Abundance, Distribution and Diversity of Chesapeake Bay Fishes: Results from chesfims



Download 43.63 Kb.
Date01.02.2018
Size43.63 Kb.
#38633
REF:[UMCES]CBL 05-140
Abundance, Distribution and Diversity of Chesapeake Bay Fishes: Results from CHESFIMS

(Chesapeake Bay Fishery Independent Multispecies Fisheries Survey)



Thomas J. Miller1, ,J. A. Nye1 , K. L. Curti1, D. Loewensteiner1 , E.D. Houde1, M. C. Christman2 J. H. Volstad3, and A. F. Sharov4.
1. Chesapeake Biological Laboratory, University of Maryland Center for Environmental Science, P. O. Box 38, Solomons, MD 20688; 2. Biometric Program, Department of Animal and Avian Sciences, University of Maryland - College Park, College Park, MD; 3 Versar Corp, 9200 Rumsey Road, Columbia, MD 21405 and 4. Fisheries Service, Maryland Department of Naural Resources, Tawes Office Building, Taylor Avenue, Annapolis, MD 21401

ABSTRACT

In support of multispecies fisheries management, we have conducted fishery-independent surveys of Chesapeake Bay fishes (CHESFIMS). This survey builds on earlier work on Trophic Interactions in Estuarine Systems (TIES) by scientists from the University of Maryland Center for Environmental Science Chesapeake Biological Laboratory. In one component of our work fish are surveyed in spring, summer and autumn each year using an 18 m2 midwater trawl deployed according to a stratified random sampling design. Together with data from TIES these survey data provide a nine-year time series on the abundance and distribution of the benthopelagic fish community in Chesapeake Bay. A second component of our work is a trawl survey conducted in shoal habitats. Our objectives have been to assess the efficiency of alternative sampling designs and to quantify patterns in the abundance, distribution, trophic demand and ecological function of the fish species that comprise the Chesapeake Bay fishery ecosystem.

We have documented changes in abundance of key species over the ten-year time series. A central feature of our results is information on the abundance, distribution and production of bay anchovy, a heretofore unsurveyed species. We also have quantified the abundance and distribution of other species important to the Chesapeake Bay fishery ecosystem including Atlantic menhaden, Atlantic croaker, striped bass, white perch and weakfish. Some species, such as white perch, are more abundant earlier in the time series than later. Other species, such as croaker, show complementary patterns in that they have increased in abundance over time. We have quantifed growth and production of these species using otolith ageing and bioenergetic models. We have quantified the diet of key species. We find temporal, spatial and ontogenetic patterns in the data. Our statistical analyses have resulted in new statistical approaches to combining complemented survey designs and to the application of geostatistical tools to fishery independent data

INTRODUCTION
The Chesapeake 2000 (C2K) Agreement commits regional jurisdictions to implement multispecies approaches to fisheries management. The potential for biological interactions and technical interactions within traditional single species management has motivated the development of multispecies approaches. Houde et al. (1998) reported the recommendations of a workshop to explore the utility and advisability of adopting multispecies approaches in Chesapeake Bay. An important conclusion of the workshop was the development of coordinated, baywide surveys to estimate key species abundances and to provide biological data on both economically and ecologically important species that are currently lacking (Houde et al. op. cit.). Several fishery-independent surveys for the assessments of important fish and shellfish stocks in the Chesapeake Bay are currently ongoing but their study design and spatio-temporal coverage limits their applicability for exploring the multispecies question directly. The workshop recommended that these surveys should permit the estimation of the temporal and spatial dynamics of key predator-prey relationships and trophic interactions (Houde et al. op. cit.).

Since 2001, we have conducted research that seeks to provide information that will be needed to design a suitable baywide multispecies survey in support of C2K commitments. This research has several objectives that directly address issues relating to the design of future fishery-independent surveys. Our research focuses on providing estimates of the mean and variability in the distribution and abundance of fishes, estimates of trophic interactions as well as the utility of alternative survey designs and analytical frameworks. Here we report the results for the second year of research, provide comparisons with data from earlier periods and compare the utility and efficiency of alternative surveys designs and analytical approaches examined to date.



