Several factors can cause short-term changes in estuarine environments and have effects on resident organisms. Among the most prominent of these are tidal fluctuations and wind events. Tidal currents in most estuaries are the primary source of energy and are therefore responsible formost short-term variations in temperature, salinity, and other habitat characteristics. However, microtidal estuaries such as those common in the northern Gulf of Mexico experience relatively low tidal currents. These habitats depend greatly upon wind events for energy input and water column mixing (Demers et al. 1987; Reed 1989; Arfi et al. 1993; Brooks and Doyle 1998; Perez et al. 2000; Chen and Valle-Levinson 2009).

Mobile Bay, Alabama, is a large (1,059-km²), shallow estuary in the northern Gulf of Mexico with a mean depth of 3 m and a mixed-semidiurnal microtidal range of only 0.4 m/d (Turner et al. 1987). However, the combination of high riverine discharge ( = 2,246 m³/s) and low winds can create stratification (Turner et al. 1987). Because tidal currents in Mobile Bay are not sufficiently strong to destratify the water column, wind events constitute the predominant mixing mechanism and are typically common enough to prevent long-term stratification in Mobile Bay throughout most of the year (DiMego et al. 1976; Turner et al. 1987; Schroeder et al. 1990; Cowan et al. 1996).

In addition to destratifying estuarine water columns, wind events frequently resuspend bottom sediments in shallow estuaries (de Jonge and van Beusekom 1995). Most estuaries along the northern Gulf of Mexico experience frequent wind-driven sediment resuspension (Denes and Caffrey 1988; Miller et al. 2005). Depending upon water depth, even low wind speeds (<4 m/s) can resuspend silt, mud, and small sand particles. Higher winds (≥4 m/s), characteristic of thunderstorms, cold fronts, and tropical cyclones, can suspend larger particles for longer periods of time (Schoellhamer 1995; Brooks and Doyle 1998; Booth et al. 2000), and cold fronts are the most frequent of such events in Mobile Bay during the fall, winter, and early spring (Miller et al. 2005). In a study of Barataria Bay in southeastern Lousiana, Booth et al. (2000) reported that between 63% and 80% of bottom sediments become resuspended during the passage of cold fronts when winds range from 4 to 10 m/s and that wind velocities ≥4 m/s occurred more than 80% of the time during the late fall, winter, and early spring. Kobashi et al. (2006) reported similar results on a shoal in the shallow Gulf of Mexico and showed that sediments become resuspended within a few hours of frontal passage and settle back to the seafloor just as quickly once winds fall below 2 m/s.

Wind-driven resuspension events introduce small organisms and organic seston into the water column that can be utilized by grazers and suspension and deposit feeders (Bell and Sherman 1980; Frechette and Grant 1991; Grant et al. 1997; Gremare et al. 1997). While a few studies have examined the relationship between wind events and production among benthic and planktonic organisms (Frechette and Grant 1991; McKinnon and Ayukai 1996; Smaal and Haas 1997; Lawrence et al. 2004), the influence of high-velocity wind on the diets and potential growth of estuarine fishes (Boisclair and Leggett 1989a, 1989b; Hewett and Kraft 1993) has been largely neglected, although the effects of turbidity on feeding by larval and juvenile fish has been studied both in laboratory and field settings (Auld and Schubel 1978; Cyrus and Blaber 1987a, 1987b; Grecay and Targett 1996; Abrahams and Kattenfeld 1997; Fiksen et al. 2002; Gadomski and Parsley 2005; Meager et al. 2005).

