Species Distribution Modelling via Feature Engineering and Machine Learning for Pelagic Fishes in the Mediterranean Sea

Abstract

In this work a fish species distribution model (SDM) was developed, by merging species occurrence data with environmental layers, with the scope to produce high resolution habitability maps for the whole Mediterranean Sea. The final model is capable to predict the probability of occurrence of each fish species at any location in the Mediterranean Sea. Eight pelagic, commercial fish species were selected for this study namely Engraulis encrasicolus, Sardina pilchardus, Sardinella aurita, Scomber colias, Scomber scombrus, Spicara smaris, Thunnus thynnus and Xiphias gladius. The SDM environmental predictors were obtained from the databases of Copernicus Marine Environmental Service (CMEMS) and the European Marine Observation and Data Network (EMODnet). The probabilities of fish occurrence data in low resolution and with several gaps were obtained from Aquamaps (FAO Fishbase). Data pre-processing involved feature engineering to construct 6830 features, representing the distribution of several mean-monthly environmental variables, covering a time-span of 10 years. Feature selection with the ensemble Reciprocal Ranking method was used to rank the features according to their relative importance. This technique increased model’s performance by 34%. Ten machine learning algorithms were then applied and tested based on their overall performance per species. The XGBoost algorithm performed better and was used as the final model. Feature categories were explored, with neighbor-based, extreme values, monthly and surface ones contributing most to the model. Environmental variables like salinity, temperature, distance to coast, dissolved oxygen and nitrate were found the strongest ones in predicting the probability of occurrence for the above eight species.

Effect of climatic oscillations on small pelagic fisheries and its economic profit in the Gulf of Cadiz

“…Their cohabitation has also been demonstrated specifically in our study area, the Gulf of Cadiz (Perez-Rubín and Mafalda 2004). The partial spatial coincidence and the same vertical distribution of larvae, as well as their morphological similarity, may result in competition for food, as has been shown by other authors (Morote et al 2008;Schismenou et al 2008;Macías et al 2014;Maynou et al 2014;Albo-Puigserver et al 2019;Effrosynidis et al 2020). The numerical abundance of European anchovy versus Round sardinella has also been shown to be one of the reasons for displacement between species, so different temperature windows caused by different climatic indices may favour the presence of one species or another (Palomera and Sabatés 1990;Raab et al 2013;Diankha et al 2015).…”

Identifying seasonal distribution patterns of fin whales across the Scotia Sea and the Antarctic Peninsula region using a novel approach combining habitat suitability models and ensemble learning methods

“…The combination of species distribution models and machine learning methods is an emerging field in ecology [e.g. (Effrosynidis et al, 2020;Beery et al, 2021)]. In our study, the combination of SDM and a random forest ensemble learning algorithm enabled us to derive information on abundance from a multi-source heterogeneous data set of mainly opportunistic sighting records and to present an additional set of abundance estimates for some of the most frequently visited research areas in the Southern Ocean.…”

Evidence from interpretable machine learning to inform spatial management of Palau's tuna fisheries

“…We then used ML‐based modeling approaches (Kuhn & Johnson, 2013), coupled with recent developments in explainable artificial intelligence (AI) tools (Lundberg et al, 2020; Qiu et al, 2022; Scholbeck et al, 2020), to derive interpretable species‐specific and spatially resolved catch predictions for pelagic longline fishing fleets that operate in the Palau EEZ. ML approaches are increasingly used in a wide range of knowledge domains including medicine, finance, geoscience, ecology, paleobiology, climatology, fisheries, marine spatial planning, and economics to derive informed predictions from data that could include spatial–temporal structures, nonlinear predictor functional form, and complex predictor interactions (Bergen et al, 2023; Dedman et al, 2017; Effrosynidis et al, 2020; Foster et al, 2022; Gerassis et al, 2021; Sokhansanj & Rosen, 2022; Viquerat et al, 2022; Yang et al, 2022). ML‐based approaches are powerful tools for applied predictive modeling and make few assumptions about data structures (Kuhn & Johnson, 2013).…”