Читать книгу Strategic Modelling and Business Dynamics - Morecroft John D. - Страница 20

Imagine you are living in a small fishing community where everyone's livelihood depends on the local fishery. It could be a town like Bonavista in Newfoundland, remote and self-sufficient, located on a windswept cape 200 miles from the tiny provincial capital of St Johns, along deserted roads where moose are as common as cars. ‘In the early 1990s there were 705 jobs in Bonavista directly provided by the fishery, in catching and processing’ (Clover, 2004). Let's suppose there is a committee of the town council responsible for growth and development that regulates the purchase of new ships by local fishermen. This committee may not exist in the real Bonavista but for now it's a convenient assumption. You are a member of the committee and proud of your thriving community. The town is growing, the fishing fleet is expanding and the fishery is teeming with cod.

Figure 1.8 shows the situation. The fish stock in the top left of the diagram regenerates just the same as before, but now there is an outflow, the harvest rate, that represents fishermen casting their nets and removing fish from the sea. The harvest rate is equal to the catch, which itself depends on the number of ships at sea and the catch per ship. Typically the more ships at sea the bigger the catch, unless the fish density falls very low, thereby reducing the catch per ship because it is difficult for the crew to reliably locate fish. Ships at sea are increased by the purchase of new ships and reduced by ships moved to harbour, as shown in the bottom half of the diagram.

Figure 1.8 Diagram of a simple harvested fishery

Figure 1.9 Interface for fisheries gaming simulator

The interface to the gaming simulator is shown in Figure 1.9. There is a time chart that reports the fish stock, new fish per year, catch and ships at sea over a time horizon of 40 simulated years. Until you make a simulation, the chart is blank. The interface also contains various buttons and sliders to operate the simulator and to make decisions year by year. There are two decisions. Use the slider on the left for the purchase of new ships and the slider on the right for ships moved to harbour. You are ready to simulate! Open the file called ‘Fisheries Gaming Simulator’ in the learning support folder for Chapter 1. The interface in Figure 1.9 will appear in colour. First of all, simulate natural regeneration over a period of 40 years, a scenario similar, but not identical, to the simulation in Figure 1.7. The only difference is that the initial fish population is 500 fish rather than 200. What do you think will be the trajectories of the fish stock and new fish per year? How will they differ from the trajectories in Figure 1.7? Would you expect any similarities? To find out, press the button on the left labelled ‘Run’ (but don't alter either of the two sliders, which are deliberately set at zero to replicate a natural fishery). You will see a five-year simulation. The fish stock (line 1) and new fish per year (line 2) both grow steadily. You can observe the exact numerical values of the variables by placing the cursor on the time chart, then selecting and holding. Numbers will appear under the variable names at the top of the chart. At time zero, the fish stock is 500 and new fish are regenerating at a rate of 63 per year. If you look carefully you will see that the catch (line 3) and ships at sea (line 4) are, as expected, running along at a value of zero, alongside the horizontal axis of the time chart. Press the ‘Run’ button again. Another five simulated years unfold showing further growth in the fish stock and in new fish per year. Continue until the simulation reaches 40 years and then investigate the trajectories carefully and compare them with the time chart in Figure 1.7. Why does the peak value of new fish per year occur so much earlier (year 10 instead of year 16)? Why is the final size of the fish stock identical in both cases?

Harvesting in Bonavista, Newfoundland – A Thought Experiment

Back to Bonavista, or at least a similar imaginary fishery, scaled to the numbers in the simulator. The fishing fleet has been growing and along with it the catch and the entire community supported by the fishery. As a member of the town's growth and development committee you want to explore alternative futures for the fishery and the simulator is one way to do so. You conjure up a thought experiment. Starting as before with an initial stock of 500 fish, you first simulate growth, through natural regeneration of fish, for a period of 10 years. The result is a well-stocked fishery similar to the one existing some 20 years ago when the hamlet of Bonavista, as it was then, began to expand commercial fishing. You know from the previous experiment that this scenario will lead to plenty of fish in the sea, but in reality you and the fishermen themselves don't know how many.

