Recent and historical climate variability effects on the population dynamics of several marine species

T T E E S S I I S S D D O O C C T T O O R R A A L L Recent and historical climate variability effects on the population dynamics of several marine species Á Á n nn n n nn n g gg g g gg g e ee e e ee e l ll l l ll l a aa a a aa a M MM M M MM M a aa a a aa a r rr r r rr rí í a aa a a aa a C CC C C CC C a aa a a aa a b bb b b bb b a aa a a aa a l ll l l ll l l ll l l ll l e ee e e ee e r rr r r rr r o oo o o oo o A AA A A AA A l ll l l ll l f ff f f ff f o oo o o oo o n nn n n nn n s ss s s ss s o oo o o oo o L LL L L LL L a aa a a aa a s ss s s ss s P PP P P PP P a aa a a aa a l ll l l ll l m mm m m mm m a aa a a aa a s ss s s ss s d dd d d dd d e ee e e ee e G GG G G GG G r rr r r rr r a aa a a aa a n nn n n nn n C CC C C CC C a aa a a aa a n nn n n nn n a aa a a aa a r rr r r rr r i ii i i ii i a aa a a aa a A AA A A AA A b bb b b bb b r rr r r rr r i ii i i ii i l ll l l ll l 2 22 2 2 22 2 0 00 0 0 00 0 1 11 1 1 11 1 1 11 1 1 11 1 D/Dª Juan Luis Gómez Pinchetti SECRETARIO/A DEL DEPARTAMENTO DE Biología DE LA UNIVERSIDAD DE LAS PALMAS DE GRAN CANARIA, CERTIFICA, Que el Consejo de Doctores del Departamento en sesión extraordinaria tomó el acuerdo de dar el consentimiento para su tramitación, a la tesis doctoral titulada “Recent and historical climate variability effects on the population dynamics of several marine species” presentada por el/la doctorando/a D/Dª. Ángela María Caballero Alfonso y dirigida por el Dr. José Juan Castro Hernández, el Dr. Unai Ganzedo López y el Dr. Ángelo Santana del Pino. Y para que así conste, y a efectos de lo previsto en el Artº 73.2 del Reglamento de Estudios de Doctorado de esta Universidad, firmo la presente en Las Palmas de Gran Canaria, a 1 de Marzo de 2011. Programa de doctorado de Oceanografía. Bienio 2006-2008. Con Mención de Calidad de la ANECA. Título de la Tesis: Recent and historical climate variability effects on the population dynamics of several marine species (Efecto de la variabilidad climática reciente e histórica en la dinámica poblacional de varias especies marinas) Tesis doctoral presentada por Dª. Ángela María Caballero Alfonso para obtener el grado de Doctor por la Universidad de Las Palmas de Gran Canaria. Dirigida por Dr. D. José Juan Castro Hernández Dr. D. Unai Ganzedo López Dr. D. Ángelo Santana del Pino El/la Director/a El/la Co-Director/a El/la Co-Director/a El/la Doctorando A mis padres, Mª Carmen y José Porque sin todo lo que me han dado, no habría llegado hasta aquí. A mi hermano, Néstor J. Por ser “el pequeño” y, sin embargo, el que siempre me marca el ritmo. Por ser mi mayor confidente, por tener siempre palabras de aliento para mí. A Fede Por su incondicional apoyo, comprensión y paciencia. Recent and historical climate variability effects on the population dynamics of several marine species      Efectos de la variabilidad climática reciente e histórica en la dinámica poblacional de varias especies marinas Binding references The images of the puzzles were obtained from: - The Ocythoe tuberculata: http://www.sealifebase.org/Summary/speciesSummary.php?id=58224&lang=laos - The Octopus vulgaris tentacles: http://panoramaacuicola.com/noticias/2010/03/18/disenan_nuevas_estructuras_de_en gorde_para_pulpo_en_cautevidad_en_espana.html - The sun image: http://www.astro-digital.com/11/sol.html - The Sardinops caeruleus: http://www.esacademic.com/dic.nsf/eswiki/1098242 - The Engraulis mordax: http://blog.pescaderiascorunesas.es/blog/el-boqueron-o-anchoa-del-cantabrico - The clouds: http://firmas.lasprovincias.es/antoniorivera/las-nubes-no-son-vapor-de-agua - The Thunnus thynnus: http://fishindex.blogspot.com/2008/08/northern-bluefin-tuna-thunnus-thynnus.html      Referencias de la portada Las imágenes del puzzle se obtuvieron de: - La de Ocythoe tuberculata: http://www.sealifebase.org/Summary/speciesSummary.php?id=58224&lang=laos - La de Octopus vulgaris tentacles: http://panoramaacuicola.com/noticias/2010/03/18/disenan_nuevas_estructuras_de_en gorde_para_pulpo_en_cautevidad_en_espana.html - La de del sol: http://www.astro-digital.com/11/sol.html - La de Sardinops caeruleus: http://www.esacademic.com/dic.nsf/eswiki/1098242 - La de Engraulis mordax: http://blog.pescaderiascorunesas.es/blog/el-boqueron-o-anchoa-del-cantabrico - Las nubes: http://firmas.lasprovincias.es/antoniorivera/las-nubes-no-son-vapor-de-agua - La de Thunnus thynnus: http://fishindex.blogspot.com/2008/08/northern-bluefin-tuna-thunnus-thynnus.html    ACKNOWLEDGEMENTS /AGRADECIMIENTOS    Han pasado los años, casi doce ya, y aún recuerdo con claridad el verano de 1999 cuando me matriculé en la licenciatura de Ciencias del Mar. Quién me iba a decir ese día que llegaría a este momento en que puedo compartirlo contigo. Por aquel entonces, ésta no era para mí Mi Profesión, yo tenía claro que quería trabajar en la rama sanitaria…Mi primera opción era Medicina, soñaba con hacer Neurología; en segundo lugar me preinscribí en Fisioterapia y como tercera alternativa, en Enfermería. De ahí para abajo, hasta completar las 8 opciones de la preinscripción para la Universidad, todo eran rellenos. No muy lejos, en cuarto lugar, estaba Ciencias del Mar; en la que me matriculé para usarla como “trampolín” para acceder luego al campo de Ciencias de la Salud. Lo que yo en ese momento desconocía era que al poco tiempo se me abriría un mar (y nunca mejor dicho) mucho más profundo de lo que creía. No eran sólo peces y cetáceos lo que me mostraban, era todo un compendio de ciencias (física, química, matemáticas, geología, biología,…) que me dieron una visión nueva de lo que nos rodea, un medio tan complejo y desconocido que estaba lleno de un encanto que las cámaras de video y fotografía jamás lograrán captar…Pero justo cuando empecé a considerar el hacer de esta carrera Mi Profesión, alguien que creía conocerme se cruzó en mi camino y tras la típica conversación protocolaria y sin interés de: “¿Cómo estás?, ¿qué estás haciendo?”, le faltó tiempo para que su rostro dibujara una sonrisa sarcástica, medio torcida, y me dijera: “¿Con lo mal que se te dan las Matemáticas?. Tú no sirves para eso, no conseguirás acabarla”. ¿Cómo me sentí? Es difícil de describir, pero creo que dolida, insegura y soberbia en proporciones iguales, es una buena combinación de adjetivos para describir los días (incluso puede que semanas) que siguieron a esa conversación. Sin embargo, tras unos primeros pasos temblorosos e inseguros, conseguí comenzar a pisar con más fuerza, logrando esforzarme por dar lo mejor de mí cada día. En ese momento aprendí que nadie tiene derecho a poner límites a lo que podemos ser, excepto nosotros mismos. En ese instante decidí que mis metas las marco yo, con mayor o menor acierto, con más o menos dificultad, pero siendo yo la que diga hasta dónde. ¿Sencillo? No, desde luego no me resultó fácil llegar a esa confianza en mí misma que aún hoy a veces se tambalea, pero como entonces, tengo a mi lado a mis padres, José y Mª Carmen, y a mi hermano, Néstor. Para quienes va mi mayor agradecimiento. Sin ellos desde luego no sería quien soy, ni podría haber llegado hasta aquí. No tendré tiempo, palabras ni sentimientos lo suficientemente grandes para poder agradecerles todo lo que me dan en cada instante, por estar siempre ahí, a mi lado, haciendo de guía cuando el camino me parecía muy complicado, por tener siempre las palabras necesarias para calmarme y la infinita paciencia que en ocasiones les ha requerido el ayudarme a sacar los pies de un bache…muchas veces sin ni siquiera ellos saberlo. Por todo eso y por todo lo que dejo sin escribir, para ellos siempre va lo mayor, mejor y más importante de mí. Casi a la vez que comencé con esta andadura, entró en mi vida una de las personas que mejor me conoce y que más me soporta, Fede. Sé que no me equivoco al decir que tú eres quien, después de 10 años, mejor sabe todo lo que estas páginas encierran y una gran parte de haber conseguido poner ese punto y final, te lo debo a tí. Gracias por tu apoyo incondicional, tu paciencia en los bajos momentos y tu mesura cuando todo parecía ser mejor de lo previsible. Gracias por estar dispuesto a seguir acompañándome en la etapa que se abre al cerrar este capítulo. Gracias por todo lo que eres. Son muchas las personas que en este tiempo han formado parte de mi vida, de forma más o menos cercana, pero siempre importante y aunque no me pare a nombrarlos, más por miedo a dejarme a alguien atrás que por ninguna otra razón, a todos les tengo algo que agradecer. A todos ellos, gracias. Sin embargo, quizás por estar más próximos actualmente, permítanme que haga una mención especial a quienes han estado conmigo en los últimos años de esta andadura: Francis (por hacerme ver que la programación es una herramienta útil y, si me apuras, hasta con cierto encanto), Anna (por ser siempre tan positiva y hacer que el momento de las comidas fuera realmente una pausa en el camino cada día), Pep, Alex, Sheila, Iván, Vero, Mine, Inma, Gara, Lidia, Marta, Claire, Ico,…(por estar ahí, por esas sonrisas entre carreras en medio de los pasillos intentando ganarle unos minutos al reloj para acabar algo o durante algún instante de calma para comer o al final del día). Y, por supuesto, a Isis, Mónica, Ari, Idaira y Borja (aunque con ustedes dos he compartido bastantes años más). Con ustedes, hay infinidad de recuerdos, pero uno que siempre me hace sacar una sonrisa es el de la famosa “noche de Métodos”: integrales, derivadas, desarrollos matemáticos y modelos de programación que aún hoy me parecen abstractos, luchando contra reloj para acabar a tiempo… y con la música del Rey León en algún momento bajo para levantar el ánimo; sin duda fueron 24 horas de encierro en la Facultad muy duras, pero lo cierto es que también lo pasé muy bien y al cabo del tiempo la única sensación que me queda es la alegría de haber compartido ese momento con ustedes. Así que gracias por hacerlo más llevadero, y gracias por la amistad, el apoyo, los ánimos y consejos que me han dado, por empatizar tan bien conmigo en todo lo que implica la realización de una Tesis y por haber estado codo con codo para los trámites burocráticos y las “luchas” con la Administración. Aunque lo cierto es que no se puede generalizar, porque han habido personas que han hecho que las cuestiones de papeleo sean más sencillas de lo que en un principio parecían ser, personas dispuestas a ayudar, muchas veces anticipándose a nuestros ruegos y preguntas. En especial me acuerdo de Alejandro, Cristina y también de Lola. Gracias por toda la ayuda y, sobre todo por el trabajo que hacen; sin ellos los trámites de matrícula y de becas seguramente habrían sido más complicados. En mi caso, tuve el privilegio de poder dedicarme exclusivamente a realizar esta Tesis por haber disfrutado de becas para ello (la del Cabildo de Gran Canaria de mayo a septiembre de 2007, la de la ULPGC de octubre 2007 a septiembre de 2008 y, finalmente desde entonces hasta mayo de 23    PRESENTACIÓN DE LA TESIS    La presente Tesis titulada Efecto de la variabilidad climática reciente e histórica en la dinámica de varias especies marinas, se realizó dentro del marco de los proyectos: (i) “Dinámica poblacional y posible estrategia de conservación del atún bonito-listado (Katsuwonus pelamis) en Atlántico Centro Oriental (PI042005/126)” financiado por la Consejería de Ecuación Cultura y Deportes del Gobierno de Canarias; y (ii) “Nuevo marco dinámico para la conservación y explotación sostenible de los stocks de pulpo (Octopus vulgaris) en las zonas de afloramiento del Noroeste africano (AGL2006-10448)”, financiado por el Ministerio de Ciencia e Innovación de España, ambos dirigidos por el Dr. D. José Juan Castro Hernández (Universidad de Las Palmas de Gran Canaria). El estudio que aquí se presenta, ha sido codirigido por el Dr. D. José Juan Castro Hernández (Universidad de Las Palmas de Gran Canaria. Departamento de Biología), el Dr. D. Unai Ganzedo López (Universidad del País Vasco. Departamento de Física Aplicada II) y por el Dr. D. Ángelo Santana del Pino (Universidad de Las Palmas de Gran Canaria. Departamento de Matemáticas). Esta Tesis está desarrollada en su mayor parte en inglés para poder optar a la Mención Europea del Título de Doctor de acuerdo a la normativa de la Universidad de Las Palmas de Gran Canaria (BOULPGC. Art.1 Cap. 4, 5 de noviembre 2008). El cuerpo general de la misma comienza con una Introducción, seguida de unos Objetivos, de la Metodología aplicada, las Contribuciones científicas originales realizadas, una Síntesis de los Resultados y la Discusión General para finalizar planteando las Líneas de Investigación que este trabajo dejan abiertas para el futuro. Las referencias que aparecen a lo largo del manuscrito, se encuentran detalladas al final del documento. También se presenta una sección en castellano, requerida por el 24 Reglamento de Elaboración, Tribunal, Defensa y Evaluación de Tesis Doctorales de la Universidad de Las Palmas de Gran Canaria (BOULPGC. Art.2 Cap.1, 5 de noviembre 2008). 25 “The human history has taken place in a constantly changing world, sometimes slowly, sometimes fast, with the long-term natural changes always darkened by the biggest oscillations of individual years. The environment will keep changing, somehow due to human activities, with their effects, desired or undesired, and somehow due to natural causes. Anything of these justify, certainly, the possibility of a standard of constant life or always in increase, in a long time-scale” (Humbert Lamb (1982); Climate, History, and the Modern World. Extracted from B. Fagan (2007); The Long Summer: how climate changed civilizations)      “La historia humana ha transcurrido en un mundo en cambio constante, a veces lento, a veces rápido, con la naturaleza de los cambios de larga escala temporal siempre oscurecida por las oscilaciones de mayor envergadura de los años individuales. El ambiente seguirá cambiando, en parte por causa de las actividades humanas, con sus efectos, tanto deseados como no deseados, y en parte por causas naturales. Nada de esto justifica, ciertamente, que sea posible un estándar de vida o bien constante o bien siempre en aumento, en el largo plazo” (Humbert Lamb (1982); Climate, History, and the Modern World. Extraído de B. Fagan (2007); El Largo Verano: de la era glacial a nuestros días) 27    CONTENTS / ÍNDICE    Abstract/Resumen..........................................................................................19 Thesis Preview/Presentación de la Tesis........................................................21 Introduction……………………………………………………………........31 Sporadic climate-induced changes in marine fauna.........................33 Brief climate description..................................................................39 Historical linkages between climate and fisheries............................46 Aims and Outline of this Thesis………………………………………….....51 Material and Methods.....................................................................................57 Background………………………………………………………...59 Data………………………………………………………………...62 Software.......................................……………………………….....65 Chapter 1 New record of Ocythoe tuberculata (Cephalopoda: Ocythoidae) in the North-east Atlantic related to sea warming..............................................67 Chapter 2 The role of climatic variability on the short-term fluctuations of octopus captures at the Canary Islands………..............................................75 Contents/Índice 28 Chapter 3 Climate and historic Bluefin tuna fluctuations in the Gibraltar Strait and Western Mediterranean............................................................................97 Chapter 4 Evaluation of climate synergies affecting Pacific sardine and Northern anchovy historical fluctuations off California (Sta. Barbara and Soledad basins).............................................................................................119 General Discussion.......................................................................................145 Conclusions..................................................................................................161 Future Research............................................................................................167 Spanish Summary/Resumen en Español…………..………………………171 Introducción………………………………………………………173 Objetivos y Guión de la Tesis……………………………………193 Material y Métodos……………………………………………….197 Resultados...………………………………………………………205 Discusión General……………………………………………..…215 Conclusiones……………………………………………………..233 Futuras Líneas de Investigación………………………………….237 Reference/Referencias..................................................................................239 List of Abbreviations/Lista de Abreviaturas................................................277 Contents/Índice 29 List of Abbreviations.....................................................................279 Lista de Abreviaturas.....................................................................281 Annexes/Anexos...........................................................................................283 Tables.............................................................................................285 Brief description of the climate indices.........................................293 Breve descripción de los indices climáticos……………………..297 I II I I II In nn n n nn nt tt t t tt tr rr r r rr ro oo o o oo od dd d d dd du uu u u uu uc cc c c cc ct tt t t tt ti ii i i ii io oo o o oo on nn n n nn n 33    INTRODUCTION    The interest of humans in marine populations is not something of the last centuries (Castro-Hernández, 2009). There are evidences that the Neanderthals (ca. 100 000 years ago) consumed fishes. Also cave paintings representing fishing activities with more than 25 000 years were found in caves in South Africa and Namibia. The first evidence of fishing activities from boats are from the Mesolithic (ca. 10 000 years ago) (Sahrhage and Lundbeck, 1992). This is known due to the discovered of cod, herring, conger eel... leftover in a site in Scotland. The oldest net dated was found in Peru; it has 8 800 years age (Castro-Hernández, 2009). Besides this, one might think that the overexploitation of the sea resources is something subsequent to the Industrial Revolution (1850s onward), Sahrhage and Lundbeck (1992) highlights that overfishing occurred due to an increase of the human population since the roman times (200-300 A. D.) and, associated with it, an increase in the fish demand. In any case, it is well known that this is a generalized problem after the mid XIX century (Pauly and MacLean, 2003). And this is not trivial, because human activities are not the only parameter affecting these populations, although nowadays it is the most important ones (Pauly, 2009). Marine ecosystems fluctuates naturally at a multitude of times scales, due to a combination of internal dynamics of populations, predator/prey and competitive dynamics and also because of the climate variability (Cushing, 1982; Laevastu, 1993; Lehodey et al., 2006; Barange et al., 2010). Climatic variations are more or less cyclical. It is known that they affect marine population’s abundances and migrations, from the planktonic to fish communities (Caballero-Alfonso, 2009; Drinkwater et al., 2010). Sporadic climate-induced changes in the marine fauna Environment conditions are determinant factors in the biodiversity and in the stability of habitats. It has been largely argued that global and