Ecosystem assessments of fisheries based only on survey data will often have to use surveys that were designed historically for special purposes, e.g. for assessing abundances of two or three target species, or for tuning VPAs. An important question then is whether the previously collected data can provide informative time series of abundance indices and other state indicators for a wider range of target and non-target species. Some potential shortcomings of existing data series are treated in this paper leading to four questions which can guide the user to determine the suitability of an existing time series: did the survey cover the stock of each species adequately? Did survey catchability vary significantly between length or age classes? Did survey catchabilities vary significantly in space or time? Was the sampling effort sufficient? Simple methods for investigating these questions are proposed and illustrated with examples.

The mean or average value of a measure or indicator variable used to monitor a fish population can be defined in various ways, each with a correspondingly different statistical estimator for use in the context of a trawl survey. When, as is typical for many species, fish are heterogeneously clustered over space, then scooped in clusters from restricted localities using a trawl, these different estimators can produce different sample mean values with contrasting variations over time, possibly leading to different inferences about the fish population. Two mean parameters and their intuitive estimators, the mean “over fish”, and the mean “over stations”, are discussed and found to present contrasting statistical properties. A third estimator based on fitting a mixed model is proposed which has intermediate properties based on the within-haul correlation. The three estimators are applied illustratively to length data for cod caught in the North Sea by the English groundfish survey from 1992 to 2007. The time series of the mean over stations was smoothest, that for the mean over fish much more variable, and the mixed mean fell between in all years. Variance estimators derived from the fitted mixed model are also put forward. Estimates made from the example suggested that the mixed mean is most efficient. The type of estimator used for the mean should always be considered carefully and mentioned when reporting indicator studies.

This paper reviews a selection of nine indicators for assessing the biological state of a fish population monitored only by using trawl surveys in which catchability is assumed constant from time to time and place to place. The intention is to inform or remind of the range of possibilities and of associated practical and statistical issues. The indicators, grouped according to relationships with abundance, length, age, weight, and reproduction, are documented in a standard format covering estimation, interpretability, sensitivity, considerations for reference points where possible, and an example of their application. Lastly, selected indicators together with year-class curves fitted to abundance indices-at-age are applied to the North Sea cod to demonstrate the feasibility of an assessment based on only one trawl survey. Results support findings of the International Council for the Exploration of the Sea (ICES) working group on the assessment of demersal stocks in the North Sea and Skagerrak.

This paper presents the spatial indicators used in the European project FISBOAT. These are statistics intended to capture spatial patterns of fish populations, using fish density data collected during scientific surveys. To handle diffuse population limits, indicators are designed not to depend on arbitrary delineation of the domain. They characterize the location (centre of gravity and spatial patches), occupation of space (inertia, isotropy, positive area, spreading area and equivalent area) and microstructure. Collocation between different populations is summarized by a global index of collocation. These spatial indicators have the potential to be used in a monitoring system to detect changes in spatial distribution. They could be helpful for relating the spatial distribution properties of fish stocks to their dynamics, their habitats, or to climate change.

Quantitative analysis of the spatial patterns in age-structured fish populations provides a useful complement to stock assessment methods. The spatial distribution of an age-structured population is characterized by indicators for location, dispersion and aggregation, which are estimated from survey data. The times series of spatial indicators are organised in a 3D data structure with dimensions along indicators, age classes and years. Multi factor analysis (MFA) is applied to quantify the reproducibility in time of the multivariate structure between age classes and indicators. MFA computes the mean annual pattern over the age classes. The deviation around that pattern in each year is also quantified. A multivariate indicator that characterise the spatial pattern over all age classes is developed. Its variation in time is evaluated with quality control methods. The multivariate indicator is shown to be $\chi ^{2}$ distributed and a CUSUM control chart is used to detect any significant changes in time. Software for implementing the method is presented and described. To illustrate this new method, North Sea cod data are analysed, and changes in the spatial patterns of the age-structured population are detected and interpreted.

A selection of ten uni- and multivariate, nonparametric statistical methods thought to be useful for assessing trends in indicators estimated from trawl surveys is described. Nonparametric methods make minimal assumptions about the data and can therefore be suitable when parametric modelling methods, as typically preferred by fisheries scientists, are not suitable. The various methods described are sensitive to a variety of specific features of a trend thereby allowing many different forms of variation over time to be identified, e.g. changes in level, sloping straight or curved lines, and differing cyclicities. In contrast, model-based investigations can be constrained in this regard if the single model which is adopted and fitted to the trend is a simplification of its form.

Statistical Process Control (SPC) methods are extensively used in manufacturing contexts to monitor production processes. This article illustrates their potential utility for monitoring the state of marine ecosystems, using survey indicators. We exemplify the use of one SPC tool, the cumulative sum (CUSUM) control chart, to detect persistent changes in the state of a system as new observations are collected, using simulated and real data. Practical guidelines are given on how the chart parameters can be tuned to achieve an acceptable compromise between the ability to detect anomalies quickly and keeping the risk of false alarms low. The common performance measures associated with control charts depend on some key assumptions being met, and the potential impact of their violation is indicated.

Fisheries research monitoring surveys provide an ensemble of measurements on fish stocks and their environment. Because the interannual variability in such survey-based indicators is high and because diagnostics on fish stocks cannot be based on noise, our concern is to make use of what is continuous in time to obtain a reliable diagnostic. In this paper, we show how min/max autocorrelation factors (MAFs) can be useful for assessing the status of a fish stock. Indeed, MAFs will allow us to (i) summarize the multivariate indicator signals into orthogonal factors that are continuous in time, (ii) select those indicators that carry the major signal in time, and (iii) forecast stock status by modelling the time continuity of the MAFs. These different potential uses of MAFs in an indicator-based approach to assessment were illustrated with North Sea cod, for which a suite of biological and spatial indicators are available over a 21-year survey series.