To replicate this fundamental uncertainty of fisheries you should ‘hide’ the trajectories for fish stock and new fish per year by colouring them grey so they blend into the background of the time chart. Some playing around with the software is necessary to bring about this change, but the result is important and worthwhile. First, press the ‘Reset’ button on the left of the time chart. The trajectories will disappear to leave a blank chart. Next move the cursor to the tiny paintbrush icon at the right of the tools bar at the top of the interface. Select and hold. A palette of colours will appear. Move the cursor to the bottom line containing greys and blacks. Select the light grey colour on the extreme left. Release the mouse button and move the cursor back onto the time chart where it will now appear as a paint brush. Select and the background of the chart will turn grey. Return to the colour palette and select the light grey colour second from the left. Now move the paintbrush cursor so that it lies exactly on top of the phrase ‘Fish stock’ at the top left of the time chart. Select and the phrase will turn from blue to grey and will, as intended, be virtually indistinguishable from the background grey. Repeat the same painting procedure for the phrase ‘New fish per year’. Your time chart is now ready.

Press the ‘Run’ button twice to recreate 10 years of natural fishery growth. At first glance the simulated chart will appear quite blank and uninteresting. That's how it should be! Now move the slider for ‘Purchase of new ships this year’ to a value of 2 by selecting, holding and dragging the slider icon until the number 2 appears in the centre box. This setting means that each simulated year two new ships will be purchased and used by Bonavista fishermen. Press the ‘Run’ button three times in succession to simulate fleet expansion for years 10–25, a period of historical growth for the imagined Bonavista fishery. Ships at sea (line 4) increase linearly from zero to 30 as you would expect from an investment policy that adds two new ships a year over 15 years. The catch (line 3) increases proportionally in a similar linear pattern. Press the ‘Run’ button once more to simulate continued fleet expansion for years 25–30. Ships at sea continue the same relentless linear expansion, but notice a dramatic change in the trajectory of the catch (line 3). In year 26, after 16 years of steady growth, the catch levels out and peaks at 786 fish per year even though new ships are being added to the fleet. (To check the numerical values move the cursor onto the time chart, then select, hold and drag.) In year 27 the catch declines for the very first time in the fishery's simulated history. At the start of year 29, the catch is down to 690 fish per year, a decline of 12 per cent from the peak. Imagine the situation in Bonavista. The town's main business is in a downturn. A community, which has become used to growth and success, begins to worry and to ask why. Perhaps the past two years have been unlucky – poor weather or adverse breeding conditions. However, year 29 sees continued decline. The catch falls below 450 fish per year while the fleet grows to 40 ships. A downturn has become a slump.

At this point you can imagine pressure building in the community to do something about the problem. But what? The fishery is in decline. Perhaps the answer is to halt the purchase of new ships and to require some ships to remain in harbour. Such measures may seem logical if you believe that overfishing is to blame. But others will argue the decline is due to a run of exceptionally bad luck and that, sooner or later, the catch will return to normal. And remember nobody knows for certain the size of the remaining fish stock or the regeneration rate. That's all happening underwater. So, as in all practical strategy development, there is scope for argument and conflict about the true state of affairs and how best to react. Moreover, it is politically and economically painful for any community or business to cause itself to shrink deliberately. There are bound to be more losers than winners.

