The Roe's abalone fishery near the Perth metropolitan area, Western Australia.

ABSTRACT The recreational Roe's abalone fishery is concentrated on easily accessible reefs that adjoin metropolitan Perth in Western Australia. There is a restricted recreational fishing season of 1.5 h a morning, for six consecutive Sundays in November and December each year. The reefs near Perth are also the focus of the commercial Roe's abalone fishery in Western Australia. The recreational catch and effort for each 10 nautical mile section of the Perth fishery is estimated from a creel survey with instantaneous counts, aerial surveys and interviews with fishers, from 1997-2003. Most of the recreational and commercial catches were taken from small areas of the fishery, with approximately 87% of the recreational catch coming from two 10 nm sections, and 98% of the commercial catch coming from these two and one additional 10-nm section. An annual quota of 36 t limits the commercial catch, whereas estimates of the recreational catch increased from 33-51 t between 1997 and 2003. Recreational catch and effort estimates from 1999 on were corroborated by an independent telephone survey, with independent estimates falling within the 95% confidence intervals.

KEY WORDS: Roe's abalone, Haliotis roei, recreational catch and effort, commercial catch, volunteer, creel survey

INTRODUCTION

Roe's abalone, Haliotis roei, is distributed between the central Victorian coast and Shark Bay on the west coast of Western Australia (WA) (Shepherd 1973). It occupies the low intertidal and shallow subtidal areas of rocky reef adjoining the coast in southwestern Australia (Shepherd 1973), and grows to a maximum size of about 120 mm (Wells & Keesing 1990). In the waters near Perth WA, Roe's abalone occupy the low intertidal and subtidal limestone reefs of the mainland and offshore islands. It also occurs on some offshore reefs topped by wave-cut platforms at about low tide level. In areas of favorable habitat, Roe's abalone attains densities approaching 200 [m.sup.-2] (Hancock 2004). Densities are highest at the seaward edge of the reef platform because of the settlement of the recruits in this habitat. Densities drop to low levels by about 1.5 m depth on the subtidal cliff at the seaward edge of the reef. Densities decline more gradually shoreward from the outer edge of the reef platform. The mean size of abalone tends to increase subtidally and shoreward from the seaward edge of the reef platform.

WA is the only state with a Roe's abalone fishery, but a feasibility study has been completed in South Australia (Preece et al. 2003). Commercial abalone fishing in the Perth area began in the early 1960s, but was small scale until 1970 when an influx of divers from other states caused a rapid increase in catches (Joll & Penn 1992). In 1970 the first of a complex array of area closures and restrictions on fishing time and daily catch were introduced for the commercial fishery in the Perth area. The Perth fishery is currently described in a series of sections of 10 min of latitude, (i.e., 10 nautical mile [nm] sections). For the commercial sector the Perth Roe's abalone fishery extends from the mouth of the Moore River in the north, to Cape Bouvard to the south (sections 36-43, Fig. 1). From 1992 to 2002 the recreational fishery extended approximately 30 nm further north, to Wedge Island (sections 33-43, Fig. 1). The commercial sector is excluded from fishing in the area around Rottnest lsland, and the Cottesloe area of the mainland coast. These areas remained open to recreational fishers until the closure of Cottesloe in 2003. Rottnest Island was not included in this study, and is not considered to be an important area of the fishery. The area around Penguin Island was closed to recreational fishers from 1996 to 1999, but remained open to some commercial fishing.

Since the summer of 1990 to 1991 the commercial catch from the Perth fishery has varied from 17.7-46.1 t whole weight, compared with a total catch from the state of approximately 108-119 t. During 1997 regulation of the commercial harvest of abalone from the Perth fishery was changed from mainly input-controlled to output-controlled by the removal of restrictions on allowable fishing times, and the introduction of quotas. Quotas were set at 3 t for each of the 12 commercial fishers (36 t total quota), for the period from October 1997 to September 1998 inclusive, reported here as the 1997 season. To align the quota periods for the different abalone managed areas around the state, the Perth season was adjusted to an April to March quota period.

Interest in the recreational fishery increased steadily from the early 1970s. By the early 1980s increasing fishing pressure had led to localized stock depletions and escalating conflict between the two user groups, resulting in a closure of the majority of the fishery for the 1983 season (Wells & Keesing 1989). Since the recommencement of fishing in 1984, there has been a series of restrictions on both sectors, including the introduction of a recreational fishing licence in 1992. Since 1995 the season has consisted of 6 Sunday mornings from 7:00 AM to 8:30 AM, starting on the first Sunday in November. The minimum legal size of abalone taken by the recreational sector is 60 mm, with a daily limit of 20. The majority of fishers wade on the reef platform to search for abalone, with a small proportion snorkeling adjacent to the platform. The use of compressed air to collect abalone from the subtidal areas by recreational fishers is not permitted.

Australian abalone fisheries, including the H. roei fishery, are currently considered to be operating at sustainable levels; however globally there are cases where high exploitation levels have been a primary cause of the collapse of major fisheries. Notable instances include the collapse of fisheries in California (Tegner et al. 1992, Parker et al. 1992, Karpov et al. 1998) and Mexico (Guzman del Preo 1992), with more recent concern for the sustainability of the South African fishery (Tan- 2000, Tarr et al. 2000).

Recreational fisheries are increasingly recognized as making up an important component of the total fishing harvest of exploited nearshore marine populations (Kearney 1995). Although commercial catches are often carefully recorded, recreational catches are not as easily quantified. Consequently there is an emphasis on measuring the recreational catch component of many fisheries. The aim of measuring the recreational catch and effort has generally been: (1) to quantify an important impact on a fished stock, in concert with commercial catch; (2) to address issues of fishery sustainability and (3) to establish a basis for informed debate regarding resource sharing between interest groups (Kearney 1995). Surveys of recreational abalone catch and effort conducted in New South Wales (Worthington et al. 1998) and Tasmania (Lyle & Smith 1998), in New Zealand (Teirney et al. 1997, Bradford 1998) and South Africa (Anon 1998, 1999), have generally concluded that catch and effort are increasing where management restrictions have not been introduced. The recreational component of these fisheries varies from over 90% in some areas of New Zealand and nearly 50% in South Africa, to as low as 3% in Tasmania.

In the Perth Roe's abalone fishery, where fishers can easily access a substantial portion of the abalone population, the recreational catch is likely to be a particularly important component of the total catch. Because Perth is also the focus of the commercial fishery, providing approximately one-third of the annual commercial catch for the state, understanding the total catch in the Perth fishery is an important component of managing the resource.

The total catch in the Perth area needs to be determined on a scale that is appropriate to the stock structure for the species. The genetic neighborhood of populations in the Perth area has been found to be less than 13 km (Hancock 2000). Therefore, catch and effort information for this intensively harvested species should be considered in units of less than 13 km, or units of semi continuous reef. The finest spatial resolution of the commercial catch records during the period of this study was 10 nm sections of coast. These sections reasonably approximate units of semi continuous reef.