This study examined short-term variations in estuarine fish feeding and caloric intake during high-wind events in Mobile Bay. Mobile Bay is a warm-temperate estuary in the northern Gulf of Mexico. In total, 337 species of fresh and saltwater fish occur in the Bay and the delta at the Bay's head (Boschung 1992). Among these is the Bay Anchovy Anchoa mitchilli, which occurs from the Gulf of Maine and Cape Cod, Massachusetts, south to Yucatan, Mexico, and throughout the Gulf of Mexico (Hoese and Moore 1977; Fives et al. 1986; Robbins et al. 1986). It is a common and often extremely abundant fish in some coastal and inshore waters of the western Atlantic Ocean (Hoese and Moore 1977) and in estuaries in the northern Gulf of Mexico. They are small (∼10 cm maximum), short-lived (3 years maximum), opportunistic life history strategists (Winemiller and Rose 1992) that are often numerical dominants in estuaries where they occur (Castillo-Rivera et al. 1994; Szedlmayer and Able 1996; Rilling and Houde 1999), including in Mobile Bay (Hoese and Moore 1977). The Atlantic Croaker Micropogonias undulatus ranges in the western Atlantic Ocean from Massachusetts and in the Gulf of Mexico to northern Mexico. It is an estuarine-dependent species that is among the most abundant fishes in northern Gulf of Mexico estuaries. It is a relatively small (>50 cm), short-lived (5 years maximum), early maturing (2 years), periodic life history strategist (Winemiller and Rose 1992) that usually occurs over mud and sandy-mud bottoms in coastal waters and in estuaries where the nursery and feeding grounds are located (Hoese and Moore 1977; Smith 1997). These two species were chosen for study both because they are abundant in Mobile Bay and elsewhere and because they occupy different feeding niches; Bay Anchovy are particulate planktivores feeding mostly upon calanoid copepods and Atlantic Croaker are benthic omnivores that feed on worms, crustaceans, and larval fish (Jackson 2004). During this study, Bay Anchovy and age-0 Atlantic Croaker were sampled during a period of frequent (every 4–7 d; Miller et al. 2005) springtime cold front passages.

METHODS

Adult Bay Anchovy and juvenile Atlantic Croaker were collected over muddy substrate in southwestern Mobile Bay (Figure 1) during twice-weekly cruises in a 6.5-m outboard motor boat from 28 February to 25 March 2002, a time of frequent cold front passage.Water depth was ≤3 m. Wind measurements were obtained from the National Data Buoy Center, Station DPIA1 (30.248N 88.073W), on the extreme eastern end of Dauphin Island, Alabama. Wind velocity and direction were recorded once every 10 minutes for the duration of the study. Mean wind velocity was calculated for the 6 hours previous to sampling and represented a wind velocity estimate for each cruise. Additionally, a plot of continuous wind velocity from February 25 to March 25 was created to demonstrate trends in wind velocity for the entire study period (Figure 2). Sampling cruises that coincided with cold front passage experienced wind velocities ≥4 m/s. Therefore, wind velocity was divided into two categories for some analyses: less than 4 m/s (quiescent) and ≥4 m/s (windy).

Water temperature (°C), salinity (psu), dissolved oxygen (DO) were measured before and after each sampling event using a YSI Model 85 DO meter. Light penetration (cm) was measured using a Li-Cor (LI-192 quantum light sensor) submersible light meter. The depth in the water column to which light penetrates is determined by the extent to which it is absorbed and scattered by dissolved compounds and suspended particles contained within the water (Kirk 1983; Oliver 1990; Kirk and Oliver 1995). In general, light intensity declines exponentially with depth as described by the Beer-Lambert equation:

FIGURE 1.

Map of Mobile Bay, Alabama. The sample location was visited twice weekly from 28 February to 25 March 2002.

FIGURE 2.

Plot of continuous wind velocity from 27 February to 26 March 2002. Wind velocity on each of the eight sampling dates is indicated by a series of eight dots. The dashed line denotes a wind velocity of 4 m/s, which represented the division between quiescent and windy days in this study.

This study was part of a larger effort that concurrently collected nutrient, primary production, and bacterial production data. Collection and analyses of these data were performed by other members of our study group. A 5-L Niskin bottle was used to collect water from the top and bottom of the water column for analysis of nitrate (NO3), ammonium (NH4), phosphate (PO4), and dissolved organic carbon (DOC). Water for analysis of chlorophyll-a concentrations was decanted from the bottle and placed immediately on ice in a closed (dark) cooler. Water for analysis of bacterial production was collected in a 20-L carboy that was placed immediately in a cooler of ambient seawater for transport to the laboratory. Resulting data were provided to us after analyses were complete.