Nevertheless, imagine Bonavista agrees a conservation policy involving a total ban on the purchase of new ships for the next five years and an effective reduction in the fleet size to be achieved by moving five ships per year into the harbour. A little mental arithmetic reveals that in its first year of operation this policy idles 12.5 % of the active fleet (5 ships out of 40), then 14.3 % in the second year (5 ships out of 35), then 16.7 % in the third year (5 ships out of 30). After five years, a total of 25 ships have been idled, which is fully 62.5 % of the original fleet – a huge reduction in a short time. Adjust the sliders to represent the implementation of this stringent conservation policy. First set the slider for the ‘Purchase of new ships this year’ to zero, either by dragging the slider icon to the extreme left or by selecting the slider's ‘Reset’ button (denoted by ‘U’) in the bottom left of the slide bar. Then, set the slider for ‘Ships moved to harbour this year’ by dragging the slider icon to the right until the number 5 appears in the centre box. Press the ‘Run’ button to see the results of the policy. You will notice that ships at sea (line 4) decline steeply as enforced idling takes place. By year 35 of the simulation, the active fleet size is 15 ships at sea, back to where it had been in the early growth heyday of the fishery almost 20 years ago in year 17. Despite the cuts and huge economic sacrifices, however, the catch has declined to less than 10 fish per year, scarcely more than 1 per cent of the peak catch in year 26. In a single decade our imagined Bonavista fishery has gone from productive prosperity to extreme hardship. Each day the community awakes to see the majority of the fishing fleet idle in its once busy harbour, and the remaining active ships returning with a dismally tiny catch. You can imagine that by now many will have lost heart and lost faith in the conservation policy.

To finish the simulation reset to zero the slider for ‘Ships moved to harbour this year’ and then press ‘Run’. In these final years it is no longer possible to enforce further reductions in the active fleet. The number of ships at sea remains constant and the catch falls practically to zero. It's a depressing story, but entirely consistent with the facts of real fisheries. Harvested fisheries are prone to catastrophic decline that nobody involved – fishermen, community leader or consumer – would wish on themselves. Yet this situation in particular, and others like it, arise from nothing more than a desire to purchase ships, catch fish and grow a prosperous community. Why? Fisheries provide but one example of puzzling dynamics that are the focus of this book. As we will see, modelling and simulation can shed useful light on why such puzzling dynamics occur and how to bring about improvement.

A Start on Analysing Dynamics and Performance Through Time

Much of the problem with managing fisheries lies in properly coordinating the number of ships at sea in relation to the number of fish. A sustainable fishery, one that provides a reliable and abundant harvest year after year, regenerates fish at about the same rate as they are being caught. Successful replenishment requires an appropriate balance of ships and fish. Balancing is easier said than done when in practice it is impossible to observe and count the number of fish in the sea, when fishing technology is advancing and when there is a natural human propensity to prefer growth and the prosperity it brings. Imagine we could reliably count the fish stock and observe the regeneration of fish through time. What new light would this new data shed on the rise and fall of Bonavista and the policy options to avoid catastrophic decline in the fish population? In our simulator we can choose to observe and report variables that, in real life, would be unobservable. Use the colour palette and paintbrush to reinstate the original coloured trajectories for the Fish stock (blue) and New fish per year (red). You will find the appropriate colours on the top row of the palette. (If you accidentally set the background colour of the chart to blue or red, which can happen if you don't align the paintbrush with the variable name, don't panic. Simply return to the colour palette, select light grey, and repaint the background. Then try again to re-colour the trajectories.) The resulting chart will look like Figure 1.10, with all the trajectories clearly visible, except that yours will be in colour.

Consider the behaviour over time of the fish stock (line 1). For the first 10 years of the simulation the number of fish grows swiftly because effectively there is a natural fishery (no ships) that is underpopulated relative to its carrying capacity. In years 10–15 commercial fishing begins and each year more ships are sent to sea (line 4). Nevertheless, the fish population continues to increase. These are the early growth years of the Bonavista community. During this entire period the catch is rising (line 3), but is always below the rate of regeneration (new fish per year, line 2). The fishery is sustainable with growing population. In years 15–20 the catch continues to rise steadily in line with fleet expansion, but the fish stock begins to decline gently as the catch exceeds the number of new fish per year (line 3 rises above line 2). This excess of catch over regeneration is not necessarily a problem for long-term sustainability because harvesting is actually stimulating the regeneration of fish, as shown by the steady increase in new fish per year. A harvested fishery, even a well-run one, will always have a fish population considerably lower than the maximum fishery size.

