In this study, we aimed to determine if tropical cyclones in the northwestern GOM differentially affected mesozooplankton community structure. We found that multivariate community structure varied between storm and non-storm years. However, among the three hurricanes, only the post-Harvey mesozooplankton communities were distinct from years where no storms occurred. Multivariate dispersion, a measure of variance within community structure, did not differ between storm and non-storm years. This result refutes our hypothesis that variability in zooplankton community structure would be higher during storm years opposed to non-storm years. Our prediction that post-storm mesozooplankton communities would differ from non-storm communities was supported, as was our expectation that mesozooplankton community structure varied among storms. We hypothesized that due to the major flooding and rainfall of Harvey, reduced salinity would likely be the main driver of mesozooplankton community differences relative to non-storm years. This expectation was partially met; differences in mesozooplankton community structure between Harvey and years with no storms were driven by salinity, stratification, and ‘distance-from-shore’ rather than solely salinity. This indicates that NWGOM zooplankton community structure varies holistically with biophysical conditions rather than being primarily driven by one or two dominating factors24. Moreover, we found that the presence of Hurricane Harvey, rather than temperature, explained the second greatest amount of variance in zooplankton community structure reflecting the importance of considering how complex disturbance mechanisms might compound and result in a unique ecological responses following a tropical cyclone. The overall PSEM showed that high salinity was directly associated with reduced fluorescence and depressed zooplankton abundance. Higher abundances were found to result in higher community dominance (i.e., lower evenness). Conversely, high salinities indirectly reduced zooplankton evenness via greater water column stratification. Lower fluorescence at higher salinities supports the spatial patterns in NWGOM phytoplankton identified by Kurtay et al.20 following Hurricane Harvey. Those authors observed declines in overall phytoplankton abundance and communities increasingly dominated by cells < 20 µm in size waters along a coastal to oceanic gradient.
Because Ike and Rita years were indistinguishable from either Harvey or non-storm years, we focused on the statistically supported contrast of Harvey and non-storm years to investigate differences in zooplankton community structure. Our expectation was that we would find a larger effect size between salinity, mesozooplankton abundance, and community evenness following Harvey. This hypothesis was partially supported. The PSEM “multigroup” analysis showed that relationships between salinity and primary production (i.e., fluorescence) switched signs and was reduced by an order of magnitude after Harvey. This reflects the post-Harvey shift in underlying phytoplankton community, towards smaller heterotrophic organisms reported by Kurtay et al.20. Finally, we see that after Harvey, higher temperatures led to lower zooplankton community evenness whereas the reverse was true for non-storm years.
We found that increased zooplankton food availability (i.e., fluorescence) was linked to a decrease in zooplankton community evenness during non-storm years but, this connection was greatly weakened and the direction changed during Harvey, likely due to the overall reduction in fluorescence20. The reduction of zooplankton evenness at high productivity during non-storm years may be explained by the “paradox of enrichment”, where excessive resource availability (e.g., high productivity) can result in specialization, leading to dominance of a few taxa, reducing community diversity and evenness37.
Notably, our model did not link fluorescence and mesozooplankton abundance in either the full or multi-group PSEM. The absence of this connection is likely due to the coarse summary of fluorescence as a metric for productivity, in which all photosynthetically active producers are grouped together. Producers are highly diverse, both taxonomically and functionally (e.g., size), with a wide range of habitat preferences. This diversity creates variation in the spatial and temporal resource environment for grazing mesozooplankton (e.g., pico- vs. nanoplankton20). Additionally, heterotrophic microzooplankton may constitute a majority of mesozooplankton diet, yet these organisms are not detected via chlorophyll fluoresced metrics.
The influence of temperature and salinity on zooplankton abundance and evenness between Harvey and non-storm years varied substantially. Confirming our hypothesis that salinity would be the major influence on zooplankton abundance, post-Harvey zooplankton abundances declined significantly more quickly with increased salinities compared to non-storm years. It is important to note that this pattern is found when spatial covariation is isolated, essentially removing the known influence of distance from shore gradients on salinity and abundance. We attribute the switch in directionality and significance of temperature on evenness between Harvey and non-storm years to the compressed temperature range exhibited during Harvey. This compression reduced the influence of temperature on zooplankton dominance resulting in the negative, non-significant relationship shown. These results indicate that cyclones like Hurricane Harvey can affect northwestern GOM zooplankton communities via changes in salinity, either through direct effects on zooplankton abundance or indirectly through the reduced influence of fluorescence on zooplankton evenness.
