Starr et a\ Movements of Sebastes pauaspinis dnd S chlomstictus in Monterey submarine canyon 



327 



(Table 1). This enabled us to use signal location and tag 

 depth data to more effectively estimate fish positions and 

 movements. Again, the most northeastern receiving zone 

 (signals only recorded by receiver 1) was labeled as zone 1; 

 zone numbers increased to the southwest (Table 1, Fig. 2). 

 Each year we placed an S-4 current meter with record- 

 ing thermometer and salinometer on a mooring near the 

 seafloor to determine if changes in current, salinity, or 

 temperature affected fish movement. In 1997, the current 

 meter was located in 100 m of water on the shelf about 400 

 m away from receiver 2 (Fig. 1 ). In 1998, the current meter 

 was moored with receiver 1, and was located in 100 m of 

 water (Fig. 2). 



Data analysis 



Receivers logged the tag number, date, time of day, acoustic 

 frequency, and tag depth each time a signal was detected. 

 Receivers also recorded signal strength, noise, gain, and 

 error messages provided by the receiver software. Data 

 collected by the moored receivers were downloaded as text 

 files and imported into a database for analysis. Tag depth 

 was plotted versus time of signal reception for each tag 

 and each receiver. Differences in fish depths by time of day 

 were analyzed with ANOVA and Scheffe's F-test post-hoc 

 analysis (Sokal and Rohlf 1997). 



In 1997, we grouped signals into half-hour time inter- 

 vals to compare signal receptions between receivers. In 

 1998, we grouped signals into hourly intei-vals. We labeled 

 each inten'al a time "bin" and standardized bin numbers 

 among all receivers. Thus, any signal recorded by a re- 

 ceiver in a given time period (bin) could be directly com- 

 pared with signals from other receivers in similar time 

 bins. In 1997, the study included 4309 half-hour time bins. 

 In 1998. the study included 2535 hourly time bins. 



Each signal was thus assigned a time bin and a receiv- 

 ing zone according to the time interval of signal reception 

 and the combination of receivers recording signals from 

 that tag number. Treating each receiving zone number as 

 a numeric rank enabled us to use a simple average of rank 

 to identify the predominant receiving zone that a fish oc- 

 cupied in the time bins that occurred during a week (336 

 weekly bins in 1997 and 118 weekly bins in 1998). A differ- 

 ence in average ranking among weeks indicated the fish 

 had moved; the value of the average indicated the direc- 

 tion in which the fish moved. For each tag, a chi-square 

 test of heterogeneity was used to test for differences in 

 average ranking among weeks. In 1997. week 2 of the 

 study was the first week for which there were transmis- 

 sions from all tagged fish; thus week 2 was used to repre- 

 sent the expected fish distribution by receiving zone for 

 the chi-square test. In 1998. we used week 1 to represent 

 the expected distribution. We also used the unplanned test 

 of homogeneity of replicates tested for goodness of fit (So- 

 kal and Rohlf, 1997) to determine if the average location 

 of a fish (average rank) was similar between weeks. This 

 method enabled us to group weeks in which a fish was in 

 a similar location. In addition to generating weekly distri- 

 butions, we used the ranking system to plot semihourly 

 (1997) or hourly (1998) movements of tagged fish. 



15 Oct 1997 



15 Nov 1997 



15 Dec 1997 



Figure 3 



Movements of tagged greenspotted rockflsh (tag-2 and 

 tag-6 fish) across study area in 1997. as depicted by 

 changes in receiving zones that were derived from pat- 

 terns of signal receptions. 



Results 



Greenspotted rockfish 



In 1997. we tagged six greenspotted rockfish, ranging in 

 total length from 35 to 39 cm (Starr et al., 2000). Lea et 

 al. (1999) reported that greenspotted rockfish in this size 

 range are 11-15 yr old and probably mature. The three 

 moored receivers recorded signals throughout the study 

 period from all tags except tag 1. Signals from tag 1 were 

 recorded for 18 hours, then not again until 67 days later. 

 The total number of transmissions recorded from each 

 tagged fish ranged from 156 to 24,132 (Starr et al., 2000). 

 Except for tag-1 fish, tagged greenspotted rockfish ex- 

 hibited two patterns of relatively small horizontal move- 

 ments. Tag-2 and tag-6 fish remained primarily in the re- 

 ceiving zone in which they were released (zone 4) and 

 exhibited few cross-zone movements (Fig. 3, Table 2). Re- 

 ceivers recorded transmissions from each of these tags in 

 99'7f of the time bins. More than 569f of the time, signals 

 originated in receiving zone 4, and 94'7f of the time sig- 

 nals originated from receiving zones 3 or 4 (Table 2), an 

 area comprising 58 ha, or a linear distance along a ledge 

 of 1200 m. Wlien the fish moved out of zones 3 or 4, they 

 most often moved southwest, towards the mouth of Soquel 

 Canyon (towards Zone 5 in Fig. 3). Chi-square and post- 

 hoc analyses indicated similarity in the pattern of signals 

 received from the tags for most weeks, as indicated by the 

 weekly average rank of receiving zone (Fig. 4). 



