Response of Acartia populations to environmental variability and effects of invasive congenerics in the estuary of Bilbao, Bay of Biscay

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Response of Acartia populations to environmental variability and effectsof invasive congenerics in the estuary of Bilbao, Bay of Biscay

Guillermo Aravena a,*, Fernando Villate a, Ibon Uriarte b, Arantza Iriarte a, Berta Ibanez c,d

a Laboratory of Ecology, Department of Plant Biology and Ecology, Faculty of Science and Technology, University of the Basque Country, PO Box 644, 48080 Bilbao, Spainb Laboratory of Ecology, Department of Plant Biology and Ecology, Faculty of Pharmacy, University of the Basque Country, Paseo de la Universidad 7, 01006 Gasteiz, Spainc Basque Foundation for Health Innovation and Research (BIOEF), 48015 Sondika, Bizkaia, Spaind CIBER Epidemiology and Public Health (CIBERESP), Spain

a r t i c l e i n f o

Article history:Received 6 March 2009Accepted 18 May 2009Available online 28 May 2009

Keywords:Acartia speciesenvironmental variabilityestuariesinvasive speciestransfer function models

a b s t r a c t

The effect of environmental factors (river discharge, water temperature and dissolved oxygen saturation)on the abundance and distribution of Acartia populations and the interactions between their congenericswas evaluated by means of transfer function (TF) models in the estuary of Bilbao during the period 1998–2005. The recorded species were Acartia clausi, Acartia tonsa, Acartia margalefi and Acartia discaudata.Acartia clausi dominated in the entire euhaline region of this estuary until 2003 when it was displacedfrom the inner part by A. tonsa. This invasive species (A. tonsa) was found for the first time in 2001 andcolonized successfully the inner (salinity 30) and intermediate (salinity 33) waters of this estuary since2003. The TF models revealed an immediate and negative effect of A. tonsa on A. clausi in the inter-mediate salinity (33) waters, where these species showed the highest spatial overlap. The results indicatethat environmental changes from 2003 influenced the abundance of Acartia species, being unfavourablefor A. clausi. The decrease of A. clausi abundance in low salinity waters was related to a significantdecrease of dissolved oxygen saturation levels, whereas the increase of temperature was linked toa significant increase of A. tonsa. Acartia margalefi and A. discaudata were scarce over the entire period,but they were found to be valuable indicators of hydrological changes, which were associated to climatefactors. These two latter species increased in abundance and expanded their seasonal distribution, and inthe case of A. margalefi also its spatial distribution, in 2002, coinciding with the period in the time-serieswhen autumn–winter rainfall and summer temperatures were lowest, and dissolved oxygen saturationlevels were highest.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

In the last decades it has been well reported that estuarineecosystems have been subject to the effects of long-term environ-mental changes and anthropogenic factors (Kennish, 2002; Leh-man, 2004; Lotze et al., 2006). The assessment of both long-termnatural and anthropogenic changes in these estuarine ecosystemsrequires monitoring programmes that can detect the cumulativequantifiable result of environmental changes, as an alteration in thestate of individuals, populations or communities (McLusky andElliott, 2004). Among estuarine plankton populations, Acartiaspecies can offer a valuable picture of the response to

environmental variability, because they dominate the zooplanktonof many temperate and subtropical estuaries, and coexistingspecies show spatial and seasonal segregation patterns associatedwith hydrological conditions, seasonal variability, trophic statusand pollution (e.g. Collins and Williams, 1981; Alcaraz, 1983; Sul-livan and McManus, 1986; Kimmerer, 1993; Lawrence et al., 2004;Uriarte and Villate, 2005). In addition, recent studies have revealedthat non-native species of Acartia are colonizing many coastal areasand estuaries by propagation or introduction (e.g. Comaschi et al.,2000; Seuront, 2005; David et al., 2007) due to the ability ofAcartiidae to cross geographic barriers, mainly as resting stages(Belmonte and Potenza, 2001). These non-native species aremodifying the status of native species which are subject tocompetitive pressure (Lakkis, 1994).

The aim of the present study was to identify the role of naturalenvironmental factors, anthropogenic effects and biological inter-actions among species in the changes of the Acartia assemblage in

* Corresponding author.E-mail addresses: [email protected] (G. Aravena), fernando.villate@

ehu.es (F. Villate), [email protected] (I. Uriarte), [email protected] (A. Iriarte),[email protected] (B. Ibanez).

Contents lists available at ScienceDirect

Estuarine, Coastal and Shelf Science

journal homepage: www.elsevier .com/locate/ecss

0272-7714/$ – see front matter � 2009 Elsevier Ltd. All rights reserved.doi:10.1016/j.ecss.2009.05.013

Estuarine, Coastal and Shelf Science 83 (2009) 621–628

Author's personal copy

the temperate and anthropogenically perturbed estuary of Bilbao(Bay of Biscay) over an 8-year period (1998–2005). This study ispart of a wider one dealing with long-term changes in thezooplankton community of the estuary of Bilbao. It is integrated ina monitoring programme of estuarine ecosystems of the Basquecoast, which provides information to detect changes in the struc-ture and function of these ecosystems.

