The sampled area includes the northeastern sector of the Gulf of Cádiz, at the northeastern Atlantic Ocean (Figure 1). The Gulf of Cádiz is a highly productive and heterogeneous region of the Iberian Atlantic shelf, shaped by the interaction of multiple physical and biogeochemical processes (Navarro and Ruiz, 2006; Ruiz and García-Lafuente, 2006; Prieto et al., 2009; Krug et al., 2017). Shelf productivity is strongly influenced by river inputs, which modulate fertility and hydrographic conditions along the eastern margin (Navarro et al., 2012; Caballero et al., 2014). Additional fertilisation mechanisms arise from intense tidal mixing near the Strait of Gibraltar, particularly associated with the interaction between strong tidal currents and submarine features such as the Trafalgar sill (Vargas-Yáñez et al., 2002; Sala et al., 2018, 2023). Wind-driven circulation, including the Huelva upwelling jet, frontal structures, and episodic eastward advection under Levante conditions, further contributes to the marked spatial heterogeneity of the shelf (García-Lafuente et al., 2006; Criado-Aldeanueva et al., 2006; Peliz et al., 2007, 2014). Together, these processes generate a dynamic hydrodynamic environment and strong biogeochemical gradients across the study area.

Monthly cruises were conducted on board the research vessel Regina Maris between March 2002 and August 2007, making a total of 59 surveys. Two distinct cruise sets of 30 stations each were conducted during the study period, with most stations overlapping between sets. A first set of cruises included bathymetries from 5 to 150 meters between the Guadiana and Guadalquivir River mouths, during the period from March 2002 to September 2004. In the second set (May 2005 to August 2007), the sampling area was extended to deeper waters (ca. 200 m) and southward to Cape Trafalgar.

Mesozooplankton samples were collected at each sampling station using a Bongo net with a 40-cm mouth diameter and 200-μm mesh size. All tows were conducted at a vessel speed of 2–2.5 knots and to a maximum altitude above the bottom of approximately 5m, and no more than 100 meters of depth at the stations with deeper bathymetry. Two independent flowmeters ‘‘General Oceanics 2030’’ fixed on the mouth of each net were used to measure the volume of water filtered. After collection, plankton samples were fixed in a formalin solution containing about 4% borax-buffered formaldehyde.

Mesozooplankton was quantified as displacement volume following standard procedures (Harris et al., 2010) and expressed as milliliters per 100 m³ of filtered water (mL 100 m^-3). No additional size-based or taxonomic separation was performed, and larger organisms were not selectively excluded during sample processing. Accordingly, in this study mesozooplankton refers operationally to the fraction of organisms retained by the 200-µm mesh, i.e. organisms larger than 200 µm.

In parallel, vertical CTD casts were also conducted at each sampling station, and surface salinity (5 m depth) was used in the subsequent analyses. Additionally, several environmental variables were measured in situ. Surface water samples (<5 m) were collected for the analysis of total Chlorophyll-a (Chla) and dissolved inorganic nutrients. Chla concentrations were determined fluorometrically after filtration through Whatman GF/F glass fiber filters (0.7 µm pore size) and extraction in 90% acetone, following the method of (Yentsch and Menzel, 1963). Anacystis nidulans (Sigma-Aldrich) was used as a calibration standard for the fluorometer (Turner Designs Model-10). Nitrate concentrations was determined using a TRAAC 800 autoanalyser following the method described by (Grasshoff et al., 1983).

The complete dataset of mesozooplankton, chlorophyll-a, salinity, and nitrate data used in this study is publicly available through the Digital CSIC repository (doi.org/10.20350/digitalCSIC/15441).

Hourly data on wind vectors, air temperature, and precipitation were obtained from a meteorological station (station code 5973) located in Cádiz and operated by the Spanish Meteorological Agency (“Agencia Estatal de Meteorología” - AEMET). Monthly mean global irradiance (kWh/m²/day) in Cádiz, averaged over the 1983–2005 period, was obtained from Sancho Ávila et al. (2012). Daily mean discharge at the Alcalá del Río dam was obtained from the Automatic Hydrological Information System (SAIH) of the Guadalquivir River Basin Authority (“Confederación Hidrográfica del Guadalquivir”, available at: https://www.chguadalquivir.es/saih/). This dam, located approximately 108 km upstream from the river mouth, constitutes the final major hydrological control point before the estuarine stretch of the lower Guadalquivir River, and is commonly used as a reference for discharge into the Gulf of Cádiz (Ruiz et al., 2015).