Introduction 34 regional climate variability affects marine ecosystems in different time scales (Cushing, 1982; Ravier and Fromentin, 2004; Ganzedo et al., 2009; Caballero-Alfonso et al., 2010; Drinkwater et al., 2010). The atmosphere and the oceans are in a continuous dynamical interaction, through an exchange of energy. The ocean has a bigger capacity of storing heat than the atmosphere, which is distributed through the ocean currents and by exchanges with the atmosphere. Initially, this controls the Sea Surface Temperature (SST) and, consequently, the rest of the ocean regime shifts and nutrients distribution. This hydrodynamic variability affects the marine ecosystems (Anadón et al., 2005). Related to this, few studies warned of the influence of the increase of the carbon dioxide (CO 2 ), mainly due to anthropogenic causes after the Industrial Revolution (Fabry et al., 2008; Drinkwater et al., 2010). This process is acidifying the oceans (which have always been slightly alkaline) (Caldeira y Wickett, 2003; Feeley y colaboradores, 2004). For some species, as the phytoplankton, this remains beneficial, but for others it seems to be highly detrimental. For example, to all organisms with structures made of calcium carbonate (CaCO 3 ), because this compound reduces in acidified mediums (Kleypas et al., 1999; Riebesell et al., 2000). This means that shells and skeletons are going to be weaker with time. On the other hand, body temperature of most of fish species is environmentally dependent (cold-blooded organisms or poikilotherm) (Jobling, 1994). The temperature plays a key role in the growth, metabolism and behaviour of those animals (Ali, 1980; Huntingford and Torricelli, 1993; Godin, 1997). This parameter varies widely in the oceans. It ranges from 0 ºC in the poles up to 26 ºC in the equator in means. Every species have a small tolerance-range of temperature, and they used to live near their limit (Harley et al., 2006). For this reason, different species are occupying different geographical areas (Wootton, 1998). The more sensitive ones are those which inhabit in tropical and subtropical latitudes. Meanwhile, boreal and polar ones support better the changes in their habitat (Poulard and Blanchard, Introduction 35 2005). For example, some tuna species (e. g.: Thunnus thynnus) inhabit waters ranging between 3 and 30 ºC, although they look for the 24 ºC or higher temperatures during the spawning phase (Ganzedo, 2005; Fromentin, 2006). It is expected that when the temperature change, this individuals are forced to move out of their distribution domain. Going further, in regions where temperate and polar water species coexists, it has been seen that with the increase of the temperature, temperate species tends to increase their abundance (or expand their distribution area) (Caballero-Alfonso, 2009). Contrary, the polar species remains stable or they diminish their relative abundance (Poulard and Blanchard, 2005). In this way, small and gradual temperature variations, produces advances and retreats of population’s limits. But if drastic changes occur, mortality is expected to be the predominant consequence for sensitive species that are not going to be able to adapt themselves (Drinkwater et al., 2010). Related to the previous statement, an increasing number of species found out of their usual distribution range are quoted in the literature (Caballero-Alfonso, 2009). Globally, in every ocean, various researchers have mentioned sights and captures of unknown or unusual species in a specific region. In other cases, a change in the behaviour of common populations has been reported. For example, in the Pacific Ocean, the Pacific cod (Gadus macrocephalus) is not being fished in the northwest of the United States as in former times. Instead, it is being caught in Canada. Even in Alaska the Pacific Ocean perk (Sebastes alutus), alaska pollock (Theragra chalcogramma) and the Pacific halibut (Hippoglossus stenolepis) have varied their abundance and a succession of them has been found in the last five decades. This has been related to a change in the south Californias water temperature (Hollowed et al., 2001). In the same way, in the Bering Sea and in the Aleutians, the abundance of Pacific Ocean perk (Sebastes alutus), Pacific herring (Clupea pallasii pallasii) and the Greenland halibut (Reinhardtius hippoglossoides) are being substitute by flat fish species from Introduction 42 solar activity is better known that in former times; it is accepted, for instance, that the sunspots fluctuate between maximums and minimums in cycles of 11 years. In the 1890s, Spörer and Maunder quoted an absence of sunspots at the end of the XVII century and beginning of the XVIII (Fagan, 2000), and this period without solar activity is the one known as the Maunder Minimum. Figure 2: The evolution of the temperatures in the northern hemisphere (y-axis) over the last millennium (x-axis) according to a simulation based on subsoil temperatures (boreholes) (Source: González-Rouco, 2003. Extracted and modified from: Uriarte, 2010). From 900 to 1300 A.D. the weather was warm and stable in summer and cold and stormy in the winters. This period of the History is known as the Climate Medieval Optimum (Fagan, 2007), where the droughts were more intense than in the previous occasions. The better register of this was found in the Santa Barbara Basin (south California) (Fagan, 2007). In a marine sediment core extracted there, it can be found that from the 450 to the 1300 A.D., the sea temperatures dropped abruptly up to 1.5 ºC under the mean Temperature (ºC) Year Spörer Introduction 43 temperature in that region for the entire Holocene (coinciding with severe droughts). However, in 500-800, 980-1250 and 1650-1750 the stronger droughts were registered. Also, from 950 to 1300 the upwelling in that coast was more intense than usual, and for this reason, the fish production was very high (Fagan, 2007). This was marked between 1150 and 1200 due to a reduction in the world volcanic activity and an intensification of the sunspots (Fagan, 2007). This created climate conditions similar to the ones that are present with La Niña events (see Annexes), with the consequent droughts in the East Tropical Pacific Ocean. After 1300, the temperature started to rise and after two centuries, the upwelling intensity decreased. This derived in a less biological production, for example in a diminished of the anchovy abundance (Soutar, 1967; Soutar and Isaacs 1969, 1974). Beside this, is noteworthy that radiocarbon ( 14 C) measures in tree-rings reveal a solar activity peak between 1100 and 1250 (middle of the Climate Medieval Optimum) (Fagan, 2000). After mid 1300s, the climate was very unstable in Europe, with warmer summers and raining springs, alternating with cold periods and strong heat waves until the XVI century (Fagan, 2000). From 1315 to 1319 there were the most raining years between 1298 and 1353, coinciding with the ‘Great Famine’ (Fagan, 2000). Something similar happened between 1399 and 1403, but less intense than the former episode (Fagan, 2000). Since 1430, a succession of extremes winters took place, including frosts of more than seven years and strong storms, associated to a high pressure system located over the Scandinavian Peninsula. Between 1460 and 1550 the solar activity reached another minimum (Spörer Minimum. Figure 2). The end of the Little Ice Age (i.e.: last decade of the XVI century and beginning of the VXII) was characterized by extreme weather conditions (Fagan, 2000). Years of hot or unusual cold, as the frost in 1607, were also registered in the XVII century, as the four periods of extreme cold (1641-1643, 1666-1669, 1675 and 1698-1699) linked to the volcanic activity occurred in those years. Introduction 44 Although none of them was as intense as the one of the 1601 summer. This is due to the fact that the associated volcanic dust produced a global temperature decrease in all cases (Fagan, 2000). Independently of these, in the Pacific Ocean four drought cycles were detected in junipers and pines trunks. They were caused due to a northward displacement of the jet stream, the first of them started in 910 and finished in 1350 A.D.; but in these measures it was also seen that between XIV and X centuries the warmed four periods occurred. The most relevant one was from 1118 to 1167. The same phenomenon occurred in 1976-1977, resulting in a severe drought in the California domain (Fagan, 2000, 2008). During the XVIII century, the unforeseeable climate variability was characterized by dry and cold winters and stormy summers; alternating with some soft and humid winters and warm summers (Fagan, 2000). Many researches suggest that this climate instability concluded with actual warming tendency, after 1860. In the 1870s, the weather was warm and from 1875 summers turned very humid. Despite this, in 1879 a cold wave settled on Europe until the end of the 1880s. Between 1890 and 1940, the North Atlantic Oscillation (NAO) index was high (in the 90s it has remained high for more years than usual) (Fagan, 2000). This entailed good weather due to low pressure systems located north Europe, contrary to what was observed at the end of the XVII century. The coldest winter in the XX century was in 1963, with a mean temperature of -2 ºC (Fagan, 2000). This value is smaller than the registered in the XVII and XVIII centuries when the Thames River was frozen (Figure 3. Fagan, 2000, 2007). Introduction 45 Figure 3: Windsorians walk on frozen Thames. A view towards Windsor Bridge photographed on 24 January 1963 (http://www.thameweb.co.uk/windsor/ windsorhistory/freeze63.html). Fagan (2000) points out that one of the reason of why the actual climate variability could be a natural phenomenon is it relations with the Sun. There are various solar processes that may be having a key role as climate controlling factor (e.g.: sunspots, solar wind...). In addition, he points that the solar irradiance is never constant. In the last 20 years, solar irradiance measures revealed 11-years cycles that coincided with the sunspots cycles. Furthermore, indirect measures from tree-rings and ice-cores corroborates the existence of those cycles and confirms that long thermal fluctuations associated to the Sun have taken place in the past centuries. However, since 1950, the solar activity has been stable, and this means that changes in the solar activity roughly explain the 50% of the warming observed in the XX century (Fagan, 2000). In this context, the shift from warm period from 1930s to 1960s to the colder 1980-1990s has been called the Atlantic Multidecadal Oscillation (AMO. See Annexes), and it is an actual example of long-term climate variability. Other examples are the decadal fluctuations as the North Atlantic Oscillation (NAO. See Annexes), the El Niño/Southern Introduction 46 Oscillation (ENSO. See Annexes) or the Pacific Decadal Oscillation (PDO. See Annexes) (Kerr, 2000; Drinkwater et al., 2010). Historical linkages between climate and fisheries Since 1870s, Spencer Fullerton Baird recognized the importance of the environment forces in the observed fish stocks fluctuations (Lehodey et al., 2006; Drinkwater et al., 2010). In this sense, Cushing and Dickson (1976) argued that there are two types of possible responses to the long-term climate influence: (i) records of periods of presence/absence or (ii) periods of high/low catches; in both cases often in scale of decades due to changes in the recruitment in the incoming year class to the stock. Examples of this is the 1904 year class of Atlantoscandian herring and its successors for six decades; or the gadoid outburst in the North Sea in 1962 which still persist beyond the mechanism that caused it: the temperature decrease (Cushing, 1982). Short-term changes (i.e.: from year to year) in fisheries have also been detected, e.g. for the cod, the herring or the sockeye salmon (Cushing, 1982). What happens is that stocks with many age groups reduce the variability in the yearly recruitment: the population fecundity with which each new year class starts represents the average of many year classes (Cushing, 1982). Sharpest responses of species to climatic variabilities are difficult to detect; but unusual catches as the ones reported in the beginning of this text are some evidences of this facts. However, these last seems no relevant because catch data are usually reported by months diminishing the importance of these punctual catches (Cushing, 1982). In the 1990s the large-scale climate variability started to be considered as a physical forcing affecting marine ecology, mostly related to atmospheric patterns and their influence on sea temperature (Forchhammer and Post, 2004; Solari, 2008; Solari et al., 2010). However, the way that climate influences individual organisms, populations and communities of the marine ecosystems still unclear. This is due to the big amount of forcing and pathways that can establish connections between climate and ecosystems. A population may react directly or with delay to an Introduction 47 external force. Forcing effects are fast in low trophic levels and short life span species, but they are longer at higher trophic levels and larger life span ones (Drinkwater et al., 2010; Ottersen et al., 2010). It is evident that all marine ecosystems are not equally sensitive to the same climate change (Beaugrand et al., 2008). One example is the replacement of two copepod species, Calanus finmarchicus by C. helgolandicus, in the North Sea due to changes in sea surface temperature (SST) and in the phytoplankton production associated to the NAO variability (Fromentin and Planque, 1996). In the same way, the cod (Gadus morhua) lives in a wide geographical range in the Atlantic Ocean (Cushing, 1982; Fagan, 2000; Drinkwater et al., 2010). It can be found from the Barents Sea up to Biscay and around Iceland and Greenland. Also it lives in North America coasts. However, it is very sensitive to temperature changes, mostly to cold waters. It has it optimum development between 2 and 13 ºC; and ranging from 4 to 7 ºC in the reproductive phase. For all these features, the water mass movement and the sea temperature variability has largely affected cod populations (Fagan, 2000). Very cold conditions, as the ones during the Little Ice Age or in the XIII century, have a detrimental effect on this species (Cushing, 1982). Nevertheless, between 1845 and 1851, the temperature in the northern latitudes increased, favouring the establishment of the cod in Greenland waters (Drinkwater, 2005; Rose, 2007). The AMO (multidecadal scale) warmed the water during 1920s and 1930s, this phenomena extended the Atlantic cod distribution approximately 1200 Km northward along the West Greenland, Iceland, Barents Sea and Spitzbergen coasts. Even more spawning sites were aumented to Norway. At a decadal scale, the NAO also influenced the cod positively, favouring the recruitment in the Barents Sea through a temperature change and the food availability (Calanus finmarchicus) (Fagan, 2000; Ottersen and Stenseth, 2001; Drinkwater et al., 2010). The cod shows a non-linear response to the temperature variability. They are related through a dome-shape relationship. This response implies a Introduction 48 regime shift were climate events trigger major changes in the ecology system (Björnsson et al., 2001). On the other hand, the Pacific (or Chinook) salmon (Oncorhynchus tshawytscha) spawns in fresh water, and juveniles move to marine realm after three years, where they stay until they get sexual maturity. Then they go back to their natal river to spawn and die. This is linked to the PDO, when the sea surface temperature is relative warm (Drinkwater et al., 2010). The European anchovy (Engraulis encrasicolus) in the Santa Barbara basin (off California) is very abundant due to the upwelling system (Fagan, 2008). But during El Niño event (see Annexes) the temperatures in the surface increased due to a diminished of the ascending water and this affected the marine productivity and consequently, the European anchovy abundance (Fagan, 2008). The marine productivity has cyclically increased during the climatic cold phases, and decreased during the warm ones in this region. These are some examples of how climate might be controlling different marine populations worldwide. With this perspective, the recruitment to the gear is a good way of measuring how climate affected a species of long-life cycles. What is being caught is the result of what the climate of a certain period did to the fishery during the larvae and juvenile stages. Cushing (1982) highlighted also the fact that low fecundity species (e.g.: the herring) were more vulnerable to recruitment failures than highfecundity ones (e.g.: the cod). If the overfishing effect on the recruitment is ignored, the gadoid stocks recruitment appears to vary steadily; meanwhile the herring and salmon stocks rise and decline with time with a variable trend (Cushing, 1982). The latest tend to be more vulnerable to recruitment overfishing. The cod-like seems to be able to resist environmental changes in a higher degree than the herring-like species (Cushing, 1982). Introduction 49 The herring (Clupea harengus), for instance, is a good example of how the appearance and disappearance of a specie can affects men’s livelihood in the Middle Ages, when an alternation between Norwegian and Sweden fisheries was observed due to the frozened of the western Baltic Sea from 1200 to the XIV century and between the XIII and XV (Cushing, 1982). In 1588 and in 1680-1730 it also moved southward from Norway in a searching of waters with 3-13 ºC (Fagan, 2008). Also in this period the fisheries in the Faeroe Islands decreased due to the southward movement of the North Pole water (Fagan, 2000), what dropped the water temperature in 5 ºC less than what is register nowadays. This phenomenon also entails storms and strong winds in the British Islands region, although it is not clear how sensitive is this specie to temperature changes. However, it has been seen that it fluctuates during the centuries besides the fishing effort influence. They were very abundant during the Warm Medieval Period (Fagan, 2000; Rose, 2007) being a big fishing industry in the North Sea during the warm centuries too. Probably, due to the fish demand, and also to the temperature decrease, the herring population collapsed since the XIV century (Fagan, 2008). From the previous statements, it can be concluded that the strength of the year class varies between species, but also that some stocks are more vulnerable to climate changes than others. For all this, variability may be considered by regions, although sometimes the influence can be seen in stocks across the ocean (Cushing, 1982). It is evident that, within the possible climate parameters, most of the attention has been paid to the temperature. Perhaps because it is the most domintan climatic parameter influencing the marine ecosystems. It controls the metabolism, migrations, spawning, etc. of the population. However this does not exempt the influence of other parameters, neither diminishes their importance in the ecosystems responses. The water column stratification and the mixed-layer depth (mainly through primary production), the sea ice (mainly through spring blooms), water mass turbulence or it advection Introduction 50 (through the larvae dispersion) are also controlling the evolution of the marine populations (Drinkwater et al., 2010). The purpose of this Thesis is to increase the comprehension on how climatic factors affect marine populations, rather than to understand the climate changes in the past, the present or the future. A AA A A AA Ai ii i i ii im mm m m mm ms ss s s ss s a aa a a aa an nn n n nn nd dd d d dd d O OO O O OO Ou uu u u uu ut tt t t tt tl ll l l ll li ii i i ii in nn n n nn n e ee e e ee e Material and Methods 61 series, which are often short, irregularly sampled, and may contain missing values (Sweldens, 1998; Cazelles et al., 2008). The idea under the Linear Model (lm. Chapters 2, 3 and 4) is to look for a simple relationship between pairs of variables (not necessarily a straight-line regression). A change in the predictor variable (e.g.: climate variables; the x i ) is assumed to produce an increase or decrease in the response variable (e.g.: captures series; the y i ). This can be summarized by: y i = β 0 +β 1 x i +ε i where β are the coefficients linked to the variations of x i and ε i is the error term (Verzani, 2005). The Stationary Bootstrap (Chapters 3 and 4) allowed having robust statistical measures beside the possible non-stationarity of the series through a re-sampling with replacement method, where persistence is preserved (Politis and Romano, 1994; Mudelsee, 2003). A common feature of all these statistical techniques (including simple regressions, wavelet and the stationary bootstrap) is that they use just a pair of variables in each comparison test. However it is known that in nature, none variable acts solely on a population. To try to improve this fact, a Multivariate moving-window regression (Casals et al., 2002), which is included in the linear models, can be applied (Chapter 4). This method allows seeing in this case which variables are potentially affecting, in different periods of the time series, to the sardines and anchovies abundance. Also this gives an idea concerning if significant correlations are found at several lags among different groups of variables. Material and Methods 62 Data The series within the present Thesis were compiled from different sources. Their length and intrinsic features are also specific of each data base. For this reason, each of them has been treated in a different way (see each chapter for more detail). However, the objective was always the same and this was bored in mind during the whole study. 