The objective of this study is to: (1) estimate the recreational catch and effort for each 10 nm section of the Perth Roe's abalone fishery and (2) compare the relative catches of the recreational and commercial sectors over the study period.

This study has been complemented by the development of an independent telephone survey designed to estimate the recreational abalone catches from the whole Western Australian coast. The survey was stratified to provide more detail for the Perth fishery (Hancock et al. 2003). Catch estimates from the Perth area derived from telephone and creel surveys provide a rare opportunity to compare independent recreational catch estimates.

MATERIALS AND METHODS

A methodology for estimating the recreational catch and effort was developed using an intensive examination of data obtained between 1997 and 1999. This technique was then used to obtain estimates for subsequent years as an on-going monitoring program. The development of the estimation procedure is described in detail and results obtained by applying the described technique to data collected from 2000 to 2003 are reported.

Spatial Stratification of the Perth Recreational Fishery

The Perth recreational abalone fishing area has been stratified into zones of high and low usage, based on extensive observations (Fig. 1). The highest levels of utilization are in areas of reef adjoining the mainland shore. Within this range the areas most targeted by recreational fishers have been identified (Fig. 2). The low usage zone is the remainder of the area between Cape Bouvard and Wedge Island (Fig. 1).

Environmental Conditions

For each day of the fishing season the prevailing fishing conditions are likely to influence the distribution of effort and the catch rate of recreational fishers. To quantify fishing condition and enable fishing conditions to be accounted for when estimating effort, a 10-point index was developed, with very good conditions equal to 0 and very poor conditions equal to 9 (Table 1). The fishery operates on the reef platform or shallow subtidal areas of an exposed coastline. Wave and tide height are thus especially important, with minor influence from light level, as determined by cloud and rain. Wave heights were obtained from the Department of Transport "Deep Channel" wave station in section 39 (31[degrees]58'40"S 115[degrees]41'12"E, Fig. 1). Recordings were divided into wind-driven wave and swell height, based on wave period. On mornings with a dominant easterly (offshore) wind, swell height was used, and on mornings with a dominant westerly (onshore) wind, the sum of swell and wind-wave height was used. Tide heights were obtained from the Department of Transport, Fremantle recording station. Cloud cover and rainfall were obtained from the Bureau of Meteorology, Mt. Lawley station.

Effort--High Usage Zone

Within the high usage zone there were 20 discrete areas from which counts of the number of fishers were taken (Fig. 2). All fishers were visible from raised vantage points within each area. The length of coast adjacent to the 20 areas ranged from 130 m to 895 m, giving a total of 6,924 m of coast in the areas counted, or 60% of the total length of reef which provides suitable habitat for Roe's abalone (11,553 m) in the high usage zone (Fig. 2). Count areas were located systematically throughout each 10-nm section to achieve an approximately even geographical spread of sampling effort, whereas the exact size and position of each area was adjusted so that unambiguous start and end points could be described from land and air, using purpose-drawn maps. Count areas represented reefs of high and low accessibility, and greater and lesser usage by fishers.

Department of Fisheries research staff and volunteer fisheries liaison officers (VFLOs) conducted instantaneous counts of the numbers of fishers. VFLOs were instructed to take a minimum of one count of fishers actively engaged in fishing from their area at 7:10 AM (710 count), and encouraged to take 3 or more counts throughout the fishing period. Research staff conducted a count every 10 min from 7:10 AM to 8:20 AM, a total of eight counts for the 1.5-h period. Effort was calculated in fisher minutes for each count area using the research samples and the VFLO samples for which there were three or more instantaneous counts. The VFLO samples were used only if a 710 count was taken, and at least one count was taken from each half hour period between 7:30 to 8:00 AM and 8:00 to 8:30 AM. The calculation of effort can be illustrated by plotting the histogram of instantaneous counts taken during the fishing period at a site, against the time the counts were taken (Fig. 3). The area of the histogram describes the effort in fisher minutes. For the volunteer samples, the area of each bar of the histogram or effort described by each count was determined by the timing of each instantaneous count. This can be described as in the following equation:

E = [C.sub.1]([t.sub.1]) + ([t.sub.2] - [t.sub.1])/2) + [f-2.summation over i=2] [C.sub.i](([t.sub.i] - [t.sub.(i-1)])/2 + ([t.sub.(i+1)] - [t.sub.i])/2) + [C.sub.f](([t.sub.f] - [t.sub.(f-1)])/2 + (90 - [t.sub.f]))

where E is the effort for the count area on a day in fisher minutes, C is an instantaneous count of fishers, t is the time the instantaneous count was taken in minutes from the start of fishing at 7:00 AM, i indicates the count number for intermediate counts between 1 and f, the final count for the fishing period.

Analysis of fisher participation was initially undertaken using data collected during the 3 years from 1997 to 1999. Counts of fisher participation were taken from a possible total of 20 areas per fishing day, for each of the six fishing days in 1997, 1998 and 1999 (Table 2). There were two kinds of measures of effort during 1997 to 1999, those calculated from the research samples with eight counts per day (41 cases), and those calculated from VFLO samples with three to five counts per day (146 cases). Each type of sample included a 710 count. Table 2 also contains two types of missing values where there were insufficient data to calculate effort for an area on a given day: those for which the available effort information was only one 710 count (58 cases), scenario a; and those for which there was no effort information at all (115 cases), scenario b. The missing values were estimated using two predictive generalized linear models that were developed with the information contained in the 187 research and VFLO samples. The use of volunteer staff limits the ability to prescribe completely the timing and number of counts taken during a fishing day and the areas counted per day, resulting in missing data for some areas.

For the first model, scenario a, the research and VFLO sampling was used to establish a relationship between the 710 count and fishing effort at an area for a given day, after accounting for variables and interactions that may affect this relationship. This model then used the 710 count to predict the missing effort estimate for the 1.5-h fishing period (Table 3). The appropriate predictive model to estimate the effort for scenario b was determined by examining the relationship between the effort calculated from research and VFLO sampling and the variables that most influence effort. The variables investigated were data source (research or VFLO), year, day, count area, and fishing condition (Table 4). As a result of increasing variances in effort with increasing 710 count levels, effort and 710 count values were log transformed. Because geometric means are calculated from log transformed data the predicted values of effort will be biased. A correction for this bias was included by multiplying all back transformed estimates by [sub.10]([S.sup.2]/2) where [S.sup.2] is the residual mean square error for the model used to make the estimate (Baskerville 1972, Caputi et al. 1979).

The models developed from the 1997 to 1999 data were applied to the counts of fisher participation for 2000 to 2003. During this period there were no research counts taken every 10 min from the same area. There were 120 possible count area/day combinations per year. For each combination there is either a minimum of three counts including the 710 count, a 710 count only or no count information (Table 2).

