CHANGES IN FISHING EFFICIENCY OF THE POLE . AND-LINE SKIPJACK TUNA FLEET BASED AT SORONG . IRIAN JAYA

Data li'onr tlre fishing operations ol' the pole-and-line fishing fleet basecl in Sorong, Irian .lava. eastet'n Intlonesia are exnnrinerl to assess the status of the fishery in that region. n (ie[er:rl l,ineal Model was usetl to rlerive st:rndardisecl indices of relative vessel efficie'cies for the baitfish trsetl lrer boat day antl t.utru r:aught per boat da.v. 'fhe nrost a;llrroltriate models in ear:h case were Ln (Bait used) = Constant * year + Month * Vessel whit:h exPlained 56'2, of the variation in the bait. usecl per boat rlay an4 Lp (Tuna catch) = Constant + Year + Month + Vessel + Ln(Bait/boat clay) which explainecl b3,2, oI.the catch'eflb|t vitliation for tuna. Palaureters for relative vcssel efficiencies we.e trsed to stlndirrrlise the ler:orrled fishine elTort. Br' 1992. lishing effort (boat tln.ys) for baitfish and tuna harl increase,l b.y approximatel.y 800'|i' sitr<'e t,ht: start of'the fishelv in l$7(i. Ilelationships between both baitfish tfu".l ar.,,l trrr" <'attght. ittrtl stlttdlrrlised I'ishing effort. were lirrear antl trrna catches were str.ongly de'enclent on the artrotttrt ol'lrait :rvailable. lhe Pole-and-line fleet has exper.iencerl cleclining bo'if 1,"ng" atttl tttnit crtt'hes pet'boat day since 1992. The reasons for.this ar" unclear froru the available rlat'[ but tltaY be related to changes in tl're operntions of thc bagan fishery that suppli"" t oitfisl, ot' a tlecline irr tlre irbtrnrlance of these fish.

The sur:<:ess of pole.-and-linefishing is totallv reliant on a regular nightlv supplv of baitfish which are <:aught in a totally separate fisherv.
Bait{ish are taken at night, in inshore areas by baitfishing units, known locally as bagans, or less commonlv by the pole-and-line vessels themselves using the "basnig" system (See Naamin & (iafa,   this volunre, Ibr a detailcd description of the fish- ing nrethtds, fishing grounds and the spe<;ies r:om- position of' baitfish).
Owners/operators of the bagans receive pay- ment based on the number of buckets of baibfish that they supply to the pole-trnd-line vessels.The captain of'the pole-and-line vessel records on a logsheet.the number of buckets of baitfish that he receives each night from each bagan.The logsheets are returned to the Usaha Mina office where the number of buckets of baitfish that have been supplied by each bagan is calculated.These clata are cornpiled on a monthly basis.
Usahtr Mina uses the same data source to com- pile details of the number of buckets of baitfish supplied to each pole-and-line vessel per month.ln addition to this, the amount of tuna landed and the number of fishing days undertaken by each pole-and-line vessel during each month are re- corded.

Purpose ofthe Study
The initial aim of this study was to review the historical catch and effort data that hacl been col- lected hy Usaha Mina in order to assess the baitfish stocks.However, the basic assumption of using catch per unit effort (CPUE) data to moni- tor the status of a fishery is that changes in CPTIE Factrlt..v of Fisheries anrl Maline Environrnent.Arrstrirlian M,,rliti,nn College.Austr:Jil accurately reflect changes in the abundance offish in the stock (King, 1995).
The data collected by Usaha Mina did not pro- vide measures of catch or effort that would allow us to meet this assumption.The catch estimates were based solely on the amount of baitfish transferred to the pole-and-line vessel.This figure would be an underestimate of the total catch as: (a) Baitfish are often held in the bagans fot 24-48 hours during which time there would be some mortalitysome estimates suggest that the mortality rate maybe as high as 30% of baitfish that are held in the bagan for a day, Usaha Mina only records the live baitfish transferred to the pole-and-line vessels and the fish that die in the bagans would not be included.
(b) Some baitfish are used for human consumption (Naamin & Gafa, this volume) and not supplied to the pole-and-line vessels.During the season of peak abundance of baitfish the amount of baitfish sold as food can be as high as 50% of the total catch.The baitfish caught for human consumption is not recorded in the Usaha Mina records and would vary between months.
The number of fishing days is used to estimate fishing effort, but the number of lifts of the bagan net during a night would be a more accurate mea- sure of effective effort.Bagans often make more than one haul of their net per night and therefore operational fishing days would be an underesti.mate of the true fishing effort.
The data could not be used to calculate a mea- sure of CPUE that could be used as an index of abundance and therefore it was not possible to assess the baitfish stocks using the available in- formation.

