Composition and distribution of diatom assemblages from core and surface sediments of a water supply reservoir in Southeastern Brazil

Fresh water biodiversity is an increasing concern due to growing human impact. Herein, we report a long-term survey (ca. 90 years) of sedimentary diatoms and the modern flora from surface sediments and their biodiversity changes along a eutrophication gradient. Study was carried out in one of the most important water supply reservoirs (Guarapiranga Reservoir) of São Paulo Metropolitan Region, Brazil. Results are based on 75 core subsamples (subfossil assemblages from core) previously dated by Pb and 14 samples from surface sediments (modern assemblages). Overall, 84 taxa were reported, belonging to 30 genera, 71 species and eight non-typical varieties, besides five probable new taxa. Results expanded two new additions for the Brazilian diatom flora (Chamaepinnularia submuscicula and Stauroneis acidoclinata) and 30 infrageneric taxa for the state of São Paulo. 47.6% of total taxa inventoried were accounted exclusively for the subfossil assemblages indicating a significant biodiversity change over time. Access to past oligotrophic conditions and to contemporary mesotrophic regions of the Guarapiranga Reservoir accounted for these new additions representing 25% of the total diatom flora. Decline in the total species number along the trophic state gradient occurred for subfossil and modern assemblages. This pattern was even clearer when considering the changes in species richness over time. Eunotia with 21 taxa was the far most represented genera particularly in the oligotrophic phase. During the transitional period (1947-1974), richness gradually declined. With the onset (in the 1970s) and the major eutrophication period (since ca. 1990) occurred a drastic reduction in richness and the replacement of oligotrophic to eutrophic species. Human management also caused abrupt changes in richness. Marked decline occurred (1933) assotiated with hydrological impacts (water discharge increase) with the initial use of the reservoir as a public water supply. Unlike, sudden increase occurred probably associated with the application of algaecide to control cyanobacterial blooms. Present findings highlight the need for surveying the diatom assemblages in protected environments or in less degraded conditions for biodiversity assessment. Furthermore, reinforce the use of paleolimnological approach as in many cases the only tool to assess biodiversity changes encompassing time scales relevant to human-induced degradation and pre-anthropogenic impacts.


Introduction
Despite covering just 0.8% of the Earth's surface, freshwater ecosystems are considered hotspots for biodiversity supporting B6% of all described species (Dudgeon et al. 2006).However, the growing anthropogenic impacts in the last century has led to growing threats to fresh water biodiversity as well as to the largely ''unknown'' diversity worldwide (Strayer & Dudgeon 2010).Decreases in biodiversity are so widespread that they are now considered a form of global change (Gregory-Eaves & Beisner 2011).Therefore, information on biodiversity changes in long timescales has become an important issue of freshwater ecology and conservation.
Given long-term community data are sparse and usually span no more than five years, the paleolimnological approach has recently been highlithed as an emerging field for biodiversity science (Gregory-Eaves & Beisner 2011).Lake sediments integrate organisms over time and space, different habitats, providing whole-lake, annual to multi-annual assemblage information more efficiently than neolimnological studies (Bennion 1995, Gregory-Eaves & Beisner 2011).They can provide valuable information about past and contemporary environmental conditions, having good records of biodiversity (Froyd & Willis 2008, Liu et al. 2012, Davidson et al. 2013) and floristic changes (Schmidt et al. 1990).
Among the biological groups preserved in the sediments, diatoms have been widely used because of their taxonomic distinction, abundance, good preservation in sediments and their rapid response to environmental changes (Reid 2005, Bennion & Simpson 2011).The use of diatoms as indicators of environmental changes require high taxonomic precision (Birks 1994), since misidentifications can modify the interpretation in obtaining reliable data on modern diatom biodiversity (Buczko ´& Magyari 2007, Wetzel & Ector 2014).Also relevant is the auto-ecological knowledge of diatom species in order to use them as modern analogues in quantitative paleo-environmental reconstruction using sediment records of past communities (Birks 1994).Moreover, past communities are very often the only available tool to provide information on natural biodiversity before human impacts, such as cultural eutrophication.For those purposes, floristic surveys in long time series are considered crucial (Schmidt et al. 1990), though rarely available in the world, especially in tropical regions There has been an increasing knowledge of the diatom flora in some regions of Brazil (e.g. Ferrari & Ludwig 2007, Melo et al. 2010, Wetzel et al. 2010, Santos et al. 2011, Bartozek et al. 2013).However, only recently taxonomical studies of surface sediments have began (Fontana & Bicudo 2009, 2012, Silva & Bicudo 2014), including the addition of new species (Wengrat et al. 2015, Almeida et al. 2015).
Despite the studies on paleoenvironmental reconstruction using diatoms (e.g.Costa- Bo ¨ddeker et al. 2012, Fontana et al. 2014), to our knowledge floristic and taxonomical studies of sedimentary diatoms in long timescale in Brazil and probably in tropical regions have not been published.We presently documented the floristic survey and the spatial-temporal distribution of diatoms from the surface sediments (modern flora) and core (ca.90 years, subfossil flora) of the Guarapiranga Reservoir along a spatial and temporal eutrophication gradient.The subfossil diatom assemblage was taxonomically studied and revised based on the ecological study of Fontana et al. (2014).The Guarapiranga Reservoir is one of the most important public water supplies for the metropolitan region of Sa ˜o Paulo.Present study expands the knowledge of biodiversity changes, taxonomical and ecological information of tropical diatoms, contributing to their use in water quality bioassessment and paleoenvironmental reconstruction.