Effort for the individual count areas was summed for each 10 nm section and then extrapolated to produce an estimate of daily effort for the whole 10 nm section. The scaling factor used was the ratio of the total number of fishers in the high usage zone, to the number of fishers in the 20 count areas. This ratio was obtained from analysis of digital video taken from light aircraft at 150 m (500 feet). Digital video was first successfully used on day 5 of the 2000 season. Previous attempts to use analog video footage to obtain this ratio were unsuccessful because of inadequate resolution. The ratio obtained for 2000 was applied to 1997 to 1999. A separate ratio was obtained from digital video taken on one day of each subsequent year, and used to extrapolate effort estimates for that year (Table 5).

Effort--Low Usage Zone

In the low usage zone information on effort was collected from aerial counts of fishers, with no land-based surveys. Aerial surveys were conducted on two days of each fishing season between 1997 and 2002, days 2 and 5 in 1997 and 1999 to 2002, days 1 and 4 in 1998, and on one day (day 2 only) for 2003. Counts were taken from 500 feet, beginning at Cape Bouvard in the south between 7:00 AM and 7:30 AM, and ending at Wedge Island in the north between 8:00 AM and 8:30 AM. Fishers on each reef were few enough to allow direct counts of abalone fishers. Direct aerial counts were not possible for the high usage zone because of the high density of fishers.

Aerial counts were assigned to the same 10- or 15-min time intervals used to calculate effort from the research counts, and an effort value in fisher minutes for that interval was calculated. To estimate effort for the 1.5-h period from this one count it was assumed that the distribution of fishing effort during the fishing period in these areas was the same as in the high usage zone, where research counts were taken. Thus,

[E.sub.s] = [n.summation over i=1] [C.sub.i]T/P 

where [E.sub.s] is the effort for the low usage zone portion of a 10-nm section, n is the number of reef counts of fishers in a section, [C.sub.i] is an aerial count of fishers on reef i, T is the 10 or 15 min period the count is allocated to and P is the mean proportion of the total effort for the research areas attributed to that 10 or 15-min period.

To estimate the effort for the low usage zone on days when aerial surveys were not conducted, the effort in each 10-nm section of this zone was assumed to be a constant proportion of the effort in the count areas of the high usage zone for each year, based on the two days with aerial survey information each year.

Catch Rate and Catch

Catch rates were obtained from interviews of fishers as they returned to the beach after fishing. Interviews were conducted by VFLOs, who recorded the time of the interview and the number of abalone taken, when the fisher commenced fishing, and how long they had been fishing. Where two or more fishers were pooling their catch, the number of abalone was divided equally between fishers. Interviews were not restricted to the 20 areas with counts, but were conducted throughout the high usage zone, and from the Mandurah area, at the southern extremity of the Perth fishery. Variations in catch rates, calculated in abalone per fisher hour, were examined using ANOVA for the years 1997 to 1999. Factors examined as possible influences on catch rate were 10 nm section, year, day and fishing condition. Catch rates were skewed toward high values and were normalized using a log (catch rate + 0.05) transformation.

Multiplying the estimates of effort and mean catch rate per day provided an estimate of catch for the recreational sector, expressed as a number of abalone.

The commercial catch in the Perth fishery is monitored in whole weight only, so the mean weight of recreationally caught abalone was investigated to allow conversion of the recreational catch to whole weight for comparison with commercial catches. Weigh stations were established at an access point, the catch of fishers leaving the beach was weighed and the number of abalone per catch recorded, giving mean abalone weight per catch. During the 1999 season catch weights were recorded from 1 site per day on days 2, 4 and 6 of the fishery. During 2000 weight measurements were taken from between 3 and 7 sites per day, from a total of 10 sites, or 26 site-day combinations, including the 3 sites sampled in 1999. The relationship of mean abalone weight and site, day and fishing condition was examined using ANOVA of data collected during the 2000 season. The relationship between mean abalone weight and site, year and fishing condition was examined using ANOVA of the data collected from the three sites sampled during the 1999 and 2000 seasons.

Commercial Sector

Commercial catch is reported for each 10 nautical mile section of the Western Australian coast from Shark Bay to the South Australian boarder. Catches in whole weight are reported on a daily basis as part of the quota management.

RESULTS

Effort--High Usage Zone

Linear regression between log effort and log 710 count, for the 187 research and VFLO samples combined, for 1997 to 1999 data, indicated a strong positive relationship ([R.sup.2] = 0.854, P < 0.00l). For scenario a, where a 710 count is available, an analysis of variance with log effort as the dependent variable, was used to consider the effect of the data source (research or VFLO samples) together with log 710 count, and other variables year, day, count area and fishing condition. The data source was significant (P = 0.028), but accounted for only 0.2% of the sum of squares after accounting for the effect of the other variables. This is a minor influence on the model where the main effects account for 92.3% of the total sum of squares. There was no significant first order interaction between the data source and the other five explanatory variables. Therefore, the research and volunteer samples were pooled for further analysis. The main effect of year was not significant (P = 0.554) and there was no interaction between year and log 710 count (P = 0.142). Therefore, year was dropped from further analyses. Day and count area were significant as main effects (P < 0.001). Fishing condition proved to have important interactions and was retained. The ANOVA was repeated after removing data source and year from the model. The main effects and the most influential first order interactions are shown in Table 3. For this model the interactions of log 710 count with day, area and fishing condition account for 1.4%, 1.3% and 0.4% of the total sum of squares respectively. This model (Table 3) was used to estimate missing effort values for areas with only the 710 count. The effect of day was positive and significant, indicating that as the season progressed, and abalone were progressively fished down, the effort required for a given number of fishers to obtain the daily catch increased. The effect of fishing condition was also positive and significant, indicating that as fishing conditions deteriorated the average individual fisher spent more time on the reef searching for their catch. This is because abalones tend to occupy the area where waves break, on the seaward section of the reef platform, and it becomes more difficult to fish as the waves become larger. This difficulty is compounded by an increase in the depth of water over the reef platform (tide height), because the majority of fishers wade without mask, snorkel or weights.

For scenario b, the ANOVA model of the five likely influences on log effort without a 710 count showed that data source was not significant as a main effect (P = 0.051) and accounted for only 0.3% of the total sum of squares. There were no significant first-order interactions between data source and the other explanatory variables. Therefore, research and volunteer samples were pooled. All other variables were significant as main effects (P < 0.001). For the remaining variables (year, day, count area and fishing condition) the dominant first-order interactions are year x count area and fishing condition x day (Table 4). These interactions account for 3.7% and 1.5% of the total sum of squares, respectively. This model (Table 4) was used to estimate the effort for areas with no counts for a given day. For this model effort increased as the season progressed, indicating that fishers continued to participate despite declining numbers of legal-sized abalone later in the season, and the time spent obtaining the catch increased after other factors were accounted for. The relationship with fishing condition was negative, indicating that on average as fishing conditions deteriorated there were fewer people fishing for the day, although the average fisher that did participate fished for longer to obtain their catch (scenario a).

