Spatio-temporal variation in the density and diversity of decapods captured with artisanal traps in an Amazon estuary

: Aim: The variability in density and species diversity of decapod crustaceans was investigated on four islands with different degrees of anthropogenic disturbance around the city of Belém, State of Pará. Methods: Samples were obtained from 15 creeks using artisanal traps, every three months between October 2013 and May 2014 on Combu, Onças, Cotijuba and Mosqueiro islands. Results: Salinity and temperature little varied, which is common for a tropical Amazon estuary strongly influenced by freshwater inflow. A total of 8,367 decapods were captured, with one record of an exotic species Penaeus monodon . In all seasons, decapod density and richness tended to increase from Combu to Mosqueiro, with increasing proximity to the sea and higher salinity and pH. Except for Combu, species richness and Margalef diversity tended to be slightly greater in the wet season at all islands, especially Onças. Eveness and Shannon diversity did not vary greatly between seasons but were lowest at Onças in the dry season and highest at Combu, decreasing to Mosqueiro, in the wet season. In general, ecological indices are similar in the dry and transition dry to wet seasons, and in the wet season, dominance occurs at Mosqueiro Island. Macrobrachium a canthurus , C. bocourti and P. gracillis were associated with the wet season, whereas M. surinamicum prefers the dry season. M. amazonicum and Macrobrachium sp. have no well-defined seasonal pattern of occurrence at all the islands. Conclusions: Despite anthropogenic disturbances and proximity to large human populations, especially on Mosqueiro Island, the density and diversity of decapod crustaceans appear to be reasonably unaffected for the moment.


Introduction
Decapod crustaceans have been widely studied on the Amazon coast (Bentes et al., 2011, Vergamini et al., 2011, Pantaleão et al., 2012, Oliveira et al., 2013Amaral et al., 2014;Lima et al., 2014;de Oliveira et al., 2016;Freire et al., 2017) where they are extensively exploited by artisanal fisheries for subsistence and commercial purposes (Bentes et al., 2011). As an example, Macrobrachium amazonicum populations invariably present signs of overfishing (Freire et al., 2012), and are not self-renewing adequately due to the capture of immature shrimp, thus making their fishing unsustainable .
The decapod genus Macrobrachium includes approximately 240 species (Wowor et al., 2009;De Grave & Fransen, 2011), and is found in tropical and subtropical fresh and brackish waters worldwide (Pereira et al., 2002;Short, 2004), as a result of its highly adaptive reproductive and life history strategy (Short, 2004;Pileggi & Mantelatto, 2010). In South America, the genus has a wide distribution in the basins of the Orinoco, Amazonas and Paraguay rivers, being abundant in waters rich in sediment and dissolved salts in the central Amazon River basin (Odinetz-Collart & Moreira, 1993). In Brazil, there are 17 species of the genus (Pileggi & Mantelatto, 2012).
The island of Mosqueiro is the largest, most urbanized and populous island around the city of Belém, with around 30 thousand inhabitants (IBGE, 2010), has access both by river and by road and has a fishing port that supplies the markets of Belém and receives an intense flow of tourists to its 24 beaches (Ribeiro et al., 2013). On the contrary, the smaller islands, Combu, Onças, and Cotijuba have little or no urbanization and are characterized by small, scattered riverine populations, which mainly carry out harvesting of fruit and small-scale artisanal fishing (Ribeiro, 2004). In this context, the present study investigates how the spatio-temporal variation, abundance and diversity of decapod crustaceans are structured in relation to human activities in Guajará Bay and Mosqueiro Island, in the eastern Brazilian Amazon.

Study area and sampling
Guajará Bay is formed by an island complex (Combu, Onças, and Cotijuba Islands) in the southern arm of the Amazon estuary, and further north is Mosqueiro Island, the largest of the four islands, which together, form a unique and dynamic aquatic environment characterized by high levels of fluvial and tidal energy and large amounts of suspended material (Dillenburg & Hesp, 2008;Souza-Filho et al., 2011;Rosa-Filho et al., 2018), limiting the penetration of sunlight (Smith, 2002;Gregório & Mendes, 2011). Semi-diurnal macrotides (up to 6 m in Pará) (DHN, 2010) flood channel banks in the region; significant saltwater entry only occurs during the driest months (July-December), when precipitation and river discharge are lower (Souza-Filho et al., 2011;Rosa-Filho et al., 2018).