Material and Methods
Guarapiranga Reservoir is a strategic public water supply located in one of the most urbanized cities worldwide, the Sa ˜o Paulo Metropolitan Region (SPMR), in the state of Sa ˜o Paulo, southeastern Brazil (23°41'S, 46°43W) (Figure 1).The SPMR is one of the most important financial, commercial and industrial centers in Brazil, and one of the most densely populated areas of the country, with nearly 20 million inhabitants (IBGE 2014).The reservoir has an area of 36.18 km 2 , mean and maximum depth of 7 and 13 m, respectively, and a water volume of 253 Â 106 m 3 (Mozeto et al. 2001).It was built in 1906-1909 for energy production, and dam construction flooded a large portion of Atlantic Forest habitat (Whately & Cunha 2006).In 1927, the city of Sa ˜o Paulo began to use the reservoir as a public water supply, and today the reservoir is the main water source for the city, supplying drinking water to about 25% of the population.Fontana et al. (2014) inferred the major ecological shifts (using diatoms and geochemical proxies) in the water body over the last 90 years related to multiple stressors, mainly the influence of forest flooding and eutrophication.The reservoir was oligotrophic from 1919 to 1947 and the onset of eutrophication occurred in the mid-1970s.By the early 1980s the reservoir had become eutrophic, in response to an explosive increase in human population in its watershed.Severe cultural eutrophication has persisted since 1990.Further information on the major shifts of the reservoir is available in Fontana et al. (2014).
Core was retrieved by divers from the northern area of the basin, close to the dam (Figure 1).The core chronology was determined by 210 Pb dating as detailed in Fontana et al. (2014).Divers collected a 75-cm core in February 2010 using acrylic tube that was sectioned at 1 cm intervals.In total, 75 subsamples (slices) were examined (subfossil diatom flora).In addition, 14 samples of surface sediments (modern diatom flora) were collected in August/2011, using a gravity corer (UWITEC), and the first 2 cm of the sediments were saved for diatom analysis.Sampling stations for surface sediments covered the trophic spatial gradient of the reservoir (Figure 1).Limnological characteristics of these sites are provided in Table 1 (according to AcquaSed database Project).Subsurface samples in the limnetic zone were taken with a van Dorn sampler in the dry (August/2011) and rainy seasons (March/2011).Water temperature (°C), pH and conductivity (mS cm -1 ) were measured in the field using standard electrodes (Horiba U-53).The analytical procedure for dissolved oxygen (DO, mg L -1 ), ammonium (N-NH 4 mg L -1 ), nitrate (N-NO 3 mg L -1 ), soluble reactive silica (SRS, mg L -1 ), total nitrogen (TN, mg L -1 ) and total phosphorus (TP, mg L -1 ) followed Standard Methods (APHA 2005).Chlorophyll a (mg L -1 ), corrected for phaeophytin, was measured using 90% ethanol (Sartory & Grobbelaar 1984).The Trophic State Index (TSI) was calculated according to Lamparelli (2004).For details, see Wengrat & Bicudo (2011).Sampling sites 1 to 5 were considered mesotrophic, and 6 to 11 and 13 to 14 eutrophic, while site 12 was classified as supereutrophic (Table 1).
Diatom samples were oxidized according to standard procedures (Battarbee et al. 2001), using concentrated hydrogen peroxide (H 2 O 2 , 35%) and hydrochloric acid (HCl 37%).Oxidized subsamples were rinsed with deionized water and permanent slides were prepared using Naphrax as mounting medium.Optical observations, measurements and micrographs were taken at a magnification of 1000 Â with a Zeiss Axioskop 2 plus microscope equipped with DIC and phase contrast, with an Axiocam ERc5s high-resolution digital camera.Micrographs were digitally manipulated and plates containing LM images were created using CorelDraw X6.Morphometric information is provided for all taxa (L: length; W: width; D: diameter; M: mantle height; S: striae; AS: apical striae; MS: median striae; F: fibulae, A: areolae; AC: alar canals; MF: mantle fultoportulae) as well as temporal and spatial distribution in the reservoir.Descriptions are presented for the new records for Brazil, and comments are provided when relevant (e.g.poorly known species worldwide or in Brazil).Taxonomy and nomenclature followed classic works and new publications (e.g., Hustedt 1950, Krammer 2000, Metzeltin et al. 2005, Lange-Bertalot et al. 2011) and the on-line catalogue of valid names (site of California Academy of Sciences 2012).The classification systems followed Medlin & Kaczmarska (2004) for supra-ordinal taxa and Round et al. (1990) for subordinal taxa, except for genera published subsequently to this work.To account for the species distribution in Brazil and the state of Sa ˜o Paulo, literature with illustration or sufficient taxonomic description of the species was considered.Sediment samples were deposited at the ''Herba ´rio Cientı ´fico do Estado Maria Eneyda P. Kauffmann Fidalgo'' (SP), Brazil.Finally, to determine species richness (Magurran 2004) diatom was quantified to standardize the analytical procedure among samples.Enumeration was made at a magnification of http://www.scielo.br/bn1000 Â using a Zeiss Axioskop 2 microscope, and at least 400 valves were counted per slide (Battarbee et al. 2001).