OBJECTIVES

Our Chesapeake Bay Fishery-Independent Multispecies Survey (CHESFIMS) builds on and expands earlier research, the NSF-funded Trophic Interactions in Estuarine Systems (TIES) program, into biological production potential and its temporal and spatial variability in the Chesapeake Bay ecosystem that had been conducted from 1995 – 2000. By using TIES as a foundation, CHESFIMS can be viewed as an ongoing decade long survey of the abundances and key trophic interactions in the Chesapeake Bay fish community. In 2004, we had four specific objectives:



  • Develop estimates of abundance, distribution and characteristics of the bentho-pelagic fish assemblage

  • Develop quantitative estimates of diet together with assessment of the size-dpendent, inter-annual, seasonal and regional variability in the diets of key species

  • Develop estimates of selectivity and relative power of different gear to permit integration of multiple surveys into a single unified assessment of abundance, distribution and characteristics of the fish assemblage in Chesapeake Bay

  • Develop multispecies and single species indices of the Chesapeake Bay fishery ecosystem with a comparison to existing single species alternatives when possible.

We completed successfully work and report progress on all four objectives.



METHODS



FIELD
The core of the CHESFIMS program is a bay-wide midwater trawl survey conducted during spring, summer and autumn months each year. The survey is conducted as a complemented fixed transect and stratified random design survey. The survey design has been unchanged for three years and involves sampling approximately 50 stations during each cruise. Twenty-nine fixed transect stations have been visited regularly since the inception of the CHESFIMS program in 2001. These stations had also been sampled during the six years prior to CHESFIMS (1995-2000) by the NSF-funded Trophic Interactions in Estuarine Systems (TIES) program. The fixed stations provide a reliable estimate of trends in the fish assemblage. In addition, since the inception of CHESFIMS an additional twenty-one stations have been added to the survey design, according to a stratified random allocation involving three salinity strata (upper bay, mid bay and lower bay). Strata were identified based on salinity, but operationally defined on latitude such that 38.668 < upper < 29.433; 37.8335 < mid < 38.667, and 36.997 < lower < 37.833. Stations were allocated according to estimated stratum volume. Figure 1 shows a typical station allocation for our Spring 2004 cruise.
FIGURE 1 NEAR HERE
Under prior NCBO support for CHESFIMS, we have developed the statistical estimators necessary for this complemented design (Volstad et al. in press).

Sampling on each survey is conducted during nighttime hours to minimize gear avoidance. Sampling is conducted using an 18m2 midwater trawl with 6 mm cod end mesh, that is fished in a single, stepped oblique tow that samples 10 equal depth intervals from the bottom to the surface. In this way the tow samples the entire water column. Comparison with other survey gear indicates that the midwater trawl effectively samples fish in the size range 30 – 300 mm TL. Thus, the CHESFIMS survey samples principally juvenile fish. Metering of the net provided information on net opening (height), depth and water temperature at the end of each tow. Net mensuration data were not available in real time, and thus deployments could not be adjusted during a tow to correct for deviations from the intend path. At each station a Seabird SBE 25 CTD was used to profile the water column and provided geo-referenced data on temperature, salinity, and fluorescence with depth.

The catch at each station was enumerated onboard the vessel. The total weight of the catch by species was recorded. Fish were identified to species, and the total weight of each species in the haul recorded. Individual fish were weighed and measured, sub-sampling when made necessary by the number of fish caught. Sub-samples of up to 100 were measured for length. A second sub-sample of the fish was immediately frozen for subsequent dietary, age and condition analysis in the laboratory. For selected species (Atlantic croaker Micropogonias undulates and striped bass Morone saxatilis), a sample of muscle tissue was taken and flash frozen in liquid nitrogen for analysis of nucleic acid levels.
LABORATORY
Size-at-age

Fish frozen in the field were defrosted and reweighed and measured. The otoliths were dissected from the cranium and prepared for ageing. Methods used for ageing varied among species, but generally thin sections (~1 mm) were cut on a South Bay Technologies Model 650 low speed diamond wheel saw. Sections were mounted to microscope slides using thermoplastic cement and viewed under a dissecting microscope.

