Planning annual allocation of fisheries surveillance effort

Planning annual allocation of fisheries surveillance effort

Fisheries Research 23 (1995) 345-360 Planning annual allocation of fisheries surveillance effort Harvey H. Millar Finance and Management Science Depa...

1MB Sizes 0 Downloads 13 Views

Fisheries Research 23 (1995) 345-360

Planning annual allocation of fisheries surveillance effort Harvey H. Millar Finance and Management Science Department, Saint Mary’s University,Hal&x, N.S. B3H 3C3, Canada

Accepted 14 September 1994

Abstract Fishery surveillance on both of the Atlantic and Pacific coasts of Canada costs the Canadian taxpayer several millions of dollars annually. Few of these dollars are recovered through tines from successfully prosecuted cases involving fishery violations. Consequently, federal fishery surveillance programs are not self-supporting-they are subsidized. In order to reduce the cost of the program to the Canadian taxpayer, it is necessary to implement a cost-effective program which maximizes the deterrent effect of surveillance effort. In this paper we present a tactical linear programming for effectively allocating annual surveillance effort to monitor and control fishing activity, and to enforce tishery regulations in accordance with resource management plans and objectives for the offshore fishery. In particular, we focus on the Atlantic fishery. The tactical model determines on an annual basis, for each of the surveillance units, the amount of effort allocated to each fishery/fishing zone in each time period while maximizing effectiveness. Once the effort allocations are made, operational tasks which involve scheduling surveillance/patrol units on a short-term basis in accordance with the annual plan and current priorities must be completed. We demonstrate the potential use of the planning model. Keywords: Surveillance effort

1. Introduction The Canadian government spends millions of dollars on fisheries surveillance on both of its coasts-the Pacific and Atlantic. Surveillance activity is necessary for the proper management of the vast fishing resource. As many as 200 000 people on the Atlantic coast alone depend directly or indirectly on the fishery for 0165-7836/95/$09.50

0 1995 Elsevier Science B.V. All rights reserved

XSD10165-7836(94)00344-O

346

H.H. Millar / Fisheries Research 23 (199s) 345-360

survival. Hence, the proper management of the resource is of paramount importance. In times of fiscal restraint, and cutbacks, the effectiveness of Canada’s surveillance program will depend on how well it is able to deploy surveillance resources. There are several complicating factors which play a major role in determining the effectiveness of a program. These include: the capability of the surveillance units, the current regulations, the distribution of fishing activity, the time of year, the types of fishery violations, the capacity of the surveillance resources, the surveillance priorities, and the deployment strategy. In order to effectively accommodate these factors in a decision process, one requires a fairly sophisticated mechanism. More than likely, this precludes a manual approach to planning. Further, a decision support tool which would enable management to assess the impact of various scenarios such as priorities and unit availability on the surveillance plan is needed. In reality, much of surveillance planning and deployment in Canada is done without the benefit of sophisticated management technology-particularly technology which allows for much experimentation over extended planning horizons, and technology which allows the exploration of ‘what if’ questions. Such a management tool could be useful in what is a highly uncertain and dynamic environment. Canada’s declaration of a 200-mile economic zone has provided the opportunity to rebuild over-fished stocks and to undertake integrated resource management. These opportunities have been complicated by the significant increase in domestic fleet activity since 1977, in both the inshore and offshore fishery. For example, the number of domestic quotas have increased dramatically from 150 to approximately 550 between 1979 and 1984. To manage the 200-mile economic zone, the Canadian government has enacted two key pieces of legislation, the Fisheries Act and the Coastal Fisheries Protection Act, which apply to both domestic and foreign fishing vessels. Consequently, managing the resource implies the enforcement of the rules and guidelines designed to ensure the sustainability of the industry. The need for regulatory control over the fisheries, particularly those involving transboundary stocks, has meant the need for effective patrol and surveillance activity. The latter must attempt to deter a range of possible violations such as: fishing without a licence, misreporting of catches, illegal dumping of fish, fishing in closed areas, use of illegal gear sizes, over-fishing quotas, and so on. Under the Coastal Fisheries Protection Act, sanctions for violators include fines, licence suspension, and forfeiture of gear and/or catch. Enforcing DFO regulations and measuring the effectiveness of enforcement strategies are both difficult tasks. Since the primary objective of Law enforcement is the prevention of crime, one may define an output of law enforcement as the number of crimes not committed owing to enforcement activity and the threat of sanctions. Unfortunately, compliance is unobservable, as is the total number of violations in the fishery. Hence proxy measures for example the probability of arrest, are used as substitutes. This probability distribution may also be difficult

