Andreas Palialexis

Modelling of essential fish habitat based on remote sensing, spatial analysis and GIS

We review the variety of existing modelling approaches applied to species habitat mapping and we discuss issues arising from the availability and nature of sampled biological data and corresponding ecological and environmental habitat descriptors, as well as the different spatial analysis approaches that are selected according to specific hypotheses. We focus on marine species habitat mapping, presenting an overview of work on modelling fish habitat carried out through a European Communities Policy-Support Action, EnviEFH ‘Environmental Approach to Essential Fish Habitat (EFH) Designation’ (2005–2008). The selection of the appropriate habitat model is dataset-specific and the resulting EFH maps are often similar in spite of using different models. Derived EFH maps are based on either environmental ranges (used as minimum and maximum environmental habitat descriptors) or probability of occurrence values. We apply model outputs to regions larger than sampled areas making use of the capacity of satellite data to cover wide areas.

Identifying essential fish habitat for small pelagic species in Spanish Mediterranean waters

Populations of small pelagic fish support important fisheries in Spanish Mediterranean waters, particularly sardine (Sardina pilchardus) and anchovy (Engraulis encrasicolus), and are distributed along the entire length of the Spanish continental shelf. Using annual acoustic survey data for the years 2003–2005, a GIS-based environmental modelling approach was used to investigate the distribution and abundance of small pelagic fish in relation to environmental variables. Multivariate analysis was applied to provide a preliminary picture of relationships between fish and environmental conditions, followed by application of Generalised Additive Models (GAMs). GAMs showed the presence/absence of fish to be related to bathymetry, sea surface chlorophyll-a concentration and sea surface temperature (SST). The strength and significance of these relationships varied spatially and temporally. High resolution Essential Fish Habitat (EFH) maps were generated for sardine and anchovy, based on the predicted probability of presence of each species. Substantial inter-annual variability in the distribution and quality of EFH was observed, particularly for anchovy. Identification of EFH is of great importance to assess and manage sardine and anchovy resources as it provides a natural link between population dynamics features and geographical scenarios.

Critical regions: A GIS-based model of marine productivity hotspots

Abstract.Marine productivity hotspots can be defined as areas of high chlorophyll concentration and low temperature distribution. Here we demonstrate how Geographic Information Systems (GIS) can be used to define areas of high productivity and we use fishery data to verify model results and reveal areas of potential fishery locations. We develop a GIS model based on spatial data integration between monthly satellite imagery of Advanced Very High Resolution Radiometer (AVHRR) sea surface temperature (SST) and Sea-viewing Wide Field-of-view Sensor (SeaWiFS) chlorophyll concentration (Chl-a) for Eastern Mediterranean waters. Data are analyzed for the production of climatology and mapping of anomaly distribution in these environmental parameters. The geographic distributions of anomalies are spatially integrated to map areas of combined persistent environmental anomaly of below average SST and above average Chl-a (lowSST/highChl-a) indicating marine productivity hotspots and potential fish feeding aggregation regions. A preliminary comparison of surveyed fisheries production data for small pelagic fish and pelagic cephalopods to lowSST/highChl-a areas is performed for the period from December 1997 to November 2001, revealing potential spatial associations among fishery resources and productivity hotspots in terms of unexploited regions as well as overexploited fishing grounds that could be characterized as candidate marine protected areas.