Fish for ageing were selected at random in proportion to the length frequency composition. These data were then used to generate age-length keys for selected species that was subsequently used to infer the age structure in the entire Chesapeake Bay.
Dietary Analysis
Fish frozen in the field were defrosted for dietary analysis. In the laboratory, total length and wet weight were recorded for each fish. Stomachs, defined here as extending from the posterior of the esophagus to the pylorus, were removed and placed in ethanol. Prior to removal of the stomach contents, full stomachs were blotted dry and weighed to obtain a full stomach weight. Stomach contents were then removed and the stomach was subsequently re-weighed to obtain an empty stomach weight. The difference between these two weights represented an estimate of the total weight of prey in the stomach. Stomachs were scored for the presence/ absence of food. A feeding incidence of 1.0 indicated the presence of food in the stomach. Stomach contents were sorted and identified to the lowest practical taxonomic level under a dissecting microscope. Individual items comprising each prey group were blotted dry and weighed to obtain an estimate of the total weight of that prey type in the stomach. For some prey taxa, the size of individual prey items was also determined using an ocular micrometer.
Statistical Analysis

The relative abundance of fish in each survey was calculated using the approach of Volstad et al. (in press), which uses optimal weightings of CPUEs from the fixed transect and the stratified random designs. We calculated CPUE for each region (upper, mid and lower Bay) and bay wide. For selected species, we followed Jung and Houde’s (2004) approach for bay anchovy (Anchoa mitchelli) to estimate absolute abundance. Briefly, absolutebay-wide anchovy spawner biomasses in June–August were estimated for 1995−2000 according to an egg production method (Rilling and Houde 1999). Comparison of the baywide estimates of anchovy spawner biomass in June−August based on the egg production method (“absolute” biomass) with estimates based on the MWT catch-per-unit-of-effort (“relative” biomass) indicated that, on average, for 1995 to 2000, s is equal to 0.20. This value was then applied to other species in the size range 30mm < TL < 250 mm.

We analyzed distributional patterns based on relative CPUE data. Following Jensen and Miller (in press), robust empirical variograms (Cressie 1993) were calculated in SAS v8.3 (VARIOGRAM procedure, SAS Institute, Inc., Cary, North Carolina). Variogram estimation was limited to stations separated by up to 40 km with a lag size of 250 m. Spherical, exponential, and Gaussian variogram models were fit to the empirical variogram (SAS v8.3, NLIN procedure), and the best fitting model was chosen except in cases where one variogram model resulted in unrealistic variogram parameters (such as a negative nugget) or failed to converge. Following variogram selection, ordinary kriging was conducted (SAS v8.3, KRIGE2D procedure) with a kriging neighborhood of the 10 nearest sample points. Fish density was mapped at a 1 km grid scale by adding the kriged predictions (residuals) to the trend at the center point of every mapped grid cell. Kriging variance maps were also created. Mapping was conducted in ArcView v8.3 (ESRI Corp., Redlands, California).

We also explored the environmental parameters responsible for the distribution of individual species by developing a two-stage Generalized Additive Model (GAM). Here, we present the results for hogchoker (Trinectes maculates). Full details of the approach are given in Curti (2005). Briefly, the first stage of the GAM model predicted presence of hogchoker at a station whereas the second stage predicts abundance of hogchoker given presence. By uncoupling the two processes, the restrictive statistical assumptions that characterized earlier analyses are relaxed. Specifically, the analysis accounted for both the large number of hauls in which no hogchoker were caught (zero-inflation) and the potentially complex, non-normal responses of hogchoker to environmental parameters. Due to their influence on growth, metabolism and distribution, average bottom salinity, temperature, oxygen and water depth were included as covariates in the full model. All GAMs were constructed using the mcgv package in R.



