). The total quantity of CPOM input to the stream was 4-fold greater in the horizontal input (HI = 1912 g m-2 year-1 or 159 g m-2 month-1 ) compared to the vertical input (VI = 512 g m2

year-1 or 42 g m-2 month-1 ) and 5-fold greater than the terrestrial input (TI = 383 g m-2 year-1 or

32 g m-2 month-1; Figure 2). The standing BS was 804 g m-2 year-1 or 67 g m-2 month -1. The VI (RM-ANOVA, F(11, 1067) = 5.1, p < 0.001) was highest in October (69 g m-2 month -1) and November (89 g m-2 month -1), whereas the highest HI (RM-ANOVA, F(11, 227) = 10.1, p < 0.001) were recorded in April (332 g m-2 month -1), May (461 g m-2 month -1) and December (637 g m-2 month -1). The highest TI (RM-ANOVA, F(11, 107) = 3.1, p = 0.001) was observed in October and November (59 g m-2 month -1 for both). However, the BS (RM-ANOVA, F(11, 125) = 0.8, p = 0.521) remained constant over the course of the study period (analysis of contrast < 0.05; Figure 3). The TI (r = 0.57, p = 0.04) and VI (r = 0.58, p = 0.04) were correlated with rainfall, that is, the highest quantity of litterfall coincided with rainy events, but HI (r = 0.32, p = 0.31) and BS (r = -0.37, p = 0.32) were not significantly correlated with rainfall. The total organic matter, leaves, reproductive parts (flowers and fruits), branches and miscellaneous other vegetational material varied significantly among inputs (HI, VI and TI) and over months, as did their interaction (Table 2). Among inputs, HI differed significantly from VI and TI, but VI was not significantly different from TI. All plant parts varied significantly over time, however, with the highest values occurring as follows: for total organic matter, in December; for leaves, from October–December; for reproductive parts, in March; for branches,

10

in May and December; and for miscellaneous material, in April, May and December (analysis of contrast < 0.05; Table 2). The PerMANOVA revealed that both the composition and proportion of the species in CPOM differed among the 3 inputs (HI, VI and TI) but not over time (Table 2). Of the 64 morphospecies identified from CPOM collected in the stream, 29 were selected as indicator species based on their contributions to total leaf material (Table 3). The plant species that were selected by the analysis of indicator species for TI were S. peltiger and W. paulliniifolia; for HI, 26 indicator species were identified, 3 of which could not be identified at the genera or family level (species A, B and D); and for the BS, Gomidesia blanchetiana was selected as the only indicator species. S. peltigera and W. paulliniifolia contributed leaf material solely to TI, and no indicator species were identified for VI. However, when compared over the months, T. obtusa and T. barnebyana were selected as the indicator species for October, and Geonoma sp. was selected for December. Regarding the mode of growth, trees contributed more than 85% of leaf material to all inputs; trees represented approximately 87% of total contribution of leaf material to VI, 89% to HI, and 91% to both TI and the BS. Lianas accounted for approximately 12%, 3%, 7% and 6% of leaf material to VI, HI, TI and BS, respectively, with shrubs and herbs making up the difference.

DISCUSSION Plant community diversity We found a positive relationship between the number of taxa and litterfall biomass, with a majority percent of leaf material (66%) originated from only five species (Bonnetia stricta, Doliocarpus elegans, Laplacea fructicosa, Richeria grandis and Vochysia acuminata). Other 11