Inclusion of year in an interaction term in the model means that it is not possible to estimate effort for a count area for any year in which there were no counts on any day during that year. For these area-year combinations, effort has been estimated using the main effects only, which still explain 87.7% of the variation. The same relationship was used to estimate the effort values for the 11 area-year combinations with no count information for 2000 to 2003.

Extrapolation of effort from the count areas per section, to the total for the high usage zone of each section, was done using the ratio of the total number of fishers, to the number within the count areas for each section. This ratio was first available from day five of the 2000 season, and was applied to the estimates for 1997 to 2000, the ratio used for extrapolation in subsequent years was available from sampling in that year (Table 5). Extrapolated estimates of effort over the total of the high usage reefs increased from 11,336 fisher-hours in 1997, to 19,216 fisher-hours in 2003 (Table 6), an increase of 70% over the 7 y.

Effort--Low Usage Zone

The mean percent of the effort from the research areas, that was attributed to each 10 or 15 min count period, ranged between 22.1% for the 710 count to 6.1% for the count at 8:20 AN (Table 7). The relationship was used to estimate the proportion of the effort for a fishing day that was described by each instantaneous count in the low usage zone. The estimates of effort indicate negligible fishing activity in sections 33, 34, 35 and 42. Between 1997 and 1999 the effort in the low usage zone decreased slightly (888-565 h), before beginning a rapid, steady rise to 1910 h in 2003. The increase in effort in the low usage zone has been most notable in sections 41 and 43 toward the southern extent of the fishery.

The total effort for the high and low usage zones combined is dominated by section 38, which received between 43% (2000) and 55% (1998) of the total effort during the study period, averaging 49%. Section 39 received between 32% (2003) and 44% (2000), averaging 38%. Sections 38 and 39 combined accounted for between 85% (2003) and 90% (1999) averaging 87%. The high usage zone accounted for an average of 94% of the effort during the 7-year study period (Table 6) varying from 91% (2003) to 97% (1999).

Catch Rate and Catch

Analysis of variance showed that year, day, section and fishing condition were significant main effects influencing catch rate between 1997 and 1999. Because of the large sample size (4,025) many of the factors and interactions are likely to be significant. Fishing condition accounted for 10.1% and day accounted for 3.1% of the total sum of squares. Section and year each accounted for less than 1% of the total sum of squares. When the most influential interaction terms are added to the model the fit to the data is slightly improved ([R.sup.2] = 0.190), with fishing condition x day accounting for 2.9% of the total sum of squares (P < 0.001), and day x section 1.4% of the total sum of squares (P < 0.001). Given the minor influence of 10 nm section on the catch rate, and the large number of interviews, the respective daily mean catch rate was considered appropriate for estimating catch in number of abalone, on a daily basis, from the estimates of effort. The average catch rates varied between 16 and 32 abalone per hour for the days fished in 1997 to 2003.

A summary of the weight data collected between 1999 and 2003 is provided in Table 8. ANOVA results for the three sites sampled in 1999 and 2000 showed that fishing condition did not have a significant effect on mean weight. Site and year were both significant as main effects with no significant interaction between the two. The model of site and year as main effects explains 28.8% of the observed variation with 5.1% being because of site and 23.8% because of year.

Results for ANOVA of the 2000 weight data indicate that fishing condition is only significant in its interaction with site (P = 0.034) when modeling mean abalone weight against site, fishing condition and the interaction of site x fishing condition. In this case fishing condition accounts for only 1.2% of the total variation explained (26.7%) and is not considered important. When the model of day, site and day x site is considered all terms are significant. The model explains 29.0% of the observed variation with 2.7% explained by the interaction of site and day, 2.7% explained by day alone and 23.6% explained by site. Examination of the mean abalone weight from all sites measured per day indicates the site x day interaction is most strongly influenced by an unusually low mean weight from Cottesloe on day 6, a mean from a relatively small sample (9 catches), which, together with the small percentage of error explained, suggest that the interaction is not of great importance and that the mean weight of abalone is predominantly influenced by site.

To account for the few sites sampled in 1999, the ratio of the 2000 value for the 3 sites that were measured in both years, over the value for all sites measured in 2000, has been used to derive the value that would have been expected if all sites were measured during the 1999 season. These seasonal means have been used to convert catch in number of abalone to catch in kilograms whole weight for 1999 to 2003 (Table 9). As no mean weight information is available for 1997 and 1998 the average of 1999 and 2000 (104.6 g) has been assumed for the conversion of the recreational catch in number of abalone to whole weight for 1997 and 1998.

The low usage zone of the Perth recreational abalone fishery represents a minor portion of the total recreational catch (Table 9), averaging only 6.3%, of the total number of abalone taken from 1997 to 2003. Total recreational catches estimated in number of abalone increased from 1997 to 2003 with a minor decrease in 2000 because of a decrease in the estimated catch from section 38 of over 50,000 abalone. The recreational catch is taken primarily from two sections, 38 and 39, which averaged 87.2% of the catch over the 7 years (85.0% to 89.5%). Catches from both sections 38 and 39 showed an upward trend over the 7-year period, with section 38 averaging 49.3% and section 39 averaging 37.9% of the total recreational catch from the Perth fishery measured as the number of abalone taken.

The effect of the variation in mean abalone weight, on the catch estimates in whole weight, is to accentuate increases in catch between 1997 and 1999 and to accentuate decreases between 1999 and 2000 (Table 9). The total recreational catch estimate increased by 36.0% between 1997 and 1999, then decreased by 34.5% between 1999 and 2000. This was driven by an increase in catch of 39.1% for section 38 and 48.0% for section 39 between 1997 and 1999, and a subsequent decrease in catch of 45.0% for section 38 and 23.8% for section 39 between 1999 and 2000. The total recreational catch increased by 51.9% between 1997 and 2003.

The telephone survey that began in 1999 (Hancock et al. 2003) provides an independent estimate of the recreational Roe's abalone catches from the Perth fishery as a whole. These estimates compare favorably with the results of the present field survey, with field survey results falling within the 95% confidence intervals of the telephone survey estimates (Fig. 4). The independent estimates corroborate the increasing trend in the recreational catch of Roe's abalone from the Perth fishery.

Combined Commercial and Recreational Catch

The commercial sector has caught its quota from the Perth fishery in recent years, the majority of the catch (97% to 99%) coming from the 10 nm sections 38, 39 and 41 during this study (Table 9). Catches from sections 38, 39 and 41 averaged 40%, 20% and 38% of the total commercial catch from the Perth fishery over the seven years. Nearly all of the commercial catch from section 41 comes from the low usage portion, an area of very low recreational catch.