We hypothesize that environmental conditions of a particularestuarine system determine the abundance and distribution ofcoexisting species and the success of arriving new species, but thatyear to year variability of climate dependent factors and anthro-pogenic factors may induce noticeable changes both in thetemporal dynamics and spatial segregation of species, through itseffect on interspecific competition. In particular, we analysed theeffect of river discharge and water temperature, as the main naturalenvironmental drivers controlling zooplankton dynamics, and theeffect of dissolved oxygen saturation, as a major factor associatedwith organic pollution in estuaries.

To assess the effect of these environmental driving forces on Acartiaspecies, and the biological interaction between native and invasivespecies, we have used transfer function (TF) models. These dynamicautoregressive models are being widely used in the last decades, notonly for forecasting purposes of water quality constituents (Bhanguand Whitfield, 1997; Worrall and Burt, 1999), but also to establishrelationships between climate variations and hydrological fluctuations(Nogueira et al., 1997; Niu et al., 1998; Aravena et al., 2009) and theirconsequent effects on both phytoplankton biomass (Lehman, 1992;Ferguson et al., 2007; Villate et al., 2008) and zooplankton assemblages(Hanninen et al., 2003; Vuorinen et al., 2003).

2. Materials and methods

2.1. Study area

Located in the inner Bay of Biscay, the estuary of Bilbao (43� 230

N 3� W) is today a physically modified system with two differentzones: a narrow artificial channel 15 km long and 2–9 m deep thatcrosses the urban-industrial area and an outer wide and deeper(10–25 m) embayment with large port developments (Fig. 1) Thissystem is classified as macrotidal and stratified, with a permanentsalt wedge and increasing salinity differences between the surfaceand bottom waters upstream (Valencia et al., 2004). Euhalinewaters (salinity� 30) constitute the main water body within theestuary, and they can reach the upper reaches below the haloclineunder average river flow conditions, whereas surface salinity is onaverage lower than 15 in the inner half of the estuary. The esti-mated time to infill the mean volume of the estuary with the meanriver flow is around 2 months (Valencia et al., 2004).

For many decades this estuary was highly polluted, but duringthe study period estuarine waters have undergone major changesdue to the recent development of urban wastewater treatmentplants (WWTP), and the decline of industrial activity in the lastdecades. Nevertheless, organic enrichment and eutrophic condi-tions are still evident in these estuarine waters. In the 1996–1998period, mean values of suspended particulate organic matterranged from around 2 mg l�1 in waters of salinity 35 to around11 mg l�1 in mesohaline waters, and a maximum of 23.1 mg l�1 wasrecorded (Cotano, pers. comm.). Regarding nutrients, ammoniaconcentrations higher than 400 mmol l�1 have been detected in theinner and middle zones (Valencia & Franco, 2004).

2.2. Environmental and biological data

Monthly environmental data and zooplankton samples werecollected at high tide in selected salinity zones of 30, 33 and �34

along the main estuary channel (Fig. 1) from 1998 to 2005.Samplings were performed during the last half of each monthduring neap tides. Vertical profiles of salinity (Practical SalinityScale is used), temperature and dissolved oxygen saturation wererecorded in situ (at one sampling site in each of the salinity zones of30 and 33 and at one or two sampling sites in the salinity zone of�34) using WTW Water Quality Meters. Zooplankton samples (oneper sampling site) were collected below the halocline (3–5 mdepth) by the horizontal towing of a 200 mm net (mouth diameter0.25 m) equipped with a Digital Flowmeter. Samples werepreserved with 4% formaldehyde buffered with methanol, andadult copepods of the genus Acartia were identified to species leveland counted under a binocular microscope.

Additionally, monthly mean data of river discharge for theNerbioi-Ibaizabal river were provided by the County Council ofBizkaia from the station of Abusu, located, nearby the estuary ofBilbao.

2.3. Statistical analyses

2.3.1. Pre-treatment of dataPrior to the statistical treatment, abundance data of Acartia

species were log (xþ 1) transformed to achieve homogeneity ofvariance and normality.

Missing values in the time-series were interpolated using theTramo-Seats package incorporated in the Demetra 2.0 interface.The method used for interpolation was the additive outlier (AO)approach with correction in the Determinantal Term of the Likeli-hood (see Gomez and Maravall, 1994). Given that the number ofmissing observations was not high (about 10%), this method gavesimilar results to the skipping approach, and the former was usedfollowing suggestions by Gomez and Maravall (1994) and Gomezet al. (1999).

Monthly values were used in the trend analysis and to visuallyinspect differences in the variability of environmental factors(water temperature and dissolved oxygen saturation) amongsalinity zones. To statistically check if these differences weresignificant, we applied the paired t-test to the monthly data cor-responding to winter (January, February and March) and summer(July, August and September) seasons. The t-test was performedusing the statistical package R 2.6.1 (library stats).

Quarterly mean (January–February–March, April–May–June,July–August–September and October–November–December)values were used to assess the relationships between variables.Although these quarterly data may produce a loss of power,because it shortens the original length of the series to one third, itallows to gain robustness in the TF and to reduce both the effect ofoutliers and the effect of interpolation to fill the gaps of missingvalues. Quarterly mean values have been previously used for TFmodelling by Hanninen et al. (2003), Vuorinen et al. (2003) andVillate et al. (2008) in similar environmental studies.