species, such as T. barnebyana (for VI) and Ilex theezans (for TI), were dominant only for specific input types. We therefore consider these species to be important components of ecosystem functioning of the stream (Gonçalves et al., 2014; Rezende et al., 2016). However, total diversity (e.g., lianas, shrubs, herbs) is also important for this ecosystem because leaf-litter quality, quantity and composition changed throughout the year (input peaks by the phenological – see below; Wantzen et al., 2008). Moreover, species that form the riparian canopy, such as B. stricta and V. acuminata, were found to contribute large amounts of CPOM to streams compared to herbaceous plants in the understory (Afonso et al., 2000). Therefore, the positive relationship between the number of taxa and litterfall highlights the importance of conserving riparian plants for maintenance of ecological processes in stream ecosystems. The high number of species (50 plant species per ha) can be explained by the ecotone nature of riparian zones, which may support much higher diversity than permanent aquatic or terrestrial habitats (Oliveira-Filho and Ratter, 1994). Temperate riparian systems average 10 plant species (Wantzen et al., 2008), but upward of 100 plant species may be found in tropical riparian systems (Kiew, 1998). Gonçalves and Callisto (2013), for example, recorded 192 plant species in a half-ha plot of a Savannah rupestrian field, whereas Gonçalves et al. (2014) identified 83 plant species in a ha plot of Atlantic Forest. However, Rezende et al. (2016), in a study of a specific Savannah stream found only 8 species in the litterfall (due to extreme soil conditions). The variation in the abundance of different taxa demonstrates the high diversity of plants in tropical riparian systems, which is associated with the environmental conditions (e.g., soil, nutrient and water availability), phytofisionomic type and altitude (Gonçalves et al., 2014; Rezende et al., 2016). This higher variation in the number of taxa, combined with the few studies in which the plant species have been identified from the CPOM of tropical systems (Table 1),

12

indicate a need for additional research concerning species identification from litterfall to ascertain a clear pattern in tropical systems. The high number of plant species of tropical riparian systems can result in differences in CPOM input directly into the stream (VI) compared to falls in riparian zones (TI). It is important to evaluate TI in tropical riparian systems because (i) the higher diversity and random distribution of plant species in riparian systems; (ii) VI and TI are not equal (quality and quantity), and, (iii) fragmented and decomposing leaves are not efficiently accounted for in the horizontal input (HI). Therefore, the riparian zones phenology are important for understanding the VI and TI (Naiman et al., 2000). In addition, the plant species identified in the study have broad phytogeographic domains, such as the Amazon Rainforest (e.g., Roucheria columbiana), Atlantic Forest (e.g., Clusia criuva), Savannah (Cerrado; e.g., Tibouchina barnebyana and Humiria balsamifera) and Dry Forest (Caatinga; e.g., Couma rigida). As such, tropical riparian vegetation should not be considered as part of a single forest domain (Méio et al., 2003; OliveiraFilho and Ratter, 1994). Therefore, the riparian vegetation is a integrated system that functions as an ecological corridor for species dispersal of different tropical physiognomies (Méio et al., 2003).

Riparian influences on the dynamics of coarse particulate organic matter The amount of CPOM in the VI (512 g m-2 year-1) is high in this stream compared to other Cerrado streams (range of 288 – 336 g m-2 year-1 in França et al., 2009; Gonçalves and Callisto, 2013; Gonçalves et al., 2006; Rezende et al., 2016), but it was in the lower part of the range reported for other tropical forests (113 – 2812 g m-2 year-1; reviewed by Abelho, 2001; 13