The commercial catches are atypical for 1998 because of a one-off reduction of the commercial quota to 24 t for the 6-month period from October 1998 to March 1999 inclusive. The commercial catch from section 38 increased during 2001 to 2003, whereas the catch from section 39 decreased during 2000 to 2003, with any difference being made up from section 41.

The commercial and recreational catches are not comparable for 1998 because of the shortened commercial season. Combined commercial and recreational catches have increased 25% between 1997 and 2003, going from 69.5 to 86.6 t respectively (Table 9). The recreational catch from the Perth fishery has increased from 47.9% to 58.4% of the combined recreational and commercial catch between 1997 and 2003, although the proportion has varied during that period. Section 38 is the most productive of the overall fishery, producing an average of 45% of the annual catch, with section 39 producing an average of 29% and section 41 producing an average of 18%.

DISCUSSION

The cost of survey personnel has commonly been a major limitation on the design of creel surveys to estimate catch and effort in recreational fisheries (e.g., Jones et al. 1990, Newman et al. 1997). In this survey the use of numerous volunteers has largely overcome this limitation. The use of volunteers introduces other limitations, the primary one being a reduced ability to prescribe the areas surveyed on a daily basis. This limitation has been addressed largely by the use of predictive GLM models to estimate the missing effort values. The use of these estimation techniques was made practical by the highly predictable patterns of recreational fishing effort within the high usage zone areas counted, with the 710 count proving to be an effective indicator of effort ([R.sup.2] = 0.951, Table 3). The estimation of effort in areas for which no counts were available was also achieved with a high degree of confidence, with [R.sup.2] values of 0.929 and 0.877 for the two models used to estimate missing effort values (Table 4).

Creel surveys have been conducted previously using roving or access point methods, requiring stratification of the fishery by time and space (Malvestuto & Knight 1991). The structure of the Perth abalone fishery makes it possible to sample the entire fishing area of the low usage zone for one third of the fishing days, requiring the assumption that the percent of effort observed during this period in the low usage zone is representative of the remaining 4 of the 6 fishing days. It is possible to obtain measures of effort from the whole fishing period for the high usage zone, leaving space as the dimension requiring stratification. This study uses direct observation of effort from 60% of the high usage zone. Extrapolating effort from count areas to total high usage zone per section, using the ratio of the total number of fishers in the high usage zone, to the number within the combined count areas from only 1 day assumes that the ratio for that day is representative of all fishing days considered. This assumption deserves further examination in the future.

The method of estimating effort also assumes that all individuals on the reef were participating in the abalone fishery. Effort may be over estimated if there were a significant number of people that were present on the reef that were not intending to collect abalone, and subsequently not interviewed later. Evidence that this was not the case is that very few people refused an interview, and that on each day there were interviews of people that had caught zero abalone. Because Roe's abalone tend to occupy the surf zone, it is unlikely that large numbers of people would venture into this environment without reason. The occasional angler present on the reef was easily identified, and not included in the count of abalone fishers.

An advantage of this method of catch estimation, over telephone or mail surveys, is that the estimate takes into account catches by unlicensed fishers on the reef, or by fishers who have returned to fish a second quota for the day. This component of illegal fishing is captured by the direct observation of effort. Much of the illegal take of more than the prescribed bag limit is also accounted for by using direct observations of catches. Other illegal activity such as out of season poaching remains unaccounted for in this study.

The low usage zone contributed between 3.5% and 9.0% of the recreational effort, and aerial surveys provide a cost effective technique for estimating this proportion. For the low usage zone, conversion of instantaneous counts to effort assumes that the distribution of fishing effort throughout the fishing period in this zone is similar to the average distribution of effort in the research areas counted in the high usage zone. Fishing is allowed throughout the fishing day in areas more than 800 m offshore. Although the amount of fishing that is conducted outside the core fishing time in offshore areas of the low usage zone is known to be low, this provides scope for an underestimation of effort. Areas affected are the low usage reefs of section 40, 42 and section 41 for 1997 only. The offshore areas of section 41 were brought under the mainland regulations in 1998. This change did not produce any discernible impact on the estimates of effort for this area, so any effects on estimates of total effort are likely to be minor.

Within the recreational sector the fishing competence and the confidence of individuals in the water vary considerably. This is reflected in the range of individual catch rates on any day, with some fishers returning to the beach with few abalone for all but the best fishing conditions and some fishers taking their bag limit of 20 abalone in as little as 5 min. It is therefore necessary to calculate mean catch rates from a large number of interviews. For this study catch rates have been calculated from between 98 and 414 interviews per day. Catch rates for the Perth recreational Roe's abalone fishery are high compared with other abalone fisheries (e.g., Lyle & Smith 1998, Worthington et al. 1998), and fishing times are low, reflecting the high densities of this species.

The recreational catches have been converted to estimates of whole weight for comparison with commercial catches. The consistent and highly significant change in the mean weight of abalone taken in the recreational fishery between 1999 and 2000 highlights the possibility that assuming a mean weight for the 1997 and 1998 seasons, on the basis of measurements made in 1999 and 2000, may introduce an appreciable bias in the estimation of catch weight for these years. Trends in catch indicated by number of abalone taken would be more reliable for these years, and comparisons with the commercial catch for 1997 and 1998 should be interpreted with added caution. The mean weights for 2000 to 2003, which have been sampled from at least 9 sites have been relatively consistent varying from 90 to 96 g. The high mean weight in 1999 may have been affected by the low level of sampling in the first year when mean weights were sampled.

The recreational catch from the high usage zone is highly dominated by sections 38 and 39, and there was a substantial decrease in fishing effort in 2000 (Table 6), particularly evident in section 38. Reasons for this decrease in effort are likely to include fishing conditions that were more challenging.

The nature of recreational catch estimates means that they are generally based on a set of assumptions that are difficult to test. This led Pollock et al. (1994) to conclude that the only real test of the accuracy of these estimates is corroboration from two or more independent estimates. Given the time and cost involved in estimating recreational catches, the availability of two independent estimates is unusual. The telephone survey was used to estimate recreational abalone participation rates, fishing effort and catch for the three exploited species of abalone in Western Australia. A stratified random sample comprising about 800 of the up to 21,000 abalone license holders were contacted by phone (Hancock et al. 2003). The close agreement between the two independent estimates of the recreational Roe's abalone catch (Fig. 4) greatly increases the confidence that each represents a valid estimate of the catch for this fishery.

In no other abalone fishery in the world is there such a focus of fishing activity adjoining the suburban area of a major city. Also, most abalone species throughout the world are predominantly sub-tidal and occur in lower densities, meaning that fishers require a higher level of expertise and equipment to access them. The more diffuse nature of most abalone fisheries tends to favor mail or telephone surveys as a means of estimating the recreational catch and effort (Weithman 1991) where recreational fishing license information can be used to target fishers.

