GULF OF MEXICO COMMERCIAL SHRIMP POPULATIONS 



357 



movement, are highly susceptible to capture b}' 

 bottom trawls of all types. The average mini- 

 mum size of shrimp retained by "standard-mesh" ^ 

 trawls sets the lower limit of what is referred to 

 herein as the fishable population. Hence the 

 Ushable biomass is that fraction (in terms of weight) 

 j of a commercial population, which comprises those 

 individuals vulnerable to capture with the gear 

 conimoidy used by the fishery. Whether or not 

 landing data include everj'thing caught b}- the gear 

 employed is a matter of vital concern. It is 

 recognized that even though standard-mesh trawls 

 may be used at all times, the minimum size of 

 shrimp selected from their catches, not the mini- 

 mum size actually caught, sets the lower limit of 

 that part of the overall population about which 

 corresponding commercial statistics can give any 

 information. The extent to which selection 

 practices prevail varies in unpredictable fashion 

 from area to area and season to season. Depar- 

 tures from the definition of fishable biomass given 

 above can also be attributed to fishing practices 

 wherein standard-mesh gear is employed, but 

 aggregations of shrimp of a specified minimum 

 size are first sought out by trial fishing. Although 

 this circumvents sorting catches predominated by 

 small shrimp, and thereby mitigates the discard 

 problem, the resulting statistics are quite restric- 

 tive as to the information they give about the 

 whole population. 



Assume now that the geographic range of a 

 given shrimp population is appro.ximately known. 

 If the manner in which commercial trawls are 

 deploj'ed over it during equivalent time increments 

 is also known, an index to the true probability 

 with which a standard unit of the fishable biomass 

 will have been removed can be derived for each 

 increment. A factor proportional to the average 

 harvestablc biomass is thus obtained when the 

 corresponding (total) commercial catch is divided 

 by this "probability-of-capturc" index. The lat- 

 ter has been termed the "effective overall fishing 

 intensity" (/) by Beverton and Holt (1957) who 

 discuss its theoretical aspects and derivation. 

 For any time interval and population, it is the 

 weighted average of all fishing intensities calcu- 

 lated for each trawling subarea included in the 



' Tho term "stantlar<i-mpsh" is defined as that sijo mesh nio.st commonly 

 used in a particular fishery, be it inshore or otishore. Botti fisheries arc 

 treated separately throuphout tliis report with IH-inch mesh beinR considered 

 tlie standard inshore, 2'i-lnch mesh the standard oflshore. A major require- 

 ment is that this average mesh size remain constant. 



population's range. The fishing intensity in any 

 subarea is simply the ratio between the amount of 

 effort expended therein and the subarea 's size. 

 Weighting factors are the subareas' corresponding 

 biomass indices. Since the ratio between catch 

 (in weight) and effective overall fishing intensity is 

 proportional to the fishable biomass, it follows 

 that the fishing intensity is also proportional to the 

 fishing mortality parameter, an important con- 

 sideration in attempts to evaluate the latter. 



To obtain biomass indices directly, Gulland 

 (1955) uses a method almost identical mathe- 

 matically to that introduced by Beverton and 

 Hold (1957). For a short interval of time, say a 

 month, catch (in weight)-effort ratios are calcu- 

 lated for each subarea within a species range. A 

 weighted average catch-effort ratio is then deter- 

 mined, the sizes of each subarea consitituting the 

 weighting factors. This ratio, the same as that 

 derived above, is theoretically proportional to the 

 size of the population's exploitable fraction, and 

 hence is termed a fishable biomass index. In 

 effect, it is a density estimator in which the elTects 

 of uneven distribution of fishing effort are ehmi- 

 nated by a process analogous to stratified sampling. 



Error and Bias 



Manj' factors, however, operate to alter the 

 theoretical utility of this index. Some of these, 

 namely error and bias associated with compiling 

 landing and effort data, have already been dis- 

 cussed. Controlling their influence entails refine- 

 ment of sample projection and data collection 

 techniques. Superimposed on compilation de- 

 fects, however, are still others which, because of 

 their inconstancy, are very difiicult to cope with. 

 Two classes may be readily distinguished. 



The first affects the comparability of effort 

 statistics and stems from differences in trawler 

 fleet composition along with nonuniformity of 

 operating conditions. All trawlers are not equally 

 powerful, are not manned b}- equally efficient 

 crews, and do not operate under identical climato- 

 logical and sea conditions. For instance, since 

 gear efficiency is directlj' related to ground speed 

 (up to some optimum point), under conditions of 

 uniform shrimp density, identically powered and 

 rigged vessels operating against the current would 

 normally be expected to make smaller catches per 

 utiit time than those operating with it. The 

 writer has observed a low resultant ground ^m^i^d 