Taxonomy and ecological preferences
Below are presented the taxonomical aspects of the species identified in this study and their ecological preferences.Taxa preceded by one asterisk represent new records for the state of Sa ˜o Paulo, and those preceded by two asterisks are first citations for Brazil.The infrageneric taxa commonly reported in Brazilian literature are shown in Table 2.
It differs from Aulacoseira granulata (Ehrenberg) Simonsen var.granulata mainly due to its greater diameter (18.0-31.0mm) and the presence of visible rimoportulae in LM in valve surface (Moro 1991).No ecological information was found in literature.This variety occurred in 2% of all samples in eutrophic conditions for subfossil and modern assemblages.This is the first report for the state of Sa ˜o Paulo.

Orthoseiraceae Ku ¨tzing
Orthoseira Thwaites Orthoseira roseana (Rabenhorst) O'Meara, Proceedings of the Royal Irirsh Academy, 2 p. 255, pl. 26, 1875. Figs. 16-18. D: 13.1-17.3This species differs from C. meneghiniana Ku ¨tzing due to the marginal area with short striae and distinct fultoportulae at every third, fourth or fifth (seldom sixth to seventh) appearing as thicker striae than the others (shadowlines), and the presence of a single,  In this study, the species occurred in 79% of all samples in modern assemblages from mesotrophic to supereutrophic conditions.First report for the state of Sa ˜o Paulo.
The species was reported in acidic and black waters in the Amazon basin (Melo et al. 2010).It occurred in 21% of modern assemblage samples in mesotrophic condition.This is the first report for the state of Sa ˜o Paulo.
It occurred in 36% of all samples from subfossil assemblages during past oligotrophic conditions.First report for the state of Sa ˜o Paulo.
Eunotia paludosa presents higher striae density (E: 18-25 in 10 mm) and more protracted ends (Lange-Bertalot et al. 2011) than the individuals observed in this study.According to these authors, this is an acidophilic species.It was reported in 7% of samples of subfossil assemblages during past oligotrophic conditions.First report for the state of Sa ˜o Paulo.Eunotia tukanorum is a planktonic species proposed by Wetzel et al. (2010) for Negro River (North Brazil), whose waters are generally oligotrophic and characterized by the presence of humic acids due to decomposition of flooded vegetation during the rainy season.In other regions of Brazil, the species has been reported as E. asterionelloides Hustedt in the plankton of oligotrophic and slightly acidic rivers (Laux & Torgan 2011), and in plankton and periphyton of a pond in the South region (Bicca et al. 2011).

Eunotia pseudosudetica
In this study, the species was reported in subfossil assemblages in oligotrophic condition, period characterized by flooded vegetation during the reservoir construction.Fontana et al. (2014) registered dominance of E. tukanorum during this phase along with the high increase in water discharge associated with the initial use of the reservoir as a public water supply.It was also reported for modern assemblages in mesotrophic conditions.This is a common species in the study area, occurring in 56% of all samples.So far, the occurrence of this species is restricted to tropical and subtropical regions and seems to be mainly associated with oligotrophic and slightly acidic environments.Although the species was cited in Fontana et al. (2014), this is the first taxonomical register for the state of Sa ˜o Paulo.
The taxon was reported in 36% of all samples in subfossil assemblages in past oligotrophic conditions.