Satellite-derived Level 4 (L4) sea surface temperature (SST) data were obtained from the CMS European North West Shelf/Iberia Biscay Irish Seas (https://doi.org/10.48670/moi-00153; last accessed: May 2025), which merges data from different European Space Agency (ESA) satellites to provide daily, gap-free means at 20 cm depth, with a horizontal resolution of 0.05° × 0.05°. These satellite observations include thermal infra-red measurements from Advanced Very High-Resolution Radiometers (AVHRRs), Along-Track Scanning Radiometers (ATSRs), the Sea and Land Surface Temperature Radiometer (SLSTR) on-board Sentinel-3, as well as microwave observations (Embury et al., 2024). Daily data were then averaged to monthly values.

Net primary production (NPP) values were derived from the Atlantic Ocean Color (Copernicus-GlobColour; https://doi.org/10.48670/moi-00289; last accessed: May 2025) multi-satellite product, which similarly merges data from various ocean color sensors (SeaWiFS, MODIS, MERIS, VIIRS-SNPP & JPSS1, OLCI-S3A & S3B) to produce daily to monthly cloud-free maps (L4) at a spatial resolution of 1 km. Copernicus-GlobColour products have been extensively validated and have exhibited improved performance in coastal waters compared to other processors, such as OC-CCI/C3S (Garnesson et al., 2019).

Inorganic suspended particulate matter (ISPM) data were obtained from the Global Ocean Color (Copernicus-GlobColour; https://doi.org/10.48670/moi-00281; last accessed: May 2025) L4 product, providing daily to monthly gap-free composites at a spatial resolution of 4 km. The decision to use the global product was simply due to the fact that SPM is not available in the previously described regional Atlantic Ocean Color product. It is important to note that, although both products are processed using the same GlobColour framework and satellite sensors, the regional dataset provides higher spatial resolution (1 km) and is specifically optimized for complex coastal optical environments. In contrast, the global product emphasizes broader spatial consistency and coverage, which may result in some smoothing of fine-scale coastal features (Garnesson et al., 2019).

Stress-equivalent wind eastward component at 10 m above the sea surface (referred to throughout the manuscript as the zonal wind component) was obtained from the Global Ocean Monthly Mean Sea Surface Wind and Stress from Scatterometer and Model dataset (https://doi.org/10.48670/moi-00181; last accessed: May 2025). This product provides surface wind and stress fields at a horizontal resolution of 0.25° × 0.25°, based on the European Centre for Medium-Range Weather Forecasts (ECMWF) ERA5 reanalysis and corrected using L3 satellite scatterometer observations (Metop-A/B/C ASCAT, QuikSCAT SeaWinds, and ERS-1 and ERS-2 SCAT).

Figure 2 presents the time series of mesozooplankton biovolume and Chla concentration across the sampling stations over the study period. These series indicate that both the temporal fluctuations and station-specific mean values of mesozooplankton biovolume and Chla exhibit marked spatial variability, reflecting pronounced spatiotemporal heterogeneity in phytoplankton and mesozooplankton dynamics across the continental shelf.