2.3.2. Trend analysisTo visually inspect the temporal variability in both environ-

mental and biological variables, we used the Locally WeightedRegression Scatter plot smoothing technique (LOWESS analysis,Cleveland, 1979) available in the R 2.6.1 package (library stats). TheLOWESS function fits a nonparametric regression by using a locallyweighted regression. Such regressions are useful for highlightinga trend in messy data or for data reduction to give some insight intoa large data set. It provides robust estimates of the trend andseasonal components that are not distorted by aberrant behaviourin the data. To choose the degree of smoothing, which is given bythe span, we followed the procedure given in Trexler and Travis(1993) who suggested to start with small span values and compare

G. Aravena et al. / Estuarine, Coastal and Shelf Science 83 (2009) 621–628622

Author's personal copy

the independence between the residuals of the LOWESS regressionand time (see also Zuur et al., 2007). A span of 0.5 was finally chosenfor the analysis, which indicates that the percentage of the dataused in the windows around each target value was 50%, leading toa trade-off between long-scale behaviour detection and relativelyshort-term changes detection.

2.3.3. Transfer function modelsTo infer causal relationships between environmental factors and

Acartia populations, and between Acartia congenerics, TF modelswere fitted using SAS 9.1 software, SAS Institute Inc. We used theBox and Jenkins methodology (Box and Jenkins, 1976) that is able toconnect one series not only with its own values but also with pastand present values of other related time-series. In particular, weused seasonal ARIMA models, the natural extension of the ARIMAmodels to cope with the seasonality that is present in the data. TFmodels relate two or more time-series using ARMA errors, and

based on the cross-correlation function on the pre-whitened series,are able to eliminate any false correlation between them. The stepof pre-whitening the series before testing for correlation is funda-mental to reveal the existence of underlying relationships apartfrom the seasonal behaviour of the data, which is already beingmodelled (see for example Box and Jenkins, 1976; Lehmann andRode, 2001; Stenseth et al., 2003).

A linear TF model can be written in the general form:

Yt ¼ C þ ðusðBÞ=drðBÞÞXt�b þ�

qqðBÞ=fpðBÞ�

at

where Yt is the output series (the dependent variable); Xt is theinput series (the independent variable); C is the constant term,us(B) and dr(B) are the polynomial delay functions; B represents thebackward shift operator and b is the delay time before Xt begins toinfluence Yt. The parameter fp is the autoregressive (AR) operator innon-seasonal series, and qq represents the moving average

Fig. 1. Map of the estuary of Bilbao showing the location of salinity zones (30, 33 and �34).

G. Aravena et al. / Estuarine, Coastal and Shelf Science 83 (2009) 621–628 623

Author's personal copy

operator, and similar operators can be included to account for theseasonal term. Here, at corresponds to the errors that are inde-pendently and identically distributed with normal distribution. Theseries of the models usually need to be appropriately transformedto achieve stationarity. For further detailed description of theterminology, see Box and Jenkins (1976). In order to take care of thenon-stationarity of the mean and variance, the response variable(output) and the explanatory variable (input) were appropriatelytransformed.

We have built the TF models in accordance to Box and Jenkins(1976) methodology, which is a three-step procedure. In the firststep, the cross-correlation function (CCF) is used to identify themodel. For the CCF to be meaningful, the input and response seriesmust be filtered with a pre-whitened model for the input series inorder to reduce the residuals to white noise. The filter is derivedfrom the univariate analysis of the input variable and is applied toboth the input and output series. In the second one, the estimationof the model is carried out using the maximum likelihood methodsand, finally, the diagnosis of the model is done by using standardtest statistics such as modified Ljung–Box test (Stoffer and Toloi,1992). For further detailed description of the methodology, see Boxand Jenkins (1976).

3. Results

3.1. Environmental variables

The seasonal pattern of water temperature was characterized bysummer maxima and winter minima, but inner waters (salinity 30)were warmer in summer and colder in winter than outer waters(salinity� 34) (Fig. 2). Maximum summer (July–August–September) averaged temperatures were 21.0 �C and 20.4 �C atsalinities of 30 and �34, respectively, and minimum winter(January–February–March) averaged values were 12.1 �C and12.6 �C. Results showed that differences in water temperaturebetween the 30 and the �34 salinities zones were significant bothin summer (t¼ 4.84, p< 0.001) and winter (t¼�3.07, p¼ 0.006). Inthe year to year variation of water temperature it is worth notingthe remarkably low temperatures registered in the anomalouslycold-summer of 2002.

Dissolved oxygen saturation (DOS) showed large seasonal andspatial variations (Fig. 2). The highest values of DOS registered insummer were 21.9% and 95.5% in average at salinities of 30 and�34, respectively, whereas the lowest values during winter werearound 50.3% and 93.1% at those salinities. These differences in DOSbetween the inner waters (salinity 30) and outer waters (sal-inity� 34) were significant both in summer (t¼�14.18, p< 0.001)and winter (t¼�6.84, p< 0.001). Whilst DOS values remainedquite constant over the entire period of study at salinities �34, inthe inner waters (salinity 30) DOS values showed a noticeableincrease until 2002 and a later decrease.

