AN INTENSIVE FISHING EXPERIMENT FOR 



THE CARIDEAN SHRIMP, HETEROCARPUS LAEVIGATUS, AT 



ALAMAGAN ISLAND IN THE MARIANA ARCHIPELAGO 



Stephen Ralston 1 



ABSTRACT 



During January 1984 an intensive fishing experiment for the deepwater caridean shrimp, Heterocarpus 

 laevigatus, was conducted near Alamagan Island in the Mariana Archipelago. Twenty standard shrimp 

 traps were set daily, producing a significant decline in the average catch rate from 3.33 to 1.82 kg/trap- 

 night over a 16-day period. This drop was associated with a removal of 776 kg of shrimp from the study 

 site. Resampling the area 4 months later showed that the catch rate remained depressed. Length-frequency 

 data demonstrate that the decrease in catch per unit effort was due to a decline in the number of shrimp 

 caught. An initial population size of 1,714 kg from 312 ha habitat is estimated, corresponding to one 

 exploitable shrimp per 51 m 2 . The estimate of catchability (0.001945 trap-night" 1 ) indicates that H. 

 laevigatus may be easily overfished by trapping. 



Intensive fishing experiments can provide the ideal 

 complement to resource surveys using catch per unit 

 effort (CPUE) to estimate the relative abundance 

 of exploitable stock. Whereas values of CPUE are 

 usually adequate for studying spatial and temporal 

 variation in resource abundance, often an absolute 

 estimate of exploitable biomass is required. This is 

 particularly true of yield assessments. Due to the 

 relative nature of CPUE statistics, a conversion fac- 

 tor is necessary to translate catch rates into absolute 

 units of biomass. This proportionality is termed 

 catchability, typically a constant parameter (but see 

 Schnute 1983; Polovina 1986) which can be esti- 

 mated from the results of intensive fishing ex- 

 periments (Ricker 1975). 



The advantages of intensive fishing over alterna- 

 tive methods of estimating the catchability coeffi- 

 cient (q) are several. Foremost is that no history of 

 either catch or effort data is needed. This character- 

 istic makes methods of fishing success (Ricker 1975) 

 or survey-removal (Schnute 1983) particularly at- 

 tractive for use in assessments involving exploratory 

 survey data, as well as for studying emerging new 

 fisheries. A second advantage is that results can be 

 obtained rapidly. Because fishing is, by definition, 

 conducted intensively over a short time period and 

 the necessary computations are quite simple, an 

 estimate of q is quickly realized. 



Southwest Fisheries Center Honolulu Laboratory, National 

 Marine Fisheries Service, NOAA, 2570 Dole Street, Honolulu, HI 

 96822-2396. 



Although these advantages recommend the ap- 

 proach, two restrictive assumptions must be made 

 in analyzing the data. One must assume, in the 

 absence of information to the contrary, that the 

 population fished is closed, or equivalently, that 

 additions exactly balance removals other than those 

 due to fishing. The basis of this assumption can be 

 strengthened if the intensive fishing site is located 

 in a naturally isolated area. For example, Polovina 

 (1986) performed an intensive fishing experiment 

 on a small pinnacle 5.5 km in circumference which 

 was isolated by 75 km of deep water from the near- 

 est similar habitat. A second assumption is that fish- 

 ing removals account for all changes in stock 

 biomass, i.e., natural mortality, growth, and recruit- 

 ment are negligible during the period of fishing. For 

 this reason, removals are carried out intensively 

 over as short a time interval as possible. If both 

 assumptions hold then q can be estimated directly 

 by the slope of the linear regression of either CPUE 

 on cumulative catch (Leslie and Davis 1939) or 

 log{CPUE} on log{cumulative effort} (DeLury 1947). 



Refinements to these two basic methods have been 

 proposed by Braaten (1969), Crittenden (1983), 

 Schnute (1983), and Polovina (1986) among others. 

 Generally, estimators have been found to be most 

 sensitive to a departure from the assumption of con- 

 stant catchability. A variety of adjustments have 

 been used to correct this and other statistical prob- 

 lems which often occur with real data. 



The work reported here is an application of the 

 intensive fishing method to estimate the catchabil- 



Manuscript accepted July 1986. 



FISHERY BULLETIN: VOL. 84, NO. 4, 1986. 



927 