A consistent spatial gradient is evident, with elevated mesozooplankton biovolume in coastal regions that progressively declines offshore. Mean mesozooplankton biovolume across the shelf ranges from values around 90 mL·100 m^-3 at the outer shelf to maxima exceeding 360 mL·100 m^-3 in nearshore areas (Figure 2A). The most prominent mesozooplankton hotspots are located in two main coastal domains. The first corresponds to the area influenced by the Guadalquivir River plume, where stations 6, 7, and 8 show consistently high mean biovolume values, over 300 mL·100 m^-3. The second hotspot is associated with the shallow banks in the vicinity of Cape Trafalgar, particularly south of the cape (station 51), where mean mesozooplankton biovolume reaches values above 360 mL·100 m^-3. The northeastern coastal band of the shelf exhibits the highest Chla concentrations across the study area (Figure 2B), with mean values ranging from ~1.8 to more than 3.5 mg·m^-3. However, this elevated Chla signal is not always matched by a proportional increase in mesozooplankton biomass. A clear example is found at stations 4 and 5, where average Chla concentrations range between 3 and 4 mg·m^-3, while mesozooplankton biovolume remains below 190 mL·100 m^-3. In fact, biovolume levels at these stations, located at the river mouth, are comparable to those recorded at offshore stations situated between the 80 and 200m isobaths, where Chla concentrations are four to seven times lower (stations 1, 11, 20, 46, and 53; Figure 2A).

The temporal component of mesozooplankton biovolume exhibits a clear seasonal pattern, with minimum values during winter (December–January) and maxima generally occurring from spring to summer (e.g. stations 3, 6, 7, 8, 28, 29, 44, and 45). In contrast, this seasonal signal is less evident at other locations (e.g. stations 4, 5, 11, 17, and 18). The large volume of data and the spatial heterogeneity of temporal responses (Figure 2) hinder the identification of consistent patterns at interannual or longer temporal scales. In complex datasets such as this, characterized by spatially structured variability and data gaps, the application of statistical approaches that reduce dimensionality and identify dominant trends is essential. In the following section, we present the results of the DFA.

DFA was employed to perform dimensionality reduction, synthesizing the entire dataset into a limited number of dynamic factors or latent trends. These trends capture the principal underlying temporal patterns shared among the time series. Specifically, four dominant trends were identified, each representing distinct components of temporal variability and enabling the classification of stations into groups with similar dynamics, thereby facilitating the spatial and temporal interpretation of the system.

Model selection was guided by the Akaike Information Criterion (AIC) and the proportion of temporal variance explained by the DFA model (Equation 1) as a function of the number of included trends. As shown in Figure 3, the model with two trends yielded the lowest AIC (3275.86), but the four-trend model, despite a slightly higher AIC (3281.22), accounted for a greater proportion of temporal variance (67.3% vs. 54.2%). Additionally, the model including the third and fourth factors captured spatial patterns that correspond to well-known oceanographic structures in the Gulf of Cádiz, as discussed in the following sections. Therefore, we selected the four-trend solution as the best compromise between model parsimony and explanatory capacity.

Among the four identified factors, Factor 1 accounted for the highest relative contribution to the model, based on the sum of squared loadings (57.1%), followed by Factors 2, 3, and 4 which explained 11%, 15.3%, and 16.6%, respectively.

All stations exhibited positive loadings for Factor 1 (Figure 4A), indicating a coherent response of the mesozooplankton across the entire shelf. However, loading magnitudes varied spatially, with lower values in certain areas where the influence of this trend appeared reduced. Such areas included the Guadalquivir River mouth, the vicinity of Cape Trafalgar, and the northern sector between the mouths of the Guadiana and Tinto-Odiel rivers, all of which showed loading values below 0.51 (Figure 4A, lighter colors).

Figure 5 illustrates the dominant periodicities associated with each dynamic factor, highlighting the most relevant temporal cycles (in months) that characterize their temporal signals. Peaks of the power spectrum (solid black line) that rise above the shaded bootstrapped confidence intervals (BCI) areas are significant at their corresponding confidence level. This indicates that the power of the major dynamic factors at these periods is greater than would be expected if the corresponding temporal signal fluctuated randomly (white noise). The periodogram for Factor 1 (Figure 5A), shows a dominant spectral peak at approximately 12 months, confirming the seasonal character of the signal.

In contrast to Factor 1, which elicits a platform-wide response of consistent sign in the spatial distribution of loadings (Figure 4A), the remaining three factors delineate spatial domains of opposing signs in their loadings (Figures 4B–D). With the exception of stations 10, 57 and 85 in Factor 3 (Figure 4C), the spatial divisions derived from the factor loadings produce well-structured regional patterns, as most stations with similar loadings cluster in contiguous areas, reflecting a clear geographic coherence in the factor values.