H.H. Millar / Fisheries Research 23 (1995) 345-360

347

to establish. As a result, it becomes necessary to establish not only effective enforcement strategies, but also useful means of measuring their effectiveness. Faced with the lack of appropriate measures of effectiveness, Blewett et al. ( 1985 ), conducted an empirical evaluation of Canada’s enforcement experience in measuring the deterrent effect of fisheries law enforcement, based on an analysis of participation in illegal activities. By establishing perceived penalties for potential violators, they were able to produce a measure of the deterrent effect of fisheries law enforcement on both of Canada’s coasts. 1.1. Current deployment practice

The DFO Offshore enforcement program is made up of three main components: aerial surveillance, surface surveillance, and observer coverage. Aerial surveillance/patrol provides the ability to cover large geographical areas, locate fishing fleets, detect violations, and maintain a constant visible presence. Aerial surveillance provides the best means to track fleet movement and hence offers the greatest deterrent to illegal entry and fishing activity. Surface surveillance/patrol provides a platform for planned inspection programs that collect statistical information, and provides an immediate means to apprehend violators. Observer coverage provides a day to day presence onboard fishing vessels to monitor compliance with legislation, and collect biological data for use in scientific assessments. Both surface and observer coverage are restricted to monitoring of activities within Canada’s 200-mile limit. Observer coverage is further restricted to domestic and foreign vessels licensed to fish inside Canadian waters. The deployment of aerial, surface, and observer surveillance resources by DFO is currently based on expert judgment of the enforcement supervisors. The deployment of enforcement effort begins with the determination of an annual plan which outlines the number of patrol hours in each enforcement class that are assigned to each NAFO division at various times of the year. The plan, which is generated manually through discussions among fishery officers, attempts to take into account several factors which include: ( 1) forecasts of anticipated fishing activity for the various divisions, (2) priorities associated with various fisheries, and (3) the historical distribution of violations. Implementation of the plan is undertaken by local supervisors, and requires short-term (daily and weekly) assignment of observer coverage to foreign fleets, the development of surface and aerial surveillance plans, amendments to existing short-term plans in the light of changes in enforcement priorities, and dispatching of response units to area violations. These activities must be carried out against a backdrop of considerable uncertainty in effort levels as well as the spatial distribution of effort. This uncertainty further complicates an already complex task. While there are many aspects of fishery surveillance deployment which deserve investigation, in this paper, we explore the potential of linear programming as a tool for planning tactical surveillance deployment. The model attempts to allocate effort in a manner which maximizes the deterrent effect of the surveillance

348

H.H. Millar / Fisheries Research 23 (1995) 345-360

activity. It incorporates several priorities as specified by the planner. Priorities for violations and/or fisheries can be specified. The model allows the manager much influence over where and how the effort is allocated by incorporating an appropriate bounding and priority structure. Perhaps the biggest potential contribution of the model, is that it facilitates ‘what if” questions, thereby allowing the planner to gain insight into the nature of the interaction between the various components of the surveillance problem. To our knowledge, no such model has been reported in the literature. Several studies on fisheries surveillance and enforcement have been conducted. Blewett et al. ( 1985) and [email protected] ( 1979, 1980) conducted reviews of Canada’s offshore enforcement program. No effective models for surveillance planning were presented. Clough ( 1980) proposed a model for predicting violations, but the model performed very poorly in future years (Crowley and Palsson, 1989). Reports by the Department of Fisheries and Oceans (DFO, 198 1, 1984) emphasize a need for improving the system. The reports recognize the complexity of the enforcement program due to the vastness of the Atlantic fishery, and the myriad of infractions committed by fishers. Further, management plans for the different types of fisheries complicate the surveillance planning process. Bergin ( 1988) discussed the issue of fisheries surveillance in the South Pacific, but offered no formal models for the deployment of surveillance effort. Sutinen ( 1988) proposed a linear integer programming model for routing of surveillance effort in the Costa Rican fishery. The routing problem is an operational one which takes place over a short-term planning horizon. Charles et al. ( 199 1) and Millar ( 1993 ) discuss several models for the deployment of land (policing) surveillance activity. Sutinen ( 1988) developed several violation indices. These measures can be used to aid the deployment of surveillance resources. However, on their own, they are insufficient. Fishery priorities, unit effectiveness on certain types of violations, violation priorities, coverage targets, and many other factors must be considered in developing a deployment plan. To date, few formal models, if any, exist for the deployment of fisheries surveillance, either at the tactical or operational levels. In fact, much of the data for the design of an effective plan is not readily available is a usable form (E. Collins, Chief Enforcement Officer, Newfoundland Region, personal communication, 1990; C. Goodwin, Chief Enforcement Officer, Scotia-Fundy Region, personal communication, 1990, 1993 ). The objective of this paper is to delineate a mathematical model for tactical surveillance planning. In the process of developing this model, we identify pertinent data requirements. The latter would provide a rational basis for data collection and analysis by the DFO surveillance division. In the next section, we outline the surveillance problem. In Section 3, we describe a linear programming model for developing the annual tactical surveillance plan. In Section 4, we present a sample problem based on a combination of real and fabricated data for the 1989 fishing season in the Newfoundland fishery. Most of the data are real. For model completeness, it was necessary to fabricate data which were not available. Computational results highlighting the usefulness of the model are presented.

