Abundance of Yellow-Phase American Eels in the Hudson River Estuary

—Fisheries for American eel Anguilla rostrata occur mostly in estuaries, yet eel abundance in large estuaries is poorly understood and the methods for estimating eel density underdeveloped. During 1997–1999, mark–recapture experiments were conducted for six consecutive days at six sites spanning the 250-km tidal portion of the Hudson River estuary, New York. Each experiment comprised 36 baited eel traps arrayed at 200-m intervals over 144-ha sampling sites. Estimates of local density were complicated by eel behavior, including trap-shy responses to marking and immigration into the experimental grid in response to bait attraction. We compared two open-population models, both modified Peterson methods: Jolly–Seber and a model created to account for eel behavior termed the mean recapture model (MRM). The biases in model outputs in response to trap-shy behavior and immigration were analyzed through simulations; the MRM showed less bias (213%) than the Jolly–Seber model (136%). Density estimates for the sampled regions ranged from 2 to 18 eels/ha for MRM and from 3 to 24 eels/ha for the Jolly–Seber model. The lowest density (1.6 eels/ha) was estimated for Albany (river km 240), but all other sites were estimated to have similar densities (5–18 eels/ha). The mean local density in the Hudson River estuary, 9.5 eels/ha, was much lower than those estimated for other systems. An overall abundance of 118,000 was calculated for Hudson River estuary eels larger than 30 cm (total length) at depths of 2–10 m. American eels Anguilla rostrata are among the most ubiquitous of North American ichthyofauna and support important fisheries throughout their range. They inhabit diverse habitats, including salinities from freshwater to oceanic water; water bodies such as lagoons, marshes, swamps, lakes, streams, and large rivers; and latitudes from Venezuela to Greenland. Despite the fact that a large proportion of the American eel fishery takes place in large rivers and estuaries (ICES 2001), little information exists about American eel population dynamics for these habitats, particularly how density varies both within and among estuaries. It is important to understand how a species reacts in response to exploitation, but it is equally important to look at the characteristics of a species in the absence of exploitation. In this study, we evaluate methods for estimating the density of American eels in the Hudson River estuary, a large middle-Atlantic estuary. The * Corresponding author: [email protected] Received July 1, 2002; accepted December 16, 2003 system is amenable to the study of eel density because it is a linear basin with well-defined gradients of salinity and depth. Unlike other middleAtlantic estuaries (e.g., the Delaware and Chesapeake bays), the Hudson River has a deep, fjordlike bathymetry that is the result of glaciation (Paul 2001). Contamination of the sediment and fauna of the Hudson River by polychlorinated biphenyls led to a ban on harvesting of American eels in 1976 that remains in place today. The closure provides a unique opportunity to study eel stock dynamics in the absence of the major exploitation that occurs in large estuaries elsewhere. The density of yellow-phase (subadult) American eels is highly variable among latitudes, watersheds, and habitat types (Table 1). In estuarine habitats, a principal hypothesis is that the growth, density, and productivity of yellow-phase eels are higher in downstream brackish habitats than in upstream freshwater locations (Helfman et al. 1987; Morrison and Secor 2003). Despite the recognized importance of estuaries as productive habitats and for eel fisheries, few studies have estimated eel density in estuaries (Helfman and Bozeman 1984; 897 YELLOW-PHASE AMERICAN EELS IN THE HUDSON RIVER TABLE 1.—Density estimates from the literature for yellow-phase American eels, along with fishing method and sizes of eels captured. Study Location Fishing method Density (eels/ha) Total length (cm) Bozeman et al. 1985 Ford and Mercer 1986 Oliveira and McCleave 2000 Oliveira 1997 LaBar and Facey 1983 This study Georgia tidal creek Massachusetts tidal creek Maine freshwater rivers Rhode Island freshwater river Vermont lake Hudson River estuary Pots Traps Electrofishing Electrofishing Electrofishing Pots 182–232 875 800–2,200 450–3,230 232–636 1–30 20–80 15–63 .10 16–74 Not stated 28–67 Bozeman et al. 1985). No study that we are aware of has compared eel density among regions within a major estuary. The lack of synoptic surveys of yellow-phase American eel density in major estuarine systems is due in part to the difficulties associated with sampling in these areas. In particular, no single gear can effectively sample all of the salinity zones, depths, and structured bottom types inherent to large estuaries. Electrofishing is not feasible in brackish water; seines, bottom trawls, fyke nets, weirs, and other trap nets are size selective and cannot be deployed across all depths and bottom types. Here we evaluate potting or trapping as a relatively easy method for catching and monitoring eels in estuaries. However, American eels possess physiological and behavioral characteristics that influence their catchability in pots. Catch rates in pots and traps are influenced by soak time, bait quality, environmental variables (e.g., temperature, salinity, and current velocity), gear saturation and escapement, interand intraspecific interactions, and behaviors related to tidal or seasonal cycles (Miller 1990). Some of these effects can be minimized by standardizing the elements related to soak time, bait quality, gear saturation, and lunar or diurnal cycles. Of principal concern in developing pot indices of density are American eel behaviors related to homing and their keen olfaction (Tesch 1977). Home ranges in yellow-phase eels depend on habitat and vary from 1 ha in a Georgia tidal creek (Bozeman et al. 1985) to 65 ha in a Vermont lake (LaBar and Facey 1983). The yellow-phase eels we studied in the Hudson River also showed limited ranges; 88% of the recaptures (N 5 58) at a freshwater tidal site occurred within a 144-ha area 1 year after the eels were marked (Morrison and Secor 2003). Displaced eels use selective tidal stream transport to precisely home to the regions from which they were removed (Parker 1995). Although yellow-phase eels have small home ranges, it is not known whether or how far an eel will travel outside its home range to follow a scent. As eels tend to feed only once every 223 d (Moriarty 1978), the catchability of an individual could change across days. Mark–recapture models often require the assumption of homogeneous capture probabilities. There are two situations that can lead to heterogeneous capture probabilities. The first occurs when capture probability is constant over time for each individual but differs among individuals (referred to as ‘‘permanent heterogeneous capture’’). The animals with high capture probabilities are captured first and recaptured more easily than those with lower probabilities, thereby biasing the results. The second type of bias occurs when animals are temporarily influenced by the act of being captured. Here, individual capture probability varies through time depending on recent capture history (referred to as ‘‘temporary trap response’’). Individuals exhibiting such responses have been called ‘‘trap-happy’’ or ‘‘trap-shy’’ depending on whether being caught increases or decreases an individual’s probability of recapture (Pollock et al. 1990). In this study, we compared two open-population methods of estimating the relative density of potted American eels at six sites in the Hudson River estuary, New York, covering the entire tidal extent of the estuary (250 km; 0–20‰ salinity; Figure 1). The critical assumptions of mark–recapture models were evaluated, particularly those related to heterogeneous capture probabilities among individuals and the ingress of eels into the study site from outside areas. Finally, density estimates were extrapolated for the entire estuary to calculate an overall abundance for eels longer than 30 cm in the Hudson River estuary.