River discharge showed noticeable seasonal variations withmaxima in autumn–winter and minima in summer, except for theautumn–winter period of 2001–2002 when values were unusuallylow, and the summer 2002 when values were unusually high(Fig. 2).

3.2. Acartia populations

Four Acartia species, Acartia clausi, Acartia tonsa, Acartia dis-caudata and Acartia margalefi were collected in the estuary of Bilbao.Acartia tonsa was registered for the first time in 2001, although inlow densities. Acartia clausi was the dominant species in the entireeuhaline region of the estuary until the successful introduction of A.tonsa, which became the dominant species in the inner (salinity 30)

and intermediate (salinity 33) zones from 2003 and 2004 onwards,respectively (Fig. 3). Acartia discaudata and A. margalefi were mostlyfound in lower densities than the former species, and showednoticeably higher abundances in late 2001 and 2002 (Fig. 3).

Our results revealed a spatial segregation of Acartia species from2002 to 2005. Once Acartia tonsa appeared in 2002, Acartia clausirestricted its dominance to the outer part of the estuary (sal-inity� 34), whereas A. tonsa became dominant in the inner waters(salinity 30). Acartia discaudata and Acartia margalefi showed anintermediate position in the spatial gradient. However, the distribu-tion of these two last species was skewed toward high and lowsalinity, respectively. A seasonal segregation was also evident withpeaks of A. clausi occurring in late winter-spring and those of A. tonsain summer. Acartia discaudata and A. margalefi were most abundant inautumn–winter. However, in the 2001–2002 period A. discaudata andA. margalefi expanded their seasonal distribution and peaked also in

Fig. 2. Monthly time-series of water temperature and dissolved oxygen saturation(DOS (%)), and monthly mean values of river discharge in the estuary of Bilbao. Trends(thicker lines) have been drawn using a LOWESS trend analysis.

G. Aravena et al. / Estuarine, Coastal and Shelf Science 83 (2009) 621–628624

Author's personal copy

spring-summer, and A. margalefi moved inward in the estuaryshowing higher abundances at 30 salinity waters (Fig. 4).

3.3. Relationships between variables

The pattern of variation of river discharge did not showa correspondence with those of Acartia populations, whereasa positive and immediate effect of water temperature on Acartia

tonsa was found in the inner and intermediate salinity zones(salinity 30 and 33) (Table 1). In these salinity zones (30 and 33)Acartia clausi showed a positive and immediate (lag¼ 0) responseto DOS (Table 1). Our TF results also revealed an immediate andnegative effect of A. tonsa on A. clausi in the intermediate salinityzone (salinity 33), where these species showed an overlap justimmediately after the successful colonization of A. tonsa in thisestuary (Fig. 3).

Fig. 3. Monthly time series of Acartia populations density (log xþ 1, ind. m�3) at the salinity sites of 30, 33 and �34. Trends (thicker lines) have been drawn using a LOWESS trendanalysis.

Fig. 4. Seasonal (left) and spatial (right) distribution of Acartia discaudata and Acartia margalefi: averaged values for the 2001–2002 period (dashed lines and white bars), andaveraged values for the rest of the years (without the 2001–2002 period) in the series (continuous lines and black bars).

G. Aravena et al. / Estuarine, Coastal and Shelf Science 83 (2009) 621–628 625

Author's personal copy

4. Discussion

Our results revealed significant relationships between envi-ronmental factors and Acartia populations as well as competitiveinteractions between Acartia clausi and Acartia tonsa. Noticeablechanges in A. clausi abundance and distribution were found in theestuary of Bilbao as a result of the presence and successful colo-nization of A. tonsa since 2003. In the absence of A. tonsa, A. clausihad a broad distribution throughout the entire euhaline region(salinity� 30); however, the introduction of A. tonsa caused thedisplacement of A. clausi, restricting its spatial distribution andsalinity niche toward the outer part of this estuary. This does notonly reflect a preference for some hydrographical conditions, but itis also the result of interspecific competition between the twospecies. The competitive interaction between A. tonsa and A. clausiwas supported by the TF models, which revealed an immediatenegative effect of A. tonsa on A. clausi in intermediate salinity (33)waters, where these species showed the highest overlap in theirabundances. Acartia tonsa is considered as an invasive species, andits introduction, successful colonization and negative effects onendemic species have been reported in other confined ecosystems(Bollens et al., 2002; Sei et al., 2006; David et al., 2007).

In estuaries with a wide range of brackish habitats, and welldeveloped populations of brackish Acartia species, salinity seems tobe the main factor accounting for the spatial segregation of Acartia