H.H. Millar / Fisheries Research 23 (199s) 345-360

349

2. The surveillance problem

In this section we provide a view of the nature of the surveillance/patrol problem in the Atlantic fishery by examining the Newfoundland fishery. The ScotiaFundy region, which is the other major part of the Atlantic fishery has a very similar structure, and need not be expounded. Much of this section is based on a deployment and utilization strategy report by the Newfoundland Enforcement Division ( 1990). The Atlantic Fishery is shown in Fig. 1. 2. I. The fishery profile The Atlantic region includes Newfoundland and the Scotia-Fundy region. In Newfoundland alone, the area of responsibility encompasses some 430 000 m* of ocean. This regions includes some of the most prolific fisheries in the world. This area extends some 2000 linear miles from Davis Straight (sub-area 0) south and west to St. Pierre Bank (Division 3Ps) and includes 400 NM of sovereign boundary that transects the Grand Banks of Newfoundland (Division 3LNO). In addition, Newfoundland includes 9000 nautical miles of coastline. There is a complex mixture of species existing in the stock areas/fisheries of Newfoundland. The fisheries can be characterized as follows. ( 1) Sub-area 0/2GH consists of large amounts of water extending north to include Davis Straight with a 100 NM boundary between Canada and Greenland. There are lucrative shrimp fishing grounds. This area has low domestic fishing activity, but much illegal fishing by unlicensed foreign fleets. Other issues include misreporting of catches or areas by licensed vessels. (2) Area 2J3K represents the northern limit of DFO’s regular enforcement effort. The area is quite vast and is subjected to high intensity fishing for Northem Cod. Problems include gear conflicts, and misreporting of catch and areas. (3 ) 3LMN0 (Continental Shelf) is a small geographic region that includes very prolific fishing grounds. The 200 mile boundary transects 3N0 and 3L. There is high foreign activity just outside the 200 mile limit, often resulting in illegal entry and fishing. Domestic vessels often misreport catch and area. (4) 3LMN0 (beyond the Continental Shelf) is a large area that is fished by both foreign and domestic swordfish and tuna longlines. (5) Recently, the area 3Ps was the center of an on-going boundary dispute between Canada and France. It is a prolific area subject to high-intensity fisheries. The main concern was over-fishing of quota allocations by the French fleet. Currently, enforcement is complex and difficult. Within the Newfoundland region, eight foreign fishing nations have access rights to a total of 56 distinct stocks which include groundfish, capelin, squid, and other pelagic species. On an annual basis, some 120 foreign vessels (including 35 Japanese tuna and squid vessels) are licensed to fish approximately 200 000 t of various allocations from Sub Area O-3.Some are licensed to purchase fish or support national fleets. Domestic vessels ranging between 45 and 65 ft have access to approximately

350

H.H. Millar / Fisheries Research 23 (1995) 345-360

66

: I

60

;

6E

: I

6~

i

6~

i

6H

Fig. 1. The Atlantic fishery.