One method to assess fish stocks using a suite of indicators is the traffic light approach. In this approach, the time series of the different indicators are mapped on a common colour scale to highlight alerts that occur when indicators cross reference limit values. Until now, however, the procedure has lacked a statistical framework. Here, we propose the cumulative sum (CUSUM) monitoring scheme as a suitable statistical framework. CUSUM is a statistical process control method that detects deviations from a reference mean, according to defined performance criteria. With the CUSUM monitoring scheme, alarm signals can be triggered when indicators cross defined in-control limits that correspond to defined probabilities of false alarm and non-alarm (i.e., precision and power of the CUSUM monitoring scheme). A table of CUSUM out-of-control deviations is constructed to serve as a diagnostics table. In this table, the deviations in the different indicators are quantitative and given in similar units of variance, which facilitates their integrated assessment. The CUSUM out-of-control table also shows how deviations accumulate over time and thus provides a view of the stock history. The procedure was applied to the North Sea cod stock to illustrate how a fishery-independent integrated assessment can be achieved using a suite of indicators derived from research survey data. The indicators used were related to the spatial distribution, abundance, length structure, length at maturity and apparent mortality. The stock was found to be outside its reference limits from 2001 and has shown a continued degradation in status since this time.

Large uncertainties in catch data (officially-reported landings and discards) are undermining the ability of scientific organisations to provide valid management advice based on the conventional approach of analytical stock assessments. There is thus an urgent need to consider alternative tools that do not depend on long series of precise age-structured catch data. This paper presents four fishery-independent assessment models developed under the EU project FISBOAT (Fishery Independent Survey Based Operational Assessment Tools). It also reports on rudimentary tests based on simulated data, using the same data sets and protocol as an evaluation study conducted by the US National Research Council in 1997. The survey-based assessment models at hand are able to reliably capture the major signal in biomass and recruitment, although they smooth out transient changes. However, they cannot provide absolute abundance estimates, only relative values on an arbitrary scale. The survey-based approaches could provide more rapid updates of the state of stocks than catch-based methods.

Consideration of control rules as part of fisheries management plans is becoming a common practice. Control rules may be simple or complex and utilise model- or observation-based indicators of the status of the stock to control exploitation through limitation of effort or catches. Rules that control harvest (HCR) often utilize commercial fishery data in the calculation of future catches. However, commercial fishery data may be inaccurate and that can compromise the effectiveness of the HCR in maintaining the stock for secure future harvesting. The idea of fishery-independent HCR has arisen because of these problems. Indicators of the status of a stock are derived from standardised fish surveys, commonly trawl surveys repeated annually. A management strategy that relies on such information also requires adoption of appropriate and effective control rules that utilise such information. Here, we draw attention to diverse HCR, including fishery-dependent examples that have been considered, tested or already adopted for fisheries management. Our intention is to stimulate ideas for HCR that are usable with, or adaptable to fishery-independent indicators.

Fisheries libraries in R (FLR) is an open-source software framework based in the R statistical software environment, with the main purpose of allowing for the efficient simulation and evaluation of fisheries management strategies. This paper details a generic (non-spatial) biological operating model used to simulate populations and their reaction to fishing, and an observation error model and methods for simulating structured sampling/observation error and bias in key fisheries assessment indices. It also contains the details of how the candidate harvest control rules, based on survey-derived information, can be characterised and implemented, and how the processes of management strategy evaluation and stock assessment can be included in the simulations.

ALADYM (age-length based dynamic model) is an age-length based simulation model designed to predict the consequences of management scenarios on a single population, in terms of different metrics and indicators. Removals are simulated as total mortality modulated using a selectivity pattern and a fishing activity coefficient. ALADYM uses the classical equations of population dynamics to create a tool that uses fishery-independent information (e.g. from trawl surveys) as the primary source. The formulation of population dynamics at a small time scale (month) allows to model the effects of harvest controls that evolve through the year, which is particularly useful for fast growing species. The possibility of specifying a vector of natural mortality by age/length makes the model more suitable for situations where stocks are exploited at an early life stage, as in many Mediterranean fisheries, while introducing variations along the time of the harvesting pattern accounts for non-equilibrium situations. Stochastic effects can be added to some key life-history traits to incorporate uncertainty in input parameters and in their relationships. ALADYM can be used for a range of applications such as comparing management strategies, evaluating indicator changes or searching for reference points.

Ideas and considerations are put forward for managing fisheries and marine populations using primarily trawl surveys to supply biological and spatial indicators of the state of stocks, and to permit catch per unit effort (CPUE)-based assessments. Trawl surveys seldom allow absolute estimates of fish population sizes but, if appropriately located, timed, and designed, can provide a broad range of information about catchable fish species and the ecosystem that supports them. This information may be more conducive to sustainable management of fisheries than the traditional focus on the abundances of selected stocks. The paper first briefly proposes how survey-based methods might supplement existing fishery-dependent stock assessments, as would be necessary during a transition phase to a more ecosystem-orientated system of management. Full survey-based management is then considered in relation to management objectives, the selection of indicators, survey design, reference periods, levels and directions, statistical aspects, CPUE-based assessments, and management responses to good and bad signals from the ecosystem. We argue that existing fishery-dependent stock assessments cannot be claimed to produce absolute estimates of stock abundance and fishing mortality because natural mortality (M) is seldom known accurately and, therefore, that they should not be presumed superior to the relative information from surveys, and an agreeable form of adaptive management.