Ecological Linkages in a Caribbean Estuary Bay

Home , Amatique Bay, Gulf of Honduras, Lolliguncula brevis

Vol. 533: 29–46, 2015 MARINE ECOLOGY PROGRESS SERIES Published August 6 doi: 10.3354/meps11342 Mar Ecol Prog Ser

Ecological linkages in a Caribbean estuary bay

H. Andrade1,2,*, J. Santos1, M. J. Ixquiac3

1Norwegian College of Fishery Science, University of Tromsø, 9037 Tromsø, Norway 2Akvaplan-niva AS, Framsenteret, 9296 Tromsø, Norway 3Centro de Estudios del Mar y Acuicultura, Universidad de San Carlos de Guatemala, Guatemala 01012, Guatemala

ABSTRACT: Central America and the western Caribbean regions form a center of freshwater and marine biodiversity that is increasingly becoming the focus of ecological and evolutionary studies. We conducted an integrated ecological study of Amatique Bay, Guatemala, a major estuary la goon connected to the low-lying Lake Izabal and to the Mesoamerican Reef System, and provide novel information for the management and conservation of similar systems across the Caribbean. Precip- itation and wind regimes constitute important environmental drivers of ecosystem functioning, which partially compensate for the weak tidal-forcing that is characteristic of the Caribbean Sea. Seasonal peaks in temperature and precipitation were strongly correlated to the reproduction of marine, catadromous and estuarine fish species, suggesting that the ensuing increase in primary production provides larval fish with an abundant food source. Increased abundances of transient marine species were observed during the dry season, which may be explained by passive transport, feeding migration, or both, considering that prey may be more abundant inshore and that environmental conditions are dominated by higher salinity and stronger onshore winds during this period. Despite being a stopover site for many long-range migrating shorebird species, the Bay serves primarily as a resting place since it lacks extensive tides and tidal flats, and thus provides limited access to invertebrate prey. Furthermore, this sheltered environment, featuring abundant freshwater, seasonally high water clarity, and low tidal amplitude, is likely to provide good habitat for abundant seagrasses and manatees. The Lake Izabal-Amatique Bay complex demonstrates a wide range of teleconnections and connectivity among terrestrial, freshwater, and marine oceanic and reef ecosystems. Understanding the evolution and ecology underlying this highly connected system is required for the management of the multi-trophic, small-scale fisheries it sustains.

KEY WORDS: Fisheries · Migratory shorebirds · Manatee · Life history · Environmental drivers · Tropical conservation · Evolution · Central America

Resale or republication not permitted without written consent of the publisher

INTRODUCTION water ecosystems by marine species (i.e. killifishes, cichlids) played a major role after a sequence of salt- The western Caribbean is highly diverse across ter- water intrusions and regressions (Hulsey & López- restrial, marine and freshwater realms, but there is a Fernández 2011). In the marine realm, the Caribbean limited understanding of linkages between its coastal Province has historically been the center of ecological assemblages of fish, birds, and mammals across these speciation and radiation of fish and many inver - systems. The paleontological and phylogenetic re - tebrate groups in the Atlantic, producing and export- cords suggest that the Central American region was ing species. but also accumulating biodiversity pro- at the core of an explosive radiation of freshwater fish duced in peripheral habitats (Briggs & Bowen 2012, species (Briggs 1984, Chakrabarty & Albert 2011). Bowen et al. 2013). The Caribbean province was also Part of the taxa were secondary freshwater fish origi- once an area of sirenian (manatee, sea cow) radiation. nally from South America, but the invasion of fresh- However, the closure of the Central American Sea -

*Corresponding author: [email protected] © Inter-Research 2015 · www.int-res.com 30 Mar Ecol Prog Ser 533: 29–46, 2015

way in the Pliocene (ca. 3 million yr ago) resulted in The estuarine areas of Central America and the mass extinctions of seagrasses and sirenian species, Caribbean are important for both conservation and and only a single species, the manatee Trichechus human utilization, especially considering their signif- manatus remains (Hunter et al. 2012, Velez-Juarbe et icance as a stop-over for long distance migrants (e.g. al. 2012, Benoit et al. 2013). The rise of the Isthmus of shorebirds), seasonal habitat for short-range Panama prompted the migration of forest birds pre- migrants, and habitat for resident taxa (Faaborg et al. dominantly in the direction south to north, which pre- 2010, Latta 2012, Somveille et al. 2013). Amatique sumably led to the high levels of bird diversity also Bay in Guatemala is connected by freshwater runoff observed in this region. For shorebirds (Order to the Mesoamerican reef, the largest barrier reef in Charadriiformes), however, understanding of their the Western Hemisphere (Soto et al. 2009), and is a original migratory behavior and home range is prob- typical Caribbean estuarine ecosystem. Upstream lematic (Weir et al. 2009, Livezey 2010, Zink 2011). (40 km) from the Bay, the low-lying Lake Izabal Many extant Arctic Charadriiformes are long-range forms the southern boundary of the Usumacinta fish migrants with northern breeding grounds, but which faunal province, a region characterized by a highly overwinter in the southern hemisphere. However, diverse continental ichthyofauna. The Bay may have several lineages of shorebirds from the southern been a major route of incursion of marine species into hemisphere are predominantly residents or short- the freshwater assemblages of Central America range migrants. Thus, the present assemblages of (Hulsey & López-Fernández 2011). To conserve this aquatic and wetland fauna are resultant of a complex complex, natural protected areas have been imple- of freshwater and marine radiations and transgres- mented across the watershed, including 2 Ramsar sions, as well as colonization by continental species. wetland sites of international importance: the Río The lack of studies examining coastal species assem- Sarstún Multiple Reserve Zone, and Punta de Man- blages and their functions, particularly in Neotropical abique Wildlife Refuge (Fig. 1). Punta de Manabique estuaries, hampers the understanding of ecological alone shelters more than 450 plant species, and 810 processes that may have driven the evolution of many faunal taxa (Jolón-Morales 2006). Several threatened taxa (Sheaves & Johnston 2009, Barletta et al. 2010, or vulnerable migratory species, including the mana- Atwood et al. 2012). tee, contribute to this biodiversity. Agriculture, herd-

Punta de Manabique

Sarstún River Amatique Bay La Graciosa Bay Livingston 15°48' N

Puerto Barrios Dulce River Santo Tomás

Protected areas

Lake Izabal Caribbean Sea

15°24'

89°53'W 88°36' Fig. 1. Map of Amatique Bay and current protected areas (dark grey). Punta de Manabique Wildlife Refuge includes a protected marine zone Andrade et al.: Ecological linkages in an estuary bay 31

ing and forestry, activities that are often preceded by of the Bay as well as in the rivers draining into it slashing and burning of existing vegetation, have (Hernández et al. 2012). The dominant species is the been identified as major sources of impact on the red mangrove Rhizophora mangle, but Avicennia wetland habitat in this area (Yañez-Arancibia et al. germinans, Laguncularia racemosa and Conocarpus 1999). However, the presence of 2 harbors receiving erectus are also common (Yañez-Arancibia et al. an excess of 1200 ships annually may also have a 1994). Seagrass beds, which are particularly abun- negative influence on the aquatic communities (Abt dant in La Graciosa Bay, cover approx. 38 km2, and 6 Associates/Woods Hole Group 2003). Fishing pres- species have been identified to date, with Thalassia sure is also high and landings account for nearly 60% testudinum as the dominant species (Yañez-Aran- of the economic value generated by fishing in the cibia et al. 1994, Arrivillaga & Baltz 1999, MacDon- Guatemalan Caribbean, supporting the livelihood of ald-Barrios 2011). Some reef structures exist, mainly more than 1000 harvesters (Ixquiac-Cabrera et al. around Punta de Manabique in the form of continen- 2008, Andrade & Midré 2011, Heyman & Granados- tal carbonate banks. These reefs are dominated by Dieseldorff 2012). Thus, the range of conservation, sedimentation-resistant coral species, such as Side - ecological and social interests to be accommodated is rastrea siderea. Live coral cover, however, is low, and broad and often conflicting. non-coralline macroalgae is highly abundant (Fon- In this study, we attempted to describe this Carib- seca & Arrivillaga 2003). The mud-dominated areas bean estuarine-marine complex with the goal of at the mouth of the Sarstún River give rise to the most identifying ecological drivers for ecosystem function- valuable shrimp fishery in the Gulf of Honduras ing and evolution in Neotropical estuaries. Ecological (Heyman & Kjerfve 2001). studies that integrate marine and freshwater ecosystems have rarely been performed in Caribbean estuaries. Our current understanding of these systems is Collection and analysis of meteorological and derived from information that is spread across data oceanographic data reports, fisheries statistics, and very specialized pub- lications. Thus, we compiled environmental and eco- Time-series of environmental data were retrieved logical information from different sources, and col- from the Guatemalan meteorological institute (IN- lected new field data on vertebrates and their SIVUHMEH, www.insivumeh.gob.gt/estacionesmet. environment. We focus on the environmental drivers, html), or extracted from NOAA or NASA (http:// seasonal rhythms, and life cycles of fish, shorebirds podaac.jpl.nasa.gov/) open internet sources available and manatees in the Amatique Bay complex, and for 1985−2010. Retrieval and treatment of these envi- suggest how these processes may link the estuary to ronmental data are described in detail in Text S1 in the riverine and marine ecosystems. Greater empha- the Supplement at www.int-res. com/ articles/ suppl/ sis is placed on the growth and reproduction cycles of m533p029_supp. pdf . These included time-series of fish, because this group has been more intensively wind speed (Wind) and direction, precipitation (Pre), and regularly sampled. This case study provides an air temperature (Tair), day length (Dayl), sea surface integrated view of an estuarine complex in the Ca - temperature (SST), tidal heights, and chl a concentra- ribbean and the Neotropics. tion. Turbidity and nutrient concentration at the outlet of Lake Izabal were measured in 2006−2007 by Quin- tana-Rizzo & Machuca (2008). We separated these MATERIALS AND METHODS measurements into 2 periods to represent the water quality; August, October and December 2006 com- Study site prised the wet season, and February, April and June 2007 the dry season. Estimates of monthly run-off Amatique Bay comprises an area of 542 km2, with were recovered from a model using land cover sce- an additional 200 km2 of associated wetlands (Fig. 1). narios for the years 2003−2004 (Burke & Sugg 2006). The bay is a diverse and complex, shallow (average depth <10 m) ecosystem consisting of coastal la - goons, seagrass meadows, reefs, mangroves, and Seasonal abundance of fish species marshes that are influenced by riverine systems in Amatique Bay (Yañez-Arancibia et al. 1999, Fonseca & Arrivillaga 2003). More than half of the 12 km2 mangrove forest Indices and maps of fish density were derived from in the Guatemalan Caribbean grows along the coast 2 sets of fishery-dependent data and one set of obser- 32 Mar Ecol Prog Ser 533: 29–46, 2015