Figure 1.10 Simulation of harvested fishery showing all trajectories

Herein lies a fundamental dilemma for fisheries management. Who is to say whether a decline in fish population is a problem or not? It could just be a sign of effective harvesting in a period of growth. Moreover, and this is vitally important to remember, nobody knows for certain how many fish of a given species are in the fishery. At best there are estimates subject to measurement error, bias and even manipulation. So it is very difficult in practice to make fish stock itself (how many fish are believed to be in the sea) the basis for investment policy (how many ships to purchase). Much more persuasive evidence comes from the catch. The simulation shows catch rising all the way through to year 25 and beyond. The temptation, even in years 20–25, is to believe that further fleet expansion is both desirable and justified. The conflicting signals from fish stock (a weak signal at best) and the catch (a strong and tangible signal of immediate economic and personal importance to fishermen and fleet operators) form the basis of the coordination problem in fisheries. Throughout year 25 and even into year 26 it is not unreasonable to continue fleet expansion even though the invisible fish population is in steady decline.

However, in year 25 something of vital significance happens under water, hidden from all but the fish themselves. The number of new fish per year (line 2) peaks and then starts to decline. This is the first evidence, a kind of early warning signal, that the fishery is being overfished. Fish density is now so low that regeneration is suppressed. The fishery teeters on the brink of catastrophe. The rate of population decline (the steepness of line 1) increases. But the catch keeps on rising throughout year 26 so no action is taken to curtail fleet expansion. In year 27 the catch itself peaks and then declines, gradually at first. This is the first tangible evidence of stock depletion underwater, but even so the signal is likely to be ignored until the trend proves conclusive and until the fishing community persuades itself to limit fishing. In the simulator, we assume that new ship purchasing continues apace until year 30. By then the fish stock has fallen to around 400, only 10 % of the maximum fishery size. The regeneration rate (new fish per year) is still in decline and far below the much reduced catch. Measures to halt investment and to idle ships in years 30 to 40, drastic though they are, are too little too late. Bonavista's fish have all but gone and with them the industry on which the community depends. By year 35 there are so few fish left (only 16!) that, even with a total ban on fishing, it would take two decades to rebuild the stock to its value in year 10 when our imagined Bonavista first began commercial fishing.

Saving Bonavista – Using Simulation to Devise a Sustainable Fishery

Now you are familiar with the gaming simulator, you can use it to test alternative approaches to growing and developing the Bonavista fishery. First press the ‘Reset’ button to obtain a new blank time chart and to re-initialise the simulator. Next, without altering either slider, press the ‘Run’ button twice in order to simulate 10 years of natural growth in the fish population so that Bonavista inherits a well-stocked fishery. Then re-simulate the same fleet expansion as before – two ships per year for years 10–25. You will find yourself back in Bonavista's heyday with a fleet of 30 ships and a history of 15 years of steady growth in the catch. Now it is your responsibility to steer the community toward a sustainable future that avoids the errors of the past. For realism you may, as before, want to ‘grey-out’ the trajectories for fish stock and new fish per year. What is happening to the fish stock underwater is difficult to know, vague and often subject to controversial interpretation. Also bear in mind the practical political difficulties of curtailing growth and of idling ships in a community that depends on fishing. Think about plausible adjustments to the two sliders at your disposal. It is a good discipline to note your intentions, and the reasoning behind them, before simulating. Imagine you first have to convince the Bonavista community and fishermen to adopt your plan. Then, when you are ready, simulate, analyse the trajectories and try to make sense of the outcome. Was the result what you expected? If not then why? If you don't like the result then try again.

Dynamic Complexity and Performance Through Time

Although in principle it is possible to create a sustainable Bonavista it is very difficult to do so in practice or even in a simulator, particularly when you inherit a fleet of 30 ships following 15 years of successful economic growth. The fisheries simulator is one example of a dynamically complex system, of which there are others in this book and many more in life. Often such systems give rise to puzzling performance through time – performance far below the achievable and, despite the best of intentions, not what people (stakeholders in the system) want. In this case, the fishery is prone to catastrophic decline when perhaps all that fishermen desire, and the fishing community wants, is growth, more and better ships, and a higher standard of living. Dynamic complexity stems from the connections and interdependencies that bind together social and business systems. When a change happens in one part of the system (e.g. more ships are purchased) sooner or later it has implications elsewhere, and vice versa. Moreover, these implications are not always obvious and are often counterintuitive (e.g. more ships can lead to a greater rate of fish regeneration, but not always).