1. Biological Data The Ocythoe tuberulata data (Chapter 1) were obtained from the fishermen of two Spanish fishing boats, the “Leporre Anaiak” and the “Oskarbi”. This is not a proper time-series, since they are only two punctual captures northwest of Spain, but they remain good enough to look for an environment effect (2 days; very small-time scale phenomena). There is not much biological information regarding this species. The Octopus vulgaris data series (Chapter 2) was gathered form the fishmonger at the Port of Mogán (south-west Gran Canaria. Canary Islands). The data series (daily catch and effort) started in 1989 and finished in 2007 (18 years; small-time scale phenomena). These series are updated monthly. The common octopus is an important target species in the small-scales trap fishery that take place in the Canary Archipelago. The Thunnus thynnus (Chapter 3) is a species that has been important since former times in the Mediterranean Sea and in the North Atlantic (Rodríguez-Roda, 1964a, 1983; López-Capont, 1997). This is why a lot of data bases can be found for the last centuries. However, it has to be taken into account that not all of them measure the same parameter (number of tuna, tons per catch, number of barrels ...). For the present study, information of 105 almadraba’s (this is the name of the trap used) in number of tuna was gathered from different sources (see Table 1 in Annexes). The Material and Methods 63 number of years with data was variable between almadraba’s. But, overall, data from 1525 to 1995 were obtained. However, the studies focused on the possible effect of climate variability on Bluefin tuna (T. thynnus); for this reason, only the time-series with more than 100 years data (11 almadraba’s; large-time scale phenomena) were finally taken into account (Table 2 in the Annexes). Finally, series of Northern anchovy (Engraulis mordax) and of Pacific sardine (Sardinops caeruleus) reconstructed from 283 to 1970 (very large-time scale phenomena) off California by Soutar (1967) and Soutar and Isaacs (1969, 1974), and obtained from Baumgartner et al. (1992), were used in the study (Chapter 4). Both, anchovy and sardine are important economic and ecological species. It is known that they fluctuate inversely in time, although the causes under it remain unclear. 2. Climatological Data Temperature data series Mean Sea Surface Temperature (SST) Satellite images from June and July 2006 were obtained from the MODIS sensor of the AQUA-satellite (MODISA Level-3 Standard Mapped Image). Also Anomalous SST images were obtained from the night tracks from the AVHRR sensor in the NOAA satellite for the same period as the previous. Both images were used in the Ocythoe tuberculata study (Chapter 1) when looking for the environmental relationship between the SST and this species. Meanwhile, in chapter 2, the series of Reynolds et al. (2002) for the SST in the North Atlantic (December 1982 - January 2007) and the Kaplan et al. (1998) local SST (28.5ºN/16.5ºW) were used. Material and Methods 64 In chapters 3 the proxy reconstructed annual air temperature at sea surface level (SLT) by Mann et al. (2009) was gathered from 1525 to 1936 but considering only the 20 º-60 ºN/60 ºW-20 ºE region. Global indices series From all the North Atlantic Oscillation (NAO) data available in the literature and on-line sites, the series from the NOAA (1950-2000) were used in chapter 2. This data set was used because it comprise the period of time wanted and because it is a tested and consistent series. El Niño/Southern Oscillation (ENSO hereafter) from 1650 up to 1970 reconstructed by McGregor et al. (2010) and the Pacific Decadal Oscillation (PDO), from 1470 until 1970 (Shen et al., 2006), were the two global indices used in chapter 4 when looking after the climate relationship with captures off California. Solar parameters series There are many parameters that have been established as proper ones to estimate the influence of the Sun on the earth ecosystems (Galactic cosmic rays, 18 O isotope, etc.). In chapter 3 the eleven years Solar Irradiance cycles Background (SIB), from 1610 to 1936 by Lean (2000) was used. Another factor that can be estimated through reconstructions is the Beryllium-10, from 1000 to 1970 (Crowley, 2000), it was considered when conducting analyses in chapter 4. Other climatic patterns River runoff can directly affect marine ecosystems, through the nutrients availability, the sediment discharge or due to changes in the coastal lines. But also it can be an indicative of a bigger climate structure. For instance, the volume of flow can be considered as an indirect proxy for the Material and Methods 65 global Intertropical Convergence Zone (ITCZ. See Annexes) movements (Chiang et al., 2000), due to the role that it has on the displacements of the Monsoons, responsible in turn of the rainy season. In chapter 4, the Sacramento River flow was considered because it is one of the most important rivers discharging in the California coast. It reflects the colder and rainier periods with an increase in the discharges. Due to this it can act as an atmospheric circulation pattern index over the North Pacific Ocean (it is an estimation of the ITCZ). This series started in 901 and ended in 1997 (Meko, 2001). Software Within this Thesis, for all the statistical analyses, the R software (a language for statistical computing and graphics) was used (http://www.rproject.org/). It is available as Free Software (Free Software Foundation's GNU General Public License). It provides a wide variety of statistical (e.g.: linear and nonlinear modelling, classical statistical tests, time-series analysis, classification, clustering...), and graphical techniques. One of R's strengths is the ease with which well-designed publication-quality plots can be produced, including mathematical symbols and formulae where needed. R has common features with functional and object oriented programming languages. Herein, functions are treated like objects that can be manipulated or used recursively (Cowpertwait and Metcalfe, 2009). Matlab software (http://www.mathworks.com/products/matlab/) was used when plotting maps (chapters 2 and 3). It is a high-level technical computing language and interactive environment for algorithm development, data visualization, data analysis, and numeric computation. It includes the graphics features (algorithms) that are required to visualize scientific data. Material and Methods 66 In chapter 4, the Ocean Data View (ODV) software was used when plotting the map (Schlitzer, 2009. http://odv.awi.de). C CC C C CC Ch hh h h hh ha aa a a aa ap pp p p pp pt tt t t tt te ee e e ee er rr r r rr r 1 11 1 1 11 1 N NN N N NN N e ee e e ee e w ww w w ww w r rr r r rr r e ee e e ee e c cc c c cc c o oo o o oo o r rr r r rr r d dd d d dd d o oo o o oo o f ff f f ff f O OO O O OO O c cc c c cc c y yy y y yy y t tt t t tt t h hh h h hh h o oo o o oo o e ee e e ee e t tt t t tt t u uu u u uu u b bb b b bb b e ee e e ee e r rr r r rr r c cc c c cc c u uu u u uu u l ll l l ll l a aa a a aa a t tt t t tt t a aa a a aa a ( (( ( ( (( (C CC C C CC Ce ee e e ee ep pp p p pp ph hh h h hh ha aa a a aa al ll l l ll lo oo o o oo op pp p p pp po oo o o oo od dd d d dd da aa a a aa a: :: : : :: : O OO O O OO Oc cc c c cc cy yy y y yy yt tt t t tt th hh h h hh ho oo o o oo oi ii i i ii id dd d d dd da aa a a aa ae ee e e ee e) )) ) ) )) ) i ii i i ii in nn n n nn n t tt t t tt th hh h h hh he ee e e ee e N NN N N NN No oo o o oo or rr r r rr rt tt t t tt th hh h h hh h- -- - - -- -e ee e e ee ea aa a a aa as ss s s ss st tt t t tt t A AA A A AA At tt t t tt tl ll l l ll la aa a a aa an nn n n nn nt tt t t tt ti ii i i ii ic cc c c cc c r rr r r rr re ee e e ee el ll l l ll la aa a a aa at tt t t tt te ee e e ee ed dd d d dd d t tt t t tt to oo o o oo o s ss s s ss se ee e e ee ea aa a a aa a w ww w w ww wa aa a a aa ar rr r r rr rm mm m m mm mi ii i i ii in nn n n nn ng gg g g gg g A. M. Caballero A. M. CaballeroA. M. Caballero A. M. Caballero - -- - Alfons AlfonsAlfons Alfons o, U. Ganzedo, G. Díez o, U. Ganzedo, G. Díezo, U. Ganzedo, G. Díez o, U. Ganzedo, G. Díez - -- - Díez, J. J. Castro Díez, J. J. CastroDíez, J. J. Castro Díez, J. J. Castro - -- - Hernández HernándezHernández Hernández Journal of Marine Biological Association of the United Kingdom Journal of Marine Biological Association of the United KingdomJournal of Marine Biological Association of the United Kingdom Journal of Marine Biological Association of the United Kingdom Biodiversity Records 2: Published on Biodiversity Records 2: Published onBiodiversity Records 2: Published on Biodiversity Records 2: Published on- -- -line (2008) line (2008)line (2008) line (2008) Chapter 1. Ocythoe tuberculata off Cantabria 69 Abstract The capture of two females of Ocythoe tuberculata during the summer of 2006, in the North-east Atlantic is reported. This pelagic cephalopod species are rare beyond subtropical waters and were caught at the sea surface by two live bait boats. The appearance of this species in the area is related with an anomalous sea warming. Introduction During the summer of 2006, two live specimens of football octopus (Ocythoe tuberculata, Rafinesque, 1814) were captured at the sea surface by live bait boats off the north-west Iberian Peninsula, at the northernmost limit of the species distribution in the North Atlantic according to Nesis (1985). The first individual was caught on the 27 June at 41º/42ºN 13º/14ºW (west of Portugal), and the second on the 18 July at 44º/45ºN 14.5º/15.5ºW (northwest of Spain). This species is known to be cosmopolitan in tropical and temperate seas (Sweeney et al., 1992; Vechione, 2002), especially in the northern hemisphere (Roper and Sweeney, 1975). In the North-east Atlantic, it has been reported off the Azores and Canary archipelagos (Cardoso, 1991) and in the western Mediterranean (Naef, 1923; Petrus and Pablo, 1993; EzzeddineNajai and El Abed, 2001). However, in the southern hemisphere, this species seems to be less common. In this way, a single catch has been reported off South Africa, associated with a storm in this area (Roper and Sweeney, 1975), and some others specimens were obtained in New Zealand and Australian waters (O’Shea, 1997; Lansdell and Young, 2007). The majority of the specimens, however, have been found in the stomach of itrs predators (swordfish, yellowfin tuna and dolphins), probably associated to their seasonal migrations. Chapter 1. Ocythoe tuberculata off Cantabria 70 Little is known about the biology and behaviour of Ocythoe tuberculata, except that it has pelagic habits and is found near the surface waters at night (Vechione, 2002). There is a strong sexual dimorphism in this species, where males are usually smaller than 3 cm in mantle length, and females are significantly larger with mantle lengths that could reach 35 cm (Roper and Sweeney, 1975; Cardoso and Paredes, 1998). Curiously, females of this species are the only known cephalopods with a swimbladder, a feature that makes them able to control their buoyancy (Packard and Wurtz, 1994), and are the only known cephalopods that give birth to live young that hatch internally (Naef, 1923). Results Morphological features The measurement of both specimens was taken according to Pickford and McConnaughey (1949), Thomas (1977), Roper and Voss (1983) and Clarke (1986). In both females, the pairs of tentacles II and III were shorter that the pairs I and IV. The mantle was muscular and strong. The ventral side was wrinkled, totally covered with hard pyramidal protuberances, but the dorsal side was smooth (Figure 1a and b, respectively). The head was more spherical than in other octopus species, and it showed two ventral pores. In Table 1 the weights and body lengths of both octopus individuals are shown. Chapter 1. Ocythoe tuberculata off Cantabria 71 Figure 1: Female of Ocythoe tuberculata captured in the Northeast Atlantic in June 2006: (A) ventral view, (B) dorsal view. Table 1: Geographical and morphological data of the two females Ocythoe tuberculata caught in the North-east Atlantic. Female 1 Female 2 Date of capture 27 June 2006 18 July 2006 Location 41-42ºN/13-14ºW 44-45ºN/14.5-15.5ºW Weight (g) 708.18 1929.0 Total Length (cm) 58.8 74.5 Mantle length (cm) 18.8 24.5 Head length (cm) 4.0 6.5 Foot length (cm) 39.0 52.4 Mantle perimeter (cm) 28.2 45.8 Mantle wide (cm) 15.4 22.9 Chapter 2. Octopus vulgaris off Canary Islands 79 identified for this specie; one from January to July with the peak in April, and the second one from October to November, with slight local variations (Guerra, 1992; Faure et al., 2000; Hernández-García et al., 2002; Katsanevakis and Verriopoulos, 2006a). In the northwest of Spain, the octopus reproductive cycle seems to be linked to the upwelling seasonality (Otero et al., 2008), while off Mauritania the relationship between recruitment and upwelling variability is seasonally dependent but not always related to the upwelling state; this is probably a reason for changes in the depth of the spawning grounds in spring and autumn (Faure et al., 2000). In accordance with this, Hernández-García et al. (2002) also pointed out that the intensity of the two annual maximum catches of octopus off the Canaries, related to the reproductive concentrations of individuals, are late winter-early spring sea water temperature dependent. In any case, this indicates the presence of, at least, two annual cohorts; generating well separated spawningcatching peaks. Of course, the relative importance of each seasonal peak is dependent on environmental conditions; although Katsanevakis and Verriopoulos (2006a) point out that the second settlement is much more environmentally dependent than the first one. Studies that evaluate the effects of environmental variability on the cephalopods, and particularly octopus, are scarce and sometimes contradictory. However, the influence of temperature on octopus abundance is always highlighted due to its importance in the first stages of its development (Mangold, 1983; Villanueva, 1995). In line with this, Sobrino et al. (2002) found that the maximum octopus abundance coincided with the minimum Sea Surface Temperature (SST) registered at the studied domain (Gulf of Cádiz). In contrast, Balguerías et al. (2002) and Moreno et al. (2002) reported for the Saharan Bank and the Portuguese coast, respectively, that maximal captures coincided with the highest SST in those domains. Chapter 2. Octopus vulgaris off Canary Islands 80 Figure 1: Canary Archipelago map and catch location (SW Gran Canaria). There is an increasing interest to understand how climate variability might affect different marine populations, mostly with the objective of predicting its possible evolution and to manage according to it. As it has been quoted previously, the common octopus varies its behaviour among localities due to regional conditions. In this sense, it is important to understand how climate is affecting octopus in the Canaries since it is one of the fisheries target species. With this aim, we hypothesis how Sea Surface Temperature (SST) and the North Atlantic Oscillation (NAO) may be controlling the O. vulgaris in the Canary domain, through a seasonal scale approach; to Chapter 2. Octopus vulgaris off Canary Islands 81 elucidate if climate is the possible main cause of the observed seasonal fluctuations in this population. Material and Methods Data set 1. Octopus fishery data The fishing data from 1989 to 2007 were obtained from the daily catch recorded by a single fishmonger who marketed the total catch obtained in the trap fishery landed in the Southwest of Gran Canaria (Port of Mogán. Figure 1), which is representative for all the islands, since it is one of the ports with the major fishery activity within the Archipelago, in fishing potential and in the number of catches landed (it represent over 25% of the total yearly captures of benthic and demersal fish landed in Gran Canaria, decreasing to the 10% if mackerel and tuna captures are included; Gobierno de Canarias, unpublished data). In relation with this, for instance, O. vulgaris is the predominant species, after seabreams (Hernández-García et al., 1998), in the fisheries between January and June, because from May to February they focus on tuna (González et al., 1991). Variations in catches through seasons are due to the life cycle of the O. vulgaris, which present two peaks as described previously, but also to the fishing objectives at each time of the year (Hernández-García et al., 1998). However, fishmongers do not exclude species at any time, so the catches are a proxy index of abundance. When analysing the effort and catches independently (results not shown), it can be seen that the fishing effort has not changed through the years in a significant way. In contrast, the octopus catches reflect variability, different from the life cycle seasonality one. Due to this, it was considered that other factors, as climatic ones, might be playing a key role in the evolution of this stock. Chapter 2. Octopus vulgaris off Canary Islands 82 The CPUE was estimated monthly from the total weight in kilograms of octopus caught per month divided by the monthly effort deployed. Afterward, seasonal means were calculated for all the years and trend removed for each new series to avoid the possible overfishing effect. We used, as an effort unit, the average number of days devoted to trap fishery per boat (sensus Hernández-García et al., 1998). 