vations made during research surveys. The first set the organisms sampled, were chosen to illustrate consists of the average monthly catch per unit effort spatial occupancy during the dry and wet seasons. (CPUE) of shrimp trawlers (kg per unit effort [effort = These species included the caitipa mojarra Diapterus number of fishing boats × month]) in the period rhombeus, lane snapper, squid, striped mojarra Eu - 2006−2010, available from the national fisheries gerres plumieri and anchovies, a group comprising directorate (DIPESCA, Guatemala). Records of catch several engraulids, including Anchoa spinifer, Anchoa and by-catch are usually pooled into coarse cate- cayorum, Anchoa colonensis and Anchoviella elon- gories that sometimes comprise several species, e.g. gata (Table S1). ‘Shrimp’ (3 Penaeid species), ‘Catfish’ (2 Ariidae species), and ‘Corvina’ (a mix of Sciaenidae and Hae- mulidae). The lane snapper Lutjanus synagris and Physiological traits of selected fish species Atlantic brief squid Lolliguncula brevis (hereafter referred to as squid) are registered as individual spe- Eco-physiological traits of fish species were inves- cies (Table S1 in the Supplement). Shrimp trawlers tigated by observations of the reproduction and operate on soft mud bottom and are typically 10- growth of lane snapper, grey snapper Lutjanus meter long vessels equipped with 120−130 hp griseus, gafftopsail catfish Bagre marinus and snook inboard engines. The trawl gear lacks otter boards Centropomus undecimalis along an annual cycle. and is retrieved by hand by a small crew (González & These species were selected due to their frequency López 2000). The legal codend mesh size is 64 mm in the catches and could be regularly sampled (stretched), but a 51 mm cover is usually employed to between March 2006 and April 2007 from the fresh improve retention of smaller sized shrimp (Ixquiac- landings in Livingston and in Puerto Barrios (Fig. 1). Cabrera et al. 2008). Snappers were usually caught with hand lines, but The second type of fishery-dependent data con- snook and the gafftopsail catfish were caught prima- sisted of the estimated monthly landings from non- rily using gillnets. The total (Wt) and gonad (Wg) trawler vessels derived from Heyman & Graham weights (±0.1 g) of the fish were recorded along with (2000) and Heyman & Granados-Dieseldorff (2012). their total lengths (L), in cm. Monthly averages of the These estimates were based on information from gonadosomatic index (GSI = 100 Wg/Wt) were used interviews with 42 experienced skippers (70% had as an indicator of the gonadal development and more than 10 years of experience) of small boats spawning seasonality (Lowerre-Barbieri et al. 2011). (dories, skiffs) performed in 1998. The most common The condition factor (CF = 100 Wt L−b) is a body-mass fishing gears were gillnets (81%), beach seines (7%), index where b is the coefficient of the length–weight small shrimp trawl nets locally known as ‘changos’ relationship (King 1995). Excluding the gafftopsail (5%), and hand lines (3%). The location of the fishing catfish, which is clearly sexually dimorphic, fish of villages and the species-distribution maps in Hey- both sexes were combined prior to analysis. This man & Granados-Dieseldorff (2012) indicate that the included an analysis of the sex-aggregated data for catches were made mostly inside the Bay. These common snook, a commercial fish species that we authors report monthly landings of many species but have previously investigated in detail (Andrade et al. we limited our analyses to those that regularly com- 2013). prised 90% of the total catch (Table S1). The oceanographic and biological observations made by Ixquiac-Cabrera et al. (2008) during 2 re- Shorebird and manatee distribution search cruises were used to map salinity profiles and fish density across Amatique Bay. The surveys were Observations of shorebirds in Punta de Manabique carried out in February and August 2008 from a fish- were available from August 2000 to June 2001 (Eiser- ing vessel equipped with a commercial shrimp trawl mann 2009). In that study, 2124 sightings were re - (Text S2 in the Supplement) and a CTD profiler. corded along beaches, coastal lagoons and river Mapping was performed after smoothing the obser- mouths, providing an index of relative abundance. vations from 11 fixed stations and categorizing data Only the most common species (n > 30 observations) into dry (February) and wet (August) seasons. The as defined by the original authors were used in the densities per square nautical mile (kg nmi−2) of some analyses, and this accounted for 97% of the birds of the most numerous species were plotted to analyze sighted and 11 out of a total of 25 species (Table S1). distribution patterns. Five out of the 11 dominant An airborne survey of manatees Trichechus manatus species (of 79 species in total), accounting for 28% of in the Izabal-Dulce-Amatique complex was per- Andrade et al.: Ecological linkages in an estuary bay 33

formed on 5 occasions between July 2006 and Febru- ture varied little along the coast (26.5 ± 1.9°C, ary 2008 by Quintana-Rizzo & Machuca (2008). How- annual mean ± SD). The SST was lowest in Novem- ever, only the sightings made in October 2006 and ber to May (~27°C), and reached a maximum in March 2007 were utilized in our study to map their September (30 ± 0.6°C, mean ± SD). Amplitude of seasonal distribution since only these 2 surveys had day duration was also small, and day length varied similar coverage and methodology (Text S3 in the from 670 min of light in December to 780 min in Supplement). June. Cross-correlation analyses showed that the

cycles of SST, Tair, and Dayl were significantly correlated (r > 0.5 and p < 0.05 in all cases) and in

Statistical analysis phase (lag zero), with the SST and Tair series pre- senting the highest correlation. The average annual Relationships between monthly average abun- precipitation in the inner part of the bay exceeded dance of selected species (fish, shorebirds) and puta- 3300 ± 615 mm in the period 1985−2010. The rainy tive explanatory variables, such as meteorological season usually starts in June, reaching peak pre- and oceanographic time-series, were analyzed by cipitation in July with about 430 mm, and remain- means of multivariate ordination with the software ing above 300 mm until November. Wind speed package CANOCO (ter Braak 1986, ter Braak & Šmi- was highest during March and April (10.3 ± 2 km lauer 2002). Direct gradient analyses were carried h−1), and lowest between September and December out by means of redundancy analysis (RDA, the con- (8.5 ± 2.7 km h−1). From January to September the strained form of principal com ponent analysis) to test wind direction was predominantly from NE, and whether species composition could be explained by from variable directions during the rest of the year. the main environmental factors SST, precipitation Overall, the yearly mean wind direction was 35° and wind. This was performed on log-transformed (circular variance 6°), i.e. straight from the mouth data after examination of the gradient lengths with of the Bay (NNE). The salinities across the Bay var- detrended correspondence analyses (DCA) (Ejrnæs, ied widely depending on the season. During the 2000). Monte Carlo permutation tests (499 permuta- dry season the increasing temperatures re-enforced tions) were employed to assess the statistical signifi- by strong onshore winds give rise to a distinct mar- cance (α = 0.05 for all statistical tests). Exploratory ine influence. Relatively high surface (18−29 ppt) analyses of the shorebird species and seasonal data (Fig. 3) and bottom (29−31 ppt) salinities were were also performed by means of RDA, with seasons observed in February, indicating relatively good expressed as categorical (dummy) environmental mixing (Ixquiac-Cabrera et al. 2008). During the variables (Šmilauer & Leps 2014). To illustrate the wet season, increased precipitation, higher run-off, cyclical occurrence of selected shorebird species, and lower wind stress lead to increased stratifica- their sightings were modeled using a generalized tion. In August, bottom salinities ranged from 23 to additive model (GAM) with season as predictor vari- 31 ppt and surface salinities from 8 to 20 ppt able. A Poisson error structure of the sightings was (Ixquiac-Cabrera et al. 2008), and were characteris- assumed and a log-link function was used, which is tically low close to the mouth of the Dulce River routine for count data (McCullagh & Nelder 1989). (Fig. 3). The tides followed a regime of damped Circular statistics (Zar 1998, Lund & Agostinelli 2014) mixed-cycles with average monthly tidal amplitude were used to calculate means and variance of of only 0.52 m with some yearly variation but no monthly wind direction. To identify linkages clear seasonal trend. between pairs of time-series while accounting for Secchi-disk measurements performed by Quin- auto-correlation, we used cross-correlation analyses tana-Rizzo & Machuca (2008) in the main channel at on ln-transformed data (El-Gohary & McNames the outlet of Lake Izabal indicate that turbidity was 2007, Wilkinson et al. 2009). highest during the rainy season at 3.0 m and lowest in the dry season in February at 4.0 m. None of these values suggests outflow of water rich in suspended RESULTS particulate matter. Nutrient concentrations were highly variable temporally and spatially within the Environmental variables lake. At the outlet of the lake, nitrate (NO−3) concentrations tended to increase from baseline levels to The time-series of the environmental variables 0.5−3.1 mg l−1 in August to October. This pattern was and chl a are illustrated in Fig. 2. The air tempera- also found for ortho-phosphates (0.13 mg l−1), which 34 Mar Ecol Prog Ser 533: 29–46, 2015

SST Dayl T were normally low and variable, or unde- 32 A air 800 tectable towards the end of the rainy season (August− November). Within Ama- 30 760 tique Bay the Secchi depth was lower at the mouth of the rivers, particularly the 28 Sarstún where it was about 0.8 m in July 720 (Carrillo-Ovalle et al. 2000). The Secchi- 26 depth increased rapidly towards the outer 680 bay where it reached 10 m, even during Day length (min) Temperature (°C) 24 the rainy season, closely mirroring the horizontal salinity gradient (Fig. 3). The 22 640 JFMAMJJASOND chl a and runoff cycles resembled that of Dry Wet precipitation, usually peaking in June− July and remaining high until October. Prec Chl 500 B 14 Cross-correlation analysis showed that the precipitation cycle was significantly 12 ) 400 –3 correlated (p < 0.05) and in phase (lag 10 zero) with the chl a cycle (r = 0.35).