Gomphonemataceae Ku ¨tzing
Gomphonema L: 25.6-34.6mm; W: 4.4-6.0mm; S: 14-16 in 10 mm.This species was previously cited for Brazil and identified as Gomphonema sp. 1 by Silva et al. (2010) and Bertolli et al. (2010).It is characterized by the presence of cuneate and slightly curved apexes, unlike Gomphonema hawaiiense Reichardt, whose apexes are attenuated.The examined population is in agreement with Kobayasi et al. (2006).Ecological information is not available in the literature.Currently, it was found in 14% of all samples in recent mesotrophic conditions.It is the first report for the the state of Sa ˜o Paulo.1994, Hofmann 1994).In this study it occurred in only 3% of all samples from subfossil assemblages in oligo to eutrophic conditions.Gomphonema lagenula Ku ¨tzing, Die Kieselschaligen Bacillarien oder Diatomeen, p. 85, pl. 30, fig. 60, 1844. Figs. 90-91.L: 16.2-23.2mm; W: 5.6-6.0 mm; S: 14-16 in 10 mm.Present in oligotrophic waters (van Dam et al. 1994).Our findings expand its range distribution, from oligo to eutrophic conditions.It occurred in 27% of all samples for subfossil and modern assemblages.Achnanthidium catenatum is very similar to Achnanthidium minutissimum Ku ¨tzing on valve view, but easily recognized in girdle view due to the "C" shaped-valves.Furthermore, A. catenatum presents a widened central portion resulting in an undulated valve margin (Hlu ´bikova et al. 2011).Bicudo et al. (2009) first registered this species in the state of Sa ˜o Paulo.However, this species was probably previously misidentified as A. minutitimum.Achnanthidium catenatum is an indicator of organic pollution (Berthon et al. 2011).In a paleoreconstruction of eutrophication of a Brazilian reservoir, this species highlighted the onset of a marked eutrophication phase (Costa- Bo ¨ddeker et al. 2012).In the present study, it was very frequent occurring in 71% of all samples from oligo to eutrophic conditions for subfossil and modern assemblages.Acording to Fontana et al. (2014), this species achived higher abundances during the major cultural eutrophication phase of Guarapiranga Reservoir.Overall, this species is probably an indicator of an environmental shift particularly associated with the eutrophication process.
Brachysira microcephala is a highly polymorphic (ranging from lanceolate to capitate forms) and cosmopolitan diatom distributed in clean and circumneutral to slightly acidic waters (Denys & Straaten 1992, Wolfe & Kling 2001).It is commonly found in periphyton and metaphyton worldwide (Czarnecki et al. 1995, Potapova & Charles 2003, Ibarra et al. 2009), and in pristine environments in paleolimnological studies (Grenier et al. 2006).In the present study it occurred in 42% of all samples, including subfossil (oligotrophic conditions) and modern assemblages (mesotrophic to eutrophic conditions).
Sellaphora capitata differs from S. pupula (Ku ¨tzing) Mereschkowsky by presenting subcapitated ends, sinuous raphe, lower striae density (16-22 in 10 mm) that are strongly radiated across the valve surface (Mann et al. 2004).This is a poorly known species in Brazil, and was probably only recorded for the southern region of Brazil by Santos et al. (2011).No ecological information is available.It presently occurred in 21% of all samples in modern assemblages from mesotrophic to eutrophic conditions.This is the species' first citation for the state of Sa ˜o Paulo.
Sellaphora rectangularis belongs to the group 'pupula' (Mann et al. 2008), but differs from S. pupula (Ku ¨tzing) Mereschkovsky by presenting linear elliptical valves, with broadly rounded poles, and parallel valve edge or slightly convex.Can be confused with Sellaphora laevissima (Ku ¨tzing) Mann, however this has grooves enclosing the raphe system and polar bars absent.
Reported in mesotrophic waters (van Dam et al. 1994).In Brazil this taxon was recorded for the Central-Western (Delgado & Souza 2007) and Southern regions (Santos et al. 2011).In our data, it was a common species occurring in 44% of all samples in subfossil assemblages during past oligotrophic conditions.This is the species' first citation for the state of Sa ˜o Paulo.