Abalone fisheries throughout the world are exploited to different extents by the commercial and recreational sectors, but the recreational catch, where it has been estimated, is generally an important component of the legal catch, with recreational fishing effort tending to increase (Teirney et al. 1997, Anon. 1998, 1999, Bradford 1998, Lyle & Smith 1998, Worthington et al. 1998). The combination of their high value and accessibility makes it important to monitor catch trends in these fisheries.

In the Perth fishery the Roe's abalone catch is high, reaching 42t for a 10 nm section of coast, and concentrated in just 30 nm (sections 38, 39 and 41). Of these, section 41 is nearly exclusively fished by the commercial sector. Sections 38 and 39 are the most heavily exploited, with the recreational sector taking between 54% and 72% of the catch from these sections during the 7-year period of this study. Examination of the population structure of Roe's abalone has demonstrated that individual stocks are separated by less than 13 km of coast (Hancock 2000). This scale is well represented by the patches of semicontinuous reef within each 10 nm section of the Perth fishery. The present results provide catch and effort information on a scale of the individual biological stocks within the fishery. Any stock assessment for this fishery should be conducted on the same spatial scale.

The recreational catch increased 65% during the period of the study, and now exceeds the commercial catch, whereas the commercial catch is fixed via quotas. This study provides the catch and effort information required to manage this intensively used fishery on the scale of the biological stocks that make up the fishery. This has been achieved despite the fine scale spatial subdivision of individual stocks.

ACKNOWLEDGMENTS 

The authors thank T. Baharthah and N. Sumner for permission to quote results of their telephone survey. This project was supported by funding from the West Coast Abalone Divers Association through the Department of Fisheries WA, the recreational fishing sector through the Recreational Fishing Advisory Committee and from the Commercial and Recreational fishing programs of the Department of Fisheries WA. This work was completed as part of the PhD thesis of B. Hancock, who thanks Drs. M. S. Johnson, R. Black and L. Joll who supervised this thesis. Field assistance was provided by numerous staff of Fisheries WA and volunteers, particularly the Volunteer Fisheries Liaison Officers, without whose help the surveys would not have been possible.

LITERATURE CITED 

Anon. 1998. Research Highlights 1997/98. Department of Environmental Affairs and Tourism South Africa, Sea Fisheries Research Institute, Publication No. 7. 68 pp.

Anon. 1999. Research Highlights 1998/99. Department of Environmental Affairs and Tourism South Africa, Marine and Coastal Management, Publication No. 8. 84 pp.

Baskerville, G. L. 1972. Use of logarithmic regression in the estimation of plant biomass. Can. J. For. 2:49-53.

Bradford, E. 1998. Harvest estimates from the 1996 national marine recreational fishing surveys. New Zealand fisheries assessment research document 98/16.27 pp.

Caputi, N., J. S. Edmonds & D. I. Heald. 1979. Mercury content from shark from South-Western Australian waters. Mar. Pollut. Bull. 10:337-340.

Guzman del Proo, S. A. 1992. A review of the biology of abalone and its fishery in Mexico. In: S.A. Shepherd, M. J. Tegner & S. A. Guzman del Proo, editors. Abalone of the world: biology, fisheries and culture. Oxford: Fishing News Books. pp. 341-360.

Hancock, B. 2000. Genetic subdivision of Roe's abalone, Haliotis roei Grey (Mollusca: Gastropoda), in south-western Australia. Mar. Freshwater Res. 51:679-687.

Hancock, B. 2004. The biology and fishery of Roe's abalone Haliotis roei Gray in south-western Australia, with emphasis on the Perth fishery. Ph.D. thesis, University of Western Australia, School of Animal Biology, Perth Western Australia.

Hancock, B., T. Baharthah & N. Sumner. 2003. Licensed recreational abalone fishery status report. In: J. W. Penn, editor. State of the Fisheries Report 2002/2003. Department of Fisheries Western Australia: Perth.

Joll, L. M. & J. W. Penn. 1992. Australian abalone fisheries and their management: Western Australia. In: S. A. Shepherd, M. J. Tegner & S. A. Guzman del Proo, editors. Abalone of the world: biology, fisheries and culture. Oxford: Fishing News Books. pp. 417-419.

Jones, C. M., D. Robson, D. Otis & S. Gloss. 1990. Use of a computer simulation model to determine the behavior of a new survey estimator of recreational angling. Trans. Ant. Fish. Soc. 119:41-54.

Karpov, K. A., P. L. Haaker, D. Albin, I. K. Taniguchi & D. Kushner. 1998. The red abalone, Haliotis rufescens, in California: importance of depth refuge to abalone management. J. Shellfish Res. 17:863-870.

Kearney, R. E. 1995. Keynote address. In: D. A. Hancock, editor. Recreational fishing: what's the catch. Australian Society for Fish Biology Workshop Proceedings, Canberra August 30-31, 1994. Australian Society for Fish Biology, Canberra. pp. 10-23.

Lyle, J. M. & J. T. Smith. 1998. Pilot survey of licensed recreational sea fishing in Tasmania--1995/96. Tasmania Dept. of Primary Industry and Fisheries: technical report No. 51. 54 pp.

Malvestuto, S. P. & S. S. Knight. 1991. Evaluation of components of variance for a stratified, two-stage roving creel survey with implications for sample size allocation. Am. Fish. Soc. Symp. 12:108-115.

Newman, S. P., P. W. Rasmussen & L. M. Andrews. 1997. Comparison of a stratified, instantaneous count creel survey with a complete mandatory creel census on Escanaba Lake, Wisconsin. N. Am. J. Fish. Manage. 17:321-330.

Parker, D. O., P. L. Haaker & H. A. Togstad. 1992. Case histories for the three species of California abalone, Haliotis corrugata, H. fulgens and H. cracherodii. In: S. A. Shepherd, M. J. Tegner & S. A. Guzman del Proo, editors. Abalone of the world: biology, fisheries and culture. Oxford: Fishing News Books. pp. 384-394.

Pollock, K. H., C. M. Jones & T. L. Brown. 1994. Angler survey methods and their applications in fisheries management. Bethesda, MD: American Fisheries Society special publication 25. 371 pp.

Preece, P. A. G., S. Mayfield & T. M. Saunders. 2003. Biology and feasibility fishing of Roe's abalone (Haliotis roei). Final report to the Abalone Industry Association of South Australia. South Australian Research and Development Institute (Aquatic Sciences): West Beach, South Australia. Publication No. RD04/0024.

Shepherd, S. A. 1973. Studies on Southern Australian abalone (Genus Haliotis). 1. The ecology of five sympatric species. Aust. J. Mar. Freshwater Res. 24:214-257.

Tarr, R. J. Q. 2000. The South African abalone (Haliotis midae) fishery: a decade of challenges and change. (abstract only). J. Shellfish Res. 19: 537.