300 8

200 6 Fish species abundance and distribution 4 The bottom trawler data suggested that Precipitation (mm) 100 2 Chlorophyll a (mg m shrimp and by-catch were associated 0 0 with the seasonal meteorological regime JFMAMJJASOND and the inflow of marine waters brought Dry Wet about by the NE winds, low precipitation, and rising SST (Fig. 4). Redundancy 2001 2006 2008 Wind analysis revealed that 23% of the varia- 11 0.75 C tion in CPUE in 2006−2010 was explained 0.7 10 largely (96%) by the 3 variables selected

9 ) in the analysis: SST, Pre and Wind. The 0.65 –1 forward selection analysis retained SST 8 and precipitation as significant variables 0.6 7 (p < 0.05). The RDA triplot shows that SST

0.55 Wind (m s and precipitation were not correlated, 6 and were the main variables determining 0.5 Avg daily amplitude (m) 5 the first and second axes, respectively. 0.45 4 The first (horizontal) axis contrasts warm JFMAMJJASOND months with higher precipitation on the Dry Wet left side, to colder and dry months on the right side. The second axis separates the 2000 D Runoff 2003–2004 months with species associated with high

× 1000) 1600 3 Fig. 2. (A) Mean ± SE sea surface temperature (SST) for 1985−2009, air temperature (T ) for 1990−2010 1200 air and day length (Dayl) for 1985−2010, (B) mean ± SE annual cycles of precipitation (Prec) for 1985−2010, 800 and chl a for 2002−2010, (C) mean ± SE wind speed for 1990−2010) and tide amplitude (predicted for 2001, 400 2006 and 2008), and (D) modeled volume of river runoff for 2003 and 2004 (Burke & Sugg 2006) in Ama- 0 tique Bay, Guatemala. Letters on x-axes represent JFMAMJJASOND months. Sources: Guatemalan Meteorological Insti- Dry Wet tute (INSIVUMEH) and NASA Modeled river runoff (m Andrade et al.: Ecological linkages in an estuary bay 35

Dry season Wet season SST at the top from the species associ- Lane snapper ated with increased precipitation at the bottom of the triplot. In contrast, precipitation and wind speed (non-significant ex planatory variables) were negatively correlated. Shrimp density and chl a concentration exhibited the strongest significant associations with the environmental variables SST and precipitation. The increased SST in Atlantic brief squid June− July was positively related to shrimp and squid abundances. Abun- dance of fish such as sciaenids, catfish and lane snapper in the bottom trawls was negatively related to precipitation, and was higher in the dry months of March−May when onshore winds tended to be stronger. The temperature and wind regimes Caitipa mojarra drive the occurrence of the different species available to dories and skiff fishers (Fig. 5). The variables included in the RDA explained 56% of the variance in the biological data, with the first and second axis ac - counting for 92% of this variation. The forward selection analysis retained SST and wind as significant Striped mojarra variables (p < 0.05). The first axis clearly contrasts warm months, on the right side, to colder and windy months, on the left side. The second axis separates the months and species according to the precipitation regime, with species predominant during the rainy season located at the top, and those indicative of dry season at the Anchovies bottom of the triplot. The Gerridae Density group, lane snapper, and grouper –2 (Kg nmi ) were positively related to precipita- 100 tion in November. Our own observa- 300 tions suggest that lane snappers 500 caught in this net and line fishery consist mostly of late juveniles and adults (mean length = 23.4 cm, size range = 13.3−40.4 cm). The shrimp Fig. 3. Surface salinity (ppt) and distribution (calculated density per square species, the tarpon Melagops atlanti- nautical mile) of lane snapper Lutjanus synagris, Atlantic brief squid Lolligun- cus and the blackbelt cichlid Para- cula brevis, Caitipa mojarra Diapterus rhombeus, striped mojarra Eugerres neetroplus maculicauda, were posi- plumieri and anchovy group (Anchoa spinifer, A. cayorum, A. colonensis tively related to SST in August− and Anchoviella elongata) in research trawls in the dry (February) and wet September. Snook was partially re - (August) seasons in Amatique Bay during 2008. No anchovies were registered in the wet season. Isolines denote salinity gradients (ppt). Source: Ixquiac- lated to both wind and precipitation Cabrera et al. (2008) in October. Crevalle jack Caranx 36 Mar Ecol Prog Ser 533: 29–46, 2015

1.0 1.0 Nov Jul Jul Jun Wind Pre

SST May Aug

Shrimp Group Apr

Squid Corvina Mar Dec Jan Lsnap Tarpon Gerr Bagre Cich Lsnap Mack Oct SST Chl Jack Snook Shrp Bagre Msnap Aug Feb Dec Apr Nov Wind Mar Pre Barra May Sep Oct Jan Jun Csnap Sep Feb Anch –1.0 –1.0 SD SD –1.0 SD 1.0 –1.0 SD 1.0 Fig. 4. Redundancy analysis (RDA) of monthly catch rates Fig. 5. Redundancy analysis (RDA) of commercial species (CPUE, kg per unit effort [effort = boat × fishing days]) of landed by dories and skiffs, using nets and hook gear, in shrimp trawlers (circles) between 2006 and 2010 in Ama- Amatique Bay during 1998. Vectors indicate the influence of tique Bay. The vectors indicate the influence of the 2 signifi- the 2 significant environmental variables: wind speed cant environmental variables precipitation (Pre) and sea sur- (Wind) and sea surface temperature (SST), as well as precip- face temperature (SST), and wind speed (dashed line), itation (Pre) (dashed line) in the same period. Triangles rep- during the same period. Triangles represent individual spe- resent individual species and species groups: Jack = cies and species groups: Lsnap = lane snapper Lutjanus crevalle jack Caranx hippos; Mack = Spanish mackerel synagris, Shrimp = penaeid species including Litopenaeus Scomberomorus maculatus; Bagre = Gafftopsail catfish schmitti, Farfantepenaeus notialis and Xiphopenaeus Bagre marinus; Gerr = ‘mojarras’ (Gerridae group) including kroyer; Squid = Atlantic brief squid Lolliguncula brevis; Diapterus rhombeus and Eugerres plumieri; Barra = bar- Bagre = catfish group constituting Bagre marinus and Ariop- racuda Sphyraena picudilla; Msnap = mutton snapper Lut- sis assimilis; Corvina = a mix of Sciaenidae and Haemulidae janus analis, Csnap = Cubera snapper group including the (probably Protosciaena bathytatos and Pomadasys corvinae- Cubera snapper Lutjanus cyanopterus and grey snapper formis and L. synagris). Chl a was also treated as a biological Lutjanus griseus; Anch = anchovies Anchoa spp.; Cich = group. Source: National Fisheries Directorate (DIPESCA) blackbelt cichlid Paraneetroplus maculicauda; Shrimp = penaeid species including Litopenaeus schmitti, Farfantepe- naeus notialis and Xiphopenaeus kroyer; Snook = common hippos, the Spanish mackerel Scomberomorus snook Centropomus undecimalis; Tarpon = tarpon Mela - maculatus, and the catfish Bagre marinus were gops atlanticus; Lsnap = lane snapper Lutjanus synagris; and Group = goliath grouper Epinephelus itajara. Time of negatively re lated to SST and were more common year (months) is represented by circles. Source: estimates in December−March coinciding with increased derived by Heyman & Graham (2000) and Heyman & Grana- onshore winds. Anchovy Anchoa spp, barracuda dos-Dieseldorff (2012) from interviews of 42 experienced Sphyraena picudilla, mutton Lutjanus analis and fishers ‘cubera’ snappers were inversely related to precipitation and were thus more common in March−May. Physiological traits of selected fish species Maps of a selection of species caught in the research surveys are shown in Fig. 3. Species such The species sampled for analysis of reproduction as the lane snapper, squid and the anchovies are and growth included lane snapper (n = 364, 23.1 ± abundant during the dry season (February) but 4.2 cm, total length ± SD), grey snapper (n = 286, 28.3 almost absent in the wet season (August). The lane ± 5.8 cm), and gafftopsail catfish (n = 169 females, snappers captured with the commercial trawl gear 46.4 ± 6.3 cm). The lane and the grey snappers dis- consisted mostly of juveniles (average length = 13.4 played similar spawning and body condition cycles cm, size range = 3.0−28.6 cm). Species such as the (Fig. 6). Their GSI showed an increasing trend from stripped mojarra were more abundant during the January, reaching peaks in March−June. Gonad wet season. The density of species like the caitipa investment was relatively low in both species com- mojarra was apparently unaffected by the seasons. pared to the snook and especially the gaftopsail cat- Andrade et al.: Ecological linkages in an estuary bay 37

2 A Lane snapper 1.3 fish, with an average maximum monthly GSI CF GSI of about 1.3%. In July−August, coin- 1.6 cident with the onset of the rainy season, 1.2 GSI decreased abruptly, suggesting that 1.2 the main spawning event was over. From 0.8 August to December, gonad investment GSI (%) 1.1 was relatively low. This pattern matched 0.4 the body condition of the fish, as both Condition factor (%) species tended to show highest CF in 0 1 MAMJ JA SOND JFMA March−June, the dry season. Contrasting 2006 2007 growth patterns were observed in other 2 Grey snapper fish species like the common snook and, B to some extent, the gafftopsail catfish. GSI CF 2.1 1.6 Both species recovered their body condition during the rainy season, from Octo- 1.2 1.9 ber to January. This seemed to trigger spawning activity earlier in the dry sea- 0.8 GSI (%) son, by March−April, as revealed by their 1.7 0.4 GSI. However, the 2 species differed strongly in their gonad investment, from a Condition factor (%) 0 1.5 maximum of 1.6% in snook (2.5% in MAM J JA SOND JFMA females) (Andrade et al. 2013), to average 2006 2007 values exceeding 10% in April for the 2 C Snook 0.2 female catfish. GSI CF 1.6 0.18 Shorebirds and manatees

1.2 The shorebirds of Amatique Bay dis- GSI (%) 0.16 played clear seasonal patterns of occur- 0.8 rence (See Table S1 in the Supplement).