2. North Atlantic Oscillation (NAO) data The NAO is a north-south dipole of anomalies, with one centre (low pressure system) located over Greenland and the other centre of opposite sign spanning the central latitudes of the North Atlantic between the Azores (Ponta Delgada) and Portugal (Lisbon) (high pressure system). This is the atmospheric predominant situation over the Atlantic, combining parts of the East and West Atlantic patterns as defined by Wallace and Gutzler (1981) for winter. It is noteworthy that the Canary Islands are in the southern limit of influence of this climatic pattern, so the impact of it on the island environment is almost undetectable (Ganzedo-López, 2005), but in certain periods it influence is strong enough (this work). The NAO index data came from the NOAA database, while monthly mean standardized 500 mb height anomalies were obtained from the CDAS, from 1950 to 2000. 3. Sea Surface Temperature (SST) data Two kinds of SST data have been used for this study: (i) the Reynolds et al. (2002) SST from December 1982 to January 2007, to plot the maps of correlations and significations; (ii) the Kaplan et al. (1998) SST for local statistical analysis (South of Gran Canaria, 28.5ºN/16.5ºW), from January 1989 to December 2007. Chapter 2. Octopus vulgaris off Canary Islands 83 Statistical analysis 1. Exploratory CPUE data analysis A boxplot analysis and a cross-autocorrelation analysis were carried out using time data to explore the existence of seasonal components and to estimate autocorrelation (Figure 3) in catches series. Statistical analysis of the series assumes stationarity. This implies that the series does not contain trends or cycles. Monthly CPUE were averaged over seasons in order to obtain a time-series of CPUE per season of the year. 2. Seasonal relations between NAO and SST To identify and to quantify the significance of the effect of the NAO index with respect to SST, Pearson's correlation maps with significance isolines (95% (0.05) and 99% (0.01)) were plotted, for points within the quadrant 20º-50º N and 45ºW to 20ºE, at a resolution of 1ºx1º. To that effect, (i) The trends have already been removed from the data set (NAO and SST data). (ii) The Pearson correlation was used between SST data (Reynolds et al., 2002) and NAO index, between 1982 and 2007. Chapter 2. Octopus vulgaris off Canary Islands 84 40˚W 40˚W 30˚W 30˚W 20˚W 20˚W 10˚W 10˚W 0˚ 0˚ 10˚E 10˚E 20˚E 20˚E 20˚N 20˚N 30˚N 30˚N 40˚N 40˚N 50˚N 50˚N 0.05 0.05 0.05 0.05 0.05 0.05 0.01 −0.8 −0.6 −0.4 −0.2 0.0 0.2 0.4 0.6 0.8 40˚W 40˚W 30˚W 30˚W 20˚W 20˚W 10˚W 10˚W 0˚ 0˚ 10˚E 10˚E 20˚E 20˚E 20˚N 20˚N 30˚N 30˚N 40˚N 40˚N 50˚N 50˚N 40˚W 40˚W 30˚W 30˚W 20˚W 20˚W 10˚W 10˚W 0˚ 0˚ 10˚E 10˚E 20˚E 20˚E 20˚N 20˚N 30˚N 30˚N 40˚N 40˚N 50˚N 50˚N 0.05 0.05 0.05 0.05 0.05 0.05 0.01 0.01 −0.8 −0.6 −0.4 −0.2 0.0 0.2 0.4 0.6 0.8 40˚W 40˚W 30˚W 30˚W 20˚W 20˚W 10˚W 10˚W 0˚ 0˚ 10˚E 10˚E 20˚E 20˚E 20˚N 20˚N 30˚N 30˚N 40˚N 40˚N 50˚N 50˚N 40˚W 40˚W 30˚W 30˚W 20˚W 20˚W 10˚W 10˚W 0˚ 0˚ 10˚E 10˚E 20˚E 20˚E 20˚N 20˚N 30˚N 30˚N 40˚N 40˚N 50˚N 50˚N 0.05 0.05 −0.8 −0.6 −0.4 −0.2 0.0 0.2 0.4 0.6 0.8 40˚W 40˚W 30˚W 30˚W 20˚W 20˚W 10˚W 10˚W 0˚ 0˚ 10˚E 10˚E 20˚E 20˚E 20˚N 20˚N 30˚N 30˚N 40˚N 40˚N 50˚N 50˚N 40˚W 40˚W 30˚W 30˚W 20˚W 20˚W 10˚W 10˚W 0˚ 0˚ 10˚E 10˚E 20˚E 20˚E 20˚N 20˚N 30˚N 30˚N 40˚N 40˚N 50˚N 50˚N 0.05 0.05 0.05 0.05 0.05 0.01 0.01 −0.8 −0.6 −0.4 −0.2 0.0 0.2 0.4 0.6 0.8 40˚W 40˚W 30˚W 30˚W 20˚W 20˚W 10˚W 10˚W 0˚ 0˚ 10˚E 10˚E 20˚E 20˚E 20˚N 20˚N 30˚N 30˚N 40˚N 40˚N 50˚N 50˚N Summer Autumn Winter Spring Figure 2: Index of correlation between NAO index and SST at each point by seasons. The black line is the p-value at 0.01 (99%) and 0.05 (95%). 3. Statistical analyses between CPUE and environmental data The relationship between the NAO teleconnection pattern, SST and the statistics of CPUE was evaluated. A traditional analysis based on Pearson's correlation coefficient between CPUE, SST (Kaplan et al., 1998) and NAO index was performed. After removing data trend, correlation coefficient has been checked against the hypothesis that it was zero with a 95% confidence level. The reduction in the degrees of freedom due to the autocorrelation of the series has been considered in the test. In order to do so, a Monte Carlo test has been used. Several realizations (500,000) of autoregressive processes (AR (1)) with the corresponding autocorrelation for each of the tested series have been created, and the correlation coefficients of segments with the same length as the tested series have been used to create an experimental histogram, which represents the distribution of correlation coefficients from the AR (1) noise processes. Values of the correlation Chapter 2. Octopus vulgaris off Canary Islands 85 coefficient under (above) the 2.5% (97.5%) percentiles in the experimental distribution of correlation coefficients obtained from the Monte Carlo analyses were considered significant. To facilitate the visualization, a matrix of scatterplots was produced. When looking for the relationship between captures (y i ) and climate descriptors (x i ), a simple linear regression model (lm) was used for describing paired data sets that are related in a linear manner (where x is the independent variable and y the dependent one). In the simple linear regression model, for describing the relationship between x i and y i , an error term is added to the linear relationship as y i = β 0 + β 1 x i +ε i . The value ε i is the error term (residual term), and the coefficients β 0 and β 1 are the regression coefficients. The data vector x is called the predictor variable and y the response variable. The coefficient of determination (R 2 ) is defined as the decomposition of the total sum of squares into the residual sum of squares and the regression sum of squares: R 2 = 1-(∑(y i -ŷ i2 )/∑(y i -Ῡ i2 )) = ∑(ŷ i -Ῡ i2 )/∑(y i -Ῡ i2 ) Here, ŷ i is the predicted term of y i , which is the original value of the captures, and Ῡ i is the mean. The R 2 is interpreted as the proportion of the total response variation explained by the regression. The 100% of the variation is explained by the regression line. The adjusted R 2 divides the sums of squares by their degrees of freedom (Wilks, 2006). The capture was considered as the response (output or y) variable in the analysis, to facilitate the analyses and visual comparison of general trends. We used the following explanatory variables (x) in the lm model: (a) SST (Kaplan et al., 1998) and (b) NAO index. With this, to evaluate the possible colinearity among climatic terms and to elucidate if the explained variance is increased when they are considered together or independently, three types of lm’s were carried out: (i) combining the influence of the NAO and SST; (ii) Chapter 2. Octopus vulgaris off Canary Islands 86 analysing the effect of the SST solely and, finally (iii) the influence of the NAO on captures. We really wanted to be sure about the constancy (in the time) of variance and normality of the residuals. To that effect: (i) The Shapiro-Wilk test (Shapiro-Wilk, 1965) was used for testing that the residuals are normally distributed (p-value > α = 0.05, confirmed the null hypothesis: normality). (ii) The Durbin-Watson function (Durbin-Watson, 1950; 1951) was used for testing whether there is autocorrelation in the residual from lm. Figure 3: Boxplot (left) and ACF of monthly CPUE (right). Results The Boxplot (Figure 3, left) shows that 20.87%, 35.82%, 17.78% and 25.52% of the total CPUE are caught in winter, spring, summer and autumn, respectively. The auto correlation function (ACF) shows autocorrelation in the series (Figure 3, right). Consequently, to avoid the inter-annual seasonality, seasonal averages of the CPUE were calculated. Thereafter, seasonal series were analysed in an independent way. Figure 4 Winter Spring Summer Autumn 0 5 10 15 20 25 Boxplot Catches(kg) 0 5 10 15 20 0.0 0.2 0.4 0.6 0.8 1.0 Lag(season) ACF Chapter 2. Octopus vulgaris off Canary Islands 87 shows the ACF of these seasonal series. This was done with the only aim of verifying our models. When the NAO index was correlated against the SST data grid (Figure 2), the relationship obtained was high and significant in spring and autumn around the Canary Islands. Significant correlations were found for SST in the local point (28.5ºN/16.5ºW) in spring (r = -0.47, p<0.05) and autumn (r = -0.48, p<0.05). While no significant correlations were obtained for winter (r = -0.16, p>0.05) and summer (r = 0.36, p>0.05). Table 1 (up) and Figure 5 show the correlations obtained between the seasonal CPUEs and the SST. It is worth emphasising that all pairs were negatively correlated. That is, when the CPUE increases, the SST decreases. Table 1 (down) shows the correlations obtained between the seasonal CPUEs and the NAO index, and both variables evolve simultaneously in spring and autumn. While in winter and summer they have an inverse behaviour almost always. Chapter 2. Octopus vulgaris off Canary Islands 88 0 2 4 6 8 10 12 −0.5 0.0 0.5 1.0 Lag(year) ACF Spring 0 2 4 6 8 10 12 −0.5 0.0 0.5 1.0 Lag(year) ACF Winter 0 2 4 6 8 10 12 −0.5 0.0 0.5 1.0 Lag(year) ACF Summer 0 2 4 6 8 10 12 −0.5 0.0 0.5 1.0 Lag(year) ACF Autumn Figure 4: ACF of seasonal CPUE. Chapter 2. Octopus vulgaris off Canary Islands 95 controlling climate factor, with a significant explained variance of 28.64% (the SST explains the 11.12%, but it remained no significance). During autumn the effect of the SST and the NAO is not such difference: 31.13% of the explained variance through the NAO and a 34.21% explained by the SST. It has also been demonstrated why more local climatic indices should be built to reach a better understanding of the interaction between climate variations and exploited marine organisms, because their sensitivity, particularly cephalopods, to environmental fluctuations is an important factor in stock assessment and management. In addition, octopus catch fluctuations may be indicators of environmental changes. Acknowledgements Authors would like to sincerely thank Dr. Jon Sáenz (University of the Basque Country) for inestimable help with the data analyses. In addition, all our gratitude goes also to the Mogán fishery association and a special mention to Leo Hernández, for her good will and for giving us the octopus catch and effort data. We also want to thank Lorena Couce for her help. A. M. Caballero-Alfonso has a scholarship from the Spanish Ministry of Science and Innovation (MICINN), and this work has been also supported by the MICINN National Research Programme (AGL2006-10448/ACU). The EKLIMAXXI project (Basque Government, Department of Industry and Basque Meteorological Service-Euskalmet, Project ETORTEK07/01 – IE07190) is acknowledged for funding the position of U. Ganzedo. C CC C C CC Ch hh h h hh ha aa a a aa ap pp p p pp pt tt t t tt te ee e e ee er rr r r rr r 3 33 3 3 33 3 A. M. Caballero A. M. CaballeroA. M. Caballero A. M. Caballero - -- - Alfonso, U. Ganzedo, E. Zorita, G. Ibarra Alfonso, U. Ganzedo, E. Zorita, G. IbarraAlfonso, U. Ganzedo, E. Zorita, G. Ibarra Alfonso, U. Ganzedo, E. Zorita, G. Ibarra - -- - Berastegi, J. Sáenz, A. Ezcurra, A. Trujillo Berastegi, J. Sáenz, A. Ezcurra, A. TrujilloBerastegi, J. Sáenz, A. Ezcurra, A. Trujillo Berastegi, J. Sáenz, A. Ezcurra, A. Trujillo- -- - Santana, A. Santana, A. Santana, A. Santana, A. Santana del Pino, J. J. Castro Santana del Pino, J. J. CastroSantana del Pino, J. J. Castro Santana del Pino, J. J. Castro- -- -Hernández HernándezHernández Hernández Submitted (2011) Submitted (2011)Submitted (2011) Submitted (2011) C CC C C CC C l ll l l ll l i ii i i ii i m mm m m mm m a aa a a aa a t tt t t tt t e ee e e ee e a aa a a aa a n nn n n nn n d dd d d dd d h hh h h hh h i ii i i ii i s ss s s ss s t tt t t tt t o oo o o oo o r rr r r rr r i ii i i ii i c cc c c cc c B BB B B BB B l ll l l ll l u uu u u uu u e ee e e ee e f ff f f ff f i ii i i ii i n nn n n nn n t tt t t tt t u uu u u uu u n nn n n nn n a aa a a aa a f ff f f ff fl ll l l ll lu uu u u uu uc cc c c cc ct tt t t tt tu uu u u uu ua aa a a aa at tt t t tt ti ii i i ii io oo o o oo on nn n n nn ns ss s s ss s i ii i i ii in nn n n nn n t tt t t tt th hh h h hh he ee e e ee e G GG G G GG Gi ii i i ii ib bb b b bb br rr r r rr ra aa a a aa al ll l l ll lt tt t t tt ta aa a a aa ar rr r r rr r S SS S S SS St tt t t tt tr rr r r rr ra aa a a aa ai ii i i ii it tt t t tt t a aa a a aa an nn n n nn nd dd d d dd d W WW W W WW We ee e e ee es ss s s ss st tt t t tt te ee e e ee er rr r r rr rn nn n n nn n M MM M M MM Me ee e e ee ed dd d d dd di ii i i ii it tt t t tt te ee e e ee er rr r r rr rr rr r r rr ra aa a a aa an nn n n nn ne ee e e ee ea aa a a aa an nn n n nn n Chapter 3. Thunnus thynnus in the Gibraltar Strait and Western Mediterranean 99 Abstract The analysis of historical capture records of bluefin tuna (Thunnus thynnus) at Western Mediterranean almadrabas reveal a decreasing trend between 1525 and 1936, with cyclical fluctuations associated to climatic factors. Analyses based on linear models of capture series emphasize the potential role of sea-level air temperature and the solar activity cycles on the bluefin population dynamic. The overall density-dependent term explains between 13.71 and 47.81% of the capture variability, while the environmental one explains between 0.37 and 17.28% (the shared terms oscillated between 5.72 and 21.87%, and the unexplained parts of the variance fluctuated between 31.61 and 65.06%). These ranges depend on the time interval and the geographic domain analysed. The cold phases of climate, normally coupled with solar activity minima, are associated to dramatic decreases of temperature and captures, after which tuna population was not able to completely recover. KeywordsAlmadraba, Atlantic Bluefin tuna, sea-level air temperature, solar activity minimum, solar irradiance, Thunnus thynnus. Introduction The Atlantic Bluefin Tuna (Thunnus thynnus, Linné 1758) (BFT hereafter) has been historically important, due to its commercial interest since Phoenicians times (ca. 900 years B.C) (López-Capont, 1997; Pairman-Brown, 2001). It is one of the oldest fisheries organised on an industrial scale (Lemos and Gomes, 2004). Fishermen early understood that this specie appears regularly near the coast, keeping the shore on their right-side and performing a cyclonic movement around the Western Mediterranean Sea. Therefore, they set up special trapnets (almadrabas) which guided hundreds or thousands of fishes into the gear (Sará, 1980), along the Western Mediterranean coasts, until present times without significant structural changes through time Chapter 3. Thunnus thynnus in the Gibraltar Strait and Western Mediterranean 100 (Ravier and Fromentin, 2001). Consequently, it is reasonable to assume that all these almadrabas exploit the same population and, therefore their fishing yield should show the same long-term fluctuations (sensus Ravier and Fromentin, 2001; Lemos and Gomes, 2004). Beside all the information that can be gathered from BFT historical data, there are uncertainties still unresolved (Fromentin, 2003). It is important to understand the fluctuations that have been taking place in the historical capture series and which variables are behind those variations, particularly those that could be environmentally forced (Lemos and Gomes, 2004; Cushing, 1982; Fromentin and Powers, 2005). This will allow us to predict the future evolution of the BFT stock and to manage them according to it. In this sense, Ravier and Fromentin (2001; 2004) found a 100-120 year periodic fluctuation in the BFT captures series from the almadrabas located around the Mediterranean and the East Atlantic. They related those fluctuations with the inverse of the temperature registered in those domains. Going further on these issues, the present study aims to shed light on the BFT fluctuations and evolution, trying to elucidate if the cause of the observed oscillations is the natural global climate variability, rather than local features and conditions. Material and Methods Data set 1. Capture Data To attain the objectives posed in the present study, historical BFT capture series (with more than 100 years data) were extracted from different sources ranging from 1525 up to 1936 (although data were gathered until 1995). This last year was selected as the most recent limit of the time series; in order to avoid distortions on the analysis due to the impact of fishing technology developed after the Second World War (i.e.: fishing capacity of Chapter 3. Thunnus thynnus in the Gibraltar Strait and Western Mediterranean 101 bait boats, purse seiners and longliners fleets) on BFT, and to elude the possible effects of overfishing. This will lead in considering only captures done with almadrabas. In addition, the Sicily series were considered after 1700 (although the earliest began in 1602) because several geological events (i.e.: earthquakes), depopulation and social instability occurred during the XVII century (Russo-Adams, 2008) caused enormous variability in captures in the Sicilian traps. Information concerning eleven almadrabas distributed along the Gibraltar Strait and the Western Mediterranean was gathered (Table 1; Figure 1). All the data considered were recorded as number of tuna caught during defined periods. These types of data might reflect properly the population dynamics because traps, like almadrabas, are not directly affected by the effort employed for captures (Ravier and Fromentin, 2001; Fromentin, 2009). Even more, the present work avoids combining different data sources, with the objective of minimizing the associated error. Time series register were unequal in time but they apparently behaved in the same way in periods when they were overlapped. For this reason (considering they were complementary) the mean yields of all captures were calculated in order to represent the total (AtlanticMediterranean) and the specific captures per region: (i) Atlantic-Gibraltar Strait: (Barril and Medo das Casas in the south of Portugal, and Zahara and Conil in the south of Spain) and (ii) Western Mediterranean (Saline, Isola Piana, Porto Scuso and Porto Paglia in Sardinia, and Bonagia, Formica and San Giuliano in Sicily). The yield was calculated as the yearly mean of the capture of all almadrabas for each year referenced to a scale of one hundred, that was in turn estimated as: Mean((a i /A i )*100), where a i is the yearly capture and A i is the highest capture obtained along the serie (in this way the highest capture in each almadraba is equal to 100). It was assumed that captures were all done with the same fishing technique: almadrabas build and operated in the same way. Chapter 3. Thunnus thynnus in the Gibraltar Strait and Western Mediterranean 102 Table 1: Almadraba BFT original data series references relations. Country Almadraba Period Referente Portugal Barril 1867-1936 Ravier, 2003; Lemos and Gomes, 2004 Medo das Casas 1776-1808 Ravier, 2003 Medo das Casas 1861-1936 Ravier, 2003 Spain Zahara 1525-1756 López-Capont, 1997; Ravier, 2003 Zahara 1910-1936 López-Capont, 1997; Ravier, 2003 Conil 1525-1677 López-Capont, 1997 Conil 1684-1756 López-Capont, 1997 Italy (Sardinia) Saline 1823-1843 Ravier, 2003 Saline 1864-1936 Ravier, 2003 Isola Piana 1820-1936 Ravier, 2003 Porto Scuso 1820-1936 Ravier, 2003 Porto Paglia 1830-1936 Ravier, 2003 Italy (Sicily) Bonagia 1700-1806 Ravier, 2003 Bonagia 1870-1936 Ravier, 2003 Formica 1700-1816 Ravier, 2003 Formica 1876-1936 Ravier, 2003 San Giuliano 1700-1804 Ravier, 2003 San Giuliano 1885-1914 Ravier, 2003 Chapter 3. Thunnus thynnus in the Gibraltar Strait and Western Mediterranean 103 Figure 1: Almadrabas locations map. 12 3 4 5 6 7 8 9 10 11 Spain North Africa Italy 1.