Changes in Fleet Structure
However, the data did allow us to examine the structure of the pole-and-line fleet based at Sorong.The number of pole-and-line vessels operating has increased to 49 in 1995 from an initial fleet size of 19 in 1976.The size composition of the pole-and- line fleet operating from Sorong has also changed (Figure 1).The original fleet was comprised of 30 GT vessels but since 1990 there has been the ad- dition of larger vessels.Kimura (1981) recognised that fishing power generally differs among vessels, and if catch per unit effort is to be proportional to abundance, ef- fort measurements must be standardised.As the size structure of the fishing fleet in Sorong has changed, the relative fishing power is likely to have increased and therefore the units of fishing effort Year Figure 1.The number of pole-and-line vessels by size class (in  would need to be standardised over time to obtain a true indication of increases in fishing effort.
The aim of this paper was to use the available data from Sorong to make comparisons between the amount of baitfish used per day by individual pole-and-line vessels and to assess the relative efficiency of pole-and-line vessels in terms of tuna caught per day.These parameters were used to provide a standardised measure of fishing effort for both baitfish and tuna.

Data Collection
Usaha Mina compiles data for the amount of baitfish used, the amount of tuna landed and the number of fishing days for each pole-and-line vessel on a monthly basis.These data were entered into a Microsoft Access database.Each record in the database contained the name of the vessel operating from Sorong, the month and year ofoperation of that vessel.and the number of buckets of baitfish used, the amount of tuna landed and the number of fishing days for the that vessel during that particular month.From these data it was possible to generate the amount of baitfish used per fishing day and the amount of tuna caught per fishing day for each vessel for each month.

Data Analysis
Analysis of catch and effort data Annual and monthly summaries of number of buckets of baitfish used, the amount of tuna landed and the number of fishing days were generated using the query routine within Microsoft Access.
Vessel comparisons of bait used per day and tuna catch per day The catch rate of vessel at a particular time (U,,), where subscript I refers to time and i refers to the vessel can be written as a statistical model as follows: l-J6 = LJl;.*r,tl .*2i&... X,rot... p9vn.qltzszt... pllnvn ... 4'ed   where U,-is the catch rate obtained by the first vessel class in the first time period, 6, is a factor that is the abundance in year I relative to year 1, 6, is the efficiency ofthe vessel d relative to vessel 1. and is a factor that accounts for the deviation between the observed Q, and the expected value for t and i.
We can obtain a linearised form of the model: log(U6) = log(Uti) + cl 1 log(x1) + cr2 log(x2)+...We can estimate the values of Ur,.o,.or...o,,.0,.0r. and F,, using the Generalised Lindai'tvtoitei (GLM)  (Hilborn & Walters 1992) routine in Systat.The raw data for baitfish used per day (Figure 2) and tuna catch per day (Figure 3) exhibited a log.normal distribution and were normalised using a natural log transformation (Figure 4 for baitfish, Figure 5 for tuna) All analyses were run using the GLM routine in Systat (Version 7).The variables Vessel, Year and Month were all put into the analysis as dummy variables.
In order to assess the best fit for combinations of variables in the model, the adjusted coefficient of determination, f o,,,was calculated from: ., 0-r2)(n-t) r.di= r_=G;r_z where z is the sample size, p'is the number or parameters in the model and / is the coefficient of determination, which is a measure of the con- tribution of the independent variable(s) in the model.
The adjusted coefficient of determination.re ,, removes the impact of degrees of freedom and giv;'3 a quantity that is more comparable than R, over models involving different numbers of parameters.