Fish were collected using a 3.9-m otter trawl with 12.7-mm mesh and 1.6-mm mesh cod end liner. Two replicate 5 or 10 min tows were conducted during each cruise. Trawl distance was measured with a mechanical flowmeter. All captured organisms from each replicate trawl were sorted by species, enumerated, and weighed, and up to 50 BayAnchovy and 50 Atlantic Croaker were randomly selected for analyses. These fish were preserved in a 70% isopropyl and 2% formalin solution until they could be analyzed.

Standard length (mm) and mass (grams dry weight) were obtained for individuals in each 50-fish subsample. Density estimates (fish per 10 m²) and biomass (grams dry weight per 10 m²) were calculated for both species from each tow. Density and biomass parameters were log10 transformed to reduce heteroscedasticity before testing for significance. F-tests (α = 0.05) were used to determine if there was a statistically significant difference in density and biomass between quiescent and windy days.

Up to 20 Bay Anchovy and 20 Atlantic Croaker were randomly selected from each subsample for stomach content analysis. Each stomach was removed and laid open with a scalpel. Contents were extracted and placed on a labeled microscope slide. When necessary, the stomach lining was rinsed to ensure that all microscopic contents were completely removed.

Prey items were identified under a dissecting microscope to the lowest possible taxonomic level and sorted. All items belonging to the same taxonomic group were enumerated and collectively placed into preweighed containers. Prey items were dried at 80°C until a constant weight was achieved. Dry weights were recorded to the nearest 0.001 mg using a microbalance.

Stomach fullness was calculated using the following formula (Duarte and Garcia 1999):

j = stomach;

PW_j = cumulative dry weight of prey in stomach j;

FW_j = dry weight of fish j;

P/F_max = greatest ratio of prey weight to fish weight.

Fullness data were normalized by arcsine transformation. F-tests (α = 0.05) were applied to determine if there was a statistically significant difference between quiescent and windy days in the stomach fullness of Bay Anchovy and Atlantic Croaker.

Three indices were used to quantify prey importance in the diets of Bay Anchovy and Atlantic Croaker. These indices were percent composition by dry weight (dw), the index of prey importance (PI), and the index of caloric importance (ICI).

Percent dw was determined for each prey category using the equation (Bowen 1996)

i = prey category;

P = number of stomachs with prey;

W_i = dw (g) of prey category;

Q = number of prey categories.

The PI was determined for each prey category using the following equation (Pope et al. 2001):

The ICI was determined for each prey category using the following equation (McCawley and Cowan 2007):

Analysis of similarity (ANOSIM), a nonparametric multivariate test, conducted at a significance level of 0.05, was used to determine whether there was a statistically significant dissimilarity between quiescent and windy days in the diets of Bay Anchovy and Atlantic Croaker. The ANOSIM was conducted using the PRIMER-6 software package based upon percent dw of all prey categories in individual stomachs. The data were log10 (x + 1) transformed to account for the zero values of prey items that may have been absent in individual stomachs. The Bray-Curtis coefficient was used to describe diet similarity between all possible pairs of stomachs in a similarity matrix. The ANOSIM then performed nonparametric permutations on the similarity matrix and produced an R-statistic that described the extent of dissimilarity in diet composition between quiescent and windy days (Clarke and Gorley 2001).

The caloric content of each prey was either obtained from literature or estimated based upon the referenced caloric value of taxonomically similar prey (Table 1). Caloric intake was defined as calories per gram of fish dw (Cj) for each fish that was analyzed and was determined by multiplying the estimated caloric content of each prey item by the dry mass of that prey in an individual stomach. Calories were then summed for all prey in that stomach and divided by the dry mass of the respective fish as described in the following formula:

Estimates of stomach fullness and caloric intake for both species were regressed (linear) on wind speed to determine if variability in either of these two metrics of food intake could be explained by changes in wind velocity (linear regression; α = 0.05; Proc Reg in SAS; SAS Institute 2006).