Tropical cyclones represent large coastal disturbances, and it is expected that years during which large storms pass through an area, community composition and structure of the biologic constituents will differ from years of no cyclone disturbance2,5,17. We found that presence of a storm significantly predicted differences in mesozooplankton community structure relative to non-storm years, supporting our original hypothesis. However, we surprisingly found that only post-Harvey zooplankton communities were statistically distinguishable from non-storm years. This lack of variation may be attributed to the unique characteristics of each storm included in our analysis. Hurricanes Ike and Rita both altered the coastal environment of Louisiana and Texas, but largely through different mechanisms than Harvey. Hurricane Ike, a Category 2 storm, pushed cooler, saltier ocean water into the coastal environment via a wave of storm surge greater than either Harvey or Rita. Hurricane Rita had similar wind speeds and storm surge to Harvey, but its precipitation rate was significantly less, falling between Harvey and Ike. Harvey had the highest recorded wind speeds as a Category 4 storm at landfall, and its rainfall totals far exceeded the precipitation of Ike and Rita, causing major flooding of the Houston area, drastically changed the salinity, turbidity, and temperature of Galveston Bay and adjacent coastal NWGOM1,17,19.
These differences in disturbance mechanism (e.g., precipitation, storm surge, mixing) and intensity may explain the variance in zooplankton communities between storm and non-storm years, but the environmental baseline remains an important consideration as antecedent conditions of a system have been shown to dictate community response after a disturbance5. While salinity values were significantly reduced relative to historic years following Harvey20 (both at the mouth of Galveston Bay and nearshore), we show that salinity values across the continental shelf do not vary considerably between Harvey and the non-storm years in our data set. Rather, we see reduced temperature and fluorescence, and increased water column stratification following Harvey. We found that salinity, fluorescence, and distance from shore (i.e., PC1), explained the most variance in community structure in our PERMANOVA, and whether or not zooplankton were collected after Harvey followed closely behind as the second strongest predictor. These conditions indicate that drivers of zooplankton community change after large cyclones like Harvey are not always obvious, and both storm characteristics and natural variability of a system must be accounted for when investigating the ecological consequences of cyclones (sensu6).
The ambiguity of which physical driver best describes zooplankton community structure in the northwestern GOM is unsurprising as these organisms are poikilothermic drifters that exhibit an incredible diversity of life history traits and environmental tolerances. This variety is reflected in the multitude of taxa found to influence differences in zooplankton community structure between Harvey and non-storm years from our SIMPER analysis. When we compare the abundances of taxa between Harvey and non-storm years, we see a general reduction across taxa such as amphipods, gastropods, cumaceans, mysids, polychaetes, and gelatinous zooplankton. An interesting result is the conspicuous absence of cladocerans following Hurricane Harvey. Cladocera taxa are often used as disturbance indicators in freshwater and coastal systems38,39 and their absence may reflect the strong environmental changes brought on by Harvey in the NWGOM as has been shown after other tropical cyclones (e.g., Hurricane Agnes40). Cladocera are known prey items for larval fish in this region (e.g., Cynoscion nothus, Thunnus thynnus41,42), and disturbances such as tropical cyclones may disrupt the temporal availability of this food source, leading to decreased fitness and recruitment.
While overall abundance often dictates the productivity of zooplankton communities, it is also pertinent to consider the size structure of taxa. Body size has been deemed the master functional trait in zooplankton ecology, as almost all ecological interactions and rates are intimately related to how large or small a zooplankter is43. Zooplankton body size may also influence the structuring of higher trophic levels through prey suitability. Larval fish predation on zooplankton is size-limited, constrained by maximum gape width or gill raker density41, while gelatinous filter feeders such as salps, doliolids and appendicularians are limited by their mesh size. While we see generally strong overlap in size distribution between Harvey and non-storm years, few taxa, such as doliolids, salps, and gelatinous zooplankton show a greater size after Harvey, though their abundance is decreased or equivocal to non-storm years (Fig. 2). Changes in the abundance of larger, soft bodied organisms such as gelatinous zooplankters may alter the available carbon within coastal food web and redirect secondary production through the microbial loop rather than transferring energy to higher trophic levels. Additionally, filter feeding gelatinous zooplankton such as doliolids and salps can process and filter greater volumes of water at increased body sizes44,45, potentially increasing their ecological success leading to positive feedback of recovering energy from the microbial loop46 and enhancing vertical carbon flux47. Kurtay et al.20 reported an increase in pico- and nanoplankton, prey items that falls within the filtering size of both doliolids and salps44,45. The increase in their available prey after Harvey may have resulted in our observed increase in gelatinous filter feeding size. Taxa specific responses to tropical cyclones are of high interest, as different organisms within the zooplankton community react differently to environmental disturbances. Understanding which disturbance mechanism relates to which specific taxa is beyond the scope of this study, but it is an important avenue to pursue if we are to better understand long-term food web consequences within coastal ecosystems.