Y199
The parameters for each pole'and-line vessel were inverse-logarithm transformed to provide a factor that represented the amount of baitfish usecl per day and the amount of tuna caught per dav compared to the performance of tlne of the origi- nal Usaha Mina 30GT pole-and-line vessels opt:r- ating in September 1976.
'lher recorclecl numller of fishing days per month lor each vessel wns nrultiplied by the factor for the corresponding vessel to rlbtain a standardised measure of fishing e{ftrrt.

Bait Caught per DaY
Therre were 611-r4 rnonthlv records of vessels thab had complcte estimtrtes of both baitfish used ancl number of fishing days.'lhese data covered thc' olterations ol'a total of ttf'r different pole-and' line vessels frorn April l97G to October l99l'r.Not all vessels were in operation over this whole pe' riocl as older vessels were decommissioned and new vessels joined the fleet.
'lhere has been tttt inc:rease in the amounb rlf baitfish used per boat day from 1976 to 1991 with a decline after 1991 (F'igure 6).The increased use of baitfish per vc'ssel t:an be atbribute<l to the larger siz,e classes of pole-and-line vessels joining the fleet which trre able to carry larger amounts of baitfish.However the decline since 1991 must be due to a retluced supplv of'baibfish as the size composition of'the fleet htrs not de<:reased during this period.
The total bait used by all vessels plotted against tlre total (uncorrected) number of operational fish' ing days (nominal lishing effort) shows that there lras been an exponential increase in bait used per boat day at the higher levels of fishing effort (Fig-  ure 7).This is due to the fact that the higher lev- els of fishing effort have occurred in the years when the larger Size classes of vessel have joined the fleet.The larger vessels have a greater capacity for carrying baitfish, therefore the amount of baitfish used per boat day has increased over time.
Vessel comparisons for bait used per boat dav The progressive improvement in the quantity of the total variation explained by the models for the amount of bait used per boat day is shown in Table l.The 'best' model including the variables vessel, year and month accounted for about 56% of the variability in the monthly bait used per ves- sel (F= 70.1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1es3 lse4 rees Year F-igtrre 6. 'lhe pumber of buckets ol'baitfish used per day per vessel for pole-and-vessels operating frotrt Sorot'tg frotn 197(i to l99l-r l'here is a linear relationship between r;or- rected annual fishing effort (effective e{T<rrt) and total baitfish used (Figure 8).Since 1976 to 1995 there has been an average of 14.7 buckets of baitfish used per boat day.
The dill'erence between the annual nominal and effective effort for baitlish is shown in Figure 9. Efl'ective fishing effort.fbr tuna inr;rcased bv 784%, from the start of the fishery in 197(i to the ef{brt recorded in l1)t)2.Since 1992 there has been a deiline in both the nominal and effective fishing efTort frrr baitfish.