Influences of environmental variability on the population structure and distribution patterns of the short-fin squid Illex coindetii (Cephalopoda: Ommastrephidae) in the Eastern Ionian Sea

Illex coindetii is the most common benthic ommastrephid squid exploited by bottom trawl in the Mediterranean Sea. A recent study examining trends in population metrics based on data from the International Bottom Trawl Surveys in the Mediterranean (MEDITS) has shown an increasing trend in average lengths of I. coindetii sampled in the eastern Ionian Sea, suggesting that this is possibly due to the thermoaline circulation reversal and the warming of the entire water column in this area after 1998 because of the Eastern Mediterranean Transient (EMT). In this study, spatial and inter-annual variability of the population structure and the distribution patterns of I. coindetii in the eastern Ionian is investigated in relation to environmental characteristics. Datasets used include demographical and biological data derived from the MEDITS surveys carried out during the summers of 1994–2006 and environmental parameters obtained through CTD sampling and remotely sensed imagery. Analyses performed include descriptive methods like the length frequency analyses and GIS mapping of standardized densities, as well as habitat modelling of presence data for different life stages of I. coindetii (based on generalized additive models and maximum entropy). An increasing trend in frequency of occurrence and density indices for all life stages has been detected during the studied period, being more evident after 1999. At the same time a temporal shift in seasonal maturation was also evidenced, coinciding to the warming of eastern Ionian waters due to EMT effects. Spawning aggregations occurred steadily on the upper slope of the relatively more protected area west of the Kefalonia Island, extending also over the lower shelf of the Patraikos Gulf in summers with lower upwelling activity. The inter-annual variability observed in the location of main recruitment areas could be related to mechanisms affecting post-hatching dispersal during late winter and spring, when maximum intensity of surface mesoscale activity has been observed. Depth was found the principal variable associated with the distribution of I. coindetii in the study area. Anomalies of surface temperature and chlorophyll-a, as well as distances from coast and thermal fronts were also selected for the final set of response variables without, however, showing any clear trend particularly in the case of recruits. The results of this study highlight the importance of a more comprehensive study of both environmental processes and I. coindetii life history in the Ionian Sea, in order to improve the understanding of its dynamics and the links with environmental variability.

Modelling the presence of anchovy Engraulis encrasicolus in the Aegean Sea during early summer, based on satellite environmental data

Acoustic and satellite environmental data as well as bathymetry data were used to model the presence of anchovy, Engraulis encrasicolus during early summer in the northern Aegean Sea (Eastern Mediterranean). Generalized Additive Models (GAMs) were used for modelling and subsequently applied in a predictive mode to identify those areas in the Greek Seas and the entire Mediterranean basin that could support species’ presence. Model results were evaluated with the estimation of Receiver Operating Characteristic (ROC)-plots as well as qualitatively, based on (a) acoustic data from concurrent studies in certain areas of the northern Aegean Sea that were not included in the estimation of the GAM model and (b) historical acoustic data from the central Aegean and Ionian Seas. Mapping the estimated environmental conditions in the Mediterranean basin indicated areas that generally agree with the known distribution grounds of anchovy, such as the straits of Sicily and coastal waters of Tunisia, areas in the Tyrrhenian Sea, the Adriatic Sea, the Gulf of Lions and the Catalan Sea.

Prediction of marine species distribution from presence–absence acoustic data: comparing the fitting efficiency and the predictive capacity of conventional and novel distribution models

The accurate representation of species distribution derived from sampled data is essential for management purposes and to underpin population modelling. Additionally, the prediction of species distribution for an expanded area, beyond the sampling area can reduce sampling costs. Here, several well-established and recently developed habitat modelling techniques are investigated in order to identify the most suitable approach to use with presence–absence acoustic data. The fitting efficiency of the modelling techniques are initially tested on the training dataset while their predictive capacity is evaluated using a verification set. For the comparison among models, Receiver Operating Characteristics (ROC), Kappa statistics, correlation and confusion matrices are used. Boosted Regression Trees (BRT) and Associative Neural Networks (ASNN), which are both within the machine learning category, outperformed the other modelling approaches tested.