Alvarez et al., 2009; Cogo and Santos, 2013; Gregório et al., 2007; Lisboa et al., 2014; Zhou et al., 2007). However, the actual inputs of organic matter available for decomposers could be higher considering the VI and HI (1912 g m-2 year-1) together, indicating that indirect inputs are the most important inputs to ecosystem functioning for our system. The material that falls on the stream margins is transported immediately into the aquatic system, and factors such as wind, rain and floods may all influence the horizontal transport of organic matter (Pozo et al., 1997). Therefore, our results demonstrate that the understory did not retain litterfall and that a large amount of the litterfall in the soil enters into the stream via horizontal inputs, which is also the case in other tropical riparian zones (França et al., 2009). The BS of organic matter in the stream was low (801 g m-2 year-1) compared to other studies of streams in the Brazilian savannah, considering the organic matter entering the stream via the VI and HI (range of 483 – 3792 g m-2 year-1 in Chara et al., 2005; França et al., 2009; Gonçalves and Callisto, 2013; Gonçalves et al., 2006). Thus, two factors can be responsible (separately or synergistically) for generating this pattern: (i) the stretch of the stream included in our study has a high transport and low retention capacity for CPOM (Li and Dudgeon, 2011; Magana and Bretschko, 2003); and/or (ii) the rate of processing organic matter via decomposition is high (as showed by Rezende et al., 2016; Sales et al., 2015). Therefore, are important additional research to examine the role of decomposition and organic matter dynamics in the functioning of tropical riparian ecosystems (Rezende et al., 2016). We believe that future research focusing on organic matter transport (inputs and outputs) and mineralization (gas fluxes, e.g., CO2) would greatly aid in better understanding both the functioning of and the anthropic impacts on tropical riparian systems.

14

As with other studies of tropical streams (Wantzen et al., 2008; Zhang et al., 2014), leaves, followed by branches, were the plant parts that made up the largest fraction of CPOM, and CPOM inputs were higher at the end of the dry season and the start of the rainy season (transition period). This finding suggests two types of leaf and branch inputs may be entering the streams (Afonso et al., 2000): (i) senescent leaves in the dry season as the result of hydric stress (Larned, 2000; Scheer et al., 2009); and (ii) green leaves and branches that are mechanically removed and fragmented by heavy rains and wind at the onset of the rainy season (Gonçalves et al., 2014; VanSchaik et al., 1993). These results also explain the higher inputs of all the inputs (HI, VI and TI) during the October to December period as well as the significant correlation between CPOM input and rainfall. However, the BS did not vary seasonally and remained constant during the study period, as was also observed by Ferreira et al. (2013) for streams in Portugal; this correlation suggests that organic matter stocks and inputs have different seasonal dynamics. Variation of the composition and proportion of plant species among the inputs over seasons explains the high number of plant species identified as indicator species. When assessed temporally, T. obtuse and T. barnebyana were characteristic of October and Geonoma sp. was characteristic of December, most likely because of greater phenological responses to precipitation. The higher number of indicator species for HI due to rain water runoff washing organic matter off the soil (bank) from the adjacent areas (rocky fields) into the stream, as well as inputs transported from tributaries. Moreover, TI and VI are unequal because the indicator plant species for HI demonstrated that there were more important species than we found in VI and TI. Thus, there is a random distribution of plant species in the riparian zone, suggesting a diversity of inputs in tropical riparian systems (Chave et al., 2010; Tank et al., 2010).

15

Therefore, we can conclude that although sampling a small stretch of riparian vegetation requires cautious interpretation, the riparian vegetation of Savannah streams are not composed of a single forest domain (i.e., it functions as an ecological corridor of Brazilian biomes). The increase in species richness change positively the CPOM input of riparian vegetation, refuting our first hypothesis. The dynamics of CPOM were strongly influenced by seasonal characteristics. Leaves, followed by branches, were the plant parts that contributed the most to the CPOM dynamics in the study stream. The inputs were higher at the end of the dry season (hydric stress) and the start of the rainy season (mechanical removal and strong winds), corroborating our second hyphothesis. We also observed a high litter input in riparian zone, but low organic matter accumulation in this riparian zone. This result highlights the importance, fragillity, and necessity for the conservation of the riparian vegetation for the functioning of Brazilian watersheds. Finally, assuming that litterfall is largely a function of climate, future changes in climate could have negative effects on the organic matter dynamics of Savannah streams.

ACKNOWLEDGMENTS This study was financed by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior CAPES through the projects PROCAD NF (process no. 306/2010) and PNADB (Process no. 517/2010) projects. Mariana Aguiar received a scholarship from CAPES and Renan S. Rezende received a Post-Doctoral Scholarship from State University of Nova Xavantina-MT (CAPES/PNPD and CNPq -number 151375/2014-3). The authors are grateful to the employees of Sempre Viva Park for their assistance with field work.