Factor 2 distinguishes two sharply contrasting zones, both with high loading magnitudes but opposite signs. A coastal strip, extending from the Tinto-Odiel estuaries to the northern Cape Trafalgar region, exhibits positive loadings (hereafter Inner-Shelf region; Figure 4B, red hues). In contrast, a more offshore domain located along the continental slope (hereafter Outer-Shelf region; Figure 4B, blue hues) displays negative loadings. These opposing loadings indicate that the temporal signals in the Inner and Outer shelf regions are in opposite phase. Stations situated between these two domains exhibit very low loadings, suggesting a minimal contribution from Factor 2 in these transitional areas. In terms of temporal variability, the power spectrum of Factor 2 (Figure 5B) shows a clear but relatively weak peak around 12 months, together with a broader increase in spectral power across the low-frequency range, particularly for periods longer than ~22 months. This spectral structure suggests enhanced variability at long periods (interannual or multiannual scales).

Factor 3 is most prominently expressed at stations 4 and 5, located at the mouth of the Guadalquivir River, with positive loadings of 0.53 and 0.67, respectively. Stations 3, 10, 85, and 57 also display positive loadings, although the strength of their association with the temporal signal of this factor is limited, as indicated by their low loading values (0.06, 0.03, 0.18, and 0.38, respectively). The power spectrum of this factor (Figure 5C) exhibits two marked peaks of comparable power, the first around 12 months and the second near 18–20 months. The dominant annual signal reflects a clear seasonal cycle, while the secondary enhancement at longer periods suggests an additional interannual modulation superimposed on the seasonal variability.

Factor 4 reveals two distinct groups of stations with opposite trends, as indicated by the contrasting signs of the loadings. The first group, which shows a strong positive correlation with this factor, is located in the northern shelf region. The second group, characterized by negative loadings, is situated in the southern sector of the study area. The periodogram of this factor (Figure 5D) displays two dominant spectral peaks, a smaller one around 15 months and a stronger one near 25 months (Figure 5D).

To investigate the mechanisms underlying the patterns captured by the four latent factors, we examined their relationships with a set of environmental variables selected for their potential influence on mesozooplankton dynamics. Environmental variables were selected based on both the spatial domains defined by each factor (Figure 4) and the spectral pattern observed in the corresponding periodogram (Figure 5), as both elements provide insight into potential environmental forcing acting on each temporal signal.

According to our results, seasonality emerges as the main driver modulating mesozooplankton dynamics on the northeastern shelf of the Gulf of Cádiz. This is consistent with expectations for mid-latitude systems such as this one (Longhurst, 1995; García Lafuente and Ruiz, 2007) where the annual cycle of irradiance, temperature, and water column stability exerts strong control over primary production, with minimum values in winter and maxima during spring and summer (Mann and Lazier, 2005). The observed correlation between the dominant latent signal and NPP (R = 0.62, p< 0.001, Table 1) suggests that mesozooplankton dynamics are largely linked to food availability and to the seasonal cycle of phytoplankton growth.

It is well established that temperature influences zooplankton, not only by modulating physiological rates such as growth and reproduction, but also by acting as a timing cue that regulates their seasonal phenology (Mackas et al., 2012). However, Table 1 shows higher correlation with NPP and irradiance, what suggests that food availability plays a more important role than temperature in shaping mesozooplankton seasonal dynamics.

This response of mesozooplankton to seasonal forcing is not homogeneous across the continental shelf. Differences among stations indicate that, although seasonality acts as a dominant driver, its expression may be modulated by specific environmental conditions in each area. In particular, coastal zones influenced by river discharge, including the mouth of the Guadalquivir River and the northern sector between the Guadiana and Tinto-Odiel rivers, as well as the southern shelf in the vicinity of Cape Trafalgar, exhibit a weaker expression of the seasonal pattern.