Habitat discrimination of juvenile sardines in the Aegean Sea using remotely sensed environmental data

Despite the importance of the recruitment process for small pelagic fish and the high economic importance of European sardine (Sardina pilchardus, Walbaum 1792) in the Mediterranean Sea, knowledge on the distribution and environmental characteristics of its nursery grounds is very limited. In the present study, we used pelagic trawl data collected during 1995–2006 to explore the spatial distribution of sardine juveniles in the Aegean Sea in early summer. Based on sardine abundance per length class, a cluster analysis was initially used to define hauls dominated by juveniles. In a subsequent step, Discriminant Function Analysis (DFA) was applied to discriminate stations with high relative abundance of juveniles using satellite environmental data and bottom depth. The parameters contributing mostly to the discrimination of juvenile grounds were sea level anomaly, photosynthetically active radiation, sea surface temperature, chlorophyll-α and bottom depth. The classification functions of DFA were finally used to post classify unsampled areas in the Greek Seas and the Mediterranean Sea in order to map grounds that meet characteristic environmental conditions for young sardine. Such areas were mostly located inshore, in semi-closed productive areas and often in proximity to river mouths, a pattern that is generally supported by existing information.

Marine GIS: Identification of mesoscale oceanic thermal fronts

A new method (the ‘sink’ method) is proposed for the mapping of productive mesoscale oceanic thermal fronts based on the combined analysis of satellite imagery for sea surface temperature (SST) distribution and chlorophyll (CHL) concentration under a Geographic Information System (GIS). In an SST lattice data array, data value sinks describe heterogeneous drops in SST distribution as this is compared to more homogeneously distributed SST in the surrounding area. Using Arc/Info GIS, these thermal discontinuities are flagged and applied on SST and CHL imagery for the calculation of differences in SST and CHL patterns (DSST and DCHL) inside and outside of these flagged areas. Spatially connected sinks that are characterized by simultaneous negative DSST and positive DCHL pattern are mapped as thermal fronts. Results include a time series of monthly front occurrence GIS maps in Eastern Mediterranean waters along with three calculated front characteristics (DSST, DCHL and bathymetry). Comparisons between the mapped fronts and isobath distributions present clearly visible front‐to‐isobath spatial and shape associations while spatial analysis between front and wind data explains 66% of the number of mapped fronts.

Contrasting multispecies patterns in larval fish production trace inter-annual variability in oceanographic conditions over the N.E. Aegean Sea continental shelf (Eastern Mediterranean)

The response of larval fish communities of the northeastern Aegean Sea (NEA) to interannual environmental changes is analyzed using data from four ichthyoplankton surveys covering the NEA continental shelf during June 1993, 1994, 1995, and 1996. Waters were significantly cooler, less saline and richer in zooplankton in 1993 and 1996 (‘cold’ years) than in 1994 and 1995 (‘warm’ years). A comparison of monthly SST series (1993–1997) between the NEA, the Marmara Sea, and the Western Black Sea revealed high correlations and similar trend components among these areas implying that oceanographic conditions over the NEA (and observed inter-annual differences) were most likely dominated by the properties and relative amount of Black Sea water inflow in the NEA. The relative composition of the larval fish community was significantly related to the ‘cold/warm’ regime and larval diversity was higher during the warm years. Larvae of the small-sized pelagic species, such as anchovy (Engraulis encrasicolus), and most mesopelagic fishes were relatively more abundant during the cold, zooplankton-rich years. Larvae of the middle-sized pelagics (Sardinella aurita, Scomber japonicus, Trachurus mediterraneaus, Auxis rochei) and certain benthopelagic species exhibited an opposite trend, i.e., they were more abundant during the warm years or absent during the cold years. Most of these species are known to be typical summer spawners (e.g., Serranus cabrilla, Lisa saliens, Trachinus draco, and Symphurus nigrescens). Co-variation in larval fish production might be indicative of similar responses among species to changing physical and/or trophic regimes.

Towards ecosystem-based management: identifying operational food-web indicators for marine ecosystems

Modern approaches to Ecosystem-Based Management and sustainable use of marine resources must account for the myriad of pressures (interspecies, human and environmental) affecting marine ecosystems. ...