16

LITERATURE CITED Abelho, M., 2001. From litterfall to breakdown in streams: a review. Scientific World Journal 1, 656-680. Afonso, A.A.d.O., Henry, R., Rodella, R.C.S.M., 2000. Allochthonous matter input in two different stretches of a headstream (Itatinga, São Paulo, Brazil). Brazilian Archives of Biology and Technology 43, 335-343. Alvarez, J.A., Villagra, P.E., Rossi, B.E., Cesca, E.M., 2009. Spatial and temporal litterfall heterogeneity generated by woody species in the Central Monte desert. Plant Ecology 205, 295-303. APGII, A.P.G., 2003. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG II. Botanical Journal of the Linnean Society 141, 399-436. Bambi, P., de Souza Rezende, R., Feio, M.J., Leite, G.F.M., Alvin, E., Quintão, J.M.B., Araújo, F., Gonçalves Júnior, J.F., 2016. Temporal and Spatial Patterns in Inputs and Stock of Organic Matter in Savannah Streams of Central Brazil. Ecosystems. Benfield, E.F., Webster, J.R., Tank, J.L., Hutchens, J.J., 2001. Long-term patterns in leaf breakdown in streams in response to watershed logging. International Review of Hydrobiology 86, 467-474. Benson, L.J., Pearson, R.G., 1993. Litter inputs to a tropical Australian rainforest stream. Australian Journal of Ecology 18, 377-383. Carvalho, E.M., Uieda, V.S., 2010. Input of litter in deforested and forested areas of a tropical headstream. Brazilian Journal of Biology 70, 283-288. Chara, J., Baird, D., Telfer, T., 2005. Allochthonous Matter Input in Five Headwater Streams in Southwestern Colombia, American Geophysical Union, Spring Meeting. Chave, J., Navarrete, D., Almeida, S., Álvarez, E., Aragão, L.E.O.C., Bonal, D., Châtelet, P., Silva-Espejo, J.E., Goret, J.-Y., Hildebrand, P.v., Jiménez, E., Patiño, S., Peñuela, M.C., Phillips, O.L., Stevenson, P., Malhi, Y., 2010. Regional and seasonal patterns of litterfall in tropical South America. Biogeosciences 7, 43–55.

17

Cogo, G.B., Santos, S., 2013. The role of aeglids in shredding organic matter in neotropical streams. Journal of Crustacean Biology 33, 519-526. Crawley, M.J., 2007. The R Book. John Wiley & Sons Ltd, England. Dufrêne, M., Legendre, P., 1997. Species assemblages and indicator species: the need for a flexible asymmetrical approach. Ecological Monographs 67, 345–366. Ferreira, V., Lírio, A.V., Rosa, J., Canhoto, C., 2013. Annual organic matter dynamics in a small temperate mountain stream. Annales de Limnologie - International Journal of Limnology 49, 13-19. França, J.S., Gregorio, R.S., de Paula, J.D.A., Gonçalves Junior, J.F., Ferreira, F.A., Callisto, M., 2009. Composition and dynamics of allochthonous organic matter inputs and benthic stock in a Brazilian stream. Marine and Freshwater Research 60, 990-998. Gonçalves, J.F., de Souza Rezende, R., Gregório, R.S., Valentin, G.C., 2014. Relationship between dynamics of litterfall and riparian plant species in a tropical stream. Limnologica - Ecology and Management of Inland Waters 44, 40-48. Gonçalves, J.J.F., Callisto, M., 2013. Organic-matter dynamics in the riparian zone of a tropical headwater stream in Southern Brasil. Aquatic Botany 109, 8-13. Gonçalves, J.J.F., França, J.S., Callisto, M., 2006. Dynamics of allochthonous organic matter in a tropical Brazilian headstream. Brazilian Archives of Biology and Technology 49, 967973. Gregório, R.S., Valentin, G., Ferreira, F.A., Aleixo, L.A., França, J.S., Gonçalves, J.F.J., Callisto, M., Batista, M.L., Gaspar, R.d.O., Rodello, C.M., 2007. Contribuição Foliar Alóctone de Espécies Vegetais num Córrego de 2ª Ordem na Estação Ambiental de Peti (CEMIG) - MG. Revista Brasileira de Biociências 5, 33-35. Ives, A.R., Carpenter, S.R., 2007. Stability and Diversity of Ecosystems. Science 317, 58-62. Kiew, R., 1998. Species richness and endemism, in: Soepadmo, E. (Ed.), The Encyclopedia of Malaysia: Plants. Archipelago Press, Singapore, pp. 14–15. Larned, S.T., 2000. Dynamics of coarse riparian detritus in a Hawaiian stream ecosystem: a comparison of drought and post-drought conditions. Journal of the North American Benthological Society 19, 215–234. Li, A.O.Y., Dudgeon, D., 2011. Leaf litter retention in tropical streams in Hong Kong. Fundamental and Applied Limnology / Archiv für Hydrobiologie 178, 159-170. 18