tonsa and Acartia clausi along axial gradients, and native brackishspecies are the worst affected by the establishment of A. tonsa. Inthe Gironde estuary, this species was found to colonize successfullythe oligo-mesohaline zone supplanting the congeneric brackishspecies Acartia bifilosa (David et al., 2007). This was not the case inthe estuary of Bilbao, where oligo-mesohaline waters do notconstitute a stable brackish habitat and native brackish species areabsent. This must have eased the colonization of the inner watersby A. tonsa. Cervetto et al. (1999) found the optimal conditions for A.tonsa between salinities of 15 and 22. Experimental studies suggestthat the negative effect of salinity decreases on the energy balanceand reproductive success of A. tonsa and A. clausi operates atsalinities of <10 and <25, respectively, whereas no negative effectswere observed for any of them at higher salinities (Calliari et al.,2006). Therefore, other environmental factors must have playeda more decisive role in the replacement of A. clausi by A. tonsa in theinner zone of the estuary of Bilbao. Previous studies reported thatthe low abundance and hatching success of A. clausi in the innerwaters of this estuary were related to the diminished oxygenconditions (Uriarte and Villate, 2005; Uriarte et al., 2005). Incontrast, A. tonsa may survive at low oxygen concentrations (DeMeester and Vyverman, 1997), and it is reported as a commonspecies inhabiting the inner hypoxic zones of estuaries (Marcuset al., 2004). TF results from the present work agreed with theseobservations and evidenced the importance of the effect of DOS onA. clausi dynamics in the inner and intermediate salinity zones (30and 33). In the inner waters DOS was found to be negatively relatedto water temperature (unpublished data). In these waters highstratification and residence time (Valencia et al., 2004) are the mainfactors limiting re-aeration during the summer, which togetherwith the temperature-enhanced demand of dissolved oxygen,cause severe hypoxia. It is well archived that the increase in watertemperature is one the main factors that exacerbate the oxygendemand during summer seasons in coastal waters (Rabalais et al.,2009) and estuarine ecosystems (Preston, 2004; Paerl, 2006). Suchsummer conditions could favour the presence and establishment ofA. tonsa in these inner waters. In fact, the abundance of this specieswas high in summer and was positively related to water temper-ature in the inner waters. An increase in water temperature hasbeen suggested as a key factor to explain the introduction of A.tonsa in other estuarine ecosystems (Kimmel and Roman, 2004;David et al., 2007). Results from the present work corroboratefindings that claim A. tonsa is an opportunistic tolerant-species,which can take advantage of eutrophic ecosystems (Brylinski, 1981;Bianchi et al., 2003; Dunbar and Webber, 2003), and of thermaladditions (David et al., 2007). Shipping activities are significant inthe port of Bilbao and the introduction of A. tonsa may have takenplace by transport through ship ballast water. It is well reportedthat in many areas the introduction of A. tonsa has been related totransport through ballast water and its widespread distribution isfrequently associated to areas where human activities are devel-oped (Ruiz et al., 1997; Cervetto et al., 1999; David et al., 2007).

The freshwater runoff had no effect on the variability of Acartiaspecies in the estuary of Bilbao. This is likely due to the fact that thedevelopment of Acartia species is restricted to the euhaline watersin this estuary. In Chesapeake Bay, linear mixed-effects modelsshowed a negative correlation between freshwater inputs andA. tonsa abundance in the oligohaline region, but not in highersalinity regions (Kimmel and Roman, 2004).

Acartia discaudata and Acartia margalefi were absent from theestuary of Bilbao in the early 1980s, when water pollution pre-vented the penetration of copepod species inside the estuary andonly the neritic species Acartia clausi was present in the outerembayment (Villate, 1991). But both A. discaudata and A. margalefiwere already present in very low numbers inside the estuary in

Table 1Transfer function modelling results of the effects of environmental variables onAcartia populations and biological interactions between congenerics in the estuaryof Bilbao. Here, W0 is the estimate of each regressor variable effect at lag (l), B4 is thestational shift operator, m is the interceptor term, and f and q correspond to ARMAparameters. All variables have been differenced with d¼ 4. Data used in the TFmodelling are quarterly means for the period 1998–2005.

G. Aravena et al. / Estuarine, Coastal and Shelf Science 83 (2009) 621–628626

Author's personal copy

1997 when zooplankton samplings restarted (Uriarte and Villate,2005). In contrast with the successful colonization of Acartia tonsa,A. discaudata and A. margalefi have remained as residual species,seasonally restricted to the cold period. Complex interactions ofclimate features with particular characteristics of each estuary maydetermine the suitability for each Acartia species. In SouthamptonWater, for example, A. discaudata and A. margalefi coexist togetherwith A. clausi and A. tonsa, but A. discaudata and A. margalefi areperennial and abundant, whereas A. clausi and A. tonsa show morerestricted spatial and seasonal distributions (Muxagata, pers.comm.). The abundance and seasonal distribution reported for A.discaudata and A. margalefi in the Rıa of Vigo (Alcaraz, 1983) are alsomuch higher than those observed in the estuary of Bilbao. It mustbe pointed out that water temperature in both Southampton Waterand the Ria of Vigo was lower than in the estuary of Bilbao. Seasonalranges of around 5–17 �C and 11–19 �C were reported for South-ampton Water and the Ria of Vigo, respectively, in the above-mentioned studies. This suggests that the temperature regime maybe a major factor determining the abundance and seasonal distri-bution of these congeneric species in a particular system. In view ofthese considerations, we hypothesize that in the estuary of Bilbaotemperature and oxygen changes in recent years have beenfavourable for those species tolerant of hypoxic conditions andwarm-water affinity such as A. tonsa, but less suitable for A. mar-galefi and A. discaudata.

