Rev Fish Biol Fisheries DOI 10.1007/s11160-010-9197-z RESEARCH PAPER Modeling population dynamics and conservation of arapaima in the Amazon L. Castello • D. J. Stewart • C. C. Arantes Received: 21 May 2010 / Accepted: 14 December 2010 Ó Springer Science+Business Media B.V. 2011 Abstract To promote understanding of fish popu- drastic changes in fishing practices in the Amazon, lation dynamics in tropical river-floodplains, we have Brazil. Model population numbers were close to synthesized existing information by developing a observed numbers, with generally low mean absolute largely empirical population model for arapaima percentage errors for juveniles (16%), adults (30%), (Arapaima sp.). Arapaima are characterized by very and catch (18%). In using the model to test ecological large bodies, relatively late sexual maturity, small hypotheses and to investigate management strategies, clutches, and large parental investment per offspring, we found the following: (1) Annual recruitment is and their populations are overexploited and even directly and positively related to spawner abundance, declining due to overfishing. We used unparalleled and it appears to be density-compensatory following time series data on growth, reproduction, catch-at- a Beverton–Holt relation (R2 = 0.85). (2) Fishing- age, and size-class abundance estimates for a popu- selectivity of arapaima caused by use of harpoons and lation that has increased several-fold and undergone gillnets can lower yield potentials dramatically through removal of the faster-growing individuals L. Castello Á C. C. Arantes of the population. That is in part because fewer Mamiraua´ Institute for Sustainable Development, Tefe´, individuals live long enough to reproduce and Amazonas, Brazil survivors take longer to reach reproductive age. (3) Arapaima populations can sustain annual catches of L. Castello Á D. J. Stewart Department of Environmental and Forest Biology, up to 25% of the number of adults in the population College of Environmental Science and Forestry, State the previous year if minimum size (1.5 m) and closed University of New York, Syracuse, NY 13210, USA season (December–May) limits are met. (4) When 25% of the number of adults in the population the L. Castello (&) The Woods Hole Research Center, Falmouth, previous year is harvested under a 1.6 m minimum MA 02540, USA size limit of catch, catches are slightly smaller but e-mail: [email protected] abundance of adults in the population is considerably greater than under a 1.5 m limit. These findings can C. C. Arantes Departamento de Cieˆncias Exatas e Tecnolo´gicas, be used in ongoing management initiatives, but Universidade Estadual de Santa Cruz, Ilhe´us, Bahia, caution is needed because of present biological and Brazil ecological uncertainty about these fishes. C. C. Arantes Instituto de Pesquisa Ambiental da Amazoˆnia, Santare´m, Keywords Brazil Á Fishing selectivity Á Floodplain Á Brazil Life-history Á Management Á Recruitment 123 Rev Fish Biol Fisheries Introduction and increased vulnerability to fishing activity (e.g., by stranding in pools; Brinkmann and Santos 1973; Fisheries in tropical river-floodplains are an impor- Halls et al. 2000; Junk 1985). tant natural resource use activity. But they are under Here, we studied the population dynamics of a increasing exploitation pressure, and most of them floodplain fish that has characteristics mainly of are being degraded rapidly, threatening biodiversity ‘‘equilibrium’’ fishes but also shares similarities with and human food security (Allan et al. 2005). Much- fishes in the ‘‘periodic’’ strategy of Winemiller and needed policy action has been hampered so far, in Rose (1992). ‘‘Periodic’’ fishes delay maturation in part because of lack of fisheries data and poor order to attain a size sufficient for production of a understanding of the dynamics of floodplain fish large clutch and to allow adult survival during populations (Bayley 1988; Bayley and Petrere 1989; periods of suboptimal environmental conditions Halls et al. 2001). We can improve our understand- (e.g., winters, dry seasons, reduced food availability). ing of these fisheries by focusing scarce scientific Arapaima (Arapaima sp.) are among the fishes of efforts on fish resources with distinct life-history greatest conservation concern (Allan et al. 2005). strategies. Arapaima gigas (Shinz) was enlisted in the IUCN Most of the existing understanding on tropical fish Red List as ‘‘vulnerable’’ in 1986, but due to lack of population dynamics in river-floodplains stems from data it now is listed as ‘‘data deficient’’ (World fishes in the ‘‘opportunistic’’ and ‘‘equilibrium’’ life- Conservation Monitoring Centre 1996). That species history strategies of Winemiller and Rose (1992). The is listed in Appendix II of the Convention on first comprehensive study of floodplain fish popula- International Trade in Endangered Species of Wild tions, by Welcomme and Hagborg (1977), was based Fauna and Flora (Castello and Stewart 2010). Others on a composite of various tilapia species in Africa, have considered the arapaima present in our study and apparently with emphasis