Lisboa, L.K., Silva, A.L.d., Siegloch, A., Junior, J.G., Petrucio, M., 2014. Temporal dynamics of allochthonous coarse particulate organic matter in a subtropical Atlantic Rainforest Brazilian stream. Marine and Freshwater Research 1, 1-14. Londe, V., De Sousa, H.C., Kozovits, A.R., 2016. Litterfall as an indicator of productivity and recovery of ecological functions in a rehabilitated riparian forest at das velhas river, southeast brazil. Tropical Ecology 57, 355-360. Machado, M.R., Rodrigues, F.C.M.P., Pereira, M.G., 2008. Produção de serapilheira como bioindicador de recuperação em plantio adensado de revegetação. Revista Árvore 32, 143-151. Magana, A.M., 2001. Litter input from riparian vegetation to streams: a case study of the Njoro River, Kenya. Hydrobiologia 458, 141-149. Magana, A.M., Bretschko, G., 2003. Retention of Coarse Particulate Organic Matter on the Sediments of Njoro River, Kenya. International Review of Hydrobiology 88, 414-426. Méio, B.B., Freitas, C.V., Jatobá, L., Silva, M.E.F., Ribeiro, J.F., Henriques, R.P.B., 2003. Influência da flora das florestas Amazônica e Atlântica na vegetação do cerrado sensu stricto. Revista Brasileira Botânica 26, 37-444. Molinero, J., Pozo, J., 2004. Impact of a eucalyptus (Eucalyptus globulus Labill.) plantation on the nutrient content and dynamics of coarse particulate organic matter (CPOM) in a small stream. Hydrobiologia 528, 143-165. Naiman, R.J., Bilby, R.E., Bisson, P.A., 2000. Riparian ecology and management in the Pacific coastal rain forest. Bioscience 50, 996–1011. Oksanen, J., Kindt, R., Legendre, P., O’Hara, B., Simpson, G.L., Solymos, P., Stevens, M.H.H., Wagner, H., 2008. Adonis function, Vegan: Community Ecology Package. R package version 1.13-1., pp. 15–10. Oliveira-Filho, A.T., Ratter, J.A., 1994. A study of the origin of central Brazilian forests by the analysis of plant species distribution patterns. Journal of Botany 52, 141–194. Pozo, J., González, E., Díez, J.R., Molinero, J., Elósegui, A., 1997. Inputs of Particulate Organic Matter to Streams with Different Riparian Vegetation. Journal of the North American Benthological Society 16, 602-611.