RESULTS

Three cold fronts passed over Mobile Bay during this study. Low winds (<4 m/s; quiescent days) were experienced on the sample dates of February 28, March 7, March 14, and March 18 and were associated with quiescent periods between fronts. High winds (≥4; windy days) associated with these fronts were experienced on the sample dates of March 4, March 11, March 21, and March 25. The highest mean wind velocity experienced during a cruise was 10.6m/s on March 21 while the lowestmean wind velocity experienced during a cruise was only 3 d prior (1.4 m/s on March 18; Figure 2), demonstrating how quickly conditions can change in response to cold front passages.

Although temperature slowly increased over the duration of the study, cold front passage did not significantly affect water temperature, salinity, dissolved oxygen, or any of the measured chemical and biological variables—NO3, NH4, PO4, DOC, chlorophyll a, and bacterial production (Kiene et al. 2003). However, light penetration to 0.5 m above the seafloor, which is affected by the quantity of suspended sediments and seston material in the water column, was significantly lower on windy days (55.0 cm) than on quiescent days (142.5 cm), which represents an increase of more than 19.6 ± 4.2 NTU in turbidity between windy and quiescent days.

TABLE 1.

Estimated caloric densities of the prey items found in the present study.Most estimates were obtained from the literature, although some estimates were approximated based on the cited caloric density of taxonomically similar prey.

Bay Anchovy ranged from 16.5 to 66.9 mm SL; mean length was 45.6 mm SL. Mean length was not significantly different between quiescent (mean ± SD, 44.6 ± 6.1 mm SL) and windy days (46.0 ± 6.7 mm SL; ANOVA: F_{1, 348} = 1.42, P = 0.234). Age-0 Atlantic Croaker ranged from 13.2 to 75.1 mm SL; mean length was 31.9 mm SL. Mean length was not significantly different between quiescent (30.9 ± 15.5 mm SL) and windy days (32.5 ± 14.7 mm SL; ANOVA: F_{1, 402} = 1.07, P = 0.302).

A total of 752 Bay Anchovy were captured during the course of this study. Of those, 713 were captured on windy days and only 39 (5%) were captured on quiescent days. Mean density was significantly greater on windy days (1.46 ± 1.09 fish/10 m²) than on quiescent days (0.09 ± 0.09 fish/10 m²; ANOVA: F_{1, 8} = 6.27, P = 0.046). Mean biomass of Bay Anchovy was also significantly greater on windy days (0.45 ± 0.38 g dw/10 m²) than on quiescent days (0.04 ± 0.05 g dw/10 m²; ANOVA: F_{1, 8} = 7.44, P = 0.034).

A total of 697 Atlantic Croaker were captured in this study. Of those, 538 were captured on windy days and 159 (23%) were captured on quiescent days. Mean density was four-fold greater on windy days (1.61 ± 1.63 fish/10 m²) than on quiescent days (0.37 ± 0.33 fish/10 m²), although this difference was not significant (ANOVA: F_{1, 8} = 2.18, P = 0.191). The mean biomass of Atlantic Croaker was greater on windy days (0.51 ± 0.57 g dw/10 m²) than on quiescent days (0.10 ± 0.07 g dw/10 m²), but this difference also was not significant (ANOVA: F_{1, 8} = 2.03, P = 0.204).

The decreased capture rate of Bay Anchovy on quiescent days resulted in many fewer stomachs being analyzed for quiescent days (n = 31) than for windy days (n = 134). However, these sample sizes were sufficient for analysis. Bay Anchovy primarily consumed calanoid copepods during both quiescent and windy days. They also ate ostracods and bivalve pelecypods in addition to occasionally preying upon barnacle nauplii, harpactacoid copepods, polychaete larvae, gammarid amphipods, euphausid shrimp, diatoms, nematodes, and crab zoea. On windy days, Bay Anchovy consumed large quantities of detritus (Table 2). Mean fractional stomach fullness declined significantly with wind speed, albeit the data are noisy (Figure 3; Fullness [fractional] = 0.233-0.0239 × Wind Speed [m/s]; R² = 0.18; P < 0.0001). There was no significant dissimilarity in diet by wind classification (ANOSIM: R=−0.12, P = 0.992), due primarily to a dietary dominance by calanoid copepods during both quiescent and windy days. However, there was a considerable increase in the contribution of detritus to diets on windy days (Figure 4). Mean caloric intake (cal/g dw) by Bay Anchovy also declined significantly with wind speed, but the data were highly variable among individual anchovy (Figure 3; Caloric Intake=20.17 [cal/g dw] - 2.124 × Wind Speed [m/s]; R² = 0.16; P < 0.0001).