The species has probably only been cited by Santos et al. (2011) for the Southern region of Brazil.Ecological information is not available in literature.In the present study, it was registered in 14% of all samples in recent mesotrophic conditions.This is the species' first report for the state of Sa ˜o Paulo.
Reported in oligotrophic waters (van Dam et al. 1994).In this study, it was registered in 14% of all samples in recent mesotrophic to eutrophic conditions.This is the species' first register for the state of Sa ˜o Paulo.L: 27.5-29.9mm; W: 7.5-8.0mm; S: 6-7 in 10 mm.Registered in oligo-mesotrophic waters (Hofmann 1994, van Dam et al. 1994).Presently distributed in 7% of all samples in past oligotrophic phase of the reservoir.
Valves eliptic-lanceolate; margin straight to slightly convex; axial area asymmetric; central area asymmetric, proximal ends in hooked shape, deflected to the same side; robust striae, radiate in the center becoming slight convergent towards the ends.L: 82.6 mm; W: 10.0 mm; S: 11 in 10 mm.
The taxon resembles Pinnularia toscana Krammer, however the later has wider valve (17.0-20.0mm) and bent raphe (Krammer 2000).It occurred in 21% of all samples for subfossil assemblages only during past oligotrophic conditions.
It differs from Chamaepinnularia mediocris (Krasske) Lange-Bertalot because the second species presents slightly swollen valves in the median portion (Metzeltin & Witkowski 1996).No ecological information was available in literature.In this study, C. submuscicola was found in 7% of all samples in recent mesotrophic conditions.This is the first report of species for Brazil.
No ecological information was found.It was rare, occurring in 1% of all samples in past oligotrophic conditions.
L: 82.8 mm; W: 10.7 mm; inconspicuous striae.Reported in eutrophic waters (Luchini & Verona 1972).Our data expanded its distribution for oligotrophic conditions.It was found in 7% of all samples in past oligotrophic phase of the reservoir.Although cited in ecological studies, this is the species' first taxonomical citation for the state of Sa ˜o Paulo.
The studied population presented larger valves dimensions and lower striae density than proposed in Werun & Lange-Bertalot (2004; L: 35-60 mm; W: 8.5-10.5 mm, 25-30 striae in 10 mm).According to these authors, S. acidoclinata was found in a fountain in Germany with low conductivity water, associated with acidophilic Eunotia species.Further ecological information was not found.In this study, the species also occurred with other abundant Eunotia species and in 24% of all samples during past oligotrophic phase of the reservoir.First citation for Brazil.
Nitzschia fruticosa Hustedt is characterized by the presence of stellate colonies.Although no colonies were found in sediment samples, further analyses of planktonic materials showed typical stellate colonies, allowing its identification.
Nitzschia fruticosa was registered in Brazil for planktonic samples in pond and rivers in southern Brazil (Moro & Fu ¨rstenberger 1993, Laux & Torgan 2011).No ecological information was available in literature.In the present study, the species was reported in 59% of all samples for subfossil assemblages in a broad environmental range from past oligotrophic to eutrophic conditions, although mainly in the eutrophic phase.This is the first taxonomical report of the species in the state of Sa ˜o Paulo.No ecological information was found.The species was widespread in modern assemblages, occurring in 93% of all samples in mesotrophic to supereutrophic conditions.*Nitzschia terrestris (Petersen) Hustedt, Abhandlungen der Ko ¨niglichen Akademie der Wissenschaften zu Berlin 8(9), p. 386, 1934.This species is not well known in Brazil, and was registered in the periphyton by Santos et al. (2011).Ecological information was not found.Currently, the species occurred in 30% of all samples in subfossil assemblages in past oligotrophic phase of the reservoir and in modern assemblages in mesotrophic conditions.Although