on species of the area to be A. gigas, but recent data suggest that it may mouth-brooding genus Oreochromis Gu¨nther, which be an undescribed species (DJ Stewart unpublished would fit within the typical ‘‘equilibrium’’ life-history data). strategy of Winemiller and Rose (1992). ‘‘Equillib- Arapaima have constituted historically important rium’’ fishes are more like the traditional K-strate- food fisheries in the Amazon Basin (Verı´ssimo 1895), gists, characterized by large eggs and parental care and they likely play key ecosystem roles because they resulting in the production of relatively small are top predators (Queiroz 2000). Government pol- clutches of relatively large-bodied juveniles. More icies of minimum size of catch (1.5 m total length, recently, a study was done by Halls et al. (2001)on TL) and closed season (December–May), and even population dynamics of the pool barb, Puntius total moratoria in three states in Brazil, have been sophore (Hamilton) from floodplains in Bangladesh, established, but they have been poorly enforced a fish that probably fits in the ‘‘opportunistic’’ (Castello et al. 2009; Castello and Stewart 2010). strategy of Winemiller and Rose (1992). ‘‘Opportu- With most of the catch being in violation of nistic’’ fishes tend to have large intrinsic rate of management policies, fishing of arapaima, which is population increase due to early sexual maturity, done through harpooning and gillnetting, is now frequent reproduction, small eggs, and rapid larval believed to have led wild populations to become growth. overexploited and even declining throughout much of These studies have shown the major distinction of the basin (see review in Castello and Stewart 2010). floodplain fish population dynamics, which is the Most populations in the wild now are limited to local influence of flooding cycles. High water levels tend to community-based management areas (Castello et al. promote their body growth and reproduction as the 2011). flooded plains typically offer abundance of food and Population recoveries of arapaima have occurred low predator density (Agostinho et al. 2004; Dudley in a management scheme developed recently based 1974; Gomes and Agostinho 1997; Nikolsky 1956; on annual population counts done at the moment of Welcomme 1979). Low water levels, on the other the species’ aerial breathing. The counts include two hand, tend to promote their mortality through dete- size classes: Juvenile (with capital ‘j’, 1–1.5 m TL) rioration of environmental conditions (e.g., anoxia) and Adult (with capital ‘a’, [1.5 m TL), and they can 123 Rev Fish Biol Fisheries be very accurate, varying by about 20% around the arapaima fishing (0.3 kg ha-1 year-1;Sa´nchez 1969) true value (Arantes et al. 2007; Castello 2004). In the and some information on their biology and ecology. management scheme, annual population counts done The arapaima in Jaraua´ have been found to migrate by local fishers in 1 year (before they harvest) are among all floodplain habitats every year (Castello used to determine catch quotas for the following year. 2008a, b). They inhabit lakes, rivers and connecting The potential effectiveness of this management channels during low water levels, at which time the scheme is evident in Jaraua´, an area within the adults form pairs and subsequently reproduce Mamiraua´ Reserve (Amazonas State, Brazil), where (Queiroz 2000). The adults build their nests in the the arapaima population was overexploited and under margins of water bodies during rising water levels, risk of collapse (Queiroz and Sardinha 1999). and there are reports that they exhibit territorial behavior (Fontanele 1948). Nest building, hatching of 1. Compliance with the minimum size policy has the young, and parental care all take about 3 months increased mean TL of the captured arapaima by (Fontanele 1948; Isaac et al. 1993), and in Jaraua´ 53 cm, from 108 cm before the management there is a one-to-one sex ratio (Queiroz and Sardinha scheme (1993–1998), to 161 cm after the man- 1999). As water levels decline, they migrate back to agement scheme was implemented (Fig. 1a). connecting channels and lakes, where they are 2. The combined population of Juveniles and generally fished. Adults of arapaima, and associated catch quotas, Body growth of arapaima in Jaraua´ has been found have increased several-fold (Fig. 1b; Castello not to follow density-compensatory theory (e.g., Rose et al. 2009; Viana et al. 2004, 2007). et al. 2001), which is not unusual in Amazonian river- These results in Jaraua´ have been largely respon- floodplains (Bayley 1988). Arapaima in Jaraua´ grow sible for the following wider implications. to 70–90 cm TL in 1 year, and mature at about 160 cm in 3–5 years (Arantes et al. 2010). But they 3. Fishing of arapaima became permitted in the were found to grow faster (by an average of 27 cm in Brazilian States of Amazonas and Acre (in 2004 median length-at-age) in the high population densi- and 2008, respectively, encompassing *26% of ties now observed in Jaraua´ than in the low popula- the Amazon basin) only to fishers that present tion densities of the un-managed situation (Fig.