Condition factor (%) Exploratory analysis of the original sight- 0.4 0.14 ing data by means of RDA detected 4 MAMJ JASOND JFMA characteristic trends of seasonality in the 2006 2007 dominant species. The most abundant 16 D Gafftopsail catfish 0.28 group by far, comprising ca. 64% of GSI CF the sightings, included sandpipers Actitis 12 0.26 macularius and Calidris minutilla, plovers Pluvialis squatarola and Charadrius semi- 8 0.24 palmatus, and the whimbrel Numenius phaeopus that had relatively short stop- GSI (%) 4 0.22 overs in March−May and August– November and comprise group I in the

Condition factor (%) RDA biplot (Fig. 7). The sighting cycle of 0 0.2 MAM J JA SOND JFMA the black-bellied plover P. squatarola is 2006 2007 shown as an example by means of a GAM Month (inset, Fig. 7). This cycle has the first clear peak in the March−May period and a sec- Fig. 6. Annual cycles of the gonadosomatic index (GSI) (black lines) and condition factor (CF) (grey lines) of (A) lane snapper Lutjanus synagris, (B) ond in August−November. In the second grey snapper L. griseus, (C) common snook Centropomus undecimalis and major group (group II in Fig. 7) the black- (D) females of gafftopsail catfish Bagre marinus, sampled from March 2006 necked stilt Himantopus mexicanus, to April 2007 in Livingston and Puerto Barrios, Guatemala. Vertical bars denote standard errors of the means of sex-pooled observations. Dashed semipalmated sandpiper Calidris pusilla, line indicates a second degree polynomial fit to the mean CF and sanderling Calidris alba accounted 38 Mar Ecol Prog Ser 533: 29–46, 2015

1.0 Sl found in Lake Izabal and were highest during the dry Mar-May Ps Ss Pb I season and lowest during the surveys conducted in Wh Swr July and October. In contrast, downstream, the high- IV est densities were found in October at the mouth of the Sarstún River, in the western Amatique Bay, Aug-Nov Srb where manatees were virtually absent during the dry Sse II season (Fig. 8). Manatees forming relatively large Dec-Febr S aggregations were detected in both seasons in Gra-

250 II III ciosa Bay, where seagrasses are abundant. However, S III the surveys covered this particular area more spo - Pc I Pc radically, making it more difficult to determine clear May-June Sw Pb seasonal patterns of abundance.

IV Counts

–1.5 Mar Feb 0 SD DISCUSSION –1.0 SD 2.0 Fig. 7. Redundancy analysis biplot of seasonal occurrence of Seasonality of primary production in the estuary the 11 most common species of shorebirds reported by Eiser- mann (2009). Species divided into 4 groups (I−IV) based on the seasonal patterns of sightings, indicated by dashed lines. The main environmental drivers of the Amatique GAM regressions were fitted to the seasonal sightings of 4 Bay ecosystem, which are most probably also impor- species representative of each group (inset). Pb = black- tant for other western Caribbean estuaries, are the bellied plover Pluvialis squatarola; Ps = semipalmated precipitation, runoff and wind regimes, combined plover Charadrius semipalmatus; Ss = spotted sandpiper Ac- titis macularius; Wh = whimbrel Numenius phaeopus; Sl = with a weak tidal forcing. Low tidal amplitudes are least sandpiper Calidris minutilla; Stb = black-necked stilt characteristic of the Caribbean Sea (Kjerfve 1981), Himantopus mexicanus; S = sanderling Calidris alba, Sse = and this reduces tidal mixing. The hydrographic data semipalmated sandpiper Calidris pusilla; Pc = collared presented show that the climate in Amatique Bay is plover Charadrius collaris; Sw = western sandpiper Calidris dominated by a marked 2-season regime. From Feb- mauri; Swr = white-rumped sandpiper Calidris fuscicollis ruary to May precipitation is low and river discharge for 18% of the total sightings. This group had a more A 1 pronounced presence in the late rainy season 2-3 (August−November), as exemplified by the sander- 4-5 Amatique ling in the GAM (inset). A third group composed of - Sarstún Bay 6 7 the collared plover Charadrius collaris and the west- River ern sandpiper C. mauri was associated with the long rainy season from May to November. This group comprised about 13% of the overall counts, and some Lake sporadic sightings were made in the dry season. The Izabal collared plover was the only species observed to breed in the area. The white-rumped sandpiper C. B fuscicollis (group IV) was the only species that was observed nearly exclusively in the late dry season Amatique Sarstún Bay (March−May), and this species accounted for 2% of River the sightings. The combined seasonal patterns of the most abundant groups of birds (I and II) explain why the majority of the sightings were made in the late Lake rainy season (52%) and late dry season (25%). Izabal According to the observations performed in aerial surveys by Quintana-Rizzo & Machuca (2008) in Fig. 8. Distribution of manatee Trichechus manatus from 2006−2007, manatees in Lake Izabal-Amatique Bay aerial counts performed in (A) October 2006 (wet season) and (B) March 2007 (dry season) in Amatique Bay, 2007. may have a local distribution related to the seasonal Adapted from Quintana-Rizo & Machuca (2008). Filled cir- precipitation regime. In these surveys, the largest cles = number of manatee sightings, black line = survey densities of manatees, both adults and calves, were tracks Andrade et al.: Ecological linkages in an estuary bay 39

is at its yearly minimum. The increase in temperature tion cycles are thought to be more tightly coupled in and evaporation give rise to higher salinities as mar- the tropics than in temperate areas, responding ine water dominates in the Bay, with reported intru- quickly (days to weeks) to the hydrological regime sions as far upstream as Lake Izabal (Brinson et al. (Hoover et al. 2006, Chew & Chong 2011, Atwood et 1974). Despite weak tidal currents resulting from low al. 2012). tidal amplitudes, a steady onshore (NE) breeze provides good vertical mixing inside the Bay. From July to December, the rainy season dominates and the Fish spawning and aggregations run-off into the Bay combined with weaker and variable sea breezes results in a distinct halocline in the In Amatique Bay, spawning of fish such as the grey water column. The precipitation cycle in Amatique snapper, lane snapper, the snook and probably the Bay preceded or was in phase with the chl a cycle, gafftopsail catfish occur just prior to or during the suggesting that primary production responds quickly rainy (and warmer) season, in the months of March− to freshwater input and/or enhanced stratification November (Fig. 9). From July to November primary (Fig. 9). The rapid linkage between runoff and nutri- production is high and may favor larval survival and ent loadings has been shown for other tropical and growth. These observations are similar to those subtropical semi-enclosed bays, including Kaneohe reported for east Africa where fish spawning is asso- Bay, Hawaii and the microtidal Patos Lagoon estuary, ciated with the monsoon and rainfall events (Blaber Brazil (Hoover et al. 2006, Abreu et al. 2010, Drupp et 2000). Increased abundance of larvae of lane and al. 2011). The validity of the remote chl a data could grey snappers has been shown to overlap with peri- be challenged (Dierssen 2010), but additional meas- ods of high chl a concentrations in other localities in urements indicate that the water flowing from the the Caribbean (Yáñez-Arancibia et al. 1993, Falfan lake has peak concentrations of nutrients and low Vazquez et al. 2008). Similarly, growth rates and sur- volumes of suspended particles at the onset of the vival of snook recruits (age < 100 d) are known to be rainy season (Carrillo-Ovalle et al. 2000, Quintana- higher for juveniles spawned during the rainy season Rizzo & Machuca 2008). This confirms that peak (Aliaume et al. 2000). While we observed 3 potential primary production remotely measured can probably spawning events for snook, low GSI values during be associated with the seasonal flooding. Further the dry season suggest the importance of the rainy studies should, however, attempt to describe this season for spawning in this species. Overall, the 2 cycle in more detail and investigate the trophic link- snappers and snook invest relatively little in gonadal age to zooplankton and zooplanktivorous larvae of mass, or have a protracted spawning period given fish and shrimp. The primary and secondary produc- their average low GSI, as suggested for other species spawning in the tropics (Longhurst & Bird group III migration Nursery area for species Pauly 1987, Houde 1989). Part of the spawned in wet season variation in gonadal investment can Late wet also be explained by a geographic Spawning estuarine season Marine species gradient, as suggested earlier for snook & freshwater fish migration into bay (Andrade et al. 2013). Thus, this spe- species cies achieves greater gonado-somatic D J N F Dry indices during a shorter spawning season Manatee migrations O M season in cooler winter waters (e.g. from lake to bay Wet Florida). Analogous reproductive stra - season S A tegies have been reported in impor- Offshore shrimp A M JJ tant Lutjanidae and Centropomidae in migration the tropical belt of the Indo-Pacific. Bird group II Bird group IV For example, in northern Australia, migration Onset wet migration the red snappers Lutjanus erythro - season pterus and L. malabaricus had more Primary production Spawning marine bloom species de fined spawning peaks in the spring− summer months than their con - Fig. 9. A conceptual model of the precipitation-driven seasonal cycle of fish, seabirds and manatees in Amatique Bay. Color depth denotes precipitation in- specifics from eastern Indonesia (Fry tensity, from low (light blue) to strong (dark blue). Composition of bird groups et al. 2009). Contrastingly, in the more is given in Table S1 in the Supplement tropical environment of Indonesia, 40 Mar Ecol Prog Ser 533: 29–46, 2015