We have quantified diets of nearly 3,000 preserved fish. Diet composition was defined according to the percent occurrence (%Oi,j – the proportion of fish of species I that have prey j in their stomachs), the percent weight (%Wi,j – the average percentage of the stomach contents by weight of fish of species I represented by species j) and the percent number (%Wi,j – the average percentage of the stomach contents by number of fish of species I represented by species j). Dietary patterns were analyzed using canonical correlation analysis (CCA) to explore dietary patterns in croaker. CCA is a special case of multiple linear regression and shares many of the assumptions of multiple regression such as linearity of relationships, homoscedasticity, lack of high multicollinearity, and multivariate normality for purposes of hypothesis testing. CCA relates two sets of variables to each other by reducing multiple variables to a fewer number of canonical variables which are linear combinations of the original variables. The first linear combination is extracted and the process is repeated for the residual data, with the constraint that the second linear combination of variables must not correlate with the first one. The process is repeated until a successive linear combination is no longer significant. Canonical Correlation Analysis (CCA) was performed on the data on 458 stomachs that could be linked to data for the following environmental variables: length of croaker (mm), depth (m), bottom temperature (C), bottom salinity (ppt), and bottom dissolved oxygen (mg/L). Diet variables were the %W values for each prey item found in each fish stomach and included: anchovy, fish, crabs, mysids, amphipods, isopods, shrimps, polychaetes, gastropods, cumaceans, and molluscs. Percent composition by weight (%W) rather than %N was used to represent consumption for croaker because many food items such as polychaetes were highly digested and count data was impossible to obtain. Percentage composition by weight also reflects the bioenergetic and nutritional quality better than a count estimate.
RESULTS
Three broad scale surveys were conducted in 2004; from April 19 - May 26 (CF0401), July 6 – 13 (CF0402) and September 13- 21 (CF0403) (Table 1). Cruise reports, which provide more details of the sampling conducted on each cruise, are available on line at http://hjort.cbl.umces.edu/cfreports.html. Samples of the fish community were collected from between 29 - 51 stations (Table 1).
TABLE 1 NEAR HERE
…………..The most striking feature of the catches in this years CHESFIMS survey was a substantial increase in diversity of the catch during the Autumn survey. This increase was particularly marked in the lower bay strata where the diversity increased from 8.12 species.tow-1 in 2003 to 13.7 species.tow-1 in 2004. This increase was caused by an increase in more marine species in the lower bay strata. Importantly, this increase in diversity was not matched by an increase in the abundances as average catch per tow decreased from 4,376 fish.tow-1 in 2003 to 77 fish.tow-1 this year.
Data summaries of catches from the CHESFIMS surveys are on line at http://hjort.cbl.umces.edu/cfdata.html . Previously, we have documented the ability of the combined CHESFIMS /TIES database to provide annual, baywide estimates of bay anchovy abundance and distribution. We now have a ten year time series of abundance of bay anchovy that does not show a clear temporal trend, but do clearly show a high degree of variability. The data indicate that there is typically an order of magnitude variation between the lowest observed stock biomass in the time series and the highest irrespective of season (Table 2).
TABLE 2 NEAR HERE
……………..In addition to our analysis of bay anchovy catches, we have also quantified time series of catches of other species, such as Atlantic menhaden (Brevoortia tyrannus – Fig 2). For menhaden, there was similarly large variation among years. We have yet to extend the CHESFIMS data with the TIES data to create a 10-year catch record, similar to bay anchovy.

We have also explored the distribution of several species of fish throughout the CHESFIMS time series. The distribution of many species exhibit clear responses to environmental forcing. For example, hogchoker (Trinectes maculates) were collected in only 21% of all hauls and were not uniformly distributed (Figure 3).


FIGURE 3 NEAR HERE
………………..Preliminary investigation of the relationships between relative hogchoker abundance and the environmental covariates did not indicate clear relationships between hogchoker relative abundance and any environmental parameter (Figure 4).
FIGURE 4 NEAR HERE
………………Hogchoker were collected in water with bottom temperatures between 8.58 and 29.40 oC, with peak catches occurring at intermediate temperatures (Figure 4a). Water salinities at which hogchoker were collected varied from 0.026 – 32.00 without apparent trend (Figure 4b). Similarly, hogchoker were caught at a depth range of 3.5 – 42.4m, with highest catches seen at intermediate depths (Figure 4c). The distribution of hogchoker catches as a function of dissolved oxygen (Figure 4d) paralleled the pattern seen with respect to temperature, reflecting the strong covariation between temperature and dissolved oxygen.

The first stage of the GAM predicts the probability of hogchoker occurrence, P, at any station. The reduced model, resulting from backward elimination of the full suite of parameters, indicated that bottom time, temperature, salinity, year, depth and the interaction of temperature with year were significant predictors of hogchoker presence. To investigate the importance of each individual covariate in the fitted stage 1 model, the conditional contribution of each covariate, given the inclusion of the remaining covariates in the model were plotted. In the first stage, the GAM plot for bottom time indicated an asymptotic relationship with the probability of hogchoker occurrence (Figure 5a).