Richness and eutrophication gradient
The subfossil and modern diatom flora of Guarapiranga comprised 84 infrageneric taxa from which 47.6% were exclusively from the subfossil assemblages.Eunotia was by far the most represented genus in species number, reaching a 3.5 times greater number than the second ranking genera Aulacoseira, Gomphonema and Pinnularia (Figure 165A), and this trend was mostly accounted by the distribution of subfossil diatoms (Figure 165B).Eunotia was mainly found during the initial oligotrophic phase of the reservoir (ca. 1919-1932), characterized by flooded vegetation (Atlantic Forest) for the reservoir construction.This early phase probably had a wellilluminated water column, oligotrophic and acidic waters due to the dissolved humic substances originated from the decomposition of vegetation (Fontana et al. 2014), favouring this commonly benthic/periphytic genus usually abundant in acidic oligotrophic waters (Krammer & Lange-Bertalot 1991, van Dam et al. 1994, Wetzel et al. 2010, Lange-Bertalot et al. 2011).Differently, modern assemblages presented the species number more uniformly distributed among the genera (Figure 165C).
Concerning diatom distribution according to trophic state range (Figure 166), the species number for the oligotrophic condition was markedly higher in the subfossil assemblages, where this phase was uniquely represented in Guarapiranga Reservoir.Furthermore, a decline in the total species number along the trophic state gradient was observed for subfossil and modern assemblages (Figure 166).This pattern was even clearer when considering the changes in species richness (number of species per sample) over time following trophic state changes in the reservoir (Figure 167).Thus, during preeutrophication period (up to ca. 1947), richness presented high values, achieving its highest figure over time and space in the reservoir history.However, a marked decline occurred around 1933 associated with physical and hydrological impacts when started the use of the reservoir as a public water supply and the water discharge had a seven-fold increase (Fontana et al. 2014).According to these authors, this change led to dominance of Eunotia tukanorum, a typical planktonic species of the genus, and disappearance of benthic species.During the transitional period with moderate cultural eutrophication (ca. 1947-1974) richness gradually declined.With the onset of eutrophication in the 1970s (Fontana et al. 2014) until 1990 there was a drastic decline in richness and the replacement of oligotrophic to eutrophic species mainly Aulacoseira granulata var.granulata, Cyclotella meneghiniana and Nitzschia fruticosa (Fontana et al. 2014).The major eutrophication period occurred particularly after ca.1990 (Fontana et al. 2014) when an unexpected increase in richness was observed.During this period, population rapidly increased in drainage basin with expansion of slum dwellings without adequate sewage treatment (Whately & Cunha 2006).Consequently, gastroenteritis infection in local population became frequent and the agency in charge of the public water supply started using copper sulphate in 1991 to control cyanobacterial blooms (Beyruth 2000).The sudden increase in richness was very probably associated with the control of cyanobacterial dominance and the abrupt opening of new resources for other algal assemblages, including diatoms.With the intensification of urbanization and eutrophication, richness continued declining over time.Considering the space gradient, richness in modern assemblages never achieved values corresponding to the oligotrophic phase of subfossil assemblages although higher values were also achieved in contemporary mesotrophic conditions of the upstream sites 01 to 05 (Fig. 167A).Towards downstream richness decreased along the trophic state gradient (Figs.167B, C, D).
Overall, the change in richness from oligotrophic to supereutrophic phases led to a sharp reduction of the oligotrophic species, which represented 23% of the total number of diatom species in the reservoir mainly represented by Eunotia (15 species and probably 3 new species).Losses in algal biodiversity due to cultural eutrophication has been scarcely reported in Brazil (e.g.Crossetti et al. 2008) given the paucity of long-term monitoring data, and the lack of information before the onset of eutrophication.In this regard, the statement by Davidson et al. (2013) that our understanding of the relationship between anthropogenic impacts and lake biodiversity changes is typically based on contemporary data sets and space-for-time substitution holds true for Brazilian ecosystems.Therephore, in many cases paleolimnological approach can offer a unique tool to assess biodiversity changes encompassing time scales relevant to human-induced degradation and since pre-anthropogenic impacts.