The absence of these scarce species (Acartia margalefi andAcartia discaudata), i.e. abundance values of 0, during many monthshas probably prevented us from finding significant relationshipswith environmental factors when using TF models. However, thenoticeable increase of A. discaudata and A. margalefi in 2002,together with the inward advance of the latter species suggest thatthese two species took advantage of the colder and betteroxygenated conditions registered during 2002. No inward advancewas observed for A. discaudata. This is likely due to salinity pref-erences because, it has been observed in other estuaries that whenthey coexist, A. discaudata occupies an intermediate positionbetween A. margalefi and Acartia clausi along the estuarine gradient(Alcaraz, 1983; Muxagata, pers. comm.). In addition, the colderconditions of the summer 2002 are the most plausible explanationfor the wider seasonal distribution of A. margalefi and A. discaudatain this year.

These results indicate that, in spite of its low quantitative rele-vance in the system, Acartia margalefi may be a valuable indicator ofwater quality changes related to organic pollution, and both A.margalefi and Acartia discaudata appear to be sensitive to changesin the seasonal pattern of hydrological factors within the estuary ofBilbao.

The year 2002 was climatologically peculiar in comparison withprevious and later years, because apart from differences in airtemperature that affected water temperature, the seasonal patternof rainfall also differed influencing runoff, and all this coincidedwith the highest positive values of North Atlantic Oscillationindices during the 1997–2005 period (Aravena et al., 2009). Thissuggests that, finally, the development of Acartia margalefi andAcartia discaudata in the estuary of Bilbao would be linked to thisclimatic phenomenon that operates in the North Atlantic Region,and responds suddenly to short-term climatic events. In fact, thepopulations of A. margalefi seem characterized by extreme fluctu-ations in abundance, either seasonal or inter-annual (Belmonte andMazzocchi, 1997).

In the same way, the mid- and long-term changes of the Acartiacongeneric might reflect the effects of climate change in the estuaryof Bilbao. Evidences of changes in the marine copepod community,with an increase of warm-water species, and a decrease of colder-water species, related to a general warming trend during the last

decades, have been reported in adjacent shelf waters of the Basquecoast and other nearby shelf areas of the southern Bay of Biscay(Villate et al., 1997; Valdes et al., 2007), as well as in wider areas ofthe North Atlantic region (e.g. Beaugrand et al., 2002; Lindley andDaykin, 2005).

Acknowledgements

This research was funded by the United Nations Educational,Scientific and Cultural Organization (UNESCO) Chair on ‘‘Sustain-able Development and Environmental Education’’ (UNESCO03/04),the University of the Basque Country (EHU06/52) and the BasqueGovernment (ETORTEK07/25). G.A. was supported by a grant fromthe PhD Scholarship Program of the University of the BasqueCountry. We thank the County Council of Bizkaia for providinghydro-meteorological data.

References

Alcaraz, M., 1983. Coexistence and segregation of congeneric pelagic copepods:spatial distribution of the Acartia complex in the Ria of Vigo (NW of Spain).Journal of Plankton Research 5, 891–900.

Aravena, G., Villate, F., Iriarte, A., Uriarte, I., Ibanez, B., 2009. Influence of the NorthAtlantic Oscillation (NAO) on climatic factors and estuarine water temperatureon the Basque coast (Bay of Biscay): comparative analysis of three seasonal NAOindices. Continental Shelf Research 29, 750–758.

Belmonte, G., Mazzocchi, M.G., 1997. Records of Acartia (Acartiura) margalefi(Copepoda, Calanoida, Acartiidae) from the Norwegian and Black Seas. Crus-taceana 70, 252–256.

Belmonte, G., Potenza, D., 2001. Biogeography of the family Acartiidae (Calanoida)in the Ponto-Mediterranean Province. Hydrobiologia 453/454, 171–176.

Bhangu, I., Whitfield, P.H., 1997. Seasonal and long-term variations in water qualityof the Skeena river at Usk, British Columbia. Water Research 31, 2187–2194.

Beaugrand, G., Reid, P.C., Ibanez, F., Lindley, J.A., Edwards, M., 2002. Reorganisation ofNorth Atlantic marine copepod biodiversity and climate. Science 296,1692–1694.

Bianchi, F., Acri, F., Bernardi Aubry, F., Berton, A., Boldrin, A., Camatti, E., Cassin, D.,Comaschi, A., 2003. Can plankton communities be considered as bio-indicatorsof water quality in the Lagoon of Venice? Marine Pollution Bulletin 46, 964–971.

Bollens, S.M., Cordell, J.R., Avent, S., Hooff, R., 2002. Zooplankton invasions: a briefreview, plus two case studies from the northeast Pacific Ocean. Hydrobiologia480, 87–110.

Box, G., Jenkins, G., 1976. Time Series Analysis: Forecasting and Control. Holden-Day, USA, 531 pp.

Brylinski, J.M., 1981. Report on the presence of Acartia tonsa Dana (Copepoda) in theharbour of Dunkirk (France) and its geographical distribution in Europe. Journalof Plankton Research 3, 255–260.

Calliari, D., Andersen, C.M., Thor, P., Gorokhova, E., Tiselius, P., 2006. Salinitymodulates the energy balance and reproductive success of co-occurring cope-pods Acartia tonsa and A. clausi in different ways. Marine Ecology ProgressSeries 312, 177–188.