Decapod sampling took place every three months between October 2013 and May 2014 at differing numbers of sites on three islands around Guajará Bay; Combu (n=2), Onças (n=4), and Cotijuba (n=2) and Mosqueiro Island (n=7) (Figure 1). To standardize the sampling locations, the number of creeks per island was decided as a function of the area of the respective island (km 2 ). Sampling periods were categorized according to the prevailing season: Dry (October to November), Transition Dry to Wet or TDW (December to January), and Wet (February to May).
Specimens were collected with artisanal traps, known locally as matapis (Figure 2), which are made from thin rods of palm fronds (Astrocaryum spp. and Atrix spp.), baited with babaçu (Orbignya speciosa) flour. Five matapis were used at each site, arranged equidistant along a local stream, between its mouth and source. The traps were set two days before the new moon, on the last low tide of the day. On the first low tide of the following day (12 hours later), the traps were retrieved (Freire et al., 2012). Due to sampling problems, data for Onças Island are not available for the transition period from dry to rainy (TDW).
Specimens of shrimps and crabs collected in the traps were identified to the lowest possible taxonomic level. Salinity, temperature (°C), dissolved oxygen (% saturation), and pH were measured with a Hanna HI 9828 digital meter at both deployment and removal of traps, from which a mean value was calculated.

Data analysis
Density (N) was the sum per island of the individuals collected in the five matapis in each stream channel (site) in each season. Since the number of streams channel per island was decided based on their respective area, the total number of matapis per island was different at each station. Due to interference or loss of sampling units, faunal replicates were absent or low in certain seasons at some islands, and thus no attempt was made to carry out a statistical analysis on the ecological indices data. For each decapod sample, species richness was determined, and the Margalef (D), equitability (J') and Shannon diversity (H') indices were calculated in the Diverse routine in PRIMER 6.0. These ecological indices were chosen because they are commonly used in studies with abundance of decapod crustacean, responding in a robust way to the comparison of the fauna among the islands.
To check if decapod species richness was adequately sampled, species accumulation curves were generated for cumulative sampling effort on each island and the total cumulative sampling effort using the jackknife1 estimator, with sites added in random order, using the specaccum() function in the vegan package (Oksanen et al., 2020) in GNU R 4.0.4 (R Core Team, 2021).
Ordination by canonical analysis of principal coordinates (CAP, Anderson & Willis, 2003) using the capscale() function in the vegan package (Oksanen et al., 2020) on a Bray-Curtis dissimilarity matrix calculated from decapod crustacean species presence and absence data, was used to verify differences in species occurrence and composition among islands and seasons, as well as determine the degree of association with abiotic factors, also using the envfit() function in the vegan package.
Differences in the multivariate pattern of variation in environmental variables were formally checked using permutational multivariate analysis of variance (PERMANOVA) with the adonis() function in the vegan package (Oksanen et al., 2020).

Abiotic data
In the dry season, salinity varied between 0.01 and 1.78 among sites, being higher at Cotijuba and Mosqueiro, whereas temperature varied from 26.28 °C to 29.55 °C, being highest at Onças. In the dry to wet transition, salinity varied from 0.05 to 0.57, and temperature from 25.95 °C to 29.32 °C, tending to be highest at Mosqueiro. In the wet season, salinity was very low (0.01 to 0.25) and varied little among sites, but was relatively higher (0.25) at Mosqueiro sites. Temperature varied relatively more in the wet season (24.6 °C to 28.6 °C), being highest at Combu and decreasing to Mosqueiro.
Dissolved oxygen (% saturation), in general, showed moderate to high values in the dry and wet seasons, between 50 and 80, however, during the dry to wet transition, some Cotijuba and Mosqueiro measurements had markedly lower values around 10%. In the wet season, values were similar among islands. Values of pH were below 7 at all locations in the wet season, whereas in the dry season and in the dry to wet transition, pH varied more, with highest values of 8.56 and 8.77 occurring in the dry to wet transition (Figure 3). In general, pH tended to be higher at Cotijuba and Mosqueiro, and lower at Combu and Onças.