spawning cycles were longer, less synchronized Seasonal abundance of fish across sampling sites and ap parently more influenced by the precipitation cycle than the tempera- Climate variables affected the landings of trawlers ture cycle. Reproduction in the barramundi Lates cal- and those of dories and skiffs differently. Trawlers carifer, an important centropomid of Asia and operate mostly where shrimp are abundant, espe- Australia, is also under strong influence of the mon- cially on soft bottoms near river mouths. The multi- soon regime (Blaber et al. 2008). Towards the end of variate analyses showed that these fish assemblages the dry season the barramundi migrate to spawning were clearly affected by precipitation and river sites where reproductive activity is secondarily mod- runoff. Increases in rainfall and temperature are ulated by the monthly tidal cycle. During the wet thought to trigger offshore migration of juvenile season, post-larvae of barramundi enter coastal penaeids (Nagelkerken et al. 2008, Nemeth 2009). In swamps under the influence of spring tides (Blaber et Amatique Bay, landings of shrimp were related to al. 2008). High rainfall and warmer temperatures increasing seawater temperatures in the months of have been related to the increased survival and June and September, at the height of the rainy sea- growth of young barramundi and other coastal spe- son (Fig. 9). Hidalgo et al. (2004) describe penaeid cies in Queensland, Australia, giving rise to in- catches in Amatique as consisting mainly of sub- creased fishing yields (Balston 2009, Meynecke & adults spawned in the previous November−Decem- Lee 2011). ber period. Thus, the increased landings of shrimp The timing of spawning of the gafftopsail catfish appear to occur during dispersal from the nursery has been associated with the increased tempera- grounds. This has also been noted in the nearby tures and the onset of the rainy season in other Celestun lagoon, Mexico (Pérez-Castañeda & Defeo tropical localities (Mendoza-Carranza & Hernán- 2001, Pérez-Castañeda & Defeo 2004). dez-Franyutti 2005, Pinheiro et al. 2006). Our obser- In contrast to the trawlers, skiffs and dories employ- vations suggest, however, that spawning may start ing hooks and lines operate in areas with rocky bot- prior to the rainy season as reflected by the in - toms or along the Punta de Manabique coast (Heyman creased GSI in March 2007 and further decrease in & Granados-Dieseldorff 2012) and their major catches April. Extensive investment in gonadal products, occurred during the cooler dry season. Occurrence large egg size (up to 19 mm in our observations), and landings of engraulids, sciaenids, catfishes, bar- and parental mouth breeding in the gafftopsail cat- racuda, jacks, mackerels and, to a lesser extent, of fish may ensure the survival of the larvae, even if mutton snapper, were greatest from December to spawning occurs markedly earlier than the onset of April, and were associated with the onshore wind the rains and the planktonic production cycle (Rim- regime and intrusion of marine waters (Fig. 9). These mer & Merrick 1982). Biogeographic studies may species are often categorized as marine stragglers help resolve discrepancies in the timing of spawning (sensu Potter et al. 2015). The engraulid fishery, and physiological adaptations across latitudinal locally known as ‘manjua,’, which may comprise up to gradients. 15 species, accounts for 20% of the total landing vol- Taken together, the fishery landings and repro - ume in the whole Gulf of Honduras, and has peak ductive observations of lane and grey snappers, catches in April (Boix-Morán 2008, Heyman & Grana- gafftopsail catfish and snook, suggest that pre- dos-Dieseldorff 2011). This occurs simultaneously spawning migrations or spawning migrations in with increased abundances of juveniles of other fish March− November either increase the catchability of species in Amatique and other estuaries of the Carib- these species or that fishers simply target them dur- bean (Ixquiac-Cabrera et al. 2008, Burgos-Leon et al. ing this time period (Fig. 9). Similarly, the formation 2009, Poot-Salazar et al. 2009). Thus, it is likely that of spawning aggregations has been used to explain catches of barracudas, jacks and mackerels are more the increased catchability of tarpon, goliath grouper directly related to active feeding migrations than to and Gerridae in other estuaries and coastal waters of passive advection. Active feeding migrations have the Caribbean (Sadovy & Eklund 1999, Rueda & also been suggested elsewhere in the Caribbean Defeo 2001, Hammerschlag et al. 2012). Although (Manjarrés-Martínez et al. 2010), an indication that fishing spawning aggregations is not always detri- these oceanic species are not merely ‘stragglers’ into mental, trade-offs between fish size and fishing effort the estuaries. On the other hand, these predators and must be analyzed to derive a simple and adequate the cubera and grey snappers form spawning concen- fishing regime in the different seasons (van Overzee trations from March to September in marine waters & Rijnsdorp 2015). nearby, including the atoll of Gladden Spit (Boom - Andrade et al.: Ecological linkages in an estuary bay 41

hower et al. 2010, Manjarrés-Martínez et al. 2010, (Ixquiac Cabrera et al. 2008). This pattern is also Granados-Dieseldorff et al. 2013). Trophodynamic observed in permanently open microtidal estuaries in studies supplemented by investigations of reproduc- temperate Australia (Valesini et al. 2014) and upper tion are needed to resolve the proximate causes of estuaries in tropical West Africa and Australia their migration, but attention must also be paid to on- (Castellanos-Galindo & Krumme 2013a). Neverthe- togenetic factors. For instance, we observed that the less, there are distinct patterns in Amatique, as well peak trawler by-catch of small lane snapper occurs as in other Neotropical estuaries. In common with the during the dry season when juveniles are abundant. microtidal Términos Lagoon in the Caribbean and In contrast, the peak catches of larger lane snappers estuaries of the western central Atlantic the families occurred in reef areas during the rainy season, coin- Ariidae, Engraulidae, Gerreidae and Tetraodontidae ciding with the main spawning event. This is in agree- are prevalent, and Clupeidae and Claroteidae are ment with the observations of Whaley et al. (2007) less important or absent (Table 1) (Blaber 2000, Bar- who found juveniles normally associated with sea- letta & Blaber 2007, Ixquiac Cabrera et al. 2008, grass and soft bottoms, and adults predominantly as- Castellanos-Galindo & Krumme 2013b, Castellanos- sociated with coral reefs or rock offshore in Char- Galindo et al. 2013). In this respect, Amatique Bay lotte Harbor, Florida. Hence, the lane snapper uses has stronger similarity to the western tropical Amatique Bay both as a nursery and spawning area Atlantic and the tropical eastern Pacific, than to the (Fig. 9) and should be classified as a marine estuarine- tropical eastern Atlantic. This is also evident from the opportunist, following Potter et al. (2015). dominance of Ariidae in terms of biomass and of Ger- Species at the extreme of physiological adaptation ridae in terms of numbers (especially in mesotidal to estuarine life are the brief squid, which probably systems). In addition, the families Lutjanidae and represents the only hypo-saline adaptation of Centropomidae that support important fisheries in cephalopods (Bartol et al. 2002), and the blackbelt Amatique are less common in the Tropical Eastern cichlid Paraneetroplus maculicauda. These types of Atlantic (Castellanos-Galindo & Krumme 2013b). In transgression processes are examples of those that contrast, the Sciaenidae are less abundant than in may have occurred many times in the evolutionary West Africa, despite similar species richness. The history of the region, as also suggested for zooplank- similarity between Amatique and the eastern Pacific ton species (Pérez et al. 2013). Abundance of the brief region must reflect somewhat similar ecological con- squid in Amatique Bay, just as in the Chesapeake Bay ditions and, in particular, the short isolation history (3 (Bartol et al. 2002), was related to increased seawater million yr). This similarity with the Pacific contrasts temperature and salinity. However, the close relation- with the observations made for the other vertebrates. ship between landings of squid and shrimp found in Amatique also suggests a targeted feeding migration by this squid, as crustaceans are the most important Shorebirds and manatees prey item in their diet (Coelho et al. 2010, Jereb & Roper 2010). The abundance of the blackbelt cichlid The majority of the shorebirds sighted in Punta de was related to rising SST in August, after a period of Manabique are visitors (Eisermann 2009). The lar- intense runoff (Fig. 2). This species is very common in gest group, comprising many sandpipers and plovers Lake Izabal and Dulce River under freshwater condi- (group I), consisted of long-distance migrants that tions (Dickinson 1974) and, similarly to other Central have summer breeding areas in the tundra of North American cichlids, is known to be tolerant of brackish America (Poole 2005). Their clear bi-modal pattern of waters and capable of crossing narrow sea barriers occurrence suggests that these are transient birds (Miller 1966, Hulsey & López-Fernández 2011). with wintering areas in South America. These shorebirds probably use Amatique Bay for short stop-overs only. Less important, but still common visitors Zoogeographic patterns of fish (groups II and III) with breeding areas in temperate to high Arctic areas of America, seem to utilize the The present observations of the occurrence of the area for somewhat longer wintering periods, nor- fish fauna in the commercial catches are consistent mally late in the rainy season. Among the more com- with some of the general zoogeographic patterns of mon species, only the collared plover Charadrius tropical estuaries. Thus, the fish communities are collaris has a regional distribution limited to the Car- dominated by marine species and both their diversity ibbean. The plover breeds in Amatique and stays for and abundances are higher during the dry season a longer period, from June to November, and proba- 42 Mar Ecol Prog Ser 533: 29–46, 2015

Table 1. Contributions of different taxa (shown at family level) to the number of species, abundance (No. of ind.) and biomass (B) of mangrove and estuarine fish assemblages from Amatique Bay (Caribbean), the tropical Eastern Pacific, Western Central At- lantic and tropical Eastern Atlantic (adapted and expanded from Castellanos-Galindo & Krumme 2013b). Data for Amatique Bay are from research surveys performed with shrimp bottom trawler (Ixquiac Cabrera et al. 2008) (see Text S2 in the Supple- ment at www.int-res. com/articles/ suppl/m533p029_ supp.pdf ), and data for the tropical Eastern Pacific, Western Central Atlantic and tropical Eastern Atlantic are from Castellanos-Galindo & Krumme (2013a) citing different sources

Family Amatique Bay (Caribbean) Tropical Eastern Pacific Western Central Atlantic Tropical Eastern Atlantic No. of No. of B No. of No. of B No. of No. of B No. of No. of B species ind. (%) species ind. (%) species ind. (%) species ind. (%) (%) (%) (%) (%) (%) (%) (%) (%)

