Regular timely survey at all landing centers will give more or less rough picture. Get the tally as to how much in being transported to hinterland centers by checking at transportation points like Railways,road transport etc. In cases where hinterland market is not accessed in fresh condition,one can get idea from drying & curing centers, as to how much that would amount to in fresh condition.In case of trawlers,for e.g in India,roughly 80% of catch belongs to trash fish,& from total trawler landings,the amount of prime fish landed can be found out. If the season of fishery is known as well as each of the size ranges,& the amount of units of nets,one can get some idea as to how much of fish is landed by gill netters.

Even illegal fish harvesters would be using some sort of gear, & one can get some idea of the amount landed.

Amount of fish caught & landed during night time is difficult to arrive at, unless the centers are surveyed even at that time.

Monitoring at all landing centers is altogether a tough task. Even with all the precautions, I agree it's not possible to get the correct figure,but still one may be say some 10% +/- or less off the mark.

At-sea fishery observers can provide good data on commercial fishing activities in terms of both catch and effort. Obviously, this is a labour intensive approach, but where data can be relayed to shore at regular intervals, it allows for targeted stocks to be modelled in real-time, which in turn facilitates dynamic management of the fishery. The Falkland Islands' fishery is a good example.

There is no one answer. The data collection must be tailored to the details of the fishery. There is a surfclam fishery in the U.S. that lands clams which have no value unless processed at one of a few specialized factories. The catches can be effectively monitored at the factories. That doesn't work with the scallop fishery in the same area because unreported scallop meats could be sold direct to consumers if catches were not monitored where and when they are landed. Observers can be a very valuable part of the system (almost the only way to get good data on discards) but 100% coverage is only affordable if the fishery uses large vessels (factory trawlers and the like). Mandatory monitoring at landing points can work well but it probably requires a complex system for hailing in some hours before landing and even then may need satellite monitoring of the boats to ensure that they are not partially offloading illegally before making their monitored landing. Very small-scale fisheries, including most recreational fisheries and many commercial ones, probably cannot be effectively monitored at all, especially if there are official limits on catches that some fishermen are trying to evade.

What might work with a factory trawler in the Bering Sea and what might work with a fleet of canoes paddled out from some tropical beach will not be the same!

It depends very much on the context/country/types of fleets you are looking at. The way you would collect catch data in iceland is different from how you would do it in sierra leone. There is no one-size-fits-all answer or approach.

It is important to provide logbooks to the fishers to fill in daily activities (including catches, positions of each tow, depth etc). Onboard observers will then verify these information on a daily basis. This way you should be able to collect fisheries-dependent catch statistics.

Thank you all very much for your answers. Obviously, there is no single answer to the problem and multiple parameters should be taken into account. However, the aspect of this issue that troubles me the most is how to trust the data. For example, the small scale fishermen may accuse the industrial fishermen that they give money to the authorities in order to avoid inspections. In addition, even if they get inspected, the trawlers may have already discarded over 60% of their catch without reporting it, and this is not accounted for in the data collection.

Irianna, i think you need to provide the basic full setting of the fisheries you are trying to collect data from, before you can attempt to come up with a coherent answer on how best to collect them. How to trust data depends on how and by whom they are being collected, what means there are to verify their truthfulness (cross-checking), and what mechanisms are in palce to sanction offenders. In high stakes fisheries with little or no effective sanctions on the books and/or law enforcement, you may safely assume that the vast body of data will be skewed in the direction that best serves the economic interests of the operators.

That is exactly the problem. My research compares Japanese and Greek fisheries and while in Japan there is a relatively reliable data-collecting system, there is no such thing in the Greek case. There is no real data-collection plan and all efforts have been sporadic and takes into no account the small-scale fisheries. In addition, corruption and personal interests make the data-collection system even more unreliable. That is the reason why I was wondering what kind of a system could maximise the possibilities of collecting reliable data in a mistrusting environment.

The attached publication might be of use to you.

Not sure what you mean by safe- I intepret your question as robust enough for science, then the only answer is get buy in from the fishers. If they acknowledge the value and use of the science, then the information flow will be stronger than any command and control approaches.The best planning and monitoring can never replace buy in. You are unlikely to rectify previous approaches in your time series and cross cutlural differences are the most challenging. Our experience with North Atlantic mackerel, one of the largest economic fisheries in Europe, has shown this. In the end we had to use stock assessment models that assumed misreporting occured in variable amounts over different periods. This is not that helpful an answer to your current problem, but you will have to use assumptions tio get around the variable data quality.

In some countries, fishermen are shy to provide truthfully their catch data because they may be used by their authorities for taxation purpose, such as income tax, or by monopolistic fish-buyers to whom the fishermen are obliged to deliver all their fish. In such cases, the best way is to create a system, by which such data are supplied to fishery statisticians anonymously.