−Barril 2.−M.Casas 3.−Zahara 4.−Conil 5.−Saline 6.−I.Piana 7.−P.Scuso 8.−P.Paglia 9.−Bonagia 10.−Formica 11.−S.Giuliano 12oW 6oW 0o 6oE 12oE 18oE 30oN 33oN 36oN 39oN 42oN 45oN Longitude Latitude 31.80’ 39.90’ Chapter 3. Thunnus thynnus in the Gibraltar Strait and Western Mediterranean 104 This data grouping entails: (i) a diminish in the number of series to work with, without losing or changing information; (ii) an accurate idea of the capturability of BFT by almadrabas in the Gibraltar Strait-Western Mediterranean domain; and (iii) robustness of the series for statistical analyses (Fromentin, 2009). 2. Climate Data Climate data series were obtained from on-line data bases. Proxy data were used because they reflect the long-term trends of the climate system and, to a certain degree, also the high frequency variations. This is not expected from integrations using GCMs unless they were performed using assimilation of observations (modelled data). The proxy series considered were: (i) reconstructed annual air temperature at sea surface level from 1525 to 1936 (SLT) by Mann et al. (2009); and (ii) eleven years Solar Irradiance cycles Background (SIB), from 1610 to 1936 by Lean (2000). Statistical analysis Once all the data were gathered, an initial exploratory analysis was carried out to elucidate the statistical procedures that should be followed to achieve the aims of the study. The entire climate series were related to the BFT capture series through Linear Models (LM) based analyses (Verzani, 2005). The rationale under it is to look for a linear relationship between pairs of variables (not necessarily a straight-line regression). It was assumed that a change in the predictor variable (x i : climate variables) produced an increase or decrease in the response variable (y i : captures series). This can be summarized by y i = β 0 +β 1 x i +ε i , where β are the coefficients linked to the variations of x i and ε i is the error term. To assume randomness in the response variable, it was considered that the ε i are independent and identically distributed, also that Chapter 3. Thunnus thynnus in the Gibraltar Strait and Western Mediterranean 105 they have a normal distribution with mean = 0 and standard deviation = σ; this means that y i distribution can be assumed as normal (Verzani, 2005). Three types of LM’s were carried out. The significant results were verified through a bootstrap analysis (Mudelsee, 2003). As non-normality is a common problem in climatological time series, the stationary bootstrap may solve this problem if necessary through a re-sampling with replacement method, where persistence is preserved (i. e.: it takes out the red-noise and makes clear the correlation among series without autocorrelation, showing the annual variations due to climate). The results considered valid from this analysis were significant at 95%. This approach was corroborated through a Monte Carlo test (Mudelsee, 2003). Also, the Akaike Information Criterion (AIC) was considered, as a measure of the goodness of the fit of the models, to finally select the best model that was significant through the bootstrap. Models were carried out for different periods of the yields series: 1525-1936; 1610-1936; and 1700-1936. This was done with the aim of having the maximum resolution in each case. (i) Temporal model (autocorrelation): As highlighted by Legendre and Legendre (1998), temporal structures in time series may incorporate other effects, such as those due to biotic, environmental and/or historical events. To take into account this complex structure, linear, second, third and fourth-degree temporal (T, in years) polynomial terms were considered in the LM in an increasingly way. Finally a fourth-degree term was selected as the best one. (ii) Environmental model: both climate parameters (SLT and SIB) were considered together against capture series in order to see their influence on BFT fluctuations in the considered region. For the period 15251936 the SLT were considered solely against captures because SIB began to be recorded after 1610. Chapter 3. Thunnus thynnus in the Gibraltar Strait and Western Mediterranean 106 (iii) Global model: Temporal and climate factors were gathered from the previous two models and considered together. Once all the models were run, the adjusted R 2 of each model was considered to obtain the explained variance (%) in order to quantify the relative contribution of environmental and temporal factors, as well as the shared proportion (i. e.: part included in the model but not assigned to any of the previous terms) and the unexplained one. That is: the value of R 2 can be interpreted as the proportion of the total response variation explained by the regression if the LM is an appropriate method to analyse the data (Verzani, 2005). When R 2 is close to 1, the model fits well to the data; if it is 0, it does not. The adjusted R 2 considers the degrees of freedom and this is useful when multiple predictors are needed to get a better R 2 values. In the present study, combined predictors were used. Results The LM (Table 2) results (through the explained variance) highlights the influence of air temperature at sea level (SLT hereafter) as quoted in previous studies (Ravier and Fromentin, 2004), but also the possible effect that the solar parameters have on this fish population, since the models results improved when adding the solar irradiance parameter. It was observed that the temporal effects explain a higher fraction of variance than what the climatic ones do. Chapter 3. Thunnus thynnus in the Gibraltar Strait and Western Mediterranean 113 Figure 7: Gibraltar Strait yields and models for the period 1700-1936 (considering SLT and SIB). Chapter 3. Thunnus thynnus in the Gibraltar Strait and Western Mediterranean 114 Figure 8: Mediterranean yields and models for the period 1700-1936 (considering SLT and SIB). On the other hand, the unexplained part of the variance was always lower than that explained by the models (environment+temporal+shared) with the exception of the 1525-1936 period for the whole area. These results highlight the accuracy of our models. Discussion The historical capture records of bluefin tuna (Thunnus thynnus) at Western Mediterranean almadrabas reveal a decreasing trend between 1525 and 1936, with cyclical fluctuations associated to climatic factors. However, Chapter 3. Thunnus thynnus in the Gibraltar Strait and Western Mediterranean 115 this decreasing tendency in captures and their quasi-periodic oscillations are not new (Ravier and Fromentin, 2001, Ravier and Fromentin, 2004, Ganzedo et al., 2009). Similar long-term fluctuations in other fish populations have been also described previously (i. e.: Cushing, 1982; Alheit and Hagen, 1997). Nevertheless, our results emphasize the relevance of the sun activity on the long-term fluctuations of the BFT at the Western Mediterranean domain, particularly during its minimum activity phases when the tuna availability to the fishery decreases drastically. This fact reflects a decline in the stock abundance due to climatic factors different from the temperature, in agreement to what has been reported in previous studies (e.g.: Ravier and Fromentin, 2004). Different mesoscale climate variables and with different periodicity, affect the tuna population dynamic as external forcings (Ravier and Fromentin, 2004; Ganzedo et al., 2009; this study), working behind the observed gross fluctuations in captures. Specifically, solar activity are longterm controlling factors (low frequency variability), while the high frequency variability is reflected by the SLT. However, an important part of the variability remains unexplained probably due to additional local-scale climatic events and socio-political circumstances (wars, famines, bankrupts, etc.). In this way, the massive recruitment of fishermen that took place to prepare the “Spanish Armada” (1588), the capture of Gibraltar by the AngloDutch forces (1704), the tsunami produced by the earthquake of Lisbon (1755) or the depopulation suffered by Sicily along the XVII century could have negative effects on the almadrabas efficiency (Archives of the Casa Medina Sidonia Foundation ; Russo-Adams, 2008). Nevertheless, social, political and economical factors are assumed to be of local and short-time impact on the almadrabas exploitation. Also, density dependent processes, reflected in the temporal term as the autocorrelation, must be important to explain the small scale temporal variations in the 4-8 years period range (Solari, 2008). For these reasons, it can be considered that climatic factors are Chapter 3. Thunnus thynnus in the Gibraltar Strait and Western Mediterranean 116 controlling low-frequency component in the BFT captures in the Gibraltar Strait-Western Mediterranean domain. The fact that climatic controlling terms (density independent processes) vary between models must be related to the predominant time scale in each of the models. However, important information could be lost when environmental parameters are considered alone, because only those with large periodicities remain significant. Thus, the SLT (Mann et al., 2009) remain noteworthy as an influencing factor on the BFT. This result is in accordance with Ravier and Fromentin (2004) observations, although they suggested that long-term fluctuations were related to the inverse of the temperature in the almadrabas domain. Moreover, it is largely known that temperature is a key parameter in tuna metabolism, migrations, reproductive behaviour (in the Mediterranean, BFT generally spawn as the SST reaches 24-25.5 ºC), larval survival and food availability, with direct consequences on recruitment (Pepin, 1991; Hazel, 1993; Polovina, 1996; Korsmeyer and Dewar, 2001; Schaefer, 2001; Graham and Dickson, 2004). Also solar parameters, concretely the solar irradiance (SIB), are also temperaturecontrolling factors. The small differences found between regions highlight the relevance of local features and environmental conditions on the observed BFT capture fluctuations. However, it is obvious that there is something else that condition the success of the almadraba fishery and that our models are not able to describe. As a main conclusion, the solar activity influences somehow the BFT population dynamic and its availability to the fishery in this domain. When adding the SIB to the models, the climate and captures oscillations look more similar than when considering the SLT as solely responsible. In this way, the Spörer (1460-1530), Maunder (1645-1715) and Dalton (17901820) minima (Gleissberg, 1958); even the ones that occurred between 1895 and 1930 in the solar activity series, coincide with a decrease in the captures recorded in the almadraba series simultaneously (Figures 2 to 8). For Chapter 3. Thunnus thynnus in the Gibraltar Strait and Western Mediterranean 117 example, the extremely cold period occurred in West Europe (the so called “Little Ice Age”) from 1560 to 1600 (Fagan, 2000) can be detected in the oldest capture series from the Gibraltar Strait’s almadrabas (Figures 5-7). An increase in captures was always observed after all these cold intervals simultaneously with an increase of the sun activity (i. e.: warmer climate conditions) produced during the three solar activity maxima occurred after 1600. However, those warmer intervals were not enough to restore BFT captures to values as the recorded previously to the Maunder minimum in the Medina Sidonia Dukedom’s almadrabas. Several reports demonstrated that the flux of energy that the sun transfers to the atmosphere by charging particles cause local warming and produce changes in the circulation patterns, influencing the troposphere, temperature, air pressure, etc (Haigh, 1996; Shindell et al., 1999). Furthermore, Egorova et al. (2000) and Neff et al. (2001) pointed out that there is a strong positive connection between solar eruptions, temperature and rainfall, and a negative relation with the air pressure. However, our models do not include all these factors, and many others which effects are still unknown, that could affect the population dynamic of BFT and/or its fisheries. Their influence is integrated in the unexplained variance. Probably many of these not-considered climatic factors have direct influence on the water temperature, as occur with the sun activity, and this last on tuna recruitment. In this way, the periods of very low temperature (i.e.: Dalton Minimum) were related to low productive periods in the almadrabas fisheries off the Gibraltar Strait and the Western Mediterranean coasts. This finding could be a consequence of repeated failures in BFT recruitment because the seawater temperature in its spawning grounds did not reach the optimum level for reproduction and larval survival during several years (Ravier and Fromentin, 2001; Carlsson et al., 2004; Gordoa et al., 2009; Ganzedo et al., 2009). Chapter 3. Thunnus thynnus in the Gibraltar Strait and Western Mediterranean 118 Acknowledgements A. M. Caballero-Alfonso has a fellowship from the Spanish Ministry of Science and Innovation (MICINN). The EKLIMAXXI project (Basque Government, Department of Industry and Basque Meteorological ServiceEuskalmet, Project ETORTEK07/01 – IE09-264) is acknowledged for funding the position of U. Ganzedo. C CC C C CC Ch hh h h hh ha aa a a aa ap pp p p pp pt tt t t tt te ee e e ee er rr r r rr r 4 44 4 4 44 4 A. M. Caballero A. M. CaballeroA. M. Caballero A. M. Caballero - -- - Alfonso, U. Ganzedo, A. Santana del Pino, Alfonso, U. Ganzedo, A. Santana del Pino, Alfonso, U. Ganzedo, A. Santana del Pino, Alfonso, U. Ganzedo, A. Santana del Pino, J. J. Castro J. J. CastroJ. J. Castro J. J. Castro- -- -Hernández HernándezHernández Hernández Submitted (201 Submitted (201Submitted (201 Submitted (2011) 1)1) 1) E EE E E EE E v vv v v vv v a aa a a aa a l ll l l ll l u uu u u uu u a aa a a aa a t tt t t tt t i ii i i ii i o oo o o oo o n nn n n nn n o oo o o oo o f ff f f ff f c cc c c cc c l ll l l ll l i ii i i ii i m mm m m mm m a aa a a aa a t tt t t tt t e ee e e ee e s ss s s ss s y yy y y yy y n nn n n nn n e ee e e ee e r rr r r rr r g gg g g gg g i ii i i ii i e ee e e ee e s ss s s ss s a aa a a aa af ff f f ff ff ff f f ff fe ee e e ee ec cc c c cc ct tt t t tt ti ii i i ii in nn n n nn ng gg g g gg g P PP P P PP Pa aa a a aa ac cc c c cc ci ii i i ii if ff f f ff fi ii i i ii ic cc c c cc c s ss s s ss sa aa a a aa ar rr r r rr rd dd d d dd di ii i i ii in nn n n nn ne ee e e ee e a aa a a aa an nn n n nn nd dd d d dd d N NN N N NN No oo o o oo or rr r r rr rt tt t t tt th hh h h hh he ee e e ee er rr r r rr rn nn n n nn n a aa a a aa an nn n n nn nc cc c c cc ch hh h h hh ho oo o o oo ov vv v v vv vy yy y y yy y h hh h h hh hi ii i i ii is ss s s ss st tt t t tt to oo o o oo or rr r r rr ri ii i i ii ic cc c c cc ca aa a a aa al ll l l ll l f ff f f ff fl ll l l ll lu uu u u uu uc cc c c cc ct tt t t tt tu uu u u uu ua aa a a aa at tt t t tt ti ii i i ii io oo o o oo on nn n n nn ns ss s s ss s o oo o o oo of ff f f ff ff ff f f ff f C CC C C CC Ca aa a a aa al ll l l ll li ii i i ii if ff f f ff fo oo o o oo or rr r r rr rn nn n n nn ni ii i i ii ia aa a a aa a ( (( ( ( (( (S SS S S SS St tt t t tt ta aa a a aa a. .. . . .. . B BB B B BB Ba aa a a aa ar rr r r rr rb bb b b bb ba aa a a aa ar rr r r rr ra aa a a aa a a aa a a aa an nn n n nn nd dd d d dd d S SS S S SS So oo o o oo ol ll l l ll le ee e e ee ed dd d d dd da aa a a aa ad dd d d dd d b bb b b bb ba aa a a aa as ss s s ss si ii i i ii in nn n n nn ns ss s s ss s) )) ) ) )) ) Chapter 4. Engraulis mordax and Sardinops caeruleus off California 121 Abstract Historical effects of environment on Pacific sardine (Sardinops caeruleus) and Northern anchovy (Engraulis mordax) were evaluated from 283 to 1970 A.D. Biomass registers are based on scale-deposition in the sediment off California. Climatic-population long-term synergies were the goal of the present study, and to face it, several types of data analysis available for time series analyses were used (wavelet, stationary bootstrap and Multivariate moving-window regression). Significant relationships were found between the abundance of fish species from the years 940 to 1011 and from 1324 to 1382, coinciding with two regime shifts in their abundances, when the environmental conditions favour both species. Our results reveal that either cold conditions or too warms ones can break the significant relationship between sardines and anchovies. Also the rainy periods as an indicative of the Intertropical Convergence Zone (ITCZ) movements (13001350 for anchovy and 1308-1324 for sardine) and the solar irradiance (16011680) had influenced those species fluctuations, probably through their effect on water temperature and food availability. This highlights the role of environment on the biomass fluctuations rather than interspecies competition (R = 0.2), mostly through the solar irradiance (R = -0.26), and through the ITCZ displacement in less proportion (-0.1 < R < -0.06). KeywordsAnchovy, climate, Engraulis mordax, Intertropical Convergence Zone (ITCZ), River runoff, sardine, Sardinops caeruleus, Solar Irradiance. Introduction Small pelagic species, such Sardinops caeruleus (Pacific sardine) or Engraulis mordax (Northern anchovy) off California, are important species from an economical point of view. But they are also, from an ecological Chapter 4. Engraulis mordax and Sardinops caeruleus off California 122 perspective. For these reasons, they have been under evaluation for decades (e.g.: Kawasaki and Omori, 1988; Silvert and Crawford, 1988; Lluch-Belda et al., 1989, 1992; Schwartzlose et al., 1999; Chavez et al., 2003). Several studies emphasized the synchronic (or not) out-ofphase in both species abundance fluctuations (Soutar and Isaacs, 1974; Daan, 1980; Silvert and Crawford, 1988; Baumgartner et al., 1992; Klyashtorin, 1998; Freón et al., 2003; Barange et al., 2009; Alheit and Bakun, 2010), although this is a question that remains unclear beside all the studies carried out. Both, anchovy and sardine, are planktivorous and utilize common food sources. This is why some studies proposed that trophic interactions might explain the inverse trends in abundance, due to a competitive replacement (Daan, 1980; Silvert and Crawford, 1988; van der Lingen et al., 2006). This might indicate a dependency on the habitat conditions (Barange et al., 2009). Beside the food competition, others mechanisms have been proposed to underlie this replacement: large atmospheric trends, as the Aleutian Low (Chavez et al., 2003) or the temperature tolerance (Takasuka et al., 2007). For this last, it has been found that cold phases favour the anchovies over sardines (Lluch-Belda et al., 1989, 1992). Instead, curl-driven upwelling would provide feeding and a better spawning habitat for sardines (Rykaczewski and Checkley, 2008). Furthermore, Fiedler et al. (1986) described a negative effect of the ENSO (El Niño/Southern Oscillation) on the Northern anchovy in the California Current Ecosystem. This was associated to an increase in nutrients from the upwelling that was detrimental for anchovies (Hsieh et al., 2009). On the other hand, the relationships between both species