Tuna Caught per Day
There were 6127 monthlv records of vessels which hld conrplete estimates of both tuna caught, baitfish usecl and number of operational fishing davs.There has been an increase in the amount of tuna r:aught lrer boat dav from lgZG to lggl with a decline al'ter 1991 (Figure 10).
The totzrl bnit used by all vessels plotted against the totzrl (un<;orrected) number of operational fishing davs (nominalfishing effort) shows that there hns been an exponential increase in tuna caught per day at the higher levels of fishing effort (see 'l'able 2. l)aranteters estimated I'rom thc dittrt firr bait veitr attd ntonth.Constant  Vessel comparisons for tuna caught per boat day The progressive impnlvement in the qulntitt' of the total variation in the amount of't,una r:a ught per boat day is shown in Table 4. The'best' tnodcl including the variables vessel, year, month an<l Ln (bait used per boat tlay) accounted for abottl 53tk of the variability in the monthly bait usctl per vessel (F = 62.4,df = 1 1l->, P<0.000I ).'l'hc pir- rameters estimated from the (]LM f or vessel.vear ancl month analysis of'the bait ust:cl pcr dav clirta are given in Tables ll and 6.'l'here is a linear relrrtionship between t:orret:ted anntutl fishing eftirrt (eflective effort) anrl total luna caught (F'igure 12).
Since 1976 to 1991-r therrc has been an averager of 8:15 kg of tuna catrght per boat dav.'l'he amount of bait used per boat rlay was an imlrortant factor in explaining the vari:ttion in the tuna r:atch per boat clav.'fhe p:trameter generabed bv bhe moclel I'or bait per boat, dav was 0.982.'lhis strggests that ttrere is an altnost linear relation- ship between tuna r:aught per boat day and the amount o{'bait used per boat dav (with slowlv de- creasing ratios of tuna caught to bait used for higher levels of bait uscrl pcr dav).The intercept :\2 firr thc model was 0.031 which equates to 3l kg of tuna cansht tor t-.very bur:ket of baitfish used.As onc bucket <ront.ainsappnrximatelv 7 kg of baitfish (Naamin & (]afa, this volume), the tuna to baitfish ratirr is :tpproximataly 4.4.
The difl'erence between the annual nominal and efl'ective eUbrt {br tuna is shown in Figure 13.Ell'e<:tive fishing ef'fbrt for tuna increased by 820% from the start of the fishery in 1976 to the effort rt--corcled in 1992.Since 1!)92 there has been a de- r:line in both the nominal and effective fishing ef- lirrt for tuna.