50 groundfish stocks, while those over 65 ft have access to 170 groundfish, 107 shrimp, seven scallop, and four clam/quahog allocations. On an annual basis, the fleet harvests some 540 000 t of various allocations in Sub Area O-3. Approximately 30 000 square miles of fishing grounds remain outside Canada’s jurisdiction. These are commonly referred to as the Nose (3L) and Tail (3NO) of the Grand Banks, and the Flemish Cap (3M). There is a significant amount of foreign activity in those regions. The transboundary stocks are exploited by nations

H.H. Millar /Fisheries Research 23 (1995) 345-360

351

Table 1 A summary of the Newfoundland

offshore fishery effort levels

Fishery

Area

Primary period

Domestic Northern cod

2J3KL

Northern shrimp Fixed gear (45- 100 ft. ) FPI American plaice Other 100 ft. (redtish, witch) Other 100 ft. Mobile gear (65- 100 ft. )

O-3K 3N0 3LN0 2+3 2+3 2+3

Apr.-Jun. Jan.-Mar. Apr.-Mar. Apr.-Nov. Jun.-Nov. Apr.-Mar. Apr.-Mar. Apr.-Mar.

50 1O-20 IO-15 20 1O-20 25 25

7000 2000 I500 3000 8000 30 000 3000

Jul.-Dec. Apr.-Jun. Jul.-Dec. Jan.-Mar. Apr.-Jun. Sep.-Mar.

10 5 30 5 5-10 5-10

1000

Foreign Far northern water turbot and cod Northern water, turbot, grenedier, redtish

0+2GH 2J+3K

St. Pierre Bank (France)

3Ps

Transboundary stock and Flemish cap

3LMN0

Apr.-Mar.

Approx. no. of vessels per day

Approx. effort (t)

50

225

Total effort

4000 1500

15000 79 000

which include Spain, Portugal, the relevant Republics of the former USSR, Cuba, Japan, Germany, Faroe Islands, and Norway. Non-NAFO members such as Korea, Cayman Islands, USA, and Panama have all increased their activity in the region. This has serious implications for the enforcement division. A summary of the major fisheries in the Newfoundland area is shown in Table 1. 2.2. The Surveillance unit and surveillance/patrol resources The surveillance unit is comprised of an enforcement coordinator, a supervisor, an operations officer, a radio officer, and some 14 fisheries officers. Support is provided by the Quota Monitoring Unit, the Observer program, and the Fishery Career Officer Program. In the case of Newfoundland, a combination of chartered service, the Department of National Defense (DND) provides aerial surveillance resources. There are two vessels which provide surface coverage. In addition, DND provides additional coverage with two other vessels. We are not sure of the exact number of observers working in the Newfoundland fishery. The number could be in excess of 100 observers. The estimated annual cost of aerial surveillance/patrol for the Atlantic region in 1989 was $3.467 million. The annual cost of both surface and observer cover-

352

H.H. Millar / Fisheries Research 23 (1995) 345-360

age exceeded $7.0. million per year. The average number of sea days per year for surface coverage is approximately 568 days. Patrol vessels cost approximately $10 000-20 000 day-’ to operate. On the basis of $10 000 day- ‘, surface surveillance cost could exceed $5.0 million year-‘. The estimated number of sea days of observer coverage allocated to the foreign and domestic fisheries for the 1990/ 199 1 period were 9 100 and 3 100 respectively for a total of 12 200 sea days. If we consider a modest cost of $160 per sea day, the estimated cost of observer coverage could be approximately $1.952 million per year. 2.3. Fisheries violations Both domestic and foreign fleets commit a range of violations. These violations may have varying degrees of impact on the management of the resource. Also, they occur with varying frequencies across lisheries and across fishing fleets. Fishery violations include: exceeding catch limit, using unlicensed, unmarked or illegal gear, fishing in a closed area, fishing in an area without a licence or outside of the fishing season for that area, a lack of co-operation with the observer, retaining prohibited species, license(s) not on board, a lack of or an invalid fisherman’s registration document, invalid fishing license, invalid vessel registration, a lack of vessel marking, inaccurate production logs, and unauthorized transshipment. The ability of the surveillance units to detect these violations varies significantly across units. Some violations cannot be detected by a given unit. For example, an observer cannot tell if a vessel has drifted into a closed zone, or aerial surveillance cannot tell if an illegal species was caught. This suggests that the deployment of surveillance/effort must be such that each fishery receives a given amount of each kind of coverage depending on the types of violations occurring in that fishery.