Differences in environmental variables among islands and seasons were significant, the most important being seasonal differences, which explained 44.7% of the variation in the data (Table 1). However, interactions among islands, seasons and sites, that is, a lack of consistency in seasonal patterns in environmental variables among islands and the creeks sampled therein, were significant explaining 24.1%, and residual unexplained variability, a further 18.7% (Table 1).

Biotic data
A total of 8,367 decapod crustaceans were captured, of which 7,371 were Macrobrachium amazonicum, the most common shrimp species   In the dry season, density was higher in Cotijuba and Mosqueiro and species richness was generally similar among islands, but lower at Mosqueiro (Figure 4). Margalef diversity was similar among islands, being slightly lower at Mosqueiro. Eveness and Shannon diversity were similar among islands, except at Onças where values were lower. In the Transition from Dry to Wet (TDW), density was highest at Cotijuba, albeit a single observation, whereas richness, Margalef, Eveness and Shannon diversity were similar among Combu, Cotijuba and Mosqueiro islands (Figure 4). In the wet season, density increased from Combu to Mosqueiro. Species richness and Margalef diversity were highest at Onças and similar at the other islands. In contrast to density, Eveness and Shannon diversity decreased from Combu to Mosqueiro in the wet season ( Figure 4). In general, ecological indices are similar in the dry and transition dry to wet seasons, and in the wet season, dominance occurs at Mosqueiro Island.
Onças Island appears to be well represented by sampling, reaching a well-defined plateau, as well as Combu and Cotijuba islands, which appear to be close to reaching a plateau. Mosqueiro Island, despite the higher sampling effort, is not close to the plateau, also reflected in the curve involving the total sampling effort at all the islands ( Figure 5).
Based on presence and absence data, the decapod fauna in Cotijuba was clearly distinct from that of Combu, whereas those of Mosqueiro and Onças overlapped considerably with the other islands ( Figure 6A). Decapod occurrence and composition in the dry season was more distinct from that of the wet season and the dry to wet transition, the latter two overlapping entirely ( Figure 6B). Macrobrachium sp., M. amazonicum and M. surinamicum were more associated with the dry season, when temperature, salinity and dissolved Of all the environmental variables, pH appeared to be least associated with the fauna by island or by season ( Figure 6B). The multivariate pattern in faunal occurrence was significantly, though weakly, associated (Envfit R 2 , P) with temperature (R 2 = 0.064, P = 0.012) and salinity (R 2 = 0.056, P = 0.034) but not with dissolved oxygen (R 2 = 0.037, P = 0.098) and pH (R 2 < 0.0001, P = 0.99).

Discussion
The Guajará bay can be considered as having the environmental characteristics of a river during almost all the year, due to discharge from the large Amazon and Tocantins/Araguia basin drainages, which effectively inhibit the advance of the oceanic saline wedge, especially during the rainy season. Guajará bay is most influenced by the mass of marine waters entering the mouth of the estuary in the dry season (Cavalcante et al., 2017).
Temperature showed relatively low variation in our data (24.63-29.55 °C). The constant water temperature is typical of the estuaries of the Amazon coast (Carvalho & Couto, 2011) and confirms the thermal stability of equatorial coastal regions. Variation in water temperature among islands did not differ from that of other nearby locations, such as Vigia, 27.5 ºC and 28.0 ºC in the wet and dry seasons, respectively (Silva et al., 2002a). Unlike temperate regions, where temperature is a seasonally limiting factor for decapod growth and distribution, in tropical regions, relatively constant favorable conditions allow for mating and larval development  throughout the year (Costa & Soares-Gomes, 2009;Brandão et al., 2011).
The low variability in salinity (0.01-1.78) differs from that observed by Lima et al. (2019) who recorded salinity from 3 to 35 in the Marapanim, an Amazon estuary further east. However, our results confirm those of Rosário et al. (2016) in Guajará Bay, which show that this estuary does not present high salinity along its longitudinal profile, since there is rapid dilution of sea water by river discharge, thereby lowering salinity to between 0 and 5. Although in an estuarine region, our sampling sites were closer to the influence of large rivers, such as the Acará, Guamá, and Maguarí rivers, and thus our salinity values were low and invariable. Abrupt changes in sea level of up to six meters during the wet season are common in the region, due to a combination of macrotides and high freshwater discharge, resulting in oligohaline waters that keep the temperature and salinity constant in this estuarine region (Barros et al., 2011).