FIGURE 5 NEAR HERE
…………………The standard error bands indicated relatively low variability near the center of the plot with increasing variability towards the extremes. Due to high variability at greater bottom time values, it was difficult to determine if the relationship was truly asymptotic or a function of the observed covariate values. The relationship between depth and occurrence was nonlinear (Figure 5b). The limited number of samples at shallow and deeper depths restricted inferences outside the 5-20m depth range. Within this range, however, hogchoker occurrence generally increased with depth up to approximately 15 meters, beyond which the probability of occurrence slightly decreased. The GAM plot for salinity indicated that occurrence was lowest at intermediate salinities and increased as the water became more and less saline (Figure 5c). The greatest probability of occurrence occurred in oligohaline and freshwater habitats. Temperature and hogchoker occurrence exhibited a linear relationship with increasing probability of occurrence as temperature increased (Figure 5d). There was more variability in the predicted spline for temperature compared to those of other covariates, including salinity and bottom time. The relationship between hogchoker occurrence and year was not significant (p = 0.67), however, year remained in the model due to its significant interaction with temperature (Figure 5e). The interaction of temperature with year generally indicated a relatively high probability of occurrence at intermediate temperatures across years (Figure 5f). This probability was greatest during early years. With the exception of the most recent years, occurrence declined as temperature progressed towards the extremes. In the most recent years, however, occurrence increased as temperature declined.

The second stage of the GAM modeled the natural log of relative hogchoker abundance, incorporating only those stations where hogchoker were present into the model. The reduced model included temperature and salinity only as significant terms. The GAM plots corresponding to the second stage portray the conditional relationship of each covariate with hogchoker relative abundance (Figure 6).


FIGURE 6 NEAR HERE
……………..Relative abundance and temperature exhibited a dome-shaped relationship with an increase in abundance with increasing temperatures up to approximately 21º Celsius, beyond which hogchoker abundance declined (Figure 6a). The peak in occurrence at 21º, however, corresponded with an area of high variability due to a small number of observed temperatures between approximately 19º and 23º. Variability also increased towards the low and high extreme temperature values. Salinity exhibited a negative linear relationship with abundance, where abundance declined with increasing salinities (Figure 6b). The rug plot indicated approximately uniform sampling coverage over the range of observed salinities.

Samples frozen during CHESFIMS surveys provide the foundation to quantify patterns of growth in all species collected. We have made substantial progress in ageing many species. For weakfish, we have estimated growth by combining estimates of average size at age for 10-d birthdate cohorts sampled longitudinally over the course of the sampling seasons, with size-at-age estimates for age-1 – age-x adults. The average weight determined for the average 40, 50, 60, 70, 80, 90, and 100-day old weakfish was 7.15g, 8.71g, 37.6g, 27.5g, 49.4g, 39.0g, and 52.5g respectively. Growth of the average age-1 weakfish over course of a season was estimated to be from 79.2 - 165.9g. Similarly, age-2 fish grew from 170.0 - 206.0g. These data have been used in a bioenergetics model to estimate the production of age-0 – age-4 weakfish in Chesapeake Bay (Figure 7). Additional age and growth work has been completed on white perch (Morone americanus), Atlantic croaker and striped bass.

FIGURE 7 NEAR HERE
Since the beginning of the CHESFIMS program we have quantified diets of 2,977 fish. Dietary analysis has focused on white perch, weakfish, hogchoker, and croaker. These studies have quantified temporal and spatial variation in diet among different seasons and years. We have also quantified ontogenetic changes in diets. More recently, we have begun to develop quantitative descriptions of the relationship between environmental drivers and observed diets.

CCA revealed seasonal and ontogenetic patterns in the diets of Atlantic croaker. Croaker diet consisted primarily of polychaetes while mysids, amphipods, and shrimp were also important. Anchovy and other fishes occured more frequently in stomachs in the summer and somewhat in the fall sampling periods, whereas mysids and other crustaceans were less prevalent in stomachs in the summer. Piscivory accounted for >60 %W of consumption for croaker in mid-Bay stations in the summer survey. The seasonal trend was strong, but there seemed to be only a slight difference between years. The effect of strata (lower, mid and upper Bay) in the analysis was not strong. Most Pearson correlation coefficients were low and the significance observed may be a result of large sample size. Of particular interest was the high negative correlation between temperature and length, suggesting that large fish tend to be found in colder water. There is also a negative correlation between temperature and DO (-0.431) and depth and dissolved oxygen (-0.296). This is to be expected because oxygen diffuses more rapidly in cold water, thus DO should be high when temperature is low. Dissolved oxygen is also affected by mixing of water layers so that at greater depths mixing is low so oxygen is low. Most of the correlations between the prey items with environmental variables occur with length, temperature, and depth.