Ariidae 3.3 14.3 31.7 6.7 4.0 19.1 7.4 31.5 32.4 4.3 1.3 3.4 Cichlidae 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 4.3 0.0 0.1 Claroteidae 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 4.3 0.5 1.5 Clupeidae 4.9 2.0 1.2 2.5 15.4 6.9 2.4 2.4 0.4 5.7 46.5 27.2 Eleotridae 0.0 0.0 0.0 2.5 0.8 0.6 0.8 0.1 0.1 1.4 0.0 0.0 Elopidae 0.0 0.0 0.0 0.3 0.1 0.0 0.9 0.0 0.0 1.4 0.3 0.4 Engraulidaee 8.2 5.9 3.4 5.4 9.5 0.0 10.0 18.2 7.4 0.0 0.0 0.0 Gerreidaee 4.9 37.4 12.5 4.1 20.4 1.6 2.9 1.6 0.8 2.9 0.1 0.1 Mugilidae 0.0 0.0 0.0 1.3 8.7 1.3 4.6 8.5 8.7 7.1 2.2 2.7 Polynemidae 3.3 0.2 0.2 0.6 0.7 0.0 0.7 0.0 0.0 4.3 1.7 3.4 Pristigasteridae 1.6 0.5 0.1 1.0 0.2 0.0 0.8 0.0 0.0 0.0 0.0 0.0 Sciaenidae 16.4 9.8 8.8 12.4 3.1 0.3 16.6 10.4 9.1 8.6 39.7 50.4 Tetraodontidaee 6.6 1.2 1.0 2.9 3.0 19.5 3.9 9.2 21.9 1.4 0.1 0.1

bly makes some limited seasonal migration there- in the Caribbean. This may help explain why sire- after. This migration pattern may represent a less nians are absent and seagrasses scarce along the Pa- common, and probably more recent, adaptation re - cific coast of Central America (Green & Short, 2003, sulting in fledging and an early growth period corre- Samper-Villarreal et al. 2014). The aerial surveys per- sponding to the productive rainy season in the Bay formed by Quintana-Rizzo & Machuca (2008) suggest (Fig. 9). Eisermann (2009) characterized the Punta de that seasonal movements related to the hydrological Manabique Wildlife Refuge as a shorebird migration cycle and to the life-cycle of manatees occur within site of secondary importance. This is in agreement the Lake Izabal-Amatique Bay complex (Fig. 9). A with the observed decline in the abundance of over- larger number of sightings in the Bay proper were wintering or migrating shorebirds in the Gulf Coast achieved during the wet season. Possible reasons for south of the Tropic of Cancer (23° 27’ N) (Withers the stronger affinity to coastal areas during the wet 2002). Furthermore, Barrantes & Chaves-Campos season include a wider access to areas with drinkable (2009) demonstrated a lower abundance of migrating freshwater, strong river currents, and increased tur- shorebirds on the east coast of Costa Rica compared bidity and subsequent loss of submerged vegetation to its Pacific coast. A likely reason for this longitudi- upstream (Auil 2004). Although the density of mana- nal contrast may be the lack of extensive tides and tees in some of the Caribbean populations may be rel- tidal flats in the western Caribbean in contrast to the atively stable, there is still much needed research Pacific coast. This may limit the access of many with regard to the environmental, behavioral and shorebirds to aquatic invertebrates, which are their physiological basis of manatee migration as essential main prey. Thus, the Amatique region most likely information for the implementation of a regional man- has greatest value as a transient resting area, rather agement plan (Harborne et al. 2006, UNEP 2010, than an important feeding or breeding ground for Castelblanco-Martínez et al. 2013). most shorebird species. Interestingly, the physical processes that may be responsible for the low abundance of shorebirds may Conclusion also have played a role in the adaptation and persist- ence of sirenian populations in the Caribbean. The Ecological connectivity can be defined as the abundance of freshwater, the sheltered environment, strength of the interactions among ecosystem compo- water clarity, and very low tidal amplitude lead to nents by movement of organisms, often at different abundant seagrass and suitable habitat for manatees stages of their life-cycles, as well as by the exchange Andrade et al.: Ecological linkages in an estuary bay 43

of nutrients and organic matter (Nagelkerken 2009, complex system, these fishers achieve reasonable Sheaves 2009). Migration to and from estuaries can levels of agreement and co-existence (Andrade & range from large-scale seasonal movements related Midré 2011). It may be that the large focus on the to reproduction, feeding and ontogeny, to short shrimp and engraulids at the lower trophic levels migrations during the twilight (Krumme 2009). The corresponds to an example of a balanced fishery most conspicuous linkage between the freshwater with output reasonably proportional to productivity system and the estuary in Amatique are the move- (Garcia et al. 2012). Future studies should inves tigate ments of snook, the blackbelt cichlid and, to some the match of size distributions in the harvest and in extent, manatees. These movements are related to the sea, as well as the consequences of fishing at the both spawning cycles and the precipitation cycle. For lower trophic levels. Another issue of interest for example, the common snook moves in and out from population management is the timing of the ro - freshwater environments to forage and spawn at sea, tational fishery closures and their suitability for thus interconnecting freshwater and marine environ- protecting spawning aggregations, especially since ments (Barbour & Adams 2012). The migrations of these closures were agreed by the fisher groups in these fish species towards the sea may reflect the a participatory process with the primary purpose general trend for catadromy observed in relict mar- of avoiding conflicts related to gear saturation ine taxa in the tropics. The more recent affinity for (Andrade & Midré 2011). The range of observed eco- freshwater may reflect an adaptation to the relatively logical linkages has played major roles in the evolu- higher food availability in freshwater than in the sea tionary processes in the western Caribbean, thus it is (Gross et al. 1988, Lucas et al. 2001). Manatees also important that they are integrated into fishery and play an important role in the recycling of nutrients conservation plans. in the western Caribbean, and they are probably responsible for a net export of nutrients to adjacent Acknowledgements. This study received financial support ecosystems downriver (Castelblanco-Martínez et al. from a Russell E. Train Education for Nature fellowship of 2012). The reverse (oceanic) input to the estuary is the World Wildlife Fund, the University of Tromsø and Akvaplan-niva AS. We thank those who helped and sup- triggered by the massive migration of several species ported us in the collection of the field samples in Livingston of penaeid shrimp, which are thought to use the man- and Puerto Barrios, including Polo, Julian Arana, and Pedro groves, and by engraulids, which probably use the Ramirez. Thanks to Mario Jolón, Regina Sanchez, and Bay for spawning. These aggregations attract a great Amely Garcia for providing extensive literature on Ama- tique Bay. Thanks to Johan Groeneveld, P. E. Renaud, J. C. number of transient coastal and oceanic predators Villagrán, I. Nagelkerken and 3 anonymous reviewers comprising, among others, the families Carangidae, for their valuable comments on early versions of the Loliginidae, Lutjanidae, Scombridae, and Sphyrae - manuscript. nidae, especially during the dry season. At this time of the year bullshark Carcharinus lecuas, and large- LITERATURE CITED tooth sawfish Pristis perotteli have also been Abreu PC, Bergesch M, Proenca LA, Garcia CAE, Ode- reported to enter Lake Izabal (Dickinson 1974). Pred- brecht C (2010) Short- and long-term chlorophyll a vari- ators like the Lutjanidae have, however, a marked ability in the shallow microtidal Patos Lagoon estuary, reef-ecosystem rather than oceanic affinity. Thus, our Southern Brazil. Estuaries Coasts 33: 554−569 observations strongly suggest an ontogenetic change Abt Associates/Woods Hole Group (2003) Gulf of Honduras. Preliminary transboundary diagnostic analysis. Inter in the utilization of different habitats in the Bay from American Development Bank soft bottom to reefs, by different life-stages of the Aliaume C, Zerbi A, Joyeux JC, Miller JM (2000) Growth of lane snapper. juvenile Centropomus undecimalis in a tropical island. By contrast, the shorebirds using Amatique Bay Environ Biol Fishes 59:299−308 Andrade H, Midré G (2011) The merits of consensus: Small- mostly depend on the Bay and estuaries as a peren- scale fisheries as a livelihood buffer in Livingston, nial refuge and contribute less to nutrient cycling. Guatemala. In: Jentoft S, Eide A (eds) Poverty mosaics: However, shorebirds represent a teleconnection Realities and prospects in small-scale fisheries. Springer, between the high-latitude systems of North America Dordrecht Andrade H, Santos J, Taylor R (2013) Life-history traits of the and those in South America that are of conservation common snook Centropomus undecimalis in a Carib- interest. At different levels of the food chain there is bean estuary and large-scale biogeographic patterns rel- a large number of exploitation strategies by fishers evant to management. J Fish Biol 82: 1951−1974 Arrivillaga A, Baltz DM (1999) Comparison of fishes and from different fleets, social and ethnic groups, indi- macroinvertebrates on seagrass and bare-sand sites on rectly involved in competing small-scale fisheries. Guatemala’s Atlantic coast. Bull Mar Sci 65: 301−319 Previous observations suggest that in an apparently Atwood TB, Wiegner TN, MacKenzie RA (2012) Effects of 44 Mar Ecol Prog Ser 533: 29–46, 2015