biomass and distribution area for the period 1978-1995, as found by Barange et al. (2009), appears very similar for both species. Authors indicated that this may be reflecting a decadal pattern of the habitat availability, rather than species pattern of occupation. None of the hypotheses are exclusive and synergistic in a theoretical framework; they suggest that both species responds oppositely to external forcing (Barange et al., 2009). Chapter 4. Engraulis mordax and Sardinops caeruleus off California 123 These types of studies might rely on catch data or on estimated biomass. Catch data are generally more accessible, concretely for last decades. They also provide better information concerning the fishery effect on populations. Is also important to consider that surveys/catch data, use to focus in one species even they report results for others apart from the targeted one. For instance, off California, surveys conducted in spring have good coverage of sardine spawning, but less for anchovy. The later spawns the year-round but with a major peak in late winter-spring (Barange et al., 2009). Unfortunately this type of data series is usually not long enough to give information about large-scale variability due to factors different for the exploitation of the stocks. In this sense, historical series (estimated biomass) are valuable tools when analysing long-term fluctuations due to environmental factors. Long series of catches are not available because most of the records have been gathered in the last half-century. For this reason, reconstructed series are so important in these evaluations. Is noteworthy the interest of scientific community in explaining the characteristics of ecological time series and the linkages between populations and environmental series in the past, to understand the present and predict the future. In particular, a variety of ecosystems and populations are driven by large-scale climatic oscillations (Jacobson et al., 2001; Cazelles et al., 2008; Hsieh et al., 2009). The California region is of particular interest in this theme because of the paleo-ecological investigations carried out by Soutar (1967) and Soutar and Isaacs (1969, 1974). They derived anchovy and sardine biomasses from the scale-deposition rate in two anoxic basins off California (Santa Barbara and Soledad basins). Anaerobic sediments conserved organic residuals, such fish scales, because anaerobic bacterium does not have the same destructive effect as the aerobic ones. In addition, they remain almost as undisturbed sediments, due to the scarce of biota in anoxic environments (only bioturbation from 1810 to 1860 approximately, as reported in Soutar and Isaacs (1974) due to an increase in dissolved oxygen). Also these substrates Chapter 4. Engraulis mordax and Sardinops caeruleus off California 130 (iii) Multivariate moving-window regression (Casals et al., 2002). This third method allows seeing which variables are, potentially, affecting in different periods of the time series to both species abundance. Also this gives an idea if significant correlations are found at several time lags among different groups of variables. The window-size (wl) in years was selected according to the longterm processes that were looked for. To see large-scale climate influence on medium-life cycle species, wide windows should be considered. Due to the different nature of the environmental variables considered, an intermediate wl was considered (wl = 50). This length allowed the detection of large-scales variability without diminishing the short-scales ones, as the ENSO punctual effect. Smaller sizes gave back too noisy results, and bigger ones remain scarce. Also the numbers of years that the window should be moved (step) must be fixed accordingly to the aims of interests, wl and the length of the series. To gain satisfactory results, 10-years step were considered. Finally, the lagged years (3 years) for the cross-correlations involved in this method were also determined considering the species life-cycles (FAO, 1985, 1988). The Pacific sardine can live for 20-25 years (FAO, 1985), however it was considered that in the first stages of development they are more environmentally influential. On the other hand, the Northern anchovy gets it standard maximum lengths with 2-3 years (FAO, 1988). For all this, a 3 to 5year lags were considered in order to evaluate climate effect on those species. To validate the results of each window analysis, a bootstrap test was conducted following a step-wise procedure for 1000 repetitions. In the x-axes of the Multivariate moving-window regression figures, are the year’s scales. In the y-axes the different considered variables within this analysis are plotted; in each bar, the dots in the vertical are representing the significant correlations lagged in time. In this sense, the lowest dot is the significant correlations within variables for a 0-year lag, the second upward is Chapter 4. Engraulis mordax and Sardinops caeruleus off California 131 the 1-year lag correlation, and so far so on. This analysis was conducted with each species and this is indicated in the first horizontal bar of the plots. Results Wavelet and Stationary Bootstrap AnalysesFrom all the analyses done just the coherence wavelet with significant results (0.95% confidence level –significant at 5 %-) inside the cone of influence (COI) are shown (Table 1). Periods that appeared as significant out of the COI are dismissed because there might be an edge effect due to the finite length of the time series (Torrence and Compo, 1998). This last was corrected in the stationary bootstrap when constructing pseudo-time series as circles. A value of zero in the phase of analysis indicates a shift from phase to out-of-phase or vice versa. With this method, significant associations between abundance of the anchovies and the sardines (Figure 2) when their behaviour was shifting (phase=0. This means that before the correlation the relation between both species was out-of-phase and that after the significant correlated period it turns to an in-phase period). The first significant relationship among the two species was from the year 940 to 1011; ranging in scale from 17.6 to 30.6 years. The second one was from 1324 to 1382 and in a scale from 15.3 to 24 years. These significant results in those two periods were tested through the stationary bootstrap correlation coefficient (R ≈ 0.2). Focusing on the significant bars at the stationary bootstrap plot, it can be appreciated that sardines and anchovies population during several years (10 negative lags and 1 positive lag) are somehow related through an external forcing or through their interspecific relationship. Chapter 4. Engraulis mordax and Sardinops caeruleus off California 132 Figure 2: Wavelet Coherence for the anchovies and sardines biomasses. The upper figure is the coherence variability and beneath it is the phase/out-of-phase relationship. If attention was paid to the effect of different climate variables in each species, it was found that the ITCZ fluctuations, through the Sacramento River runoff has a slightly significant influence on the anchovy population as seen in the wavelet analysis (Figure 3). It was appreciated that there were some orange-red zones from 1300 to 1350 and in 1600, denoting a high relation among the considered variables. Going further, these relations coincided when the behaviour was close to a swift between phase and out-ofphase processes. The stationary bootstrap evidenced that these were really negative significant relationships, but they were not very strong (R < -0.1). Chapter 4. Engraulis mordax and Sardinops caeruleus off California 133 Figure 3: Wavelet Coherence for the anchovies biomass against the Sacramento River flow. The upper figure is the coherence variability and beneath it is the phase/out-ofphase relationship. As happened with the anchovy, the abundant rainfall, which derived in a high Sacramento River flow, also affected the Pacific sardine through the wavelet analysis (Figure 4) between the years 1308-1324, during 2.4 years (4.1-6.5). Small significant regions (red) are appreciated alternating with no significant ones (blue). The stationary bootstrap slightly reflected this negative relation (R=-0.06) with a 23-24-years lag. However, it was not highly significant (significant R close to the interval of confidence). Chapter 4. Engraulis mordax and Sardinops caeruleus off California 134 Figure 4: Wavelet Coherence for the sardines biomass against the Sacramento River flow. The upper figure is the coherence variability and beneath it is the phase/out-ofphase relationship. The Beryllium-10 affected the Pacific sardine population in a highly significant way (Figure 5) from the year 1601 to 1680, in a 2.6 scales of years (5.1-7.8). It is an out-of-phase relationship. The stationary bootstrap reflected this negative influence with a lag of 10 years onward (R ranging roughly from -0.22 to -0.28). Chapter 4. Engraulis mordax and Sardinops caeruleus off California 135 Figure 5: Wavelet Coherence for the sardines biomass against the Beryllium-10. The upper figure is the coherence variability and beneath it is the phase/out-of-phase relationship. The remainder environmental variables did not have consistent significant results in the wavelet analyses, neither in the stationary bootstrap test. Multivariate Moving-window Regression AnalysisAfter testing different parameters sizes, the wl=50, steps=10 and 3-lagged was considered the right size within the length-of-influence of the different climate parameters (not to wide either to short) in the correlations with anchovies (Figure 6) and sardines (Figure 7) off California. Repetitions of the analysis Chapter 4. Engraulis mordax and Sardinops caeruleus off California 136 were done 1000 times. Is important to remember the years in which each series begun: enso (1650-1970), PDO (1470-1970), Be (1000-1970) and river (901-1970). Both species series started at the 283 and finished in 1970 A.D. Figure 6: Multivariate window-moving regression considering anchovies as the dependent variables. All the climatic parameters, as well as the sardines data are the independent ones. The wl=50, the step=10 and a maximum lag=3. Chapter 4. Engraulis mordax and Sardinops caeruleus off California 137 Figure 7: Multivariate window-moving regression considering sardines as the dependent variables. All the climatic parameters, as well as the anchovies data are the independent ones. The wl=50, the step=10 and a maximum lag=3. In general, what this analysis highlighted was that environment variables effect is not persistent in time but do affect the Northern anchovy and the Pacific sardine. And that they do not act solely. Also, that all the variables show a different wide-cyclical significant-no significant interactions with both species. Chapter 4. Engraulis mordax and Sardinops caeruleus off California 138 Table 1: Summary of the wavelet coherent and bootstrap analyses significant results. Pairs of variables Correlation period (Year) Length of correlation (Year) Correlation coefficient Anchovy-Sardine 940-1011 17.6-30.6 1324-1382 15.3-24 0.2 Anchovy-Sacram. River 1300-1350 1600 * < -0.1 Sardine-Sacram. River 1308-1324 4.1-6.5 -0.06 SardineBeryllium10 1601-1680 5.1-7.8 -0.26 *This significant correlation was verified through the bootstrap analysis, although in the wavelet method it seems that it might be a spurious correlation. This is why it was not possible to determine the length of the years of correlation. However, the result validated through the bootstrap was relevant enough to be highlighted. For the anchovy (Figure 6), it was appreciated that the abundance of the next year was dependant on the year before for the whole period. Some effect was detected with 2 and 4-years lags, but the proportion of significant results increased for lags=3 and 5. Sardines seemed to predominate on the Northern anchovy only for certain stretches of the period analyzed. There were only two relative long periods (650-920s and 1000-1200s) without sardine linkages recorded. Also the influence of the cold weather indicated by an increase in the Sacramento River runoff, can be detected after the year 1200, 300 yeas after the beginning of the series. The PDO remains the less influential variable on anchovies; just some influence can be detected in the Chapter 4. Engraulis mordax and Sardinops caeruleus off California 139 mid XVI century and beginning of XX. The Be seems to control the anchovy population every ~200 years during cycles of 15 years. When evaluating the sardine abundance (Figure 7) it can be seen that the Be was cyclically well correlated with it during the whole period. However, it did not reflect relative long periods without affecting the sardine biomasses, as happened with the anchovy. The ITCZ index, through the Sacramento River, gave significant relations with sardines only during the XII century period and few isolated years afterward. Discussion We have presented the results for a large-scale evaluation of the effect that the environment must be having on the Pacific sardine and on the Northern anchovy off California. It has been largely argued if both species live in a permanent competition for their feeding resources and the habitat they occupied (Soutar and Isaacs, 1974; Daan, 1980; Silvert and Crawford, 1988; Baumgartner et al., 1992; Klyashtorin, 1998; Freón et al., 2003; Barange et al., 2009). In this evaluation of the system, it was found a worth significant relationship between anchovies and sardines (Figures 2, 6 and 7). The first relation was obtained from 940 to 1011; this coincides with the beginning of the Medieval Warm Period (950-1220 in California. Li et al., 2000). At the year 1000, a temperature maximum was reached which seems to be detrimental for this relation. The second significant correlation period (13241382) coincides with the end of the A. D. 1300 Event and the beginning of the Little Ice Age; just after this period, in 1400 there were another minimum that broke the two species relationship. These results are an indirect measure of the effect of temperature on both, sardine and anchovy. They corroborate the hypothesis of an Optimal Growth Temperature, as an extent of the Optimal Environmental Window Theory, proposed by Takasuka et al. (2007). 147    GENERAL DISCUSSION    “The Earth functions as a unique and auto-regulated system, formed by physical, chemical, biological and human components. The interactions and fluxes of information between the former parts are complex and highly variable in multitude of temporal and spatial scales” (Lovelock, 2007) The actual loss and alteration of the biodiversity is a worrying issue. Concerning this, enormous attention has been paid to the climate change caused by anthropogenic actions and to the resource’s overexploitation. However, the influence of the climate goes further, but to understand this, studies have to go back in time as far as possible. Although historical studies are limited by the quality of the documents used (even more as going further back in time) this kind of information become scarce and harder to interpret (Cushing, 1982). The apparent response of marine ecosystems to short-term and decadal scales atmospheric variations suggests that the large scale climate/ocean processes may be the principal responsible for the fluctuations observed in marine populations (Poulard and Blanchard, 2005). Concerning this issue, the knowledge that the biosphere and the atmosphere are linked is not new (Bernal, 1951; Wigner, 1961; Lovelock and Margulis, 1973). In 1972 James E. Lovelock proposed the Gaia Theory based on the idea that “life is one member of the class of phenomena which are open or continuous reaction systems able to decrease their entropy at the expense of free energy taken from the environment and subsequently rejected in a degraded form”. This means that the biosphere and the physical Earth component are integrated, forming a complex system. How can they be isolated? It is impossible to understand what is happening in one of them without knowing General Discussion 148 the behaviour of the other. Related to this, it has been stated that, the observed punctual (chapter 1) and medium/long-term (chapters 2-4) changes in the marine ecosystems are, in a high proportion, due to the climate variability. Largely in the literature the importance of local condition is highlighted (Cushing, 1982; Caballero-Alfonso et al., 2010; chapter 3), but large-scale atmospheric processes are gaining the interest of the scientific community. Marine ecosystems are complex adaptive systems (Alheit and Bakun, 2010), and characterized by the non-linearity of their dynamics (Ottersen et al., 2004; Alheit and Bakun, 2010). This implies that small changes in large-scale climate patterns (e.g.: ENSO or NAO) may produce large effects at various species and/or trophic levels (Ottersen et al., 2010). Due to this, it is essential to determine the environment’s degree of influence (Levin, 1998; Ottersen et al., 2001). The goal of this Thesis was to contribute to the comprehension on how climate has affected marine ecosystem. Firstly, an extensive review of the historical knowledge on the evolution of the marine ecosystems, as well as of the climate was carried out. After that, how some species response to climate variables changes was posed. As an approximation to the proposed objective, three types of marine groups (with different biological mechanisms and possibly different responses to natural climate variability) were considered: relatively short (cephalopods), medium (clupeids) and large (bluefin tuna) life span species. Marine organisms show a broad range of responses to environmental changes, due to the nature and intensity of the acting force, but also associated to the life features of each population. This complexity in the responses may explain the patterns of recruitment (Cushing, 1982; Caddy and Gulland, 1983) and also the biomass variations (Spencer and Collie, 1997) observed in marine populations. The biological changes can be seen in multiple space and time scales. An evidence of how fast marine population can evolve related to environment changes are all the unusual events gathered General Discussion 149 in the Introduction, which are only an small representation of what is reported in the literature (Quéro, 1998; Stebbing et al., 2002; among others). A recent example is the caught, out of its distribution range, of two specimens of football octopus (Ocythoe tuberculata) off Cantabria (Northwest of Spain) related to an anomalous warming of the SST (chapter 1). In the same way, the Octopus vulgaris (chapter 2) is highly temperature dependent in its embryonic development, paralarvae stage, benthic settlement and spawning-catching peaks (Boyle, 1983). Cephalopods are short-life cycle species (Boyle, 1983, 1987; Hernández-López et al., 2001) and this implies that their responses to environmental changes are fast (Hernández-García et al., 2002; Pierce et al., 2008). If the environmental force is strong enough, this influence can be detected with days of delay between the forcing and the anomalous capture or sighting (chapter 1). However is more common to appreciate the effect in a seasonal scale as reported for the Canary region with the common octopus (Octopus vulgaris) catches (chapter 2). Here the highest catches took place in spring (35.82%) and autumn (25.52%) coinciding with the spawning peaks (Figure 3 –left-, chapter 2). Figure 3 (left) of the chapter 2: Boxplot of the monthly