Cervetto, G., Gaudy, R., Pagano, M., 1999. Influence of salinity on the distribution ofAcartia tonsa (Copepoda, Calanoida). Journal of Experimental Marine Biologyand Ecology 239, 33–45.

Cleveland, W.S., 1979. Robust locally weighted regression and smoothing scatter-plots. Journal of American Statistical Association 74, 829–836.

Collins, N.R., Williams, R., 1981. Zooplankton of the Bristol channel and severnestuary. The distribution of four copepods in relation to salinity. Marine Biology64, 273–283.

Comaschi, A., Acri, F., Bianchi, F., Bressan, M., Camatti, E., 2000. Temporal changes ofspecies belonging to Acartia genus (Copepoda: Calanoida) in the Northern Basinof the Venice Lagoon. Bollettino del Museo Civico di Storia Naturale Venezia 50,189–193.

David, V., Sautour, B., Chardy, P., 2007. Successful colonization of the calanoidcopepod Acartia tonsa in the oligo-mesohaline area of the Gironde estuary (SWFrance) – natural or anthropogenic forcing? Estuarine, Coastal Shelf Science 71,429–442.

De Meester, L., Vyverman, W., 1997. Diurnal residence of the larger stages of thecalanoid copepod Acartia tonsa in the anoxic monimolimnion of a tropicalmeromictic lake in New Guinea. Journal of Plankton Research 19, 425–434.

Dunbar, F.N., Webber, M.K., 2003. Zooplankton distribution in the eutrophic King-ston Harbour, Jamaica. Bulletin of Marine Science 73, 234–260.

Ferguson, C.A., Bowman, A.W., Scott, E.M., Carvalho, L., 2007. Model comparison fora complex ecological system. Journal of the Royal Statistical Society: Series A170, 691–711.

Gomez, V., Maravall, A., 1994. Estimation, Prediction, and Interpolation forNonstationary Series with the Kalman Filter. Journal of the American StatisticalAssociation 89, 611–624.

G. Aravena et al. / Estuarine, Coastal and Shelf Science 83 (2009) 621–628 627

Author's personal copy

Gomez, V., Maravall, A., Pena, D., 1999. Missing observations in ARIMA models:skipping approach versus additive outlier approach. Journal of Econometrics 88,341–363.

Hanninen, J., Vuorinen, I., Kornilovs, G., 2003. Atlantic climatic factors control decadaldynamics of a Baltic Sea copepod Temora longicornis. Ecography 26, 672–678.

Kennish, M.J., 2002. Environmental threats and environmental future of estuaries.Environmental Conservation 29, 78–107.

Kimmel, D.G., Roman, M.R., 2004. Long-term trends in mesozooplankton abun-dance in Chesapeake Bay, USA: influence of freshwater input. Marine EcologyProgress Series 267, 71–83.

Kimmerer, W.J., 1993. Distribution patterns of zooplankton in Tomales Bay, Cal-ifornia. Estuaries 16, 264–272.

Lakkis, S., 1994. Coexistence and competition within Acartia (Copepoda, Calanoida)congeners from Lebanese coastal water: niche overlap measurements. Hydro-biologia 292/293, 481–490.

Lawrence, D., Valiela, I., Tomasky, G., 2004. Estuarine calanoid copepod abundancein relation to season, salinity, and land-derived nitrogen loading, Waquoit Bay,MA. Estuarine, Coastal and Shelf Science 61, 547–557.

Lehman, P.W., 1992. Environmental factors associated with long-term changes inchlorophyll concentration in the Sacramento-San Joaquin delta and Suisun bay,California. Estuaries 15, 335–348.

Lehman, P.W., 2004. The influence of climate on mechanistic pathways that affectlower food web production in northern San Francisco Bay estuary. Estuaries 27,311–324.

Lehmann, A., Rode, M., 2001. Long-term behaviour and cross-correlation waterquality analysis of the river Elbe, Germany. Water Research 35, 2153–2160.

Lindley, J.A., Daykin, S., 2005. Variations in the distributions of Centropages chier-chiae and Temora stylifera (Copepoda: Calanoida) in the north-eastern AtlanticOcean and western European shelf waters. ICES Journal of Marine Science 62,869–877.

Lotze,H.K., Lenihan, H.S., Bourque, B.J.,Bradbury, R.H., Cooke, R.G.,Kay, M.C.,Kidwell, S.M.,Kirby, M.X., Peterson, C.H., Jackson, J.B.C., 2006. Depletion, degradation, and recoverypotential of estuaries and coastal seas. Science 312, 1806–1809.

Marcus, N.H., Richmond, C., Sedlacek, C., Miller, G.A., Oppert, C., 2004. Impact ofhypoxia on the survival, egg production and population dynamics of Acartiatonsa Dana. Journal of Experimental Marine Biology and Ecology 301, 111–128.

McLusky, D.S., Elliott, M., 2004. The Estuarine Ecosystem. Oxford University Press,Oxford, 214 pp.

Niu, X.F., Edmiston, H.L., Bailey, G.O., 1998. Time series models for salinity and otherenvironmental factors in the Apalachicola estuarine system. Estuarine, Coastaland Shelf Science 46, 549–563.