This attenuation of the seasonal signal may be attributed to the influence of processes operating on non-seasonal timescales. In coastal zones influenced by rivers, the continuous inputs of nutrient- and organic-rich freshwater can sustain relatively stable trophic conditions year-round (Mann and Lazier, 2005; García Lafuente and Ruiz, 2007). This may contribute to the weakening of the expected seasonal signal in plankton dynamics in these areas, particularly since the seasonality of river flows in southern Spain is highly modified by human regulation.

This mechanism is particularly evident at the mouth of the Guadalquivir River, as indicated by the low loading values at stations 4, 5, and 6 (Figure 4A). The strong regulation of the river flow by dams and other hydraulic infrastructures (Vilas et al., 2009) alters the natural discharge pattern and partially decouples freshwater inputs from precipitation-driven seasonality. As a result, trophic conditions may become less tightly linked to the seasonal cycle, contributing to a weaker expression of the seasonal signal in mesozooplankton dynamics.

In the area surrounding Cape Trafalgar, the attenuation of the seasonal pattern appears to be influenced by local physical processes that differ from the broader seasonal forcing. The interaction between the steep bathymetry and tidal currents in this region generates intermittent pulses of local upwelling that enrich the water column episodically and not necessarily in a seasonal manner, as documented in previous studies (Vargas et al., 2003; Bolado-Penagos et al., 2020; Sala et al., 2023).

Beyond the dominant seasonal signal, mesozooplankton variability across the northeastern shelf also exhibits a clear spatial organization. In particular, two spatial patterns emerge across the study area, a latitudinal differentiation between northern and southern sectors and a contrast between coastal and offshore stations. These spatial patterns are examined in detail in the following sections.

One of the main spatial axes structuring mesozooplankton variability on the northeastern shelf of the Gulf of Cádiz is a latitudinal contrast between the northern and southern sectors, revealing the presence of regionally differentiated dynamics superimposed on the common seasonal framework described above. This north–south pattern is consistently expressed in the spatial structure of Factor 4, whose loading distribution clearly separates northern and southern stations, indicating that a significant fraction of mesozooplankton variability might be modulated by large-scale regional processes operating across the shelf.

This interpretation must, however, be considered in light of a methodological limitation. Because northern stations were sampled mainly in 2002–2007, whereas southern stations correspond to 2005–2007, part of the observed north–south contrast may reflect interannual variability in addition to the regional gradient. Even so, the spatial coherence of the pattern—evidenced by the consistent behavior of neighboring stations—together with the latitudinal species-level zonation previously documented in the region (Llope et al., 2020), supports the view that the signal most likely reflects a genuine ecological gradient rather than an artefact of the sampling scheme.

Under this latitudinal contrast, two distinct regional contexts can be distinguished (Figure 4D). The northern sector is characterized by a stronger continental influence, associated with the presence of several river systems that supply freshwater, nutrients, and suspended material, contributing to a more complex and variable hydrographic environment (García Lafuente and Ruiz, 2007; Navarro and Ruiz, 2006; Ruiz et al., 2013; Gomiz-Pascual et al., 2021). In contrast, the southern sector exhibits a more distinctly oceanic character, with limited direct fluvial influence and a dynamic largely governed by regional circulation and the proximity of the Strait of Gibraltar (Vargas et al., 2003; Peliz et al., 2014; Bolado-Penagos et al., 2020).

Among the different forcings that may contribute to these regional differences, the wind regime stands out as a particularly relevant factor at the regional scale, given its role in structuring circulation and hydrographic properties in the Gulf of Cádiz (Vargas et al., 2003; Criado-Aldeanueva et al., 2006; García Lafuente and Ruiz, 2007). The shelf displays a marked zonal wind pattern with a strong latitudinal component. In the southern sector, easterly winds (Levante) are more frequent and intense due to the proximity of the Strait of Gibraltar, whereas towards the north their influence weakens and westerlies progressively gain importance (Peliz et al., 2014; Gomiz-Pascual et al., 2021; Mulero-Martinez et al., 2024). This transition in the relative influence of easterly and westerly winds is located approximately between the Bay of Cádiz and the mouth of the Guadalquivir River, based on the mean wind fields derived from Copernicus data for the study period (Figure 10). That figure clearly shows two distinct regions to the north and south of the shelf, each with markedly different wind-forcing characteristics, whose spatial distribution closely matches the loading map of Factor 4 (Figure 4D). Notably, the transition between these two wind regimes coincides spatially with the boundary between the zones delineated by Factor 4, suggesting that the signal captured by this factor may reflect a latent mesozooplankton response to differences in wind forcing between these two regions.