3. The tactical planning model In this section, we present a linear programming model for developing a tactical enforcement plan. The model determines the enforcement effort in days for each surveillance unit, in each NAFO zone, in each time period. The model attempts to maximize a ‘weighted value’ of the effort as it relates to fishery violation priorities, violation distribution, fishing activity distribution. The model assumes that areas with high violation rates, high violation priorities, high fishing density, and high management priorities, should receive the most attention. Further, resources should be allocated in, such a way, that their effectiveness in detecting and deterring crime is maximized. Hence the coupling between a unit’s ability to detect certain violations and the zonal distribution for the violations will influence where the way resources are allocated. While we assume a linear objective function, other (nonlinear) functions may be investigated. We employ linear programming as a tool so as to allow the use of

H. H. Millar /Fisheries

Research 23 (1995) 345-360

353

Notation the set of all periods the set of all zones the set of all the types of surveillance units the set of all violations the set of all fisheries the total surveillance capacity of unit k (in days) an upper bound on the surveillance capacity (in days) of unit k in period t a lower bound on the surveillance capacity (in days) of unit k in period t an upper bound on the annual surveillance effort allocated to zone j a lower bound on the annual surveillance effort allocated to zone j an upper bound on the total surveillance effort allocated in period t a lower bound on the total surveillance effort allocated in period t the cost for one day of surveillance effort by unit kin period t the surveillance budget for unit k the relative effectiveness of surveillance unit k in zone j the priority associated with fisheries r in period t the probaiblity of violation v taking place in fisheries r the effectiveness of unit k detecting violation v a binary variable indicating if fisheries r is in zonej the expected inspection rate per sea day for fisheries r an upper bound on the surveillance effort in fisheries r in period t a lower bound on the surveillance effort in fisheries r in period t an upper bound on the total surveillance effort for fisheries r a lower bound on the total surveillance effort for fisheries r a minimum target on the expected number of sinpections in fisheries r the priority weight for violation v the expected number of vessels per day in fisheries r in period t a penalty cost per under-realized boarding in fisheries r a penalty cost per day of under-utilized surveillance capacity for resourcek

sensitivity analysis to gain insight into the problem structure. Given the complexity of the fishery and the level of uncertainty, the tactical plan will inevitably be modified during implementation. A linear programming model will provide a good starting solution. 3.1. The constraints The constraints for the linear programming model are as follows.

3. I. 1. Period coverage ( 1) The total surveillance effort allocated in period t must be within bounds

354

H.H. Millar / Fisheries Research 23 (1995) 345-360

(2) The total surveillance effort allocated to zone j must be within bounds

3.1.2. Resource capacity utilization (3) The total effort allocated for surveillance unit k in period t must be within bounds

(4) The total surveillance effort allocated over the year for unit k must not exceed its capacity

3. I .3. Fisheries coverage (5) The surveillance effort allocated to fisheries r in period t must be within

bounds

(6) The total annual surveillance effort allocated to fisheries r must be within bounds

3.1.4. Inspections coverage (7) The expected number of boardings plus the shortage for fisheries r must be

equal to or exceed a lower bound

3.1.5. Annual budget (8 ) The total surveillance budget expended for unit k must not exceed an up-

per bound

There are additional constraints that may be considered. These include constraints on the effort allocated to a fishing ground in a given period, constraints on the type of surveillance effort allocated to a given fishing ground, constraints on the type of surveillance effort allocated to a given fishery in a given period,

H.H. Millar / Fisheries Research 23 (1995) 345-360

355

and several others. These constraints are easily incorporated if they are deemed necessary by the appropriate authorities. 3.2. The objectivefunction The objective of the model is to maximize the total ‘value’of the surveillance effort allocated to the fishing grounds. We note that the objective function will tend to allocate surveillance effort in proportion to the likelihood of detecting violations on the fishing grounds weighted by priorities associated with the lishing grounds as well as the expected number of vessels active in the various fisheries.

Placing penalties on under-realized boardings, the model will be forced to give priority to satisfying boarding requirements. The penalty on under-utilized capacity has the effect of forcing the model to allocate all the capacity. The penalties may reflect the cost of idle capacity. The model also allows for allocating different penalties to the different resources. For example, a large penalty on aerial resources will force all of it to be utilized unless some constraint, such as the budget, precludes that possibility. 4. Computational experience

We demonstrate the potential of the model by solving a problem based on the Newfoundland surveillance program. Some of the data had to be fabricated owing to unavailability of the data required by the model. Hence, the results described here do not reflect an actual surveillance plan for use by the Department. It simply shows the results for the particular set of data used. For the sake of brevity and so as not to detract from the essence of the paper, only pertinent data are presented here. 4.1. Problem data