hydrological forcing on the structure of a tropical estuar- watershed analysis for the Mesoamerican Reef. World ine food web. Oikos 121:277−289 Resources Institute, Washington, DC Auil NE (2004) Abundance and distribution trends of the West Carrillo-Ovalle L, Salaverría A, Pacas-Martínez L, Martínez- Indian manatee in the coastal zone of Belize: implications Mencos J (2000) Evaluación del recurso pesquero y for conservation. MSc thesis, Texas A&M University oceanografía del atlántico guatemalteco, durante el año Balston J (2009) Short-term climate variability and the com- 2,000. Centro de Estudios del Mar y Acuicultura, Univer- mercial barramundi (Lates calcarifer) fishery of north- sidad de San Carlos de Guatemala, Guatemala east Queensland, Australia. Mar Freshw Res 60:912−923 Castelblanco-Martínez DN, Barba E, Schmitter-Soto JJ, Barbour A, Adams A (2012) Biologging to examine multiple Hernandez-Arana HA, Morales-Vela B (2012) The life stages of an estuarine-dependent fish. Mar Ecol Prog trophic role of the endangered Caribbean manatee Ser 457: 241−250 Trichechus manatus in an estuary with low abundance of Barletta M, Blaber SJM (2007) Comparison of fish assem- seagrass. Estuaries Coasts 35:60−77 blages and guilds in tropical habitats of the Embley Castelblanco-Martínez DN, Padilla-Saldivar J, Hernandez- (Indo-west Pacific) and Caeté (Western Atlantic) estuar- Arana HA, Slone DH, Reid JP, Morales-Vela B (2013) ies. Bull Mar Sci 80: 647−680 Movement patterns of Antillean manatees in Chetumal Barletta M, Jaureguizar AJ, Baigun C, Fontoura NF and oth- Bay (Mexico) and coastal Belize: A challenge for regional ers (2010) Fish and aquatic habitat conservation in South conservation. Mar Mamm Sci 29:E166−E182 America: a continental overview with emphasis on Castellanos-Galindo G, Krumme U (2013a) Tidal, diel and neotropical systems. J Fish Biol 76:2118−2176 seasonal effects on intertidal mangrove fish in a high- Barrantes G, Chaves-Campos J (2009) Birds in coastal and rainfall area of the Tropical Eastern Pacific. Mar Ecol marine environments. In: Wehrtmann I, Cortés J (eds) Prog Ser 494: 249−265 Marine biodiversity of Costa Rica, Central America. Castellanos-Galindo G, Krumme U (2013b) Mangrove fish Springer, Dordrecht assemblages from data-sparse regions and the measure- Bartol IK, Mann R, Vecchione M (2002) Distribution of the ment of ecological equivalence: Comment on Sheaves euryhaline squid Lolliguncula brevis in Chesapeake (2012). Mar Ecol Prog Ser 474:299−302 Bay: effects of selected abiotic factors. Mar Ecol Prog Ser Castellanos-Galindo G, Krumme U, Rubio E, Saint-Paul U 226: 235−247 (2013) Spatial variability of mangrove fish assemblage Benoit J, Adnet S, El Mabrouk E, Khayati H and others composition in the tropical eastern Pacific Ocean. Rev (2013) Cranial remain from Tunisia provides new clues Fish Biol Fish 23: 69−86 for the origin and evolution of Sirenia (Mammalia, Chakrabarty P, Albert J (2011) Not so fast: a new take on the Afrotheria) in Africa. PLoS ONE 8:e54307 great American biotic interchange. In: Albert JS, Reis RE Blaber S (2000) Tropical estuarine fishes: ecology, exploita- (eds) Historical biogeography of neotropical freshwater tion, and conservation. Blackwell, Oxford fishes. University of California Press, Berkeley, CA Blaber SJM, Milton DA, Salini JP (2008). The biology of bar- Chew LL, Chong VC (2011) Copepod community structure ramundi (Lates calcarifer) in the Fly River system. In: and abundance in a tropical mangrove estuary, with Bolton BR (ed) The Fly River, Papua New Guinea: Envi- comparisons to coastal waters. Hydrobiologia 666: ronmental studies in an impacted tropical river system. 127−143 Elsevier, p 409−424 Coelho LI, Muto EY, Marian J, Soares LSH (2010) Contribu- Boix-Morán JL (2008) La pesquería de la Manjúa en el tion to the knowledge on the diet, feeding activity, and Caribe de Guatemala, impactos biológicos y socioe- reproduction of Lolliguncula brevis (Blainville, 1823) in conómicos para las poblaciones pesqueras y de consumo the coastal region off Santos (Sao Paulo state). Bol Inst en el Altiplano de Guatemala. Proyecto FODECYT No Pesca Sao Paulo36:225−236 100-2006. Centro de Estudio del Mar y Acuicultura, Dickinson JC (1974) Fisheries of Lake Izabal, Guatemala. USAC, Guatemala Geogr Rev 64:385−409 Boomhower J, Romero M, Posada J, Kobara S, Heyman W Dierssen HM (2010) Perspectives on empirical approaches (2010) Prediction and verification of possible reef-fish for ocean color remote sensing of chlorophyll in a chang- spawning aggregation sites in Los Roques Archipelago ing climate. Proc Natl Acad Sci USA 107:17073−17078 National Park, Venezuela. J Fish Biol 77:822−840 Drupp P, De Carlo EH, Mackenzie FT, Bienfang P, Sabine Bowen BW, Rocha LA, Toonen RJ, Karl SA (2013) The ori- CL (2011) Nutrient inputs, phytoplankton response, and gins of tropical marine biodiversity. Trends Ecol Evol 28: CO2 variations in a semi-enclosed subtropical embay- 359−366 ment, Kaneohe Bay, Hawaii. Aquat Geochem 17: Briggs JC (1984) Fresh-water fishes and biogeography 473−498 of Central America and the Antilles. Syst Zool 33: Eisermann K (2009) Abundance of shorebirds on the Carib- 428−435 bean Coast of Guatemala. Waterbirds 32: 337−344 Briggs JC, Bowen BW (2012) A realignment of marine bio- Ejrnæs R (2000) Can we trust gradients extracted by geographic provinces with particular reference to fish detrended correspondence analysis? J Veg Sci 11: distributions. J Biogeogr 39:12−30 565−572 Brinson MM, Brinson LG, Lugo AE (1974) Gradient of salin- El-Gohary M, McNames J (2007) Establishing causality with ity, its seasonal movement, and ecological implications whitened cross-correlation analysis. IEEE Trans Biomed for Lake Izabal-Rio-Dulce ecosystem, Guatemala. Bull Eng 54:2214−2222 Mar Sci 24:533−544 Faaborg J, Holmes RT, Anders AD, Bildstein KL and others Burgos-Leon A, Pérez-Castaneda R, Defeo O (2009) Dis- (2010) Recent advances in understanding migration sys- cards from the artisanal shrimp fishery in a tropical tems of New World land birds. Ecol Monogr 80:3−48 coastal lagoon of Mexico: spatio-temporal patterns and Falfan Vazquez EE, Ordonez Lopez UU, Ornelas Roa MM fishing gear effects. Fish Manag Ecol 16:130−138 (2008) Spatial variation of snapper and grouper larvae in Burke L, Sugg Z (2006) Hydrologic modeling of watersheds Yucatan shelf. Hidrobiologica 18:69−76 discharging adjacent to the Mesoamerican Reef. On the Fonseca AC, Arrivillaga A (2003) Coral reefs of Guatemala. Andrade et al.: Ecological linkages in an estuary bay 45

In: Cortés J (ed) Latin American Coral Reefs. Elsevier, graphy of neotropical freshwater fishes. University of Amsterdam California Press, Berkeley, CA Fry G, Milton DA, Van Der Velde T, Stobutzki I, Andamari Hunter MEM, Mignucci Giannoni AA, Tucker KPK, King R, Badrudin, Sumiono B (2009) Reproductive dynamics TLT, Bonde RKR (2012) Puerto Rico and Florida mana- and nursery habitat preferences of two commercially tees represent genetically distinct groups. Conserv important Indo-Pacific red snappers Lutjanus erythrop - Genet 13: 1623−1635 terus and L. malabaricus. Fish Sci 75:145−158 Ixquiac Cabrera MJ, Corona Avalos A, García BR, Salazar Garcia SM, Kolding J, Rice J, Rochet MJ and others (2012) M, Tejeda CA, Rodas R (2008) Evaluación y manejo de Reconsidering the consequences of selective fisheries. las comunidades demersales de la Bahía de Amatique Science 335:1045−1047 Izabal, Guatemala. Estimación de los impactos pes- González P, López D (2000) Desarrollo de herramientas queros. Centro de Estudios del Mar y Acuicultura, Uni- económicas para la preparación de políticas sostenibles versidad de San Carlos de Guatemala, Guatemala. http:// en el sector pesquero del Golfo de Honduras. Tomo I: digi. usac. edu. gt/ bvirtual/informes/ puirna/ INF-2008-090. Caracterización de las pesquerías de camarón y lan- pdf gosta, aproximación conceptual para la gestión pes- Jereb P, Roper CFE (2010) Cephalopods of the world. An quera. ICSED-PROARCA/CAPAS. http: // pdf. usaid. gov/ annotated and illustrated catalogue of cephalopod spe- pdf_ docs/PNACK091. pdf cies known to date, Vol 2. Myopsid and Oegopsid Granados-Dieseldorff P, Heyman WD, Azueta J (2013) His- Squids. FAO Species Catalogue for Fishery Purposes, tory and co-management of the artisanal mutton snapper Rome (Lutjanus analis) spawning aggregation fishery at Glad- Jolón-Morales MR (2006) Plan de conservación de área den Spit, Belize, 1950−2011. Fish Res 147: 213−221 2007-2011, refugio de vida silvestre Punta de Mana- Green EP, Short FT (2003) World atlas of seagrasses. Univer- bique. FUNDARY-CONAP-PROARCA-TNC, Guate mala. sity of California Press, Berkeley, CA. www. conap. gob.gt/ index. php/sigap/ planes-maestros. Gross MR, Coleman RM, McDowall RM (1988) Aquatic pro- html? download=367:plan-maestro-refugio-de-vida- ductivity and the evolution of diadromous fish migration. silvestre-punta-de-manabique Science 239:1291−1293 King M (1995) Fisheries biology, assessment and manage- Hammerschlag N, Luo J, Irschick DJ, Ault JS (2012) A comment. Fishing News Books, Blackwell, Oxford parison of spatial and movement patterns between sym- Kjerfve B (1981) Tides of the Caribbean Sea. J Geophys Res patric predators: bull sharks (Carcharhinus leucas) and 86: 4243−4247 Atlantic tarpon (Megalops atlanticus). PLoS ONE 7: Krumme U (2009) Diel and tidal movements by fish and e45958 decapods linking tropical coastal ecosystems. In: Nagel- Harborne AR, Mumby PJ, Micheli F, Perry CT, Dahlgren kerken I (ed) Ecological connectivity among tropical CP, Holmes KE, Brumbaugh DR (2006) The functional coastal ecosystems. Springer, Dordrecht value of Caribbean coral reef, seagrass and mangrove Latta SC (2012) Avian research in the Caribbean: past con- habitats to ecosystem processes. Adv Mar Biol 50: 57−189 tributions and current priorities. J Field Ornithol 83: Hernández B, Cherrington EA, Bosarreyes A, Gildardo- 107−121 Gálvez J, Oyuela M, Sempris E (2012) Clasificación de Livezey BCB (2010) Phylogenetics of modern shorebirds cobertura actual del mangle de las costas del Pacífico y (Charadriiformes) based on phenotypic evidence: analy- Atlántico de Guatemala, a través de técnicas de percep- sis and discussion. Zool J Linn Soc 160:567−618 ción remota. CATHALAC, Panama Longhurst A, Pauly D (1987) Ecology of tropical oceans, Vol Heyman WD, Graham RT (2000) La voz de los pescadores de 389. International Center for Living Aquatic Resources la costa atlántica de Guatemala. Trinational Alliance for Management, Manila the Conservation of the Gulf of Honduras. FUN- Lowerre-Barbieri SK, Ganias K, Saborido-Rey F, Murua H, DAECO/TIDE, Puerto Barrios Hunter JR (2011) Reproductive timing in marine fishes: Heyman WD, Kjerfve B (eds) (2001) The Gulf of Honduras, variability, temporal scales, and methods. Mar Coast Vol 144. Springer, New York, NY Fish: Dynam Manag Ecosys Sci 3:71−91 Heyman WD, Granados-Dieseldorff P (2011) The Manjua Lucas MC, Baras E, Thom T, Duncan A, Slavík O (2001) fishery of Guatemala’s Caribbean: balancing the needs Migration of freshwater fishes. Blackwell, Oxford for regional ecosystem productivity and national food Lund U, Agostinelli C (2014) http://cran.r-project. org/ web/ security. Focus Geogr 54:45−50 packages/CircStats/CircStats.pdf Heyman WD, Granados-Dieseldorff P (2012) The voice of MacDonald-Barrios BD (2011) Análisis de la diversidad de the fishermen of the Gulf of Honduras: improving las praderas de pastos marinos en la laguna marino- regional fisheries management through fisher participa- costera de la Bahía La Graciosa, Izabal, Guatemala. tion. Fish Res 125−126: 129−148 Licenciatura, Universidad San Carlos de Guatemala, Hidalgo H, Villagrán E, De Campo E (2004) Evaluación de la Guatemala sostenibilidad de la captura del camarón en la bahía de Manjarrés-Martínez LM, Gutiérrez-Estrada JC, Mazenet- Amatique y propuestas para su adecuado manejo y recu- González J, Soriguer MC (2010) Seasonal patterns of peración. FUNDAECO, Guatemala three fish species in a Caribbean coastal gill-net fishery: Hoover RS, Hoover D, Miller M, Landry MR, DeCarlo EH, Biologically induced or climate-related aggregations? Mackenzie FT (2006) Zooplankton response to storm Fish Res 106: 358−367 runoff in a tropical estuary: bottom-up and top-down McCullagh P, Nelder JA (1989) Generalized linear models. controls. Mar Ecol Prog Ser 318:187−201 Chapman & Hall, London Houde ED (1989) Comparative growth, mortality, and ener- Mendoza-Carranza M, Hernández-Franyutti A (2005) getics of marine fish larvae: temperature and implied lat- Annual reproductive cycle of gafftopsail catfish, Bagre itudinal effects. Fish Bull 87:471−495 marinus (Ariidae) in a tropical coastal environment in the Hulsey CD, López-Fernández H (2011) Nuclear Central Gulf of Mexico. Hidrobiologica 15:275−282 America. In: Albert JS, Reis RE (eds) Historical biogeo - Meynecke JO, Lee SY (2011) Climate-coastal fisheries rela- 46 Mar Ecol Prog Ser 533: 29–46, 2015