DISCUSSION
F ishing vessol characteristics such as size, tonlrirge.or spced <:nn in.rpact greatly on catch rates (Hilborn & Walters 1992).The general linear model (Gl,M) provides a very powerful, consistent methocl for examining the effects of vessel differences on trends in tuna and baitfish abundance.
However. if catch rate by a specific vessel type in a specific area is not proportional to abundance, then thc mean cat.ch rates estimated from GLM will not be proportion:rl to abundant:e (Hilborn &   Walters 1992)   1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 19B7 1988 1989 1990 199't 1992 1993 1994 Year l'igurc 1).'l'hc trrrnrber of uncorrected opcrational fishing days (nominal effort) and the number of cort'ucl.e<lopcrational fishing days (eff'ective effort) for baitfish for pole-and-line vessels opt,rtt.ingf'ronr Surong f'rom 197(i to 1991-  1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 Year F'igure l0' Thc'trnrount of tuna caught per clay per vessel for pole-and.vesselsoperating from Sorong frcnr 1976 to lggl-r.There are situations where CPUE do not arx:uvessel characteristics of the fleet have changed and rately reflect changes in the abundance of the fish so therefore has the capacity to take baitfish and stock.These situations usually relate to the meathe catch of tuna per day.This is exactly the situ' sure of fishing effort that is recorded and the way at,ion when CPUE would not accurately reflect fishing effort may actually be changing ber:ause of' <:hanges in abundance of the fish stock.increasing efficiency (King 1995).'[.he estimates of catch rate of baitfish for each In Sorong, the measure of fishing eflbrt is the lxrle-and-line vessel based at Sorong is not pro- operational fishing day.However over time the ;xrrtional to fish abundance for the reasons already Effective effort (boat days) F-igure 12.'l'healpunl antountoftunacaughtbypole-and-linevesselsoperatingfromSorongversus the annual nurnber of correct,ecl operational tishing days (effective effort) by these vessels I'ronr 1!)7G to l99l-r.1978 1979 1980 1981 1982 1983 1984 1985 1986   Year '1990 1991 l'igure I ll.The nnrnber ol'uncorrectecl operational fishing days (nominal effort) and the number of r:orrectetl eperlt,ignal fishing clavs (ef'{'ective effort) for tuna ftrr pole-and-line vessels operating {ronr Sorotrg ltonr 197(i to 1995. with differences in the use ol'bait between poleand-line vessels.However.this is important in order to understand the changes in demand for baitfish hy the pole-and-line fleet and standardising {ishing effort over time. The results show that the increased demand f'or baitlish over time is primarily due to the lurger size classes of vessel that are entering the pole- and-line fleet based at Sorong.However there must be a con(:ern that the amount of bait usecl by eleh pole.ancl-linevc.ssel reached a maximum in l99l and eince then has been declining.The amount of baitfish used per fishing da.y by each individual pole-and-line vessel will be inlluenced by: (a) The eapacity of the live-bait tanks on the pole- and-line vesselthe larger the vessel the larger the bqit tank capacity and therefore the greater the demand for baitlish by the pole-and-line vessel.In most cireumstances the pole-and-line vessel captain will try to maximise the amount ol'baitlish that is carried before steaming to the tuna fishing grounds.
(b) Availability o{'baitfish to the pole-and-line vessels which will be dependent on: (i) the abundance of baitfish at anv particular time.ancl (ii) the amount of baitfish supplied bv the bagans.This will be influenced by the number of bagans that are operating and the demancl ltrr bait{ish b.v other users of the resoun)es e.g.dried salterl fish, If it becomes unprolitable to supply baitfish to the pole-and-line fleet.bagan operators will either stop working or look {trr better prices fot the baitlish.If this situation (xi(:urs.then the pole-and-line vessels will not be able to take their maximum requirernent of baitfish.In this situation, low usage of baitfish will not be an indication of low abundance nf baitfish but the inabilitv of the bagan operators to meet the demand frrr baitfish b.y the pole-and-liner fleet.
IFR Journal Vil.IV No.2. 1998 'lherefbre the decline in bait used per boat day since l99l could be attributable to a decline in abundance of the baitfish but could also be to due a reduced capacity of the bagans to supply the baitlish.There may even have been a conscious decision bv pole-and-skippers to reduce the amount ol'baitfish they carry as higher stocking densities within the bait tanks may have increased the mortalitv of baitfish while steaming to the tuna fishing grounds.Any of these options or combina- tion of'options could be a possible reason for the rlecline in the amount of baitfish taken per day.
Only with rnore detailed information would it pos- sible to make the correct conclusions.
F'rom the analysis of the tuna caught per day it is apparent that there is a strong link between baitfish used per day and the amount of tuna r:aught per dav.The declining amount of bait used per day'since 1991 has led to reduced catch rates of tuna per day from 1.64 tonnes per day in lggl to 1.05 tonnes per day in 1995.
This trend must be a concern for Usaha Minu and pole-and-line vessels working in the fishery.
It is important that the reasons for these trends are identified as soon as possible.Without accu- rate catch and effort data recorded from each of the baitfishing locations it is impossible to detect whether there is a declining abundance of baitfish at any of the baitgrounds commonly used by the pole-and-line fleet.
It is therefore imperative that accurate records ofactual catch oftuna and baitfish are kept.These records must include baitfish used for human con- sumption.baitfish that die in the bagans before being transferred to the pole-and-line vessels and actual baitfishing (hauls of the bagan net) are re- corded from each location that supplies baitfish.Only then will it be possible to assess the abun- dance of baitfish based on fishery-dependent data.Without accurate catch and effort data.the abun- dance of baitfish at each baitground can only be assessed by fisheries independent techniques such as the egg production method (Milton et al., this  volume).These techniques can only provide an estimate of baitfish abundance at the time of sam- pling and will not show trends over time unless the sampling is conducted on a regular basis. ;lg

Figure 2 .
Figure 2. Frequency distribution of monthly values of baitfish used per boat day for each vessel from 1976 to 1995.

Figure
Figure I l ' l'he annual antount of tuna caught bv pole-and-line vessels operating from sorong versus the annual number of uncorrected operational fishing .lays(nominal effort) by these vessels from lg76 to lg9l.r.
. the GLM can only provide us gross tonnage) operating from Sorong from 1976 to 1995.
Variables such as year, month and ves.sel are all categorical variables. .