Holistic approaches towards understanding disturbance mechanisms, such as community structure, provide a complex picture of taxa-specific responses and compounding environmental predictors. However, when we collapsed zooplankton community variability into a univariate metric (i.e., zooplankton abundance and evenness), a clearer relationship between environmental drivers and our zooplankton emerged. Mesozooplankton evenness was largely influenced by stratification and mesozooplankton abundance. Additionally, salinity indirectly impacted evenness through water column stratification. As salinity decreases, stratification increases, reducing the overall evenness of zooplankton. Zooplankton abundance was negatively related to both salinity and zooplankton evenness. This relationship is highlighted through a major shift in the influence of salinity on mesozooplankton abundance after a high precipitation storm event (i.e., Hurricane Harvey). The overwhelming influence of salinity on mesozooplankton abundance following Harvey, but not during non-storm years, emphasizes the importance of unique storm characteristics and the difference in their mechanistic impact on coastal ecosystems. Within the current climate regime, larger, wetter storms are likely to occur in the Northern Hemisphere25. As we show here, increased precipitation from tropical cyclones can lead to a local change in zooplankton abundance in the northwestern Gulf of Mexico. Storms may also have far-reaching impacts on coastal zooplankton communities through their impacts on freshwater discharge upstream of coastal estuaries (e.g., Atchafalaya and Mississippi River). High precipitation events can modify alongshore transport and current patterns of coastal waters off the coast of Texas15,21 altering the environment and distribution of coastal zooplankton taxa. Consequently, larger, wetter cyclones can modify coastal food webs not only through salinity-controlled zooplankton abundances (Fig. 4) or compositional shifts (e.g.4), but via physical displacement potentially causing a spatial mismatch between plankton and their predators.
We show that Harvey affected northwestern GOM zooplankton communities for up to a month following landfall, longer than has been reported for other storms48,49. The extended temporal span of Harvey’s impacts on zooplankton, relative to Rita and Ike, likely has to do with the duration of post-storm coastal ocean conditions in the northwestern GOM. After Hurricane Rita, elevated water levels returned to pre-storm values within 48 h9. Similarly, after Ike, elevated water heights caused by the intrusion of 5 m of storm surge were recorded for 3 days after landfall12, and sea surface temperatures and elevated chlorophyll-α remained for 2 days before returning to pre-Ike levels13. In contrast, water levels of Galveston Bay were elevated for up to a week after Harvey, and salinities within the Bay did not return to pre-storm levels for over 2 months (62 days1). It is important to note that the residence time of Galveston Bay is around 1 month50. After the freshening event from Harvey’s rain, the bay acted like a low salinity reservoir to the NWGOM, exchanging saltwater via tidal pumping and extending the duration of low salinity intrusion into coastal waters1. This scenario is supported by prolonged levels of elevated water velocity at the Galveston Bay channel recorded until 6 September 2017 and elevated sediment flux out of the Bay until 26 September 2017, almost an entire month after Harvey made landfall1,19.
As energetic expressways between primary production and higher consumers, changes in the abundance and evenness of zooplankton communities impact energy transfer in coastal food webs. Large increases in coastal zooplankton abundance, like we see following Harvey22, can represent a flood of food resources for planktivorous consumers in the area (e.g., forage fish). Zooplankton abundance is often closely linked to larval fish recruitment success, with higher zooplankton abundances representing greater availability of prey in the water column51, but can also represent increased top-down control by zooplankton predating on larval fish52. We show that increases in zooplankton abundance drove down evenness, which may restrict energy transfer within the food web through selectivity effects. Plankton communities where only a few taxa are numerically dominant can become energetically unfavorable for trophic transfer if the dominant taxa are unpalatable for consumers53. Changes in taxa dominance can result from mechanistic selectivity issues for predators if prey selection is determined by gape or gill raker size (e.g., larval fish41). Additionally, behavioral changes of the dominant taxa (e.g., distribution within water column, diel vertical migration) could lead to an ecological mismatch in capture efficiency by predators. For fish that spawn between September and October when cyclones activity is prevalent (e.g., Cynoscion nothus, Larimus fasciatus54), the abundance and composition of zooplankton in the northwestern Gulf of Mexico play a vital role in overall fitness of the organism41. Further connection between storm impacts on zooplankton size, distribution, and its downstream effect on larval fish is warranted to understand how cyclones can impact commercially important fish populations.
Source: Ecology - nature.com