Intense mixing of waters of differing salinity allows coexistence of both marine and freshwater species according to seasonality (Bentes et al., 2011). The results of the present study re-emphasize the influence of relatively low salinity levels on the distribution of the decapod species found in this region, which, according to Cavalcante et al. (2017), would account for the occurrence of freshwater crab species -Sylviocarcinus devileii and Sylviocarcinus pictus, especially at Belém and Combu Island -where salinity was virtually zero throughout the year -as well as the presence of the marine-estuarine genera Uca and Callinectes on Mosqueiro Island. The wet season pH was slightly more acidic, as rainwater has a pH of around 5.5, and the influence of rainwater runoff may decrease pH. Variation in pH is the result of the interaction of several biogeochemical factors and is a primary indicator of water quality (Blume et al., 2010), however, regardless of the seasonal period, our values were close to those found by Montes et al. (2009) also in Guajará Bay, with a minimum of 5.10 and a maximum of 6.08, showing that anthropogenic factors did not appear associated with changes in pH in this period.
Relatively successful catches of species of the Macrobrachium genus seem to be a characteristic of Amazon estuary. However, the species M. amazonicum presented a higher density than the other species of the same genus, this difference may be related to the ecological flexibility of this species (Bentes et al., 2014). High densities of M. amazonicum were recorded at all sites, reflecting the tolerance of the species to environmental variation, which occurs worldwide in tropical and subtropical fresh and brackish waters, inhabiting a variety of ecosystems (Pileggi & Mantelatto, 2010;Vergamini et al., 2011) from continental to coastal environments with different salinity gradients; (Short, 2004;Freire et al., 2017).
The presence of the exotic species Penaeus monodon in the Brazilian Amazon has been previously reported. Silva et al. (2002b) registered the presence of two specimens from the continental shelf of the State of Amapá. Cintra et al. (2011) reported the capture of an individual by an industrial fishing vessel operating on the continental shelf of the State of Pará and Cintra et al. (2014) recorded the capture of three more individuals in 2013 also by an industrial fishing vessel operating on the Amazon continental shelf in the State of Pará. The species P. monodon, which may have been introduced into Brazilian waters by ballast water from ships (Farrapeira et al., 2007), may be a threat to local Amazonian species, as the region seems to represent an ideal environment for their survival (Cintra et al., 2014, Lutz et al., 2015. The introduction of exotic species into different ecosystems may affect the survival of native species (Rodríguez, 2001), which may lead to competition for resources, especially food, between species, generate changes in the balance of predator-prey relationships, and significantly impact trophic webs (Snyder & Evans, 2006;Baum & Worm, 2009;Lord et al., 2019).
The type of trap used in the present study may account for the reduced proportion of crabs (marine and freshwater) in the samples, although this had been expected, given that local coastal communities traditionally use these traps for the harvesting of shrimp (Odinetz-Collart & Moreira, 1993). At higher salinities -closer to the sea -density and the number of species increased. Mosqueiro Island is closest to the sea and is located within an area of intense mixing of saline and fresh waters, resulting in a nutrient-rich environment (Smith, 2002), which may favor the density of decapods.
The CAP analysis identified an association between M. acanthurus and the wet season, which is consistent with Díaz Herrera et al. (1998), who recorded the tolerance of the species for freshwater or low salinity, with adults capable of osmoregulation in fresh and brackish water at temperatures above 15 °C. However, during the larval stage, M. acanthurus has low tolerance to salinity (Choudhury, 1971).
The composition of the decapod fauna in the present study was variable and related to both location and season. The resulting temporal and spatial mosaic of salinity favors, besides shrimps, other crustaceans such as crabs. M. jelskii is classified as a species restricted to freshwater (Melo, 2003), as evidenced by a greater association with the inner estuary in Onças and Combu islands that receive a greater riverine inflow and less influence from saline coastal waters. In addition, M. jelskii has abbreviated larval development (Magalhães, 2000), where females carry only a few large eggs (Anger, 2013). Palaemonidae species with abbreviated larval development occupy strictly freshwater habitats (Rabalais & Gore, 1985).