As noted in previous studies, variability in the influence of the dominant wind regime in the region may differentially modulate the system’s dynamic and hydrographic characteristics, such as water column stratification, sea surface temperature, or local circulation (Peliz et al., 2014; Bolado-Penagos et al., 2020; Gomiz-Pascual et al., 2021), all of which are key factors in shaping the structure and dynamics of the planktonic community (Mann and Lazier, 2005). In fact, Figures 9A, B suggest that the most pronounced declines in the temporal signal of Factor 4 may coincide with easterly wind events exceeding 40 km/h, raising the possibility of a mesozooplankton response to episodes of intense easterly winds. Previous studies have indicated that strong easterly wind events can negatively affect the development of early life stages of pelagic species in the Gulf of Cádiz by promoting their displacement toward less favorable areas (Ruiz et al., 2006; Ruiz and Rincón, 2018).

Based on these findings, the spatiotemporal pattern associated with Factor 4 may reflect a regional-scale response of the mesozooplankton community to wind-driven forcing, particularly in relation to the contrasting dynamics between the northern and southern sectors of the shelf.

Beyond the latitudinal pattern, another spatial organization emerges when examining latent Factor 2. Its loading distribution suggests a coastal–offshore gradient that appears to follow the underlying bathymetric structure, distinguishing the shallow inner shelf from the deeper outer shelf (Figure 4B).

However, in the southern sector of the shelf, the coast–offshore gradient associated with Factor 2 exhibits a structure in which negative loadings are not restricted solely to the most oceanic slope stations; in addition, positive loadings along the coastal band are notably weaker than in the northern sector. This configuration reflects the strong impact of the instabilities generated by the exchange of water masses between the Atlantic Ocean and the Mediterranean Sea in the vicinity of the Strait of Gibraltar (Vargas et al., 2003; Bolado-Penagos et al., 2020; Sala et al., 2023). Along the easternmost transect (stations 51, 52 and 53; Figure 1), the influence of these processes is so pronounced that it completely blurs the coast–offshore differentiation, resulting in negative Factor 2 values across the entire profile, including the innermost station (station 51).

In contrast, in the northern sector, the particularly high positive loadings of Factor 2 along the coastal band situated between the mouths of the Tinto–Odiel and Guadalquivir rivers (Figure 4B), with values increasing toward the mouth of the latter, indicate that the variability represented by this component reflects not only a coast–slope gradient, but also the direct influence of the Guadalquivir’s fluvial discharge.

This coastal band between the mouths of the Tinto–Odiel and Guadalquivir rivers, defined by the positive loadings in the northern sector (Figure 4B), has been widely characterized in previous studies as a region of elevated primary productivity. This has been attributed to the combined effects of tidal dynamics, warm temperatures and the continuous inflow of freshwater enriched with nutrients and organic matter from the river (Navarro and Ruiz, 2006; Ruiz and García-Lafuente, 2006; Prieto et al., 2009; González-Ortegón et al., 2018; Caballero et al., 2014; Gomiz-Pascual et al., 2021). It has also been described as an important spawning and nursery area for various meroplanktonic species, owing to suitable conditions for larval development and survival (García-Isarch et al., 2003; Baldó et al., 2006; Ruiz et al., 2006). This is consistent with our observations, which show that this coastal sector supports the highest mesozooplankton biovolumes on the shelf, particularly for stations located at an intermediate distance from the river mouth (stations 6, 7, 8; Figure 2A).