The problem under consideration involves I3 fishing grounds and 11 fisheries. The fisheries and their corresponding grounds are listed in Table 1. There are three types of surveillance resources, vessels, planes, and observers. We assume four planning periods (quarters), January-March, April-June, July-September, and October-December. There are 16 types of violations that occur throughout the various fisheries. The annual budgets for the vessel surveillance, the aerial surveillance, and the observer coverage programs are $5 million, $3.467 million, and $2.1 million respectively. Resource capacities are 695 vessel days for surface surveillance, 123 days (2870 h ) for aerial surveillance, and 10 000 observer days. We fabricated unit costs for the use of each resource type: $5000 per vessel day, $45 600 per flying day ($1900 per flying hour), and about $160 per observer day.

356

H. H. Millar /Fisheries Research 23 (1995) 345-360

4.2. Results

The following tables present the results obtained from the model. Tables 2-4 show the allocation of the various types of surveillance resources to each of the grounds in each period. Table 5 gives a summary of how the effort was allocated for each type of resource in each period. There appears to be a reasonable good balance in the allocation of effort across periods. Table 6 shows the allocation of effort to the grounds. Allocation limits of 3000 surveillance days were set for each ground. This prevents the allocation of all the Table 3 Patrol vessel effort allocations (vessel days) Ground

Quarter 1

2

3

4

2J+3K 3LN0 (out) 3Ps

120

120

35 106 50

265

Total

120

120

191

265

Table 4 Aerial surveillance effort allocations (flying days) Ground

Quarter 1

2

3

4

0+1

31

26

10

10

Total

31

26

10

10

Table 5 Observer effort allocations (observer days) Ground

Quarter 1

0+1 2GH 2J+3K 3LN0 (in) 3LN0 (out) Total

2 120

3 120

1960

2380

190 1526 740 44 0

2500

2500

2500

2380

2500

4 265

275

H.H. Millar / Fisheries Research 23 (1995) 345-360

357

Table 6 Summary of effort allocations for the resource units Surveillance unit

Quarter

Total

Percent of limit

100 62.6 100

1

2

3

4

Vessels Aerial Observers

120 31 2500

120 26 2500

190 10 2500

265 10 2500

695 77 10000

Total

2651

2646

2700

2115

10772

Table 7 Summary of effort allocations (in days) to the fishing grounds Ground

Total

Quarter 1

0+1 2GH 2J+3K 3LN0 (in) 3LN0 (out) 3Ps Total

2 151

4

3 146

200 1526 175 44 105 50

_ 2380 120

2500

2651

2646

Percent of limit

2700

215 2225 275

2175

172 1526 3000 2424 3000 50

25.1 50.9 100 80.8 100 1.67

10772

Table 8 The amount of surveillance exposure received by each fishery Fisheries

Northern cod Northern shrimp Fixed gear (45-100 ft) FPI American plaice 100 ft (redfish, witch) Other 100 ft. 65 ft (mobile) Far northern water (turbot and cod) Northern water (turbot, grenedier, redfish ) St. Pierre Bank Transboundary stock

Total

Quarter

Percent of limit

1

2

3

4

2500 151

2500 146

275 2500

151 2500 120

146 2500 2500

149 2500 50 715 2500 149 105

2225 2500 275 275

5424 5297 50 3000 5291 5424 3000

2500

2500

2500

2500

10000

100

2500 2500 2500

2500 2500 2500

2500 200 2500

2500 275 2500

10000 5475 10000

100 54.75 100

54.24 52.91 0.50 30 52.91 54.24 30

358

H.H. Millar / Fisheries Research 23 (199.5) 345-360

Table 9 The potential number of boardings in each fishery owing to the surveillance allocations Fisheries

Minimum

Potential

Northern cod Northern shrimp Fixedgear (4SlOOft.) FPI American plaice 100 ft. (redfish, witch) Other 100 ft. 65 ft. (mobile) Far nothem water (turbot and cod) Northern water (turbot, grenedier, red&h) St. Pierre Bank Transboundary stock

