254 



Fishery Bulletin 104(2) 



and small lobsters are more likely to appear in the 

 inshore survey. This is true for both sampling seasons, 

 fall and spring. Two possible hypotheses can be devel- 

 oped to explain such patterns: one is that large differ- 

 ences in size compositions exist in lobsters inhabiting 

 the areas covered by the inshore and NMFS survey pro- 

 grams, and the other is that the observed differences in 

 size compositions between the DMR inshore and NMFS 

 surveys result from differences in gear selectivity for 

 lobsters of different sizes. The first hypothesis implies 

 that lobsters have size-dependent inshore-offshore distri- 

 bution, namely that large lobsters tend to inhabit deep 

 waters and small lobsters inhabit shallow waters. The 

 second hypothesis implies that although lobsters may 

 have no size-dependent inshore-offshore distribution, the 

 NMFS survey gears are selective for large lobsters and 

 the inshore survey gears are selective for small lobsters 

 in the population. Such a difference in gear selectivity 

 between the two programs may result from differences 

 between the two sampling programs not only in sam- 

 pling gears but in towing speed and duration. 



In order to determine which hypothesis is more plau- 

 sible, we evaluated size composition for a depth stratum 

 covered by both the DMR inshore and NMFS surveys. In 

 the 2003 spring inshore survey, the fourth depth stra- 

 tum had an average depth of 125 m which overlapped 

 depth ranges in the NMFS surveys. The average size of 

 lobsters in the fourth depth stratum in the 2003 spring 

 inshore survey was almost identical to the average size 

 of lobsters in the NMFS survey (Table 1 and Fig. 3). 

 This finding indicates that the inshore survey gears 

 can, like the NMFS survey gears, catch large lobsters 

 if they are present in the areas covered by the survey 

 program. Likewise, the differences in the average size 

 between the first three strata and the fourth stratum 

 (Table 1) in the 2003 spring inshore survey are likely 

 to result from a lack of large lobsters in the first three 

 strata in the inshore surveys, rather than from gear 

 selectivity. Thus, the differences in size composition of 

 lobsters between the DMR and NMFS surveys are likely 

 the result of the differences in the size composition of 

 the lobster population between the areas covered by the 

 inshore and NMFS surveys, rather than the result of 

 sampling methods. Although our study could not exclude 

 the impacts of possible differences in gear selectivity on 

 the survey size compositions, the results of our study 

 seem to support the first hypothesis that the lobsters 

 have a size-dependent inshore-offshore distribution, 

 where large lobsters are more likely to be found in deep 

 waters and small lobsters are to be found in shallow 

 inshore waters. Thus, the two sampling programs tend 

 to cover different segments of the stock in the Gulf of 

 Maine. In order to have an adequate representation of 

 the lobster population, it is necessary to include data 

 from both sampling programs to describe the lobster 

 population dynamics in the Gulf of Maine. 



The temporal changes in the abundance index in the 

 spring surveys were rather consistent between the DMR 

 and NMFS surveys and for data on the original and log 

 scales. However, for the fall survey, the delta mean of 



the NMFS survey was consistent only with the mean 

 of log-transformed data, but not with the mean of the 

 original data and the mean retransformed data from 

 the mean of log-transformed data (Table 2). This find- 

 ing raises an interesting question regarding the type of 

 data transformation we should choose and the potential 

 impacts of each type on detecting temporal changes in 

 the population abundance. A different choice of data 

 transformation methods may lead to different interpre- 

 tations of temporal variations in stock abundance. 



If the size-dependent inshore-offshore distribution 

 pattern is not taken into consideration in making stock 

 assessments, the entire stock will not be managed. We 

 may need a separate set of abundance indices and size 

 compositions for the lobsters in the inshore and offshore 

 waters to describe the population dynamics of lobsters 

 of different size groups more accurately. The inclusion 

 of only NMFS survey data or inshore survey data in the 

 stock assessment may lead to errors in determining the 

 status of the lobsters if the population dynamics of large 

 lobsters is different from that of small lobsters. The 

 blending of DMR and NMFS data into a single set of 

 data may be a solution. However, whether such a set of 

 data can describe the dynamics of the whole population 

 depends upon whether the lobster population dynamics 

 are consistent among different size classes and whether 

 there is a large difference in gear selectivity and in 

 catchability of the trawl gear. Large differences in the 

 population dynamics of large and small lobsters and in 

 the gear selectivity and catchability of DMR and NMFS 

 trawls may make such blended data less desirable for 

 describing the population dynamics with DMR-NMFS 

 blended data. For example, a significant increase or 

 decrease in the abundance of lobsters of prerecruiting 

 sizes may not be well defined with the NMFS surveys 

 or DMR-NMFS blended data. We recommend both sets 

 of data from the DMR inshore and the NMFS surveys 

 be used separately in the stock assessment to define 

 the population dynamics of the lobsters in the Gulf of 

 Maine more accurately. 



Many fish species in the world exhibit a size-depen- 

 dent distribution pattern similar to that of the Ameri- 

 can lobster in the Gulf of Maine. Their population 

 abundance and size structure are often assessed by 

 fisheries-independent survey programs. The informa- 

 tion is then used in the assessment and management of 

 these fish species. Because it is often difficult with large 

 trawls to survey inshore areas where the majority of 

 stocks are present, the fishery-independent survey may 

 not provide adequate coverage of inshore areas that are 

 productive and critical, in particular, to recruitment. 

 Thus, it is highly likely that fish in inshore areas are 

 not adequately represented in an offshore-focused sur- 

 vey, and this misrepresentation would lead to errors 

 in data on stock-size structure. Such an error may, in 

 turn, result in large errors in a stock assessment if an 

 age- or length-based stock assessment model is used. 

 We suggest that two sampling programs with different 

 spatial focuses may help identify the problems associ- 

 ated with a sampling program that does encompass fish 