A striking feature is that stations located directly at the Guadalquivir mouth (stations 4 and 5) show low mesozooplankton biovolume values (Figure 2A), despite elevated chlorophyll-a concentrations (Figure 2B), which is a persistent feature of the estuary mouth throughout the year (García Lafuente and Ruiz, 2007). In agreement with our results, previous studies have shown that the Guadalquivir estuary mouth differs markedly from both the adjacent marine area (Baldó et al., 2006; Llope et al., 2020) and the inner estuary (Taglialatela et al., 2014; Reyes-Martínez et al., 2024), with mesozooplankton biomass and abundance remaining low, a pattern also reported in other estuary–shelf systems (Diadychko and Kharytonova, 2024).

This apparent decoupling suggests that river-associated processes constrain trophic transfer from phytoplankton to mesozooplankton in the continental shelf at the immediate vicinity of the river mouth, highlighting the complexity of this area.

This is reflected in the temporal evolution of the cross-shelf mode (Factor 2), which shows that periods of enhanced freshwater input are not systematically associated with higher mesozooplankton values, suggesting that river influence may exert both positive and negative effects depending on the prevailing conditions. The 2003 anomaly is particularly illustrative, as intense rainfall episodes produced high discharges at the Alcalá del Río dam, associated with elevated NPP and Chla (Figures 7B, C), yet the second factor declined that year (Figure 7A) and remained high during the 2004–2006 drought period (García-Herrera et al., 2007). The 2003 decline coincides with persistently low salinity and high ISPM (Figures 7A, C), consistent with sustained fluvial influence that may have driven mesozooplankton outside its tolerance range, counteracting the benefits of increased food availability associated with river fertilization (Navarro and Ruiz, 2006; Caballero et al., 2014; Gomiz-Pascual et al., 2021). Such a decline is likely driven by the fact that low salinity can impair mesozooplankton feeding, metabolism, and reproduction, or trigger avoidance behaviors (Lance, 1962; Calliari et al., 2006; Smyth and Elliott, 2016; Oghenekaro and Chigbu, 2019; Podbielski et al., 2022).

In addition to low salinity, rivers may bring suspended solids and turbidity to adjacent shelf waters. Guadalquivir estuary exhibits exceptionally high suspended solids compared with other rivers worldwide (Navarro et al., 2011; Diez-Minguito et al., 2014; Ruiz et al., 2015; Megina et al., 2023) and its plumes strongly reduce light penetration in the water column of adjacent coastal areas (Caballero et al., 2011; Gomiz-Pascual et al., 2021). It is known that these conditions affect negatively to mesozooplankton feeding and swimming behavior (Newcombe and Macdonald, 1991; Bilotta and Brazier, 2008; Arruda et al., 1983; Sew et al., 2018; Liu et al., 2020). Horizontal advection also adds another layer of complexity to this framework since high discharges could transport estuarine waters and their planktonic communities away from this area (Gomiz-Pascual et al., 2021).

This interpretation, according to which estuarine influence may exert a negative effect on mesozooplankton at stations close to the river mouth, is reinforced by the spatial structure of Factor 3. This factor clearly differentiates the river mouth from the rest of the shelf, pointing to additional local processes not captured by the cross-shelf gradient (Factor 2). While coastal areas of the shelf typically exhibit a biomass peak during the summer (Llope et al., 2020; Mafalda et al., 2007), this temporal pattern is absent at the Guadalquivir mouth. The temporal signal of Factor 3 shows a marked negative correlation with temperature (Table 1), with its latent trend reaching minimum values during the warmer months. This behavior suggest that the processes captured by Factor 3 and that are associated to temperature signal prevent stations 4 and 5 from developing the seasonal maximum observed elsewhere in summer. Consequently, this localized inhibition during the most productive season leads to the lower annual average biovolume values observed at the estuary mouth.

In summary, the same freshwater inputs that fertilize the system may also limit mesozooplankton development, with this negative influence being strongest at the estuary mouth but potentially extending offshore during periods of higher river discharge. At the river mouth, this limitation is further reinforced by local processes that disrupt the seasonal development of mesozooplankton, preventing the emergence of the typical summer maximum and leading to persistently low annual biovolume despite high phytoplankton availability. Together, these results highlight the estuary mouth as an ecologically distinct domain within the shelf, where enhanced primary production do not necessarily translate into increased mesozooplankton biomass.