FISHERY BULLETIN: VOL. 82, NO. 3 



The effects of season and location on crab density 

 were examined by analysis of variance (SPSS 

 ANOVA procedure). The sampling year was di- 

 vided into two seasons: spring-summer (March- 

 August) and fall-winter (September-February). 

 The navigation channel was divided into two 

 areas: the outer estuary (stations 2, 3, and 4: sta- 

 tion 1 deleted due to lack of winter data points), 

 and the inner estuary (stations 7, 8, and 9). A 

 two-way ANOVA was performed with these two 

 seasons and two station groups as the independent 

 variables, and crabs/ha as the dependent variable. 



Population Estimation 



Two basic assumptions were made concerning 

 the trawl data: 1) Sampling efficiency of the net 

 was not 100% and varied for each age class of 

 crabs. Efficiency was estimated to be 0.33 for the 

 0+ age class during summer, and 0.25 in winter, 

 based on comparisons between net catches and 

 visual counts of young instars on mudflats at low 

 tide (see Discussion). Sampling efficiency was es- 

 timated to be 0.50 for all other age groups in accor- 

 dance with Gotshall (1978a). 2) Sampling effi- 

 ciency was assumed to remain constant and not to 

 vary as a function of changes in crab behavior 

 (e.g., burial or diel activity variations). 



Data on crab densities were used from a 12-mo 

 period, June 1980 to May 1981, which was divided 

 into three "seasons": summer (June-August 

 1980), fall-winter (September 1980-February 

 1981), and spring (March-May 1981). Population 

 estimates were made for three age groups (0+, 1 +, 

 and 2+, the latter including all 3+ animals which 

 were identifiable only in summer 1980) in each of 

 the three defined seasons. A stratified random 

 technique was used, using the following variables 

 (see Cochran 1953): 



h 



<*ih 

 n h 



Xih 



s2 (x h ) 



= stratum of harbor 



= catch of crabs in tow /', stratum h 



= area (ha) covered by tow /, stratum h 



= number of tows in stratum h 



= individual estimates of crabs ha -1 , 



from tow i, stratum h 

 = mean catch (crabs ha -1 ) in stratum h 



for a given season 

 = area of harbor in stratum h (ha) 

 = variance of mean x^ in stratum h 

 = total number of crabs in harbor. 



that would have led to complications in the deter- 

 mination of confidence intervals, but only minor 

 changes in the resultant mean densities of crabs. 

 Mean crab density in each stratum was calculated 

 for each age group and season by 



Xh 



I 



t=i (c lh /a ih ) 



n h 



The total number of crabs in each stratum was 

 calculated as T^ = A^ (x h ), and the total for the 

 harbor by the sum of all stratum totals, 



15 



T = !T h . 



h=l 



For strata 1-9, the variance of each stratum total 

 was calculated by 



V(T h ) = 



A%s 2 (x h ) 



n 



Data used for population estimates were not trans- 

 formed as done for ANOVA comparisons because 



and the variance of the total was calculated by 

 summing the individual variances 



9 



V(T) = 2 V(T h )  

 h=l 



Confidence intervals for T were approximated at 

 the 95% level by 



T± *(df, 0.05) ^(T) • 



Crab abundances in strata which were not sam- 

 pled by trawl (10, 11, 14, and 15) were calculated 

 using mean density values from nearby strata of 

 similar ecological characteristics. Data were used 

 from strata 6 (for 10 and 11), 5 (for 14), and 3 (for 

 15). Totals by age group and season for those strata 

 were added to totals for strata 1-9, to obtain totals 

 for the entire estuary. The estuary totals were 

 divided by the estimated trawl efficiency factors to 

 obtain final corrected estimates of crab abundance 

 by age group and season. Confidence intervals for 

 these final estimates could not be computed. 



RESULTS 



Temperature-Salinity Profile 



Grays Harbor has a strong horizontal salinity 

 gradient (Fig. 2). Temperature and salinity were 



472 