225 12 15 50 20 10 25

345 179 15 210 210 207 242

10

417

5 5 20

695 395 695

Table 10 Summary of budget utilization Surveillance unit

Annual budget

Amount consumed

Percent of total

Vessels Aerial Observers

5000000 3467000 2100000

3187500 3467000 1600000

63.74 100 76.19

effort to a single ground. In a real application, the planner would set appropriate limits for each ground. Table 7 shows how the effort is allocated from the point of view of the fisheries. Since several fisheries may occupy the same zone, effort allocated to a zone with multiple fisheries will be viewed as giving exposure to all the fisheries within that zone. In detailed deployment, the effort will have to be disaggregated among the fisheries in the zone. Table 8 shows the number of boardings possible in a given fishery due to the presence of surface surveillance. Once again, in actual surveillance deployment, the actual number of boardings conducted in each fishery would have to be determined. Finally, Table 9 shows how the budgets for each resource was consumed. The aerial surveillance budget is a limiting factor in the model. Only 77 of the available 123 days were deployed. The surface surveillance and the observer coverage budgets were not completely utilized even though the corresponding capacities were fully utilized. 5. Conclusions We have presented a linear programming model for assisting surveillance officers in developing annual plans for the deployment of surface, aerial, and ob-

H.H. MilIar /Fisheries Research 23 (I 995) 345-360

359

server surveillance effort. While the model can be prescriptive, its intent is to assist the officer in exploring various scenarios. There is great opportunity for simulating the effects of priority structures, allocation limits, and resource budgets on the nature of the deployment of effort and the relative effectiveness of the effort deployed. While the objective function does not, measure the deterrent capabilities of the plan directly, its structure implies the maximization of the deterrent effect. In fact, the objective function allows for more than just the deterrent effect. It allows the officer or user to exert their own biases by manipulating fishery priorities and the effort allocation limits. The sensitivity analysis capability of the model is a major advantage for the decision maker. S/he can explore the impact of increase or reduced budgets, altered capacities, or changes in the surveillance priorities. Given that there is a lack of tools for assisting decision makers in fishery surveillance, the model delineated in this paper can be viewed as a necessary and important step in improving the planning capabilities of the decision makers. The value of this model is further amplified by the reality of tight fiscal times. Acknowledgments

The author is extremely grateful to Carl Goodwin, Ernie Collins and Margo Frayling of the Federal Department of Fisheries and Oceans for the discussions and material on the Newfoundland and Scotia-Fundy fisheries surveillance programs. However, the author claims responsibility for all errors. The author is also grateful to Mona Kiragu for assistance with the computational work, the National Science and Engineering Research Council of Canada for its partial support of this research through grant OGPIN020, the Saint Mary’s University Senate Research Committee for providing partial funding for the research project, and the anonymous referees for their valuable comments. References Bergin, A., 1988. Fisheries Surveillance in the South Pacific. Ocean Shoreline Manage., 1 I : 467-49 1. Blewett, E., Furlong, W. and Toews, P., 1985. Canada’s experience in measuring the deterrent effect of fisheries law enforcement. Department of Fisheries and Oceans, Program Evaluation Branch. Charles, A.T., Cross, M.L., Mazany, L. and Millar, H.H., 1991. Fisheries Law Enforcement: Applications of Economic Analysis and Operations Research Models. Oceans Institute of Canada, Halifax, N.S. Clough, D.J., 1979. Performance and effectiveness of offshore fisheries surveillance, 1977-1978, and options to 1985. Department of Fisheries and Oceans. Clough, D.J., 1980. Optimization and implementation plan for offshore fisheries surveillance. Department of Fisheries and Oceans. Crowley, R.W. and Palsson, H.P., 1992. The application of operations research models to offshore fishery regulation enforcement in Canada. Am. J. Math. Manage. Sci., 12: 153-190. Department of Fisheries and Oceans, 198 1. Effectiveness of enforcement and surveillance. Rep. No. 41.

360

H.H. Millar / Fisheries Research 23 (1995) 345-360

Department of Fisheries and Oceans, 1984. Effectiveness of Enforcement and Surveillance. Rep. No. 41. Millar, H.H., 1993. An overview of models for the planning and deployment of fisheries surveillance/ patrol effort in the Canadian Atlantic offshore fishery. Working Pap. FM 04-93, Finance and Management Science Department, Saint Mary’s Urilversity, Halifax, Nova Scotia. Newfoundland Enforcement Division, 1990. An annual surveillance plan for the Newfoundland fishery. Department of Fisheries and Oceans, Newfoundland. Sutinen, J., 1988. Enforcement economics in exclusive economic zones. Geojoumal, 16: 273-28 1.