In contrast to Bay Anchovy, the number of Atlantic Croaker stomachs available for analysis on quiescent days (n = 81) and windy days (n = 107) were similar. Age-0 Atlantic Croaker primarily consumed calanoid copepods, mysid shrimp, and detritus. They also ate polychaetes, gammarid amphipods, harpactacoid larvae, ostracods, and clupeiform larvae in addition to occasionally preying upon barnacle nauplii, euphausid shrimp, crab megalopae, pelecypods, and foraminiferans (Table 3). Stomach fullness increased slightly with increasing wind speeds, but this difference was not significant (Figure 5; R² = 0.07; P > 0.05). Diets were significantly dissimilar between wind classifications (ANOSIM: R = 0.17, P = 0.001), principally due to the varying dietary importance of calanoid copepods, mysid shrimp, and detritus between quiescent and windy days (Table 3; Figure 6). Caloric intake (cal/g dw) increased with increasing wind speed (m/s), but the data are variable among individual Atlantic Croaker stomachs (Figure 5; Caloric Intake = 7.82 (cal/g dw) + 2.46 × Wind Speed; R² = 0.05; P = 0.051). That said, mean caloric intake was twice as high on windy days (23.9 ± 29.9 cal/g dw) than on quiescent days (12.3 ± 11.9 cal/g dw).

TABLE 2.

Percent dry weight (DW), index of caloric importance (ICI), and index of prey importance (PI) for stomach contents of Bay Anchovy collected on quiescent (wind <4 m/s) and windy (wind ≥4 m/s) days in Mobile Bay during resuspension cruises in spring 2002.

FIGURE 3.

Regression relationship for Bay Anchovy between stomach fullness (fractional) and wind speed (m/s) (top panel) and the regression of caloric intake (cal/g dw) on wind speed (bottom panel).

FIGURE 4.

Bubble volume denotes mean percent dry weight of prey in the diets of Bay Anchovy captured on quiescent (wind < 4 m/s) and windy (wind ≥4 m/s) days.

DISCUSSION

Both species examined in this study appeared to show shifts in density and biomass between windy and quiescent sampling days. One possible cause of this result is gear avoidance. Several trawling studies have reported catch increases in trawls during nighttime hours that are thought to occur because of reduced visual detection of the trawl (Walsh 1988; Casey and Myers 1998; Hjellvik et al. 2002). However, it is also possible that the fishes we studied changed their distribution in response to frontal passages. Our sampling efforts were constrained because of the size of the vessel (6.7 m) that was available for our use, given the sea states in Mobile Bay when wind velocities are high.

Bay Anchovy had fuller stomachs on quiescent days than on windy days but diet composition was not significantly dissimilar by wind category, as calanoid copepods dominated Bay Anchovy diets under both wind conditions. However, there was an increase in the mean contribution of detritus to Bay Anchovy diets on windy days (Figure 4). Calanoid copepods, while small (0.02 g per adult copepod), have an estimated caloric density of 6,626 cal/g dw, whereas detritus has a considerably lower caloric density of 2,229 cal/g dw (Table 1). Increased consumption of low-quality food has also been observed for shellfish during wind-driven resuspension events. Although high wind velocity increased seston availablilty for shellfish feeding, the fraction of quality seston was negatively correlated with wind velocity (Frechette and Grant 1991; Grant et al. 1997; Smaal and Haas 1997). In the present study, relatively empty stomachs and increased ingestion of low-energy food resulted in significantly lower mean caloric intake with increasing wind speed.

TABLE 3.

Percent dry weight (DW), index of caloric importance (ICI), and index of prey importance (PI) for stomach contents of Atlantic Croaker collected on quiescent (wind < 4 m/s) and windy (wind ≥4 m/s) days in Mobile Bay during resuspension cruises in spring 2002.