CPUE. Winter Spring Summer Autumn 0 5 10 15 20 25 Boxplot Catches(kg) General Discussion 150 It was also seen that the temperature has an inverse relation with the octopus Capture per Unit of Effort (CPUE). More striking was the relationship with the NAO, which was direct for both seasons with the octopus abundance when the NAO index was changing phase. In spring it explained the 28.64% of the fluctuation variance; meanwhile in autumn it accounted for a 31.13%. These results were also found by Polanco et al. (2011) using a wavelet analysis that might correct the error associated to the lineal models applied within chapter 2; this finding is giving consistency to our results. As Cushing (1982) posed, the “fish stocks are sufficiently remote from climatic events to restrict the number of physical factors common to both”. This idea can be extended to all marine populations, but also the interest has to focus on the details of which variables are influencing each species population in higher or less proportion. In this sense, data sets compiling and computerization, as well as the statistical methods are improving and representing in a better way with time what is taking place in the nature. However, larger and more truthfulness climatic and biological series have to be gathered and available for the scientific community. Besides, factors as the wind, the temperature or the solar activity variability still predominant in the determination of the rates of natural increase or decrease of the marine stocks through their effect in larvae and juvenile stages (Cushing, 1982; chapters 1-4). This is also observed in the octopus biomass, where the temperature and the NAO affected octopus paralarvae survival, growth rates, age of juvenile benthic settlement and the timing of the reproductive peaks; that is, on recruitment to the fishery (chapter 2). Previous works, since the 1990s (Ottersen et al., 2001), also highlighted the influence of this atmospheric pattern, as the ENSO, the NAO or the ITCZ displacement, on other type of marine populations. For instance, Fromentin and Planque (1996), stands out its influence on two Atlantic zooplankton species, Calanus finmarchicus and C. helgolandicus. Both General Discussion 151 species have different spatial and temporal patterns in the eastern North Atlantic and in the North Sea (Planque and Fromentin, 1996) due to different responses to environment and because of other biological factors. Their longterm trends are opposite. On the other hand, positive NAO values are related with low C. finmarchicus abundances and vice versa. Fromentin and Planque (1996) posed that the relationship between NAO and C. helgolandicus is less evident (only the 18% of the variability is explained oppositely to the 58% observed with the other copepod species), but it remains as a positive relationship. However extreme NAO indices caused drastic changes in Calanus species. The differences in the way the NAO influence these two species are associated to the intermediate factors such the wind stress, the temperature or even the competition between them. Overall, the NAO seems to be the main responsible controlling the Calanus replacement. A NAO effect can be also detected when evaluating higher trophic levels. For example, a positive NAO in the north of Europe (implies cold phases in the Mediterranean basin) translates in warm years that favoured higher growth rates and survival of the cod (Gadus morhua) in the Arctic, Norwegian, west Greenland and off Canada waters. This derived in a rise in the capelin (Mallotus villosus) consumption in the Barents Sea by cod. The same is observed with the North Sea cod (Ottersen et al., 2001). It has been seen that the NAO caused changes in the thermal habitat of the Atlantic salmon (Salmo salar) too (Ottersen et al., 2001; Drinkwater et al., 2010), because its distribution range decreases with positive NAO values (warm conditions in northern latitudes) and expands when negative conditions occur. This behaviour has also been detected for the cod in the Barents Sea, the herring (Clupea harengus) and for sardines (Sardina pilchardus) in the northern of Europe (Ottersen et al., 2001). In the last two cases, the herring is favoured with negative NAO indices, meanwhile the sardines prefers positive values (Brander, 1995; Alheit and Hagen, 1997; Ottersen et al., 2001). These are more or less direct responses to NAO, but complex ones might involve General Discussion 152 physiological responses that modulate population’s dynamics and interactions (Ottersen et al., 2001). The NAO affects different species in different ways when it is in positive or negative phase. In chapter 2, the NAO affected the common octopus when changing from one phase to another in the Canary Islands. All this results highlight the need of evaluating the effects of climatic oscillations at individual, population and community level (Caballero-Alfonso et al., 2010; chapter 2). This is due to the fact that such patterns affect the reproduction, abundances, distributions, species competition… responses. Also a bigger comprehension of the intermediate mechanisms, between the NAO and the ecosystems response, should be gathered (Ottersen et al., 2001). Beside the previous results, no significant influence of the NAO has been detected in other species such the bluefin tuna (Thunnus thynnus) (Ravier and Fromentin, 2004; Ganzedo-López et al., 2009). Although Santiago (1998) reported that the relationship between winter NAO index and bluefin tuna was positive, but negative with albacore (T. alalunga). Contrary, the temperature has been largely considered as the main controlling factor in the northward bluefin tuna migration up to their spawning regions in the Mediterranean Sea (e. g.: Ravier and Fromentin, 2001; Ravier and Fromentin, 2004). This is not trivial since it is the variable that controls biological processes as the reproduction phase or the spawning moment (Fromentin, 2006; Goldstein et al., 2007). However, in chapter 3 the importance of the solar activity (measured as irradiance. Lean, 2000) was highlighted as a possible cause under the observed long-term abundance fluctuations, together with other climatic factors. In the Gibraltar Strait (1610-1936) a 68.02% explained variance of the Bluefin tuna fluctuations was observed; while for the western Mediterranean (1700-1936) the explained variance was of 46.5%. More over, those abundances minima coincides with the minimum values of environmental variables (Figures 6 and 8 of the chapter 3). These minima corresponds with the solar activity minima described in the literature, which General Discussion 153 in turns controls the temperature (Gleissberg, 1958; Bond et al., 2001). However, an important part of the variability in the conducted analyses remains unexplained probably due to additional local-scale climatic events and socio-political circumstances (e.g.: wars, famines, bankrupts, etc.) that historically affected the Mediterranean countries and this traditional fishery. Figure 6 of the chapter 3: Gibraltar Strait yields and models for the period 1610 to 1936 (considering the SLT and the SIB). General Discussion 154 Figure 8 of the chapter 3: Western Mediterranean yields and models for the period 1700 to 1936 (considering the SLT and the SIB). For the bluefin tuna (chapter 3) local features seem to be playing a key role in the abundance of the stocks. Small differences were found between regions considered in chapter 3 (i.e.: Gibraltar Strait and Western Mediterranean). Concretely, the fluctuations of the Mediterranean domain seem to be less explained by global parameters than the Gibraltar Strait ones. Also, in cephalopod species local features are important because of their rapid response to environmental changes. In chapter 2 it was highlighted the differences between regions for the same species, the common octopus. For example, concerning the temperature, Sobrino et al. (2002) found that the maximum octopus abundance coincided with the minimum sea surface temperature (SST) registered at the studied domain (Gulf of Cadiz), as happened in chapter 2. In contrast, Balguerías et al. (2002) and Moreno et al. General Discussion 155 (2002) reported for the Saharan Bank and the Portuguese coast, respectively, that maximal captures coincided with the highest SST in those domains. These are clear examples of the effect of climatic forcing on the recruitment when controlling populations abundance (chapter 2 and 3), but some times adults can move out of their habitat looking for better conditions if the environment conditions favoured it (chapter 1). In this sense, Drinkwater et al. (2010) explains that observed changes in marine ecosystems are a consequence of the local conditions. Also large-scale climatic indices often account for significant portions of this local or regional variability. This is because the local climate changes are often a response to the large-scale processes, and they are related to several physical elements (air and ocean temperatures, sea ice, winds...). Due to this, they can be more representative of the climate forcing than any single local variable (Stenseth et al., 2003). Also, local indices can vary significantly while largescale indices are smoother, with less signal of noise (Stenseth et al., 2003). Furthermore, a variety of ecosystems and populations are driven by largescale climatic oscillations (Jacobson et al., 2001; Cazelles et al., 2008; Hsieh et al., 2009). Drinkwater et al., (2003) point out that “in some cases largescale indices can account for as much, or at times more, of the variance of the ecosystem elements than the local indices”. The advantage of using largescale climate indices is the possibility of linking climate-induced ecological dynamics over a range of trophic levels, species, and geographical locations (Drinkwater et al., 2010). To diminish the uncertainty concerning how natural climate (largescale) variability affects different populations, time-series should have, at least, more than 100 years data (as concluded from the conducted time-series analyses within this Thesis). Global marine ecosystems long-time series are not common in literature either in data bases. However, the California basin is one of the most documented regions concerning the climate effect on fishes General Discussion 156 due to the California Cooperative Oceanic Fisheries Investigations (CalCOFI) program. Pelagic ecosystems vary largely in spatio-temporal scales (Hayward, 1997; Ottersen et al., 2010). From chapter 3 and 4 it can also be concluded that the biological response to environmental variability differs between species. In the mid-1970s, the sea surface temperature and the upwelling intensity off California increased, although this is contrary to what is expected (Hayward, 1997). However, there was also a decrease in the macrozooplankton community (Hayward, 1997) lasting until 1995. A decline in zooplankton and anchoveta (Engraulis rigens) biomasses was also observed in the Peru Current system due to the same physical pattern as off California (Hayward, 1997); although, this condition favoured the sardines (Sardinops sagax). Finally, mackerels (Scomber japonicus and Trachurus murphyi) seem to be unaffected (Hayward, 1997). A relevant investigation within this issue is the one carried out by Soutar (1967) and Soutar and Isaacs (1969, 1974). They reconstructed time series of the biomass of Northern anchovies (Engraulis mordax) and Pacific sardines (Sardinops caeruleus) in the Santa Barbara and Soledad basins (California) from scales deposited in the sediment. From an economical and ecological point of view small pelagic species, such Sardinops caeruleus (Pacific sardine) or Engraulis mordax (Northern anchovy), are interesting (Barange et al., 2009). Therefore, they have been evaluated for decades (Barange et al., 2009; chapter 4), with most of the studies focused on the fluctuation causes and on the alternation of both species (Daan, 1980; Silvert and Crawford, 1988; Kawasaki and Omori, 1988; Lluch-Belda et al., 1989, 1992; Baumgartner et al., 1992; Klyashtorin, 1998; Schwartzlose et al., 1999; Chavez et al., 2003; Freón et al., 2003; van der Lingen et al., 2006; Takasuka et al., 2007; Barange et al., 2009). Concerning the latest, in chapter 4 it was concluded that the alternation in both species seem to be linked to climatic conditions more than to behavioural features or Conclusions 164 fact that local and biological features have to be considered too, and also due to the complex and non-linear interactions within variables of the system that are difficult to completely consider in the models. In addition to the previous conclusion, also the sea level temperature remains as a key factor influencing the fluctuations of the bluefin tuna. 3. The solar and the Intertropical Convergence Zone influences have to be highlighted as important environmental factors controlling Engraulis mordax and Sardinops caeruleus fluctuations off California. However, when applying a multivariate moving-window regression it was found that other variables were playing a key role in some moment on the marine populations evaluated (even those that remain no-significant when considered solely). In this sense, the present study contributes to reduce the uncertainty of the relationships between climate and the small pelagic fluctuations. However, more variables should be taken into account and a more precise multivariate regression test should be done, as well as non-linear analyses that involve the complexity of the interactions between the different parts of the system. In this evaluation, it was found an important significant relationship between anchovies and sardines. The results reveal that either cold conditions or too warm ones can modified this relationship between both species. It was assumed that this co-existenting relation was provoqued more by climatic factores than by competitive strategies for food resources. 4. In conclusion, minima solar activity seems to be an important factor in the control of the long-term fluctuation of the bluefin tuna, Pacific sardines and Northern anchovies. On the other hand, the North Atlantic Oscilation is known as an important atmospheric pattern on the North Atlantic Ocean, but in this case, in the short-scale processes. This last issue Conclusions 165 highlights the need of developing more local atmospheric indices which explain regional variations in a higher proportion. However, little is known about the solar effect on marine populations due to the short length of the fisheries series available. Finally, a data set longer than 100 years would be needed to see this type of influence (whereas series have to be more complete to be able to apply more accurate analyses than linear models). F FF F F FF Fu uu u u uu ut tt t t tt tu uu u u uu ur rr r r rr re ee e e ee e R RR R R RR Re ee e e ee es ss s s ss se ee e e ee ea aa a a aa ar rr r r rr rc cc c c cc ch hh h h hh h 169    FUTURE RESEARCH    In the present Thesis the role of the solar activity has been strongly highlighted. But also large-scale atmospheric patterns as the El Niño/Southern Oscillation or the North Atlantic Oscillation remain crucial when describing marine population’s behaviour. However, there are several issues that remain unsolved and that are open doors for future researches. Firstly, it is important to gather long-time abundance series that can be compared to the climatic reconstructions ones. The solar activity seems to be relevant in the environment-ecosystems interaction, but to see it influence more than one century of data is needed. Also, the reconstruction of the global and regional past climate must continue because the same physical process can act in a different way around the globe. For example, it has been shown that during the Little Ice Age, the tropical Pacific was under relative warm and stable conditions (Fagan, 2000). Secondly, and related to the previous statement, local atmospheric indices should be developed. An example of the relevance of having local indices is the Western Mediterranean Oscillation Index (WeMOi) described by Martín-Vide and López-Bustins (2006). It is defined by means of a dipole composed by a high pressure over Azores and a low pressure over Liguria. It was built to study local phenomena, and something similar is required in regions with specific environmental conditions as, for instance, the Canary Archipelago. In conclusion, more complex environmental-ecosystems non-linear analyses should be conducted, including as much variables as possible to achieve a better understanding of the surrounding system. S SS S S SS S p pp p p pp p a aa a a aa a n nn n n nn n i ii i i ii i s ss s s ss s h hh h h hh h S SS S S SS S u uu u u uu u m mm m m mm m m mm m m mm m a aa a a aa a r rr r r rr r y yy y y yy y R RR R R RR Re ee e e ee es ss s s ss su uu u u uu um mm m m mm me ee e e ee en nn n n nn n e ee e e ee en nn n n nn n E EE E E EE Es ss s s ss sp pp p p pp pa aa a a aa añ ññ ñ ñ ññ ño oo o o oo ol ll l l ll l Spanish Summary/Resumen en Español 173    INTRODUCCIÓN    El interés del ser humano en las poblaciones marinas no es algo de los últimos siglos (Castro-Hernández, 2009). Existen pruebas de que los Neandertales (ca. 100 000 años) consumían pescado. Del mismo modo, hay pinturas rupestres de más de 25 000 años, en las que se representan actividades pesqueras en Sudáfrica y Namibia. Por otra parte, las primeras muestras de pesca desde embarcaciones datan del Mesolítico (ca. 10 000 años) (Sahrhage y Lundbeck, 1992), lo que se sabe por el hallazgo de restos de bacalao, arenque, congrios,… en un asentamiento en Escocia. En Perú se encontró la red más antigua del mundo hasta el momento, con 8 800 años de antigüedad (Castro-Hernández, 2009). Al margen de todo esto, cabría pensar que la sobreexplotación de los recursos marinos es también consecuencia de la Revolución Industrial (desde 1850); sin embargo, Sahrhage y Lundbeck (1992) destacan que desde la época de los romanos (200-300 A. D.) ha existido la sobreexplotación de este medio debido al aumento poblacional, que a su vez vino asociado con una mayor demanda de pescado. Aunque esto no deja de ser más que un hecho puntual, pero destacable, se sabe que es un mal generalizado después de la mitad del siglo XIX (Pauly y MacLean, 2003). No se trata de algo sin importancia, ya que las actividades antropogénicas no son la única variable que afecta a las poblaciones marinas, aunque hoy se revela como el factor más importante (Pauly, 2009). Los ecosistemas marinos fluctúan de forma natural en múltiples escalas de tiempo debido a una combinación de la dinámica interna de la propia población, de las interacciones predador/presa y a la competencia dinámica; pero esto también se debe al efecto de la variabilidad climática (Cushing, 1982; Laevastu, 1993; Lehodey y colaboradores, 2006; Barange y colaboradores, 2010). Las variaciones climáticas son más o menos cíclicas y se sabe que afectan a las abundancias y migraciones de las poblaciones marinas, desde las Spanish Summary/Resumen en Español 174 comunidades de plancton hasta la de peces (Caballero-Alfonso, 2009; Drinkwater y colaboradores, 2010). El clima induce cambios esporádicos en la fauna marina Las condiciones ambientales son un factor determinante en la biodiversidad, así como en la estabilidad de los hábitats. Mucho se ha hablado de cómo afecta la variabilidad climática global y local a los ecosistemas marinos en las distintas escalas de tiempo (Cushing, 1982; Ravier y Fromentin, 2004; Ganzedo