The results of the CCA produced four significant canonical correlation functions that were significant (Table 3).
TABLES 3 and 4 NEAR HERE
……………..Wilk’s Lambda statistic was significant (F60,2068.8=4.21, P<0.0001) and thus we rejected the null hypothesis that there is no discriminating power in the data set. The first Canonical Correlation (CC1) was positively correlated with mysids and negatively correlated with polychaetes (Table 4). CC1 is also positively correlated with fish and anchovy, but not as highly as mysids. Of the environmental variables, CC1 was correlated highly with bottom temperature (Table 5). Based on the graphs of diet composition (Figure 8) CC1 may represent the change in diet from polychaetes to fish and mysids in warmer months (summer and early fall).
FIGURES 8 and 9 NEAR HERE
……………..The second CC (CC2) was positively correlated with anchovy, but negatively correlated with polychaetes (Table 3). CC2 was highly positively correlated with length of the fish (0.7043, Table 4). When CC2 scores for prey items (V2) and environmental data (W2) were plotted against each other and coded due to three arbitrarily chosen length classes a pattern emerged (Figure 9). Fish smaller than 150mm were clearly separated from larger fish. Fish smaller than 150mm are typically fish age-1 or younger and they tend to occur in low salinity areas.
CONCLUSIONS
In the fourth and final year of CHESFIMS, we completed three broadscale surveys. We also made considerable progress on the interpretation and analysis of data collected in previous years The results from the different surveys provide a solid foundation from which to address important questions relevant to multispecies management.
1) Our surveys provide reliable indices of abundance and distribution of ecologically and economically important finfish species in Chesapeake Bay. The survey documented changes in abundance and in the distribution of important Chesapeake Bay fishes during 2001-2004. In combination with previous CHESFIMS surveys and with the data from the TIES program, we are now capable of developing ten-year times series of estimated abundances. Importantly, these data provide estimates of not only commercially important species, but also of ecologically important species, such as bay anchovy, on which many of the commercial species rely. The abundance estimates we develop will be important to the development of a range of multispecies assessment models including EwE, multispecies production models, and MSVPA.
2) Our sampling will provide important information on the trophodynamics of key components of the Chesapeake Bay fish community. As regional agencies begin to explore multispecies management models, such as EwE, the need for diet data, collected coincidentally with abundance estimates will become acute. We are developing spatially and temporally explicit diet matrices for key species. We have also developed statistical models to quantify spatial, temporal and ontogenetic changes in diets of key species.
3) On going efforts with regard to statistical analysis of the data offer the opportunity to optimize current survey designs. We have continued our analysis of survey designs, and have greatly expanded the use of geostatistical techniques to estimate abundance (and estimates of error).
4) We have embarked in a database project to make the QA/QC’ed data available over the web.
LITERATURE CITED.

(Bold type indicates publications resulting from NCBO funding of CHESFIMS)


Cressie, N. A. C. 1993. Statistics for Spatial Data, Revised edition. Wiley Interscience, New Tork.

Curti, K. L. 2005. Patterns in the distribution, diet and trophic demand of the hogchoker, Trinectes maculatus in the Chesapeake Bay, USA. MS. University of Maryland, College Park.

Houde, E. D., M. J. Fogarty, and T. J. Miller. 1998. Prospects for Multispecies Management in Chesapeake Bay. 98-002, Scientific and Technical Advisory Committee of the Chesapeake Bay Program, Edgwater, MD.



Jensen, O. P., and T. J. Miller. in press. Geostatistical analysis of blue crab (Callinectes sapidus) abundance and winter distribution patterns in Chesapeake Bay. Transactions of the American Fisheries Society 00:000-000.

Jung, S., and E. D. Houde. 2004. Recruitment and spawning-stock biomass distribution of bay anchovy (Anchoa mitchilli) in Chesapeake Bay. Fishery Bulletin 102:63-77.

Rilling, R. C., and E. D. Houde. 1999. Regional and temporal variability in distribution and abundance of bay anchovy (Anchoa mitchilli) eggs, larvae, and adult biomass in the Chesapeake Bay. Estuaries 22:1096-1109.



Volstad, J. H., M. C. Christman, and T. J. Miller. in press. Design efficiencies of transect and stratified random trawl surveys. Fishery Bulletin.







Download 43.63 Kb.

Share with your friends:




The database is protected by copyright ©ininet.org 2024
send message

    Main page