This scenario implies that periods of high wind velocity (>4 m/s) in Mobile Bay could have an adverse effect upon energy uptake by resident Bay Anchovy. Such energy deficits might accumulate during seasons of frequent frontal passage, or other wind events, and ultimately have a negative impact on the growth rate (Boisclair and Leggett 1989a, 1989b; Hewett and Kraft 1993) and subsequent biomass production of Bay Anchovy in Mobile Bay. An examination of historical wind data from the National Data Buoy Center Station DPIA1 revealed that during each of the four years between 2001 and 2004, wind velocity in Mobile Bay exceeded 4 m/s for at least 8 h on 81, 66, 76, and 70 d, respectively. This indicates that on 18% to 23% of the days in a year conditions existed that could have a negative impact on Bay Anchovy growth and production at vulnerable locations in Mobile Bay. During late fall (November), winter, and early spring (March), the percentage of days that winds exceeded 4 m/s was much higher, averaging ≥70%, and was similar to the results reported by Booth et al. (2000). We believe Bay Anchovy in warm-temperate and subtropical estuaries may balance capital breeding, in which compensatory feeding takes place in advance of breeding so that reproduction initially may be financed from stored energetic capital, with income breeding, in which reproduction is financed using current energetic income during long spawning seasons in which females spawn successively over many days (Thomas 1988; Stearns 1992). If this is true, it is during the early spring period that overwintering females must obtain sufficient nutrition to prepare for spawning once waters in the estuary reach 20°C, which occurs in late April in Mobile Bay (Leak and Houde 1987; Zastrow et al. 1991; Peebles et al. 1996). These results are compelling if one considers that Bay Anchovy stomachs examined during 12 monthly sampling cruises in Mobile Bay under quiescent conditions contained little or no detritus (Jackson 2004).

Atlantic Croaker stomach fullness did not differ significantly with wind speed, although there was a weak trend toward fuller stomachs on windy days. However, Atlantic Croaker diets were significantly dissimilar between quiescent and windy days. This dissimilarity was driven by less reliance on calanoid copepods and detritus, coupled with increased dietary importance of larger mysid shrimp (Mysidopsis spp.; mean individual mass = 0.35 g dw), during windy days. As mentioned, calanoid copepods have a much higher caloric density than detritus. Additionally, mysid shrimp have a caloric value of 7,533 cal/g dw, which was the highest caloric value of any prey identified in this study (Table 1). Despite the dissimilarities in diet, caloric intake by Atlantic Croaker did not differ significantly with wind speed, but this result was likely due to high variance in the data because the mean caloric intake on windy days was twice as high as it was on quiescent days. The variance was caused by Atlantic Croaker feeding habits, particularly with regard to mysid shrimp. When they ate mysid shrimp, Atlantic Croaker either ate them in large quantities or hardly at all. We suspect that consumption could be enhanced either (1) by uncovering and exposing burrowing mysids to foraging Atlantic Croaker or (2) by increased turbidity that makes mysids more susceptible to predation because low light levels similar to that experienced during nighttime elicits a foraging response by the predatory marine mysids, causing them to exit their burrows to feed (Lasenby and Langford 1973; Siegfried and Kopache 1980; Fulton 1982). Consequently, individual Atlantic Croaker with mysids in their stomachs consumed a diet of extremely high caloric density while those that did not feed on mysid shrimp had a considerably lower caloric intake.

FIGURE 5.

Regression relationship for Atlantic Croaker between stomach fullness (fractional) and wind speed (m/s) (top panel) and the regression of caloric intake (cal/g dw) on wind speed (bottom panel).

FIGURE 6.

Bubble volume denotes mean percent dry weight of prey in the diets of age-0 Atlantic Croaker captured on quiescent (wind < 4 m/s) and windy (wind ≥4 m/s) days.

This scenario implied opposite consequences for age-0 Atlantic Croaker, which are benthic omnivores, than it did for Bay Anchovy, which are particulate planktivores. Periods of high wind velocity (≥4 m/s) in Mobile Bay may potentially have a beneficial effect on energy uptake by Atlantic Croaker. Enhanced uptake could accumulate during seasons of frequent frontal passage, or other wind events, and ultimately have a net positive impact on the growth rate and production of age-0 Atlantic Croaker on their nursery grounds in Mobile Bay.