tionships and their spatial variation in Queensland, Aus- ➤ Sheaves M, Johnston R (2009) Ecological drivers of spatial tralia. Fish Res 110: 365−376 variability among fish fauna of 21 tropical Australian ➤ Miller RR (1966) Geographical distribution of Central Amer- estuaries. Mar Ecol Prog Ser 385:245−260 ican freshwater fishes. Copeia 773−802 Šmilauer P, Lepš J (2014) Multivariate Analysis of Ecological Nagelkerken I (2009) Introduction. In: Nagelkerken I (ed) Data Using CANOCO 5. Cambridge University Press, Ecological connectivity among tropical coastal ecosys- New York, NY tems. Springer, Dordrecht ➤ Somveille M, Manica A, Butchart SHM, Rodrigues ASL ➤ Nagelkerken I, Blaber SJM, Bouillon S, Green P and others (2013) Mapping global diversity patterns for migratory (2008) The habitat function of mangroves for terrestrial birds. PLoS ONE 8:e70907 and marine fauna: a review. Aquat Bot 89:155−185 ➤ Soto I, Andrefouet S, Hu C, Muller-Karger FE, Wall CC, Nemeth R (2009) Dynamics of reef fish and decapod crus- Sheng J, Hatcher BG (2009) Physical connectivity in the tacean spawning aggregations: underlying mechanisms, Mesoamerican Barrier Reef System inferred from 9 years habitat linkages, and trophic interactions. In: Nagelker- of ocean color observations. Coral Reefs 28:415−425 ken I (ed) Ecological connectivity among tropical coastal ➤ ter Braak CJF (1986) Canonical correspondence-analysis: a ecosystems. Springer, Dordrecht new eigenvector technique for multivariate direct gradi- ➤ Pérez L, Lorenschat J, Massaferro J, Pailles C and others ent analysis. Ecology 67: 1167−1179 (2013) Bioindicators of climate and trophic state in low- ter Braak CJF, Šmilauer P (2002) CANOCO Reference man- land and highland aquatic ecosystems of the Northern ual and CanoDraw for Windows user’s guide: software Neotropics. Rev Biol Trop 61:603−644 for canonical community ordination (version 4.5). Micro- ➤ Pérez-Castañeda R, Defeo O (2001) Population variability of computer Power, Ithaca, NY four sympatric Penaeid shrimps (Farfantepenaeus spp.) UNEP (2010) Regional management plan for the West in a tropical coastal lagoon of Mexico. Estuar Coast Shelf Indian manatee (Trichechus manatus) (compilers: Quin- Sci 52:631−641 tana-Rizzo E, Reynolds J III). CEP Technical Report No. ➤ Pérez-Castañeda R, Defeo O (2004) Spatial distribution and 48. UNEP Caribbean Environment Programme, Kingston structure along ecological gradients: penaeid shrimps in ➤ Valesini FJ, Tweedley JR, Clarke KR, Potter IC (2014) The a tropical estuarine habitat of Mexico. Mar Ecol Prog Ser importance of regional, system-wide and local spatial 273: 173−185 scales in structuring temperate estuarine fish communi- ➤ Pinheiro P, Broadhurst MK, Hazin FHV, Bezerra T, Hamilton ties. Estuaries Coasts 37: 525−547 S (2006) Reproduction in Bagre marinus (Ariidae) off Per- ➤ van Overzee HJ, Rijnsdorp A (2015) Effects of fishing during nambuco, northeastern Brazil. J Appl Ichthyol 22: the spawning period: implications for sustainable man- 189−192 agement. Rev Fish Biol Fish 25:65−83 Poole A (2005) The birds of North America online. http:// ➤ Velez-Juarbe J, Domning DP, Pyenson ND (2012) Iterative bna.birds.cornell.edu/BNA/ (accessed July 2013) evolution of sympatric seacow (Dugongidae, Sirenia) ➤ Poot-Salazar A, Pérez-Castaneda R, Vega-Cendejas ME, assemblages during the past similar to 26 million years. Defeo O (2009) Assessing patterns of ichthyofauna dis- PLoS ONE 7:e31294 carded by an artisanal shrimp fishery through selectivity ➤ Weir JTJ, Bermingham EE, Schluter DD (2009) The great experiments in a coastal lagoon. Fish Res 97:155−162 American biotic interchange in birds. Proc Natl Acad Sci ➤ Potter IC, Tweedley JR, Elliott M, Whitfield AK (2015) The USA 106: 21737−21742 ways in which fish use estuaries: a refinement and ex - ➤ Whaley SD, Burd JJ, Robertson BA (2007) Using estuarine pansion of the guild approach. Fish Fish 16: 230−239 landscape structure to model distribution patterns in Quintana-Rizzo E, Machuca O (2008) Monitoreo científico nekton communities and in juveniles of fishery species. para la conservación del manatí y su habitat. Fundación Mar Ecol Prog Ser 330:83−99 Defensores de la Naturaleza-FONACON F10/2006, Gua - Wilkinson L, Balasanov Y, Adinarayana Murthy SN, Kumar temala R (2009) Time series. Systat Software, Chicago, IL Rimmer MA, Merrick JR (1982) A review of reproduction ➤ Withers K (2002) Shorebird use of coastal wetland and bar- and development in the fork-tailed catfishes (Ariidae). rier island habitat in the Gulf of Mexico. Scientific World Proc Linn Soc N S W 107: 41−50 J 2: 514−536 ➤ Rueda M, Defeo O (2001) Survey abundance indices in a ➤ Yáñez-Arancibia A, Lara-Domínguez A, Day J Jr (1993) tropical estuarine lagoon and their management implica- Interactions between mangrove and seagrass habitats tions: a spatially-explicit approach. ICES J Mar Sci 58: mediated by estuarine nekton assemblages: coupling of 1219−1231 primary and secondary production. Hydrobiologia 264: Sadovy Y, Eklund A (1999) Synopsis of biological data on the 1−12 Nassau grouper, Epinephelus striatus (Bloch, 1792), and Yañez-Arancibia A, Zárate-Lomelí D, Terán-Cuevas A the jewfish, E. itajara (Lichtenstein, 1822). NOAA Tech (1994) Evaluation of the coastal and marine resources of Rep NMFS 146. US Department of Commerce, Seattle, the Atlantic coast of Guatemala. CEP Tech Rep No. 34. WA UNEP Caribbean Environment Programme, Kingston ➤ Samper-Villarreal J, Bourg A, Sibaja-Cordero JA, Cortés J ➤ Yañez-Arancibia A, Lomeli DZ, Cruz MG, Orantes RG, (2014) Presence of a Halophila baillonii Asch. (Hydro- Fandino VS (1999) The ecosystem framework for plan- charitaceae) seagrass meadow and associated macro- ning and management the Atlantic coast of Guatemala. fauna on the Pacific Coast of Costa Rica. Pac Sci 68: Ocean Coast Manage 42: 283−317 435−444 Zar JH (1998) Biostatistical analysis. Prentice-Hall, Upper ➤ Sheaves M (2009) Consequences of ecological connectivity: Saddle River, NJ the coastal ecosystem mosaic. Mar Ecol Prog Ser 391: ➤ Zink RMR (2011) The evolution of avian migration. Biol J 107−115 Linn Soc 104:237−250

Editorial responsibility: Ivan Nagelkerken, Submitted: May 26, 2014; Accepted: May 11, 2015 Adelaide, South Australia, Australia Proofs received from author(s): July 28, 2015