Acta Limnologica Brasiliensia, 2022, vol. 34, e9 The relative abundance of M. amazonicum during the dry season reflects the population dynamics of this species, which tends to be rare during the wet season (Bentes et al., 2011). The increased volume of water during the wet season may facilitate the dispersal of many species of crustacean and fish (Odinetz-Collart & Moreira, 1993). Densities may thus be higher during the dry season, when river volume decreases and many individuals move out of the floodplain lakes, facilitating trapping. Variation in shrimp catches in the Amazon depends on the intensity of the floods, which interfere with survival and growth. In Amazon basin floodplain lakes, catches are higher during the dry season and lower during the wet season, due to the wide spatial dispersion of individuals with higher water volume and flow (Odinetz-Collart & Moreira, 1993). Although M. surinamicum is more associated with the dry period in our CAP analysis, Cavalcante et al. (2017) found that its distribution was not influenced by either salinity or temperature in the Guajará estuary, possibly since these factors did not change very much in the region, also evidenced in our results (salinity 0.01-1.78, temperature 24.63-29.55 °C).
Macrophyte wet season growth reduces shrimp intra-species competition, through increased microhabitat availability in the early stages of life (Nurminen et al., 2007) and introduction of suspended material increases productivity and energy flow (Paiva et al., 2006). Tides are the major natural disturbance in mangrove areas along the Pará coast, directly affecting abundance of zooplankton, crustaceans, and fish, as well as promoting the entry and exit of wastes and nutrients (Dittmar & Lara, 2001;Schories et al., 2003;Krumme & Liang, 2004).
The highest shrimp density was observed on Mosqueiro Island, which is the largest and most populous island studied in present work (IBGE, 2010). Thus, a larger human population may be related to higher amounts of suspended organic matter that is a food source for shrimp. Eutrophication processes due to phosphorous and nitrogen-rich compounds from domestic and industrial waste cannot be recognized as a positive scenario (Viana et al., 2010), and causes long term deleterious effects in the benthic fauna (Dorgham, 2014). Eutrophication of aquatic environments via sewage generates poor water quality, represented by a higher biochemical oxygen demand and consequently low dissolved oxygen concentrations due to the proliferation of microorganisms. Miranda et al. (2016) observed values above the allowed (5mg/L) by CONAMA (Conselho Nacional do Meio Ambiente) Resolution No. 357/2005 (Brasil, 2005) for biochemical oxygen demand and low values of dissolved oxygen on Mosqueiro Island, relating the disposal of urban effluents and vessels such as pollution sources. However, our values of dissolved oxygen were similar between the islands, especially during the wet and dry seasons, which does not seem to be a limiting factor for both species diversity and abundance.
Studies of density and diversity of species between areas that receive different levels of environmental impacts are of vital importance for understanding how biotic and abiotic dynamics react to these factors of pollution of estuarine and coastal waters. Basic physical-chemical parameters and ecological indices can be especially valuable for the understanding of the current status of an ecological community, as well as future impacts (Mason, 1991). While ecological indices are often criticized due to their limitations for understanding community dynamics, they can provide a useful comparative tool for the analysis of impacts on community structure (Morris et al., 2014), especially in the long-term, given that impacts on species diversity may often be extremely subtle or masked by the dynamic restructuring of the community (Kay et al., 2018).
Aquatic organisms may be exposed to multiple stressors, such as the discharge of pollutants into the water, overfishing, habitat modifications and so on (Scholz et al., 2012;Ward et al., 2013;Hook et al., 2014). Due to their short life cycle, presence of a larval phase, and their importance as components in trophic webs, crustaceans are often used as bioindicators by water quality monitoring programs (Ghisi et al., 2017;Kumar et al., 2017). The islands in Guajará Bay and Mosqueiro Island are heterogeneous environments, with shipping and ports, domestic and industrial effluents, and fishing and aquaculture operations affecting primarily its eastern margin (Mosqueiro Island and the islands around Belém). Our results show that although this region suffers from greater anthropogenic disturbances, mainly because it has a large human population, the current diversity and density of shrimp and crabs appear not to be greatly affected. However, the current species diversity found within the study area may change significantly over time, leading to ecological imbalances, which may ultimately affect the area's traditional artisanal fisheries.