19

Rezende, R.d.S., Graça, M.A.S., Santos, A.M., Medeiros, A.O., Santos, P.F., Nunes, Y.R., Junior, J.F.G., 2016. Organic Matter Dynamics in a Tropical Gallery Forest in a Grassland Landscape. Biotropica 1, xx-xx. Sales, M.A., Goncalves, J.F., Jr., Dahora, J.S., Medeiros, A.O., 2015. Influence of Leaf Quality in Microbial Decomposition in a Headwater Stream in the Brazilian Cerrado: a 1-Year Study. Microbial Ecology 69, 84-94. Scheer, M.B., Gatti, G., Celina Wisniewski, Mocochinski, A.Y., Cavassani, A.T., Lorenzetto, A., Putini, F., 2009. Patterns of litter production in a secondary alluvial Atlantic Rain Forest in southern Brazil. Revista Brasileira de Botânica 32, 805 - 817. Sutherland, W.J., Freckleton, R.P., Godfray, H.C.J., Beissinger, S.R., Benton, T., Cameron, D.D., Carmel, Y., Coomes, D.A., Coulson, T., Emmerson, M.C., Hails, R.S., Hays, G.C., Hodgson, D.J., Hutchings, M.J., Johnson, D., Jones, J.P.G., Keeling, M.J., Kokko, H., Kunin, W.E., Lambin, X., Lewis, O.T., Malhi, Y., Mieszkowska, N., Milner-Gulland, E.J., Norris, K., Phillimore, A.B., Purves, D.W., Reid, J.M., Reuman, D.C., Thompson, K., Travis, J.M.J., Turnbull, L.A., Wardle, D.A., Wiegand, T., Gibson, D., 2013. Identification of 100 fundamental ecological questions. Journal of Ecology 101, 58-67. Tank, J.L., Rosi-Marshall, E.J., Griffiths, N.A., Entrekin, S.A., Stephen, M.L., 2010. A review of allochthonous organic matter dynamics and metabolism in streams. Journal of the North American Benthological Society 29, 118-146. Valone, T.J., Barber, N.A., 2008. An empirical evaluation of the insurance hypothesis in diversity–stability models. Ecology 89, 522-531. Vannote, R.L., Minshall, G.W., Cummins, K.W., Sedell, J.R., Cushing, C.E., 1980. River Continuuum Concept. Canadian Journal of Fisheries and Aquatic Sciences 37, 130-137. VanSchaik, C.P., Terborgh, J.W., Wright, S.J., 1993. The Phenology of Tropical Forests: Adaptive Significance and Consequences for Primary Consumers. Annual Review of Ecology and Systematics 4, 353-377. Wantzen, K.M., Yule, C.M., Mathooko, J.M., Pringle, C., 2008. Organic matter processing in tropical streams, in: Dudgeon, D. (Ed.), Tropical Stream Ecology. Elsevier Inc, Amsterdam, Netherlands, pp. 43-64.

20

Webster, J.R., Meyer, J.L., 1997. Organic matter budgets for stream: a synthesis. In: Webster JR, Meyer JL. Stream organic matter budgets. Journal of the North American Benthological Society 16, 3-161. Zhang, H., Yuan, W., Dong, W., Liu, S., 2014. Seasonal patterns of litterfall in forest ecosystem worldwide. Ecological Complexity 20, 240–247. Zhou, G., Guan, L., Wei, X., Zhang, D., Zhang, Q., Yan, J., Wen, D., Liu, J., Liu, S., Huang, Z., Kong, G., Mo, J., Yu, Q., 2007. Litterfall Production Along Successional and Altitudinal Gradients of Subtropical Monsoon Evergreen Broadleaved Forests in Guangdong, China. Plant Ecology 188, 77-89.

21

Figure 1. Regression of the natural logarithm of the proportion of species richness and litterfall biomass over the sampling time period (circle = dry season; triangle = rainy season).

Figure 2. Estimated biomasses of the vertical, horizontal and terrestrial inputs and benthic standing stocks (mean ± SE) for the Boiadeiro stream over the 12-month (January to December 2011) study period.