y colaboradores, 2009; CaballeroAlfonso y colaboradores, 2010; Drinkwater y colaboradores, 2010). La atmósfera y el océano se encuentran en una continua interacción dinámica mediante el intercambio de energía, y sin embargo, los océanos poseen una mayor capacidad de almacenar calor que la atmósfera. Este calor se distribuye por los océanos mediante las corrientes oceánicas y por el intercambio que realiza con la propia atmósfera. Inicialmente, esto controla la Temperatura Superficial del Mar (TSM) y, en consecuencia, el comportamiento del resto del océano y la distribución de nutrientes. Esta variabilidad hidrodinámica afecta a los ecosistemas marinos (Anadón y colaboradores, 2005). A colación con esto, algunos estudios advierten de la posible influencia que el incremento del dióxido de carbono (CO 2 ) antropogénico, que se está dando desde la Revolución Industrial, puede estar teniendo en este entorno (Fabry y colaboradores, 2008; Drinkwater y colaboradores, 2010) debido a la acidificación de los océanos (que siempre han tendido a ser ligeramente alcalinos) (Caldeira y Wickett, 2003; Feeley y colaboradores, 2004). Para algunas especies, como las fitoplanctónicas, esto puede resultar beneficioso, pero para otras es muy perjudicial; por ejemplo, para todos aquellos organismos con estructuras hechas de carbonato cálcico (CaCO 3 ), porque este compuesto se reduce en medios ácidos (Kleypas y colaboradores, 1999; Riebesell y colaboradores, 2000), lo que significa que las conchas y los esqueletos serán cada vez más frágiles. Spanish Summary/Resumen en Español 175 Por otra parte, la temperatura corporal de la mayoría de las especies de peces depende de la del ambiente en que se encuentran (organismos de sangre fría o poiquilotermos) (Jobling, 1994), ya que ésta juega un papel fundamental en el crecimiento, metabolismo y comportamiento de esos animales (Ali, 1980; Huntingford y Torricelli, 1993; Godin, 1997). Este parámetro varía mucho en los océanos, oscilando en media desde los 0 ºC en los polos a los 26 ºC en la región ecuatorial. Cada especie posee un pequeño rango de tolerancia térmica y, además, suelen vivir cerca de su límite (Harley y colaboradores, 2006). Por esta razón se encuentran diferentes especies viviendo en distintas áreas geográficas (Wootton, 1998); las más sensibles son aquellas que habitan en latitudes tropicales y subtropicales. Mientras que las boreales y polares soportan mejor los cambios en su hábitat (Poulard y Blanchard, 2005). Por ejemplo, algunas especies de túnidos (p.e.: Thunnus thynnus) se encuentran en aguas que oscilan entre los 3 y los 30 ºC, aunque se desplazan en busca de los 24 ºC o más durante los períodos de puesta (Ganzedo, 2005; Fromentin, 2006). Es de esperar que cuando la temperatura cambia, los individuos se vean forzados a moverse fuera de su dominio habitual de distribución. Por otro lado, especies de zonas templadas y polares pueden coexistir en sus rangos de distribución. En este sentido, se ha observado que con el aumento de la temperatura, las especies de aguas templadas tienden a aumentar su rango de distribución así como sus abundancias (Caballero-Alfonso, 2009). Por otro lado, las especies polares son más estables o incluso puede que disminuyan un poco sus abundancias relativas (Poulard y Blanchard, 2005). Así, pequeños y graduales cambios en la temperatura, producen avances y retrocesos en los límites poblacionales, pero un cambio brusco puede producir la muerte en aquellas especies sensibles a las variaciones, ya que no serán capaces de adaptarse (Drinkwater y colaboradores, 2010). En relación a lo anterior, se ha encontrado un aumento en el número de especies halladas fuera de su rango habitual de distribución (Caballero- Spanish Summary/Resumen en Español 176 Alfonso, 2009). Globalmente, en todos los océanos, diversos investigadores han citado avistamientos o capturas de especies hasta ahora desconocidas o poco frecuentes en regiones concretas. En otros casos, lo que se ha registrado, son cambios en los comportamientos de las poblaciones. Por ejemplo, en el Océano Pacífico, el bacalao del Pacífico (Gadus macrocephalus) ha dejado de capturarse en el Noroeste de Estados Unidos como sucedía antes y, en cambio, se está capturando en las costas de Canadá. Incluso en Alaska se han observado alteraciones en las abundancias de la perca del Pacífico (Sebastes alutus), el abadejo de Alaska (Theragra chalcogramma) y en el fletán del Pacífico (Hippoglossus stenolepis) y, además, se ha observado una sucesión en las apariciones en esa costa de estas especies en las últimas cinco décadas. Estos hechos han sido asociados a un cambio en la temperatura del agua en el sur de California (Hollowed y colaboradores, 2001). Del mismo modo, en el Mar de Bering y en las Aleutianas, la perca del Pacífico (Sebastes alutus), del arenque del Pacífico (Clupea pallasii pallasii) y del fletán negro (Reinhardtius hippoglossoides) están siendo reemplazados por peces planos típicos de aguas más cálidas (Hunt y colaboradores, 2002). Estos últimos autores también han citado cambios en los ecosistemas pelágicos, en el salmón y en poblaciones de peces bentónicos y de cangrejos. En un principio estos cambios se vincularon a un aumento en la presión que los predadores, como el abadejo (Pollachius pollachius), ejercían sobre estas especies, pero estudios recientes han demostrado que se deben a causas climáticas. Concretamente, en el mar de Bering, el aumento de la temperatura y la disminución de la cobertura de hielo pueden haber favorecido el desplazamiento hacia el norte de especies como el abadejo o el bacalao del Pacífico (Gadus macrocephalus), el lenguado del Pacífico (Lepidopsetta bilineata) y el fletán del Pacífico (Hippoglossus stenolepis) debido a la alta productividad que se está generando en el norte de esta región (Drinkwater y colaboradores, 2010). Spanish Summary/Resumen en Español 177 En el Océano Atlántico, también se está dando un desplazamiento hacia latitudes mayores de especies aguas cálidas. La mayoría de los casos, se han registrado en la costa suroeste del Reino Unido (Cornwall) y de Irlanda (Cork) (Poulard y Blanchard, 2005). En este sentido, en las aguas británicas, el área de distribución de las especies marinas locales se ha visto restringida en un 57-84% debido al calentamiento (Genner y colaboradores, 2004). Con esta tendencia, se espera que el bacalao (Gadus morhua) desaparezca del Mar Céltico y de las aguas de Irlanda para el 2100 (Drinkwater, 2005); también dentro de este escenario, en el sur del Mar del Norte y en las aguas de ‘George Bank’, esta especie disminuirá su abundancia pero sin llegar a desaparecer. Por el contrario, en Groenlandia, y en los mares de Barents y del Labrador (como sucedió para el período cálido de mediados del siglo XX), el bacalao aumentará su rango de distribución, llegando incluso a la plataforma Ártica debido a la disminución de la cobertura de hielo. La ventaja que esta especie presenta es que se trata de una de las más estudiadas a nivel mundial debido al gran interés comercial que posee (Fagan, 2000; Drinkwater y colaboradores, 2010). Se ha estimado que el 30% del bacalao del Mar del Norte se ha perdido debido a un aumento de 0.25 ºC en la temperatura del mar (Clark y colaboradores, 2003) y también, su tasa de crecimiento en el Mar del Labrador se ha visto afectada como consecuencia también de la variación de la temperatura del agua (Brunel y Boucher, 2007). Al margen de esto, se ha detectado un cambio en la composición de los animales marinos y en sus abundancias en las costas de Portugal, donde se han cogido especies Mediterráneas y del Noroeste de África que ahí eran desconocidas (p.e.: la barriguda Mediterránea –Parablennius incognitos-, la aceia lusa – Microchirus boscanion-, el roncador –Pomadasys incisus-, el tordo – Symphodus ocellatuso el tapaculo –Bothus podas-) (Cabral, 2002; Brander y colaboradores, 2003). Como especies nuevas en la costa sur de Portugal, podemos citar a las chuclas (Spicara flexuosa y S. maena) y lo que es más, los góbidos (Gobius couchi y Pomatoschistus pictus) han aumentado su rango de distribución (Arruda y Azevedo, 1987; Brander y colaboradores, Annexes/Anexos 287 Las Torres 1902-1923 Ravier, 2003 1933 Piccinetti and Omiccioli,1999 Reina Regente 1913-1940 Ravier, 2003 Sancti Petri 1552-1606 López-Capont, 1997 1917-1929 Ravier, 2003 1930-1961 Rodríguez-Roda, 1964a 1962-1971 Rodríguez-Roda, 1973 Tarifa 1743-1755 López-Capont, 1997 1914-1924 Piccinetti and Omiccioli,1999 1927-1928 Ravier, 2003 1929-1961 Rodríguez-Roda, 1964a 1962-1971 Rodríguez-Roda, 1973 1972-1985 Rey et al., 1987b Torre Atalaja 1914-1934 Ravier, 2003 Zahara 1525-1756 López-Capont, 1997 1910-1936 Ravier, 2003 Conil 1525-1756 López-Capont, 1997 Castil Novo 1525-1622 López-Capont, 1997 Carboneros 1743-1754 López-Capont, 1997 Río Terrón 1741-1756 López-Capont, 1997 1914-1920 Piccinetti and Annexes/Anexos 288 Omiccioli, 1999 La Línea 1952-1961 Rodríguez-Roda, 1964a 1962-1972 Rodríguez-Roda, 1973 Las Cabezas 1914-1928 Piccinetti and Omiccioli, 1999 Nuestra Sñra. de la Cinta 1914-1928 Piccinetti and Omiccioli, 1999 La Higuera 1914-1928 Piccinetti and Omiccioli, 1999 Torre del Puerco 1919-1925 Piccinetti and Omiccioli, 1999 Morocco Punta Negra 1936-1957 Lozano-Cabo, 1958 1958-1960 Ravier, 2003 Garifa 1927-1957 Lozano-Cabo, 1958 Las Cuevas 1935-1954 Lozano-Cabo, 1958 Los Cenizosos 1940-1957 Lozano-Cabo, 1958 Jolot 1947-1954 Lozano-Cabo, 1958 Es-Sahel 1948-1954 Lozano-Cabo, 1958 Algiers Aguas de Ceuta 1940-1960 Ravier, 2003 Tunisia Bordj Khadidja 1903-1929 Ravier, 2003 Conigliera 1896-1929 Ravier, 2003 Monastir 1894-1926 Ravier, 2003 Pricipe 1940-1960 Ravier, 2003 Ras el Ahmar 1905-1941 Ravier, 2003 Sidi Daoud 1863-1960 Ravier, 2003 Annexes/Anexos 289 1986-1994 Piccinetti and Omiccioli, 1999 El Aouaria 1907-1930 Piccinetti and Omiccioli, 1999 Ras El Mihr 1908-1911 Piccinetti and Omiccioli, 1999 Ras Marsa 1907-1918 Piccinetti and Omiccioli, 1999 Kuriat 1904-1927 Piccinetti and Omiccioli, 1999 Ras Zebib 1923-1930 Piccinetti and Omiccioli, 1999 Libya Sliten 1925-1934 Piccinetti and Omiccioli, 1999 Gargaresch 1920-1928 Piccinetti and Omiccioli, 1999 El Mongar 1924-1927 Piccinetti and Omiccioli, 1999 Mongar El Chebir 1924-1931 Piccinetti and Omiccioli, 1999 Marsa Marrecan 1927-1928 Piccinetti and Omiccioli, 1999 Marsa Zuaga 1921-1931 Piccinetti and Omiccioli, 1999 Marsa Sabrata 1921-1931 Piccinetti and Omiccioli, 1999 Marsa Sorman 1924-1927 Piccinetti and Omiccioli, 1999 Marsa Dila 1927-1928 Piccinetti and Omiccioli, 1999 Ras Lahmar 1920-1929 Piccinetti and Omiccioli, 1999 Annexes/Anexos 290 Ras Urir 1926-1927 Piccinetti and Omiccioli, 1999 Gebbana Sidi Mahfud 1921-1927 Piccinetti and Omiccioli, 1999 Sidi Abdul Gelil 1920-1930 Piccinetti and Omiccioli, 1999 Sidi Sbeh Lahmar 1923-1931 Piccinetti and Omiccioli, 1999 Sidi Bu Mefta 1927 Piccinetti and Omiccioli, 1999 Dzeira 1920-1931 Piccinetti and Omiccioli, 1999 Punta Lebda 1921-1924 Piccinetti and Omiccioli, 1999 Ras El Mseu 1921-1926 Piccinetti and Omiccioli, 1999 Mellaha 1924 Piccinetti and Omiccioli, 1999 Marsa Hamra 1924 Piccinetti and Omiccioli, 1999 Marsa Beltar 1923-1925 Piccinetti and Omiccioli, 1999 Italy Saline 1823-1960 Ravier, 2003 Isola Piana 1820-1960 Ravier, 2003 Flummentorgiu 1824-1912 Ravier, 2003 Porto Scuso 1820-1960 Ravier, 2003 Porto Paglia 1830-1960 Ravier, 2003 Carloforte 1868-1995 Piccinetti and Omiccioli, 1999 Asinelli 1904-1929 Ravier, 2003 Annexes/Anexos 291 Bonagia 1603-1950 Ravier, 2003 Castellamare 1909-1931 Ravier, 2003 Cofano 1670-1796 Ravier, 2003 Favignana 1600-1796 Ravier, 2003 1929-1930 Piccinetti and Omiccioli, 1999 Formica 1602-1959 Ravier, 2003 Magazzinazzi 1891-1933 Ravier, 2003 Marzamemi 1874-1931 Ravier, 2003 Oliveri 1903-1960 Ravier, 2003 1961-1964 Piccinetti and Omiccioli, 1999 Pizzo 1876-1932 Ravier, 2003 San Cusumano 1903-1940 Ravier, 2003 San Elia 1909-1929 Ravier, 2003 San Giorgio 1880-1960 Ravier, 2003 San Giuliano 1601-1914 Ravier, 2003 San Nicola 1909-1929 Ravier, 2003 Scopello 1909-1960 Ravier, 2003 Secco 1874-1933 Ravier, 2003 Solanto 1909-1928 Ravier, 2003 Tonno 1884-1960 Ravier, 2003 1961-1963 Piccinetti and Omiccioli, 1999 Trabia 1909-1929 Ravier, 2003 Annexes/Anexos 292 Table 2: Register of the final almadraba’s data sources for each period of time considered in the study. Location Almadraba Period Reference Portugal Barril 1867-1933 Ravier, 2003 1934-1966 Lemos and Gomes, 2004 Medo das Casas 1776-1969 Ravier, 2003 Spain Zahara 1525-1756 López-Capont, 1997 1910-1936 Ravier, 2003 Conil 1525-1756 López-Capont, 1997 Italy Saline 1823-1960 Ravier, 2003 Isola Piana 1820-1960 Ravier, 2003 Porto Scuso 1820-1960 Ravier, 2003 Porto Paglia 1830-1960 Ravier, 2003 Bonagia 1603-1950 Ravier, 2003 Formica 1602-1959 Ravier, 2003 San Giuliano 1601-1914 Ravier, 2003 Annexes/Anexos 293 Brief description of the climate indices AMO -Atlantic Multidecadal OscillationIt is defined by the annual mean of SST averaged detrended over the North Atlantic. A positive AMO index is associated with higher than normal SST throughout most of the North Atlantic and if the SST is cooler than normal, it is a negative AMO index. It has varied between positive and negative indices with a periodicity of about 60-80 years since the late 1880s. The temperature changes linked to the AMO can obscure and exaggerate the anthropogenic induced warming (Latif et al., 2004). It correlates (Sutton and Hodson, 2005), particularly, with the air temperature and precipitations over North America and Europe (extracted from Drinkwater et al., 2010). El Niño El Niño events are characterized by a significant warming of the equatorial surface waters; they are a change in the marine current pattern movements in the intertropical region. During an El Niño event, the sea level atmospheric pressure tends to be higher than usual over the western tropical Pacific and lower than usual in the eastern part. The westerlies are intensifies, provoquing a displacement of the warm and shallower waters to the east, to the American coast. At the same time the thermocline in the west goes deepest than in normal situations (Fagan, 2008). ENSO -El Niño/Southern OscillationSO -Southern OscillationIt is an irregular fluctuation that causes a shift in the surface atmospheric pressure over the western and eastern tropical Pacific Ocean. When the surface pressure is high in the east, it is low in the west and vice versa. In general, this affects simultaneously the sea temperatures. This type Annexes/Anexos 294 of shift modifies the rainfall regimes and the wind in both regions (Fagan, 2008). ENSO In 1960, the meteorologist Jacob Bjerknes, established a relation between the SO and El Niño events. That relation was named El Niño/Southern Oscillation (ENSO), which is the term commonly used in the literature (Fagan, 2008). Warm ENSO events are those in which both, a negative SO and an extreme El Niño occur together; this produce an increase of the upper layer sea temperatures in the central and eastern tropical Pacific, with an associated shift in the rainfall pattern. The interval between ENSO events are around 2-7 years (extracted and modified from Drinkwater et al., 2010). ITCZ -Intertropical Convergence ZoneThe Easterlies winds from the northeast and southeast converge near the Equator, forming a zone of low pressure and displacing upward the humidity. This air mass that elevates, gets cold, condensing the water vapour and generating a band of abundance rainfall that moves with the seasons to where the solar irradiance and the surface temperature are higher. It is localize in the South Hemisphere from September to February and it moves to the North Hemisphere during it summer. El Niño events influence this pattern, displacing it to the tropical Pacific, where the SST are warmer (extracted from Fagan, 2008). La Niña It is characterize by cold temperatures and dryer conditions over the central equatorial Pacific than normal due to a weakening of the atmospheric jet stream. These events, as well as El Niño, produce changes in the atmospheric winds over the tropical Pacific Ocean, including an increase of Annexes/Anexos 295 the easterlies over the surface eastern Pacific and in the western Pacific in the upper layers of the atmosphere. In the ocean this is translated as a reactivation of the upwelling in the east coast and in a cooling of the surface water (Fagan, 2008). NAO -North Atlantic OscillationIt is the atmospheric pattern over the North Atlantic Ocean composed by a low pressure sistem over Iceland (subpolar) and a high pressure sistem over Azores (subtropic), that tend to intensify and weaken at the same time (Barnston and Livezey, 1987). The most commonly used index is the avarage of the surface atmospheric pressure difference between Iceland and Lisbon (Portugal) from December to March (winter NAO). The positive NAO phase is when the Low pressure over Iceland is weakener than normal and the Azores High is stronger than usual. This intensifies the westerlies winds across the North Atlantic and produce warmer temperatures and higher precipitations than normal in the eastern coast of the United States and northern Europe and cooling in the eastern Canada, Greenland and in the Mediterranean Sea. The opposite occurs during the negative phase. The NAO has strong influence on the atmosphere and hence in the physical oceanography of the North Atlantic. (extracted from Drinkwater et al., 2010). PDO -Pacifical Decadal OscilationIt is based on the observed monthly SST anomalies in the North Pacific north of 20ºN, after removing the long-term warming. It is a persistent fluctuation of the Pacific Ocean. During the cold phases, it can be observed a decrease of the SST in the western equatorial Pacific; meanwhile, during the warm phases, the western Pacific warms and the eastern cools. These changes in the water temperature affect the atmospheric jet stream, moving it northward in the cold phases. The PDO phases fluctuate in a range of 20-30 years (extracted from Fagan, 2008). Annexes/Anexos 297 Breve descripción de los indices climáticos El Niño Los eventos de El Niño se caracterizan por un calentamiento significativo en la zona ecuatorial del agua superficial; son cambios en el patrón de corrientes marinas de la región intertropical. Durante un evento de El Niño, la presión atmosférica a nivel del mar tiende a ser mayor de lo normal en la parte oeste del Pacífico tropical y más baja en la vertiente este. Los vientos alisios del oeste se intensifican, provocando un desplazamiento del agua cálida hacia el este, a la costa Americana, al tiempo que la termoclina en el oeste se profundiza más de lo normal (Fagan, 2008). ENOS -El Niño/Oscilación SurOS -Oscilación SurEs una fluctuación irregular que causa un cambio en la presión superficial en el oeste y este del Pacífico tropical. Cuando la presión es mayor en la parte oeste, es bajo en el este y viceversa. En general, esto afecta simultáneamente a las temperatures del mar. Este tipo de cambio de regimen modifica el patron de lluvias y los vientos en ambas regions (Fagan, 2008). ENOS En 1960, el meteorólogo Jacob Bjerknes, encontró una relación entre la OS y los eventos de El Niño, que es lo que se conoce como El Niño/Oscilación Sur (ENOS), que es el término que más se usa en la literatura (Fagan, 2008). Eventos cálidos del ENOS son aquellos en los que coinciden valores negativos de la OS y un evento extreme de El Niño, esto produce un aumento de la temperatura de las capas superficiales del mar en el Pacífico tropical este y central, lo que lleva asociado un cambio en el patrón