Nogueira, E., Perez, F.F., Rios, A.F., 1997. Modelling thermohaline properties inan estuarine upwelling ecosystem (Ria de Vigo: NW Spain) using Box-Jenkins transfer function models. Estuarine, Coastal and Shelf Science 44,685–702.

Paerl, H.W., 2006. Assessing and managing nutrient-enhanced eutrophication inestuarine and coastal waters: interactive effects of human and climaticperturbations. Ecological Engineering 26, 40–54.

Preston, B.L., 2004. Observed winter warming of the Chesapeake Bay Estuary(1949–2002): implications for ecosystem management. Journal EnvironmentalManagement 34, 125–139.

Rabalais, N.N., Turner, R.E., Dıaz, R.J., Justic, D., 2009. Global change and eutrophi-cation of coastal waters. ICES Journal of Marine Science. doi:10.1093/icesjms/fsp047.

Ruiz, G.M., Carlton, J.T., Grosholz, E.D., Hines, A.H., 1997. Global invasions of marineand estuarine habitats by non-indigenous species: mechanisms, extent, andconsequences. American Zoology 37, 621–632.

Sei, S., Invidia, M., Gorbi, G., 2006. Near anoxia and sulfide as possible factorsinfluencing the spatial distribution of Acartia tonsa and Acartia clausi:comparative evaluation of egg tolerance. Journal of Experimental MarineBiology and Ecology 337, 121–130.

Seuront, L., 2005. First record of the calanoid copepod Acartia omorii (Copepoda:Calanoida: Acartiidae) in the southern bight of the North Sea. Journal ofPlankton Research 27, 1301–1306.

Stenseth, N.C., Ottersen, G., Hurrell, J.W., Mysterud, A., Lima, M., Chan, K.S.,Yoccoz, N.G., Adlandsvik, B., 2003. Studying climate effects on ecology throughthe use of climate indices: the North Atlantic Oscillation, El Nino SouthernOscillation and beyond. Proceedings of the Royal Society B 270, 2087–2096.

Stoffer, D.S., Toloi, C.M.C., 1992. A note on the Ljung–Box–Pierce Portmanteaustatistic with missing data. Statistics and Probability Letters 13, 391–396.

Sullivan, B.K., McManus, L.T., 1986. Factors controlling seasonal succession ofcopepods Acartia hudsonica and A. tonsa in Narragansett Bay, Rhode Island:temperature and resting egg production. Marine Ecology Progress Series 28,121–128.

Trexler, J.C., Travis, J.,1993. Nontraditional regression analyses. Ecology 74,1629–1637.Uriarte, I., Villate, F., 2005. Differences in the abundance and distribution of cope-

pods in two estuaries of the Basque coast (Bay of Biscay) in relation to pollution.Journal of Plankton Research 27, 863–874.

Uriarte, I., Cotano, U., Villate, F., 2005. Effects of estuarine conditions and organicenrichment on the fecundity and hatching success of Acartia clausi in contrastingsystems. Journal of Experimental Marine Biology and Ecology 320, 105–122.

Valdes, L., Lopez-Urrutia, A., Cabal, J., Alvarez-Ossorio, M., Bode, A., Miranda, A.,Cabanas, M., Huskin, I., Anadon, R., Alvarez-Marques, F., Llope, M., Rodrıguez, N.,2007. A decade of sampling in the Bay of Biscay: what are the zooplankton timeseries telling us? Progress in Oceanography 74, 98–114.

Valencia, V., Franco, J., 2004. Main characteristics of the water masses. In: Borja, A.,Collins, M. (Eds.), Oceanography and Marine Environment of the BasqueCountry. Elsevier Oceanography Series, vol. 70. Elsevier, Amsterdam, p. 616.

Valencia, V., Franco, J., Borja, A., Fontan, A., 2004. Hydrography of the southeasternBay of Biscay. In: Borja, A., Collins, M. (Eds.), Oceanography and Marine Envi-ronment of the Basque Country. Elsevier Oceanography Series, vol. 70. Elsevier,Amsterdam, p. 616.

Villate, F., 1991. Annual cycle of zooplankton community in the Abra harbour (Bay ofBiscay): abundance, composition and size spectra. Journal of Plankton Research13, 691–706.

Villate, F., Moral, M., Valencia, V., 1997. Mesozooplankton community indicatesclimate changes in a shelf area of the inner Bay of Biscay throughout 1988 to1990. Journal of Plankton Research 19, 1617–1636.

Villate, F., Aravena, G., Iriarte, A., Uriarte, I., 2008. Axial variability in the relationshipof chlorophyll a with climatic factors and the North Atlantic Oscillation ina Basque coast estuary, Bay of Biscay (1997–2006). Journal of Plankton Research30, 1041–1049.

Vuorinen, I., Hanninen, J., Kornilovs, G., 2003. Transfer-function modelling betweenenvironmental variation and mesozooplankton in the Baltic Sea. Progress inOceanography 59, 339–356.

Worrall, F., Burt, T.P., 1999. A univariate model of riverwater-nitrate time series.Journal of Hydrology 214, 74–90.

Zuur, A., Ieno, E., Smith, G., 2007. Analysing Ecological Data. Springer, 680 pp.

G. Aravena et al. / Estuarine, Coastal and Shelf Science 83 (2009) 621–628628