Tarr, R. J. Q., A. J. Williams, A. J. Mackenzie, E. Plaganyi & C. Moloney. 2000. South African fishery independent abalone surveys. (abstract only). J. Shellfish Res. 19:537.

Tegner, M. J., J. D. DeMartini & J. D. Karpov. 1992. The California red abalone fishery: a case study in complexity. In: S. A. Shepherd, M. J. Tegner & S. A. Guzman del Proo, editors. Abalone of the world: niology, fisheries and culture. Oxford: Fishing News Books. pp. 370-383.

Teirney, L. D., A. R. Kilner, R. B. Millar, E. Bradford & J. D. Bell. 1997. Estimation of recreational harvests from 1991-1992 to 1993-1994. New Zealand Fisheries assessment research document 97/15.43 pp.

Weithman, A. S. 1991. Telephone survey preferred in collecting angler data statewide. Am. Fish. Soc. Symp. 12:271-280.

Wells, F. E. & J. K. Keesing. 1989. Reproduction and feeding in the abalone Haliotis roei Gray. Aust. J. Mar. Freshwater Res. 40:187-197.

Wells, F. E. & J. K. Keesing. 1990. Population characteristics of the abalone Haliotis roei on intertidal platforms in the Perth metropolitan area. J. Malac. Soc. Aust. 11:65-71.

Worthington, D. G., R. C. Chick, P. A. Blount, Brett & P. Gibson. 1998. A final assessment of the NSW abalone fishery in 1997. NSW Fisheries fishery resource assessment series, 5. 74 pp.

BOZE HANCOCK (1,2) AND NICK CAPUTI (1)

(1) Department of Fisheries, Western Australia Fisheries and Marine Research Laboratories, PO Box 20, North Beach, Western Australia 6920, Australia; (2) Department of Zoology, University of Western Australia, Netherlands, Western Australia 6907, Australia

Corresponding author. E-mail: [e-mail]

TABLE 1.

Environmental measures included in an index of fishing conditions
for the Perth recreational Roe's abalone fishery. Category values
are summed across measures (columns), by choosing the
corresponding value for the category from each column to give an
overall index on a scale of 0 to 9. Wave height is the mean of two
recordings at 7:00 and 8:00 AM, tide height is the mean of four
half-hourly recordings taken from 7:00 AM, cloud cover refers to the
mean of two observations at 7:00 and 8:00 AM, and rainfall refers to
the period from 6:00 to 9:00 AM. Category values for the four
environmental variables are summed to derive the fishing
condition index.

               Wave       Tide
Category      Height     Height     Cloud
 Value         (m)        (cm)      Cover     Rainfall

   0        0.00-0.30    to 30     0/8-6/8       Nil
   1        0.31-0.60    31-60     7/8-8/8      Rain
   2        0.61-0.90    61-90        --         --
   3        0.91-1.20     >90         --         --
   4          >1.20        --         --         --

TABLE 2.

Number of effort counts of different types for 1997-2003.

        Research      Volunteer
         Counts        Counts         Count at        Number of
Year    8 per Day    3-5 per Day    7:10 AM Only    Missing Counts

1997       17            33              27               43
1998       12            52              24               32
1999       12            61               7               40
2000                     44              14               62
2001                     73               5               42
2002                     65               6               49
2003                     39              12               69

TABLE 3.

ANOVA model of the dependent variable log effort used to estimate
effort for areas with only one 710 count (scenario a).

                                            Mean
             Source                  df    Square    F Ratio       P

log 710 count *                       1     0.846     104.5     <0.001
Day *                                 1     0.185      22.8     <0.001
Area counted                         19     0.015       1.9      0.024
Fishing condition *                   1     0.041       5.0      0.026
[log.sub.10] 710 count x day          1     0.120      14.8     <0.001
[log.sub.10] 710 count x area
  counted                            18     0.016       2.0      0.013
[log.sub.10] 710 count x fishing
  condition                           1     0.090      11.1      0.001
Residual                            144     0.008

Asterisk indicates covariates. [R.sup.2] = 0.951.

TABLE 4.

ANOVA model of log effort, used to estimate effort for areas with
no 170 count, for areas with at least one instantaneous count taken
each year (scenario b).
                                             Mean
        Source              df    Square    F Ratio       P

Year                         2     0.260     19.6      <0.001
Day                          4     0.033      2.5       0.045
Area counted                19     0.771     58.1      <0.001
Fishing condition *          1     0.023      1.7       0.191
Day x fishing condition      4     0.081      6.1      <0.001
Area counted x year         28     0.031      2.3      <0.001
Residual                   127     0.013

Asterisk indicates covariate. [R.sup.2] = 0.929.

TABLE 5.

Percentage of the total number of fishers that were within the count
areas used to estimate effort for the high usage zone of each section
of the Perth recreational Roe's abalone fishery.

                   % in Count Areas

           2000,    2001,    2002,     2003,
Section    Day 5    Day 2    Day 2     Day 2

  37       100.0    100.0    100.0     100.0
  38        62.9     64.9     67.1      63.6
  39        90.4     93.3     86.8      94.7
  40        75.0     70.0    100.0    Closed
  41        75.0    100.0     87.5      64.3

TABLE 6.

Estimates of effort (fisher hours) for the low (LUZ) and high (HUZ)
usage zone portion of each 10-nm section of the Perth recreational
abalone fishery. Effort is from the aerial survey counts on the days
indicated. LUZ % is the mean of the aerial survey effort expressed as a
percent of the total effort for the HUZ count areas for that day. LUZ
effort is the sum of the two effort values derived from aerial surveys
and the effort for the remaining 4 days estimated as a proportion (mean
%) of the total effort for the HUZ count areas per day. HUZ is the
total extrapolated HUZ effort per section, and Total is the sum of the
LUZ and HUZ effort.

                     1997                         1998
 Year
Section     LUZ      HUZ      Total      LUZ      HUZ      Total

  33          32                  32        0                   0
  34           0                   0       67                  67
  35          66                  66        0                   0
  36         152                 152      129                 129
  37         175       254       429      188       523       712
  38          40     6,115     6,155       60     8,498     8,558
  39           0     4,311     4,311        0     5,187     5,187
  40          37       483       520        0       504       504
  41         165       172       337       45       173       218
  42          55                  55        7                   7
  43         166                 166      123                 123
 Total       888    11,336    12,223      619    14,884    15,503

                     1999                         2000
 Year
Section     LUZ      HUZ      Total      LUZ      HUZ      Total

  33           0                   0       14                  14
  34          14                  14       10                  10
  35           0                   0        0                   0
  36         120                 120      126                 126
  37          55       605       660      107       723       830
  38         107     8,189     8,296      115     6,693     6,808
  39           0     6,243     6,243        0     7,042     7,042
  40          49       404       453       41       316       357
  41          98       245       343      240       165       405
  42           6                   6       27                  27
  43         117                 117      198                 198
 Total       565    15,686    16,251      877    14,939    15,817