Final remarks
This survey encompassing subfossil diatom (ca.90 years) and modern assemblages in surface sediments along an eutrophication gradient allowed the following conclusions: Overall, 84 infrageneric taxa were reported belonging to 71 species and eight non-typical varieties, besides five identified only to the genus level.From those, 47.6% were accounted exclusively for the subfossil assemblages indicating a significant biodiversity change over time.Our results expanded two new additions for Brazilian diatom flora (Chamaepinnularia submuscicola and Stauroneis acidoclinata), 30 infrageneric taxa for the state of Sa ˜o Paulo and four probable new species.Access to past oligotrophic conditions and mesotrophic regions of Guarapiranga Reservoir allowed a significant number of new additions to the Brazilian and the state of Sa ˜o Paulo diatom floras, which accounted for 25% of the total reservoir flora.Human-induced eutrophication led to a sharp decline in the oligotrophic species number, mainly of Eunotia and to a drastic reduction in species richness mainly since the major eutrophication period in ca.1990.Abrupt changes in richness were also caused by human management as the increase in the reservoir's water discharge and the application of algaecide to control cyanobacterial blooms.Present findings highlight the need for surveying diatom assemblages in protected environments or in less degraded conditions (considering space and time) for biodiversity assessment, and reinforce the use of paleolimnological approach in many cases as the only tool to access baseline conditions of degraded fresh waters (Smol 2008, Gregory-Eaves & Beisner 2011).These issues are crucial given freshwater ecosystems have been experiencing far greater declines in biodiversity than those recorded in terrestrial ecosystems and constitute a valuable natural resource (Dundgeon et al. 2006).Furthermore, cultural eutrophication is considered a common scenario worldwide and one of the most pervasive environmental issues affecting freshwater ecosystems (Smol 2008, Davidson & Jeppesen 2013, Cumming et al. 2015).

Figure 1 .
Figure 1.Map showing the location of the state of Sa ˜o Paulo in Brazil and the city of Sa ˜o Paulo metropolitan region (SPMR) with location of Guarapiranga Reservoir.Enlarged map of the reservoir with sampling sites: solid circles for surface sediments (1 to 14) and solid red square for core (C 1 ).

Figure 166 .
Figure 166.Taxa number according to trophic state gradient exclusively found in subfossil assemblages, modern assemblages and in both assemblages.

Table 1 .
Water chemistry data for Guarapiranga Reservoir (water subsurface mean values for summer and winter), according to AcquaSed database Project.Abbreviation in Material and methods.

Table 2 .
Valve dimensions and trophic state range distribution based on literature and in this study for the species recorded in Guarapiranga Reservoir that are commonly reported in Brazil.