22

Figure 3. Monthly values of precipitation (rainfall in mm) and biomasses (mean ± SE) of the (A) vertical inputs, (B) horizontal inputs, (C) terrestrial inputs (January to December 2011) and (D) benthic standing stocks (April to December 2011) for the Boiadeiro stream.

23

Table 1 Comparison of litterfall (direct input in g m-2 year-1) production, % of Leaf (%L) and plant richness (PR) in riparian vegetation in various tropical and subtropical forest types of the world.

Location Northeast Brazil

Central of Brazil Central of Brazil Central of Brazil Southeastern Brazil Southeastern Brazil

Climat e Tropic al Tropic al Tropic al Tropic al Tropic al Tropic al

Litterfal P l %L R Reference 6 Brazilian Savannah 512 72 4 This study 1 60- 1 Brazilian Savannah 1236.5 80 2 Bambi et al., 2016 60- 7 Brazilian Savannah 715 80 0 Bambi et al., 2016 60- 2 Brazilian Savannah 147 80 9 Bambi et al., 2016 Rezende et al., Brazilian Savannah (Veredas) 158 63 8 2016 9 Atlantic rainforest 361 75 9 França et al., 2009 1 Brazilian Savannah 9 Conçalves & (Rupestrian fields) 302 73 2 Callisto, 2013 Brazilian Savannah 1 Conçalves et al., (Rupestrian fields) 24 50 4 2006 2 Gregorio et al., Atlantic rainforest 167 73 7 2007 8 Conçalves et al., Atlantic rainforest 697 60 4 2014 Carvalho & Atlantic rainforest 1218 100 Uieda, 2010 Atlantic rainforest (Open 3 Afonso et al., Area) 86 79 5 2000 Atlantic rainforest (Closed 3 Afonso et al., Area) 713 73 1 2000 1406 to 56 to Tropical Rain Forest 2812 72 - Chara et al., 2005 Vegetation

Tropic al Tropic Southeastern Brazil al Tropic Southeastern Brazil al Tropic Southeastern Brazil al Tropic Southeastern Brazil al Tropic Southeastern Brazil al Tropic Southeastern Brazil al Southwestern Tropic Colombia al Rift Valley Tropic Province of Kenya al Savannah (Closed canopy site) Rift Valley Tropic Province of Kenya al Savannah (Open canopy site) Tropic Southeastern Brazil al Brazilian Savannah Tropic Northeast of Australian al Notophyll vine forest Tropic North of Hawaii al Tropical forest Southeastern Brazil

286

40

-

Magana, 2001

128

40

840

65

484

70

1058

80

Magana, 2001 2 7 Londe et al., 2016 2 Benson & Pearson 9 1993 5 0 Larned, 2000 24

Central Monte desert of Argentina Central Monte desert of Argentina Central Monte desert of Argentina

Southern Brazil Southern Brazil Southern China Southern China Southern China

Subtro pical Subtro pical Subtro pical Subtro pical Subtro pical Subtro pical Subtro pical Subtro pical

Semiarid to arid (Exposed areas) Semiarid to arid (Under Larrea vegetation) Semiarid to arid (Under Prosopis vegetation)

Alvarez et al., 2009 Alvarez et al., 2009 Alvarez et al., 2009

53

27

5

189

4

5

354

62

Atlantic rainforest

389

58

5 1 2 2

Semideciduous forest

957

61

-

Lisboa et al., 2014 Cogo & Santos, 2013

Pine forest Mixed Pine and Broadleaved forest Monsoon evergreen broadleaved forest

356

75

-

Zhou et al., 2007

861

68

-

Zhou et al., 2007

849

57

-

Zhou et al., 2007

25

Table 2. Results of comparisons among categories of organic matter inputs (factorial two-way RMANOVA), species detritus composition (PerMANOVA) and analysis of contrast (p < 0.05) for the fluxes and study months. (Vertical-VI, Horizontal-HI, and Terrestrial-TI inputs).

d.f.