                     2001                         2002
 Year
Section     LUZ      HUZ      Total      LUZ      HUZ      Total

  33           0                   0        0                   0
  34           0                   0       21                  21
  35           0                   0       84                  84
  36          63                  63      146                 146
  37           7     1,037     1,044      130       667       797
  38          25     7,930     7,955      131     8,316     8,447
  39           0     7,392     7,392        0     7,400     7,400
  40          37       192       229       64        61       124
  41         701       118       819      584        90       674
  42          20                  20       45                  45
  43         204                 204      391                 391
 Total     1,059    16,668    17,727    1,594    16,534    18,127

                     2003
 Year
Section     LUZ      HUZ      Total

  33         134                 134
  34           0                   0
  35           0                   0
  36         313                 313
  37         213     1,047     1,260
  38         162    11,008    11,170
  39           0     6,804     6,804
  40          77         0        77
  41         535       357       892
  42          18                  18
  43         459                 459
 Total     1,910    19,216    22,126

TABLE 7.

The mean percentage of effort per area counted per day, described
by each time interval used to calculate daily effort for the research
areas counted from 1997 to 1999. *

Time Interval    Mean % Effort    [+ or -] SE

7:00-7:15 AM         22.30            0.91
7:16-7:25 AM         16.10            0.44
7:26-7:35 AM         15.85            0.37
7:36-7:45 AM         13.62            0.28
7:46-7:55 AM         11.09            0.33
7:56-8:05 AM          8.54            0.40
8:06-8:15 AM          6.31            0.37
8:16-8:30 AM          6.17            0.54

* Standard error. n = 41 (R in Table 3.2).

TABLE 8.

Summary of weigh station data collected from the Perth
Recreational Roe's abalone fishery 1999 to 2003.

                                                              Mean
                                               Number        Weight/
         Areas         Day/Area     Catches      of          Abalone
Year    Surveyed    Combinations    Weighed    Abalone    (Weighted, g)

1999       3                3           133      2,320        * 119.4
2000      10               26           762     14,516           89.8
2001      14               47         1,854     35,838           91.6
2002      11               36         1,702     31,909           89.7
2003       9               29         1,563     29,760           96.4

* Weighted mean from 3 sites sampled, adjusted by the ratio of the mean
weight for those 3 sites to the mean from the 10 sites using the 2000
data.

TABLE 9.

Estimates of recreational catch (kg) in the high (HUZ) and low (LUZ)
usage zones of each 10 nm section, with combined recreational catch
(Rec.) commercial catch (Com.) and the total catch from the Perth
abalone fishery, 1997-2003. Blanks indicate no HUZ, or that these
sections lie beyond the northern boundary of the Perth fishery for the
commercial sector (sections 33 to 35). Catch weights for 1997 and 1998
have been derived using a mean weight of 104.6g obtained for 1999-2000.

                               1997
 Year
Section     HUZ       LUZ      Rec.      Com.     Total

  33                    84        84                  84
  34                     0         0                   0
  35                   178       178                 178
  36                   415       415       436       851
  37          702      475     1,177       354     1,531
  38       16,698      108    16,806    14,184    30,990
  39       11,662        0    11,662     9,925    21,587
40 Is.      1,342      101     1,443       207      1650
  41          454      454       908    11,045    11,953
  42                   150       150         0       150
  43                   455       455         0       455
 Total     30,858    2,421    33,279    36,151    69,430

                               1998
 Year
Section     HUZ       LUZ      Rec.      Com.     Total

  33                     0         0                   0
  34                   166       166                 166
  35                     0         0                   0
  36                   319       319       188       507
  37        1,334      467     1,800         0     1,800
  38       20,973      147    21,121    11,562    32,683
  39       12,812        0    12,812     5,878    18,690
40 Is.      1,235        0     1,235         0     1,235
  41          440      111       551     6,563     7,114
  42                    17        17         0        17
  43                   304       304         0       304
 Total     36,794    1,532    38,326    24,190    62,516

                               1999
 Year
Section     HUZ       LUZ      Rec.      Com.     Total

  33                     0         0                   0
  34                    40        40                  40
  35                     0         0                   0
  36                   341       341        17       358
  37        1,636      159     1,795       269     2,064
  38       22,934      301    23,235    12,242    35,477
  39       17,282        0    17,282     9,259    26,541
40 Is.      1,112      136       248       153       289
  41          685      276       961    13,924    14,885
  42                    16        16         0        16
  43                   337       337       228       565
 Total     43,649    1,607    45,256    36,091    81,347

                               2000
 Year
Section     HUZ       LUZ      Rec.      Com.     Total

  33                    26        26                  26
  34                    19        19                  19
  35                     0         0                   0
  36                   236       236       363       599
  37        1,362      201     1,563       641     2,204
  38       12,573      217    12,790    10,516    23,305
  39       13,178        0    13,178     5,075    18,254
40 Is.        588       77       665         0       665
  41          309      451       760    19,718    20,478
  42                    52        52         0        52
  43                   371       371       197       568
 Total     28,010    1,649    29,660    36,509    66,169

                               2001
 Year
Section     HUZ       LUZ      Rec.      Com.     Total

  33                     0         0                   0
  34                     0         0                   0
  35                     0         0                   0
  36                   158       158         0       158
  37        2,625       18     2,644       258     2,902
  38       19,739       63    19,802    15,430    35,232
  39       18,342        0    18,342     6,232    24,574
40 Is.        478       93       571       210       781
  41          292    1,745     2,037    13,277    15,314
  42                    51        51                  51
  43                   507       507                 507
 Total     41,476    2,636    44,112    35,406    79,518

                               2002
 Year
Section     HUZ       LUZ      Rec.      Com.     Total

  33                     0         0                   0
  34                    40        40                  40
  35                   160       160                 160
  36                   280       280                 280
  37        1,254      258     1,512       491     2,003
  38       16,768      258    17,026    16,038    33,064
  39       14,550        0    14,550     4,305    18,855
40 Is.        125      125       249       761     1,010
  41          177    1,159     1,337    14,370    15,707
  42                    86        86                  86
  43                   774       774                 774
 Total     32,875    3,141    36,016    35,965    71,981

                               2003
 Year
Section     HUZ       LUZ      Rec.      Com.     Total

  33                   320       320                 320
  34                     0         0                   0
  35                     0         0                   0
  36                   749       749        74       823
  37        2,521      511     3,032       266     3,298
  38       26,262      387    26,649    15,668    42,317
  39       16,347        0    16,347     6,054    22,401
40 Is.          0      185       185     1,198     1,383
  41          860    1,280     2,140    12,601    14,741
  42                    42        42                  42
  43                 1,099     1,099       145     1,244
 Total     45,990    4,574    50,564    36,007    86,571