Dynamic complexity does not necessarily mean big, detailed and complex, involving hundreds or thousands of interacting components. Indeed, as the fisheries simulator shows, dynamic complexity and puzzling performance can arise from only a few interacting components. What matters is not so much the raw number of components but the intricacy with which they are bound together.

Such intricacy involves time delays, processes of stock accumulation (such as the accumulations of ships and of fish), non-linearities (such as the hump-shaped relationship between fish density and fish regeneration), and closed feedback loops (such as the reinforcing relationship between fish stock, fish density, fish regeneration and fish stock). These special terms, the language of feedback systems thinking, will become clearer later. For now it is sufficient to appreciate that dynamic complexity stems from intricate interdependencies of which there are many, many examples in our increasingly interconnected world. Sometimes it is possible to reduce dynamic complexity by making interdependencies less entwined and more understandable. Indeed, this goal of simplification is really the ultimate aim of policy design in system dynamics – redesigning social and business systems so that, despite their complexity, normally-competent people can run them successfully.

Why are fisheries so dynamically complex? What changes would make them less prone to sudden and catastrophic decline? Herein lies the whole area of fisheries policy involving fishermen, fishing communities, governments, marine scientists, consumers and fish themselves. There is a lot that could be modelled about the interactions among these stakeholders and arguably a serious fisheries policy simulator would be much bigger and would involve many more variables and relationships than those in our small Bonavista model. Nevertheless, at the heart of any such model will be a representation of the factors – biological, economic, political and social – that determine the balance of ships at sea and fish in a commercial fishery.

A vital part of dynamic complexity in fisheries lies in the relationship between the catch and fish density. Not surprisingly, if the fish density is very low then it is difficult for fishermen to locate fish and the catch is lower than normal. But the relationship is non-linear as shown in Figure 1.11. Here, fish density is measured on a scale from zero to one, where one is the highest possible density (the number of fish is equal to the carrying capacity) and zero is the lowest (there are no fish). The vertical axis shows the effect of fish density on catch per ship, also on a scale from zero to one. In our imagined Bonavista, the normal catch per ship is 25 fish per ship per year – remember this is a scale model. The actual catch per ship is obtained from the product of normal catch (25) and the effect of fish density.

Figure 1.11 Relationship between catch per ship and fish density

When the fish density is high, in the range between 0.7 and one, the catch per ship is stable at 25 because there is little or no depressing effect from fish density. The sea is full of fish and they are easy to find and catch. When the fish density is lower, in the range 0.4 to 0.7, the catch is still very close to normal (25). The assumption, borne out empirically in real fisheries, is that fish are still quite easy to find even when there are fewer, because they tend to cluster. Only when the fish density falls very low, in the range between zero and 0.4, does scarcity make fishing more difficult. In this narrow range the effect of density falls swiftly from 0.9 (almost normal) to zero.

The non-linearity, the sudden depressing effect of density on the catch, makes fisheries management difficult. You can appreciate why if you imagine the argument between a marine biologist and a fisherman about the need to conserve stocks. When the fish population falls to half the maximum (fish density equal to 0.5) the marine biologist argues that stocks are too low. But the fisherman reports (accurately) there is no difficulty catching fish, so what's the problem? In all likelihood, the fisherman thinks the fish stock is actually higher than the marine biologist's estimate. The biologist is exaggerating the problem, or so it seems to someone whose livelihood depends directly on the catch. When the fish population falls to one-quarter of the maximum (fish density equal to 0.25) the marine biologist is frantic and even the fisherman is beginning to notice a reduction in the catch, down by about one-third relative to normal. That outcome, though worrying, is not obviously fatal. Perhaps with a bit more effort and luck the poor catch can be rectified, and why believe the marine biologist now, when he/she was seemingly so wrong and alarmist before? The non-linearity creates confusion in the attribution of causality – what causes what in the system – and such confusion is a typical symptom of dynamic complexity.