Sum Sq (%)

Error: collector

1

1.7

Fluxes

2

7.6

87.1

Month

11

4.5

9.4

Fluxes x Month

22 140 3

25.2

26.2

Error: collector

1

1.2

Fluxes

2

4.2

43.2

Month

11

6.6

12.2

Fluxes x Month

22 140 3

19.7

18.4

Error: collector

1

0.0

Fluxes

2

3.3

29.2

Month

11

8.4

13.6

Fluxes x Month

22 140 3

9.5

7.6

78.8

Error: collector

1

1.6

Fluxes

2

6.0

59.1

Month

11

3.4

5.9

Test Factorial two-way RMANOVA

F

p

Analysis of Contrast (p > 0.05)

Total Organic Matter

Residuals

< 0.001 < 0.001 < 0.001

TI = VI < HI Other Months < April = May < December

61.0

Leaves

Residuals

< 0.001 < 0.001 < 0.001

TI = VI < HI Other Months < October = November = December

68.3

Flowers and Fruits

Residuals

< 0.001 < 0.001 < 0.001

TI = VI < HI Other Months < April = December < Frebruary < March

Branches < 0.001 < 0.001

VI = TI < HI Other Months < May = December

26

Fluxes x Month

17.8

Error: collector

1

1.8

Fluxes

2

9.1

111. 5

Month

11

5.0

11.6

Fluxes x Month

22 140 3

26.6

29.4

< 0.001 < 0.001 < 0.001

Fluxes

11

25.4

5.7

< 0.001

Residuals

24

74.6

Month

2

40.0

1.5

0.06

Residuals

33

60.0

Residuals

15.8

< 0.001

22 140 3

71.3

Miscellaneous

Residuals

TI = VI < HI Other Months < April = December = May

57.5

PerMANOVA of Plant Species TI ≠ VI ≠ HI

27

Table 3. Plant species selected by analysis of indicator species with their respective importance into fluxes (Vertical-VI, Horizontal-HI, and Terrestrial-TI inputs) and stock (Benthic Stock-BS), indicator values (VI) (% of perfect indication, based on combining the values for relative abundance and relative frequency) and P-values.

Indicator Species

Fluxes and Stock

IV

p

Gomidesia blanchetiana

BS

0.37

0.010

Humiria balsamifera

HI

0.95

< 0.001

Blechnum scrrultum

HI

0.93

< 0.001

Chamaecrista cytisoides

HI

0.81

< 0.001

Baccharis retusa

HI

0.76

< 0.001

Myrcia hiemals

HI

0.74

< 0.001

Retinifillum laxiflora

HI

0.65

< 0.001

Alchornea triplinervia

HI

0.65

0.032

Stilingia saxatilis

HI

0.62

< 0.001

Illex affinis

HI

0.61

0.002

Pagamea guianensis

HI

0.60

0.006

Cyathea delgadii

HI

0.59

< 0.001

Persea aurata

HI

0.56

0.011

Baccharis oblongifolia

HI

0.55

0.003

Myrsine umbellata

HI

0.49

0.024

Laplacea fruticosa

HI

0.47

0.007

Vocysia acuminata

HI

0.46

0.044

Myrcia guianensis

HI

0.41

0.008

Bysonimia morii

HI

0.37

0.015

Acritopappus heterolepis

HI

0.33

0.016

Sapium hiematospermim

HI

0.29

0.026

Species A

HI

0.6

0.003

Species B

HI

0.87

< 0.001

Species D

HI

0.46

0.004

Stipecoma peltigera

TI

0.75

< 0.001

Weinmannia paulliniifolia

TI

0.69

0.002

0.42

0.005

Months Tapirira obtusa

October

28

Tibouchina barnebyana Geonoma sp.